DOI: 10.1038/s41467-024-55050-y

Resource: RRID:BDSC_48925

Curator: @scibot

SciCrunch record: RRID:BDSC_48925

What is this?

DOI: 10.1038/s41467-024-55050-y

Resource: RRID:BDSC_68251

Curator: @scibot

SciCrunch record: RRID:BDSC_68251

What is this?

DOI: 10.1038/s41467-024-55050-y

Resource: RRID:BDSC_49840

Curator: @scibot

SciCrunch record: RRID:BDSC_49840

What is this?

DOI: 10.1038/s41467-024-55050-y

Resource: RRID:BDSC_57617

Curator: @scibot

SciCrunch record: RRID:BDSC_57617

What is this?

DOI: 10.1038/s41467-024-55050-y

Resource: RRID:BDSC_46382

Curator: @scibot

SciCrunch record: RRID:BDSC_46382

What is this?

DOI: 10.1038/s41467-024-55050-y

Resource: RRID:BDSC_7010

Curator: @scibot

SciCrunch record: RRID:BDSC_7010

What is this?

DOI: 10.1038/s41467-024-55050-y

Resource: RRID:BDSC_48880

Curator: @scibot

SciCrunch record: RRID:BDSC_48880

What is this?

DOI: 10.1038/s41467-024-55050-y

Resource: RRID:DGRC_1470

Curator: @scibot

SciCrunch record: RRID:DGRC_1470

What is this?

DOI: 10.1038/s41467-024-55050-y

Resource: RRID:BDSC_39351

Curator: @scibot

SciCrunch record: RRID:BDSC_39351

What is this?

DOI: 10.1007/s11571-024-10185-y

Resource: (Thermo Fisher Scientific Cat# A-11034, RRID:AB_2576217)

Curator: @scibot

SciCrunch record: RRID:AB_2576217

What is this?

DOI: 10.1007/s11571-024-10185-y

Resource: (Promega Cat# W4011, RRID:AB_430833)

Curator: @scibot

SciCrunch record: RRID:AB_430833

What is this?

DOI: 10.1007/s11571-024-10185-y

Resource: (Thermo Fisher Scientific Cat# A-21443, RRID:AB_2535861)

Curator: @scibot

SciCrunch record: RRID:AB_2535861

What is this?

DOI: 10.1007/s11571-024-10185-y

Resource: (Promega Cat# W4021, RRID:AB_430834)

Curator: @scibot

SciCrunch record: RRID:AB_430834

What is this?

DOI: 10.1007/s11571-024-10185-y

Resource: (MBL International Cat# M186-3, RRID:AB_10897859)

Curator: @scibot

SciCrunch record: RRID:AB_10897859

What is this?

DOI: 10.1007/s11571-024-10185-y

Resource: (Abcam Cat# ab124913, RRID:AB_10975154)

Curator: @scibot

SciCrunch record: RRID:AB_10975154

What is this?

DOI: 10.1007/s11571-024-10185-y

Resource: (Abcam Cat# ab56416, RRID:AB_945626)

Curator: @scibot

SciCrunch record: RRID:AB_945626

What is this?

DOI: 10.1007/s11571-024-10185-y

Resource: (Abcam Cat# ab18258, RRID:AB_444362)

Curator: @scibot

SciCrunch record: RRID:AB_444362

What is this?

DOI: 10.1007/s11571-024-10185-y

Resource: (Abcam Cat# ab65783, RRID:AB_1658870)

Curator: @scibot

SciCrunch record: RRID:AB_1658870

What is this?

DOI: 10.1007/s11571-024-10185-y

Resource: (Cell Signaling Technology Cat# 3495, RRID:AB_1903911)

Curator: @scibot

SciCrunch record: RRID:AB_1903911

What is this?

http://sea-entomologia.org/PDF/BSEA45ARACNO/B45173.pdf

La transcription de la conférence ne se concentre pas sur les actions spécifiques que la famille, l'établissement et l'institution scolaire peuvent mettre en place pour favoriser la réussite des élèves français.

Le sujet principal est la réussite scolaire des élèves d'origine asiatique.

Cependant, la transcription aborde certains facteurs qui peuvent influencer la réussite scolaire de tous les élèves, y compris les élèves français.

À partir de ces informations, on peut déduire quelques pistes d'actions possibles.

Actions de la famille:

• Croire aux bienfaits de l'éducation et transmettre cette conviction aux enfants:

La transcription mentionne que "la forte croyance des parents dans les bienfaits de l'éducation" est une hypothèse pour expliquer la réussite des élèves d'origine asiatique.

• Fixer des exigences élevées et encourager l'éthique de travail, l'effort, l'application et la persévérance :

La transcription note que ces éléments sont également présents chez les familles d'origine asiatique.

• Maintenir une communication ouverte avec l'école et s'impliquer dans la scolarité des enfants: La transcription souligne l'importance des relations famille-école.

L'existence d'une association de médiation scolaire pour les familles chinoises suggère que la communication peut être un défi pour certaines familles.

• Favoriser la transmission intergénérationnelle et maintenir des liens familiaux forts :

La transcription note que des liens familiaux fragilisés peuvent affecter la réussite scolaire.

Actions de l'établissement scolaire:

• Mettre en place des programmes de médiation scolaire pour faciliter la communication entre les familles et l'école :

L'exemple de l'association franco-chinoise Pierre Duer montre que ce type de programme peut être bénéfique.

• Prendre en compte les difficultés liées à la ségrégation résidentielle et scolaire et proposer des solutions pour y remédier :

La transcription mentionne la ségrégation comme un obstacle à la réussite.

• S'adapter aux besoins spécifiques des élèves issus de l'immigration, notamment en ce qui concerne la langue et la culture :

La transcription aborde la question des pratiques culturelles et linguistiques au sein des familles.

Actions de l'institution scolaire:

• Améliorer les outils statistiques pour mieux comprendre les facteurs de réussite scolaire des élèves issus de l'immigration et adapter les politiques éducatives en conséquence :

La transcription souligne les limites des enquêtes statistiques actuelles.

• Encourager la recherche sur les pratiques éducatives des familles d'origines diverses et diffuser les résultats auprès des professionnels de l'éducation :

La transcription note le manque de recherches sur ce sujet en France.

• Promouvoir l'égalité des chances et lutter contre les discriminations liées à l'origine sociale et ethnique :

La transcription rappelle que l'origine socio-économique est un facteur important de réussite scolaire.

Cette transcription d'une conférence au Collège de France explore les stratégies éducatives familiales, mettant l'accent sur la socialisation culturelle plutôt que sur l'éducation scolaire.

L'intervenant analyse trois piliers de cette socialisation : les objets culturels, les interactions parents-enfants, et l’exemple parental.

Il souligne les inégalités et les discriminations liées à l’accès à la culture, notamment en ce qui concerne l’usage des écrans et les pratiques culturelles extra-scolaires.

Enfin, il compare deux modèles éducatifs contrastés, la concerted cultivation et le natural growth, pour illustrer la diversité des approches parentales et leur impact sur le parcours scolaire des enfants.

Sommaire minuté des points forts:

Introduction (0:00 - 2:30):

• Le sujet abordé est celui des pratiques culturelles des enfants et comment elles éclairent les stratégies éducatives des parents.
• La perspective adoptée s'éloigne de l'éducation institutionnelle pour se concentrer sur la socialisation, un ensemble de pratiques incluant l'éducation mais qui la dépasse.
• Il y a une confrontation entre le "métier d'enfant" et le "métier d'élève", qui ont des impératifs et des cadres différents.
• L'exposé se centrera sur la socialisation culturelle familiale, notamment le rôle des parents.

Les piliers de la socialisation culturelle (2:30 - 4:45):

• Trois piliers sont généralement distingués: le système des objets (culture matérielle), les interactions éducatives avec les parents et les exemples parentaux.
• L'accès aux objets culturels et les autorisations d'accès données aux enfants sont des éléments importants des stratégies éducatives.
• Les interactions éducatives incluent les dynamiques relationnelles, les pratiques d'occupation du temps et les représentations de l'enfant et de la "bonne parentalité".
• Les exemples parentaux, bien que moins documentés, définissent un champ des possibles pour les enfants.
• Ces trois piliers construisent des "climats familiaux", qui peuvent être compris comme des stratégies de fait, sans intention éducative prégnante.

Contexte actuel et particularités (4:45 - 7:30):

• Le contexte actuel se caractérise par une forte emprise scolaire, avec une centralité de l'école dans la validation des compétences et le classement des individus.
• Cette centralité crée un paradoxe douloureux dans un contexte de crispation économique et d'inflation des diplômes.
• Il y a une forte pression à la "bonne parentalité", alimentée par un nombre croissant de rapports et de normes sociales.
• Les discours de légitimation pour les loisirs, notamment culturels, avancent masqués et mettent en avant l'épanouissement personnel pour masquer des objectifs de bénéfices scolaires.

Exemples d'activités et analyse (7:30 - 19:00):

• Enfants et écrans:

Les discours sur les écrans sont à charge et mettent l'accent sur les risques.

L'usage réel des écrans est très différent des normes institutionnelles, servant à la régulation des temps et à l'intégration linguistique et sociale.

• Langage et lecture:

L'oralité est privilégiée dans les catégories peu diplômées, tandis que les catégories plus diplômées favorisent le livre.

Les pères accentuent les différences dans les pratiques. Les enfants d'immigrés se mettent à lire moins malgré une plus grande exposition à l'oralité.

• Pratiques en amateur:

Les clubs et associations sont choisis pour doter les enfants de ressources éducatives, développer des passions et créer des liens sociaux.

Ces activités sont très genrées et les parents les plus investis sont ceux qui ont le plus de capitaux scolaires.

• Ecrans (suite):

Les enfants des fractions intellectuelles accèdent aux écrans numériques plus tard, illustrant une stratégie d'effet retard.

L'entrée à l'école renforce les normes institutionnelles et impacte différemment les fractions de la population.

Conclusion (19:00 - 21:00):

• Il faut déconstruire la normativité présente dans les discours sur les stratégies éducatives.
• Les stratégies éducatives des familles populaires, même si elles diffèrent des normes institutionnelles, sont des tentatives positives d'émancipation.
• Il faut distinguer deux modèles éducatifs: la "concerted cultivation" (culture concertée) et le "Natural growth" (croissance naturelle).
• Ces deux modèles produisent des enfants qui ne s'autorisent pas de la même manière à parler aux adultes dans les institutions, avec des effets sur la scolarité.

Points clés à retenir:

• Différences de stratégies éducatives selon le milieu social:

Les pratiques culturelles et l'accès aux objets culturels varient fortement selon le capital culturel et la position sociale des familles.

• Rôle des exemples parentaux: Les pratiques des parents, même non intentionnelles, ont un impact majeur sur les trajectoires des enfants.

• Impact des normes institutionnelles:

L'école joue un rôle central dans la validation des compétences et la diffusion de normes, ce qui influence les pratiques familiales. * Diversité des stratégies éducatives:

Il n'existe pas de "bonne" stratégie universelle, chaque famille met en place des pratiques qui répondent à son contexte et à ses aspirations.

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

• ???? -¿Qué motivó a De Broglie a proponer su hipótesis? La idea de que la materia, al igual que la luz, podría tener propiedades duales de onda y partícula. (Einstein, 1905)

• Planck sobre solids heat capacity???

• fue Einstein y luego Debye:

• Einstein, A. (1907). Die Plancksche Theorie der Strahlung und die Theorie der spezifischen Wärme. Annalen der Physik, 327(1), 180-190.

• Debye, P. (1912). Zur Theorie der spezifischen Wärmen. Annalen der Physik, 344(14), 789-839.

• Fue un martillo, no una moneda
• CHECK si se probo antes de grabar

• La "velocidad" no se puede "ver"

• REFERENCES NEEDED

-

-

-

• CHECK

• CHECK

• se refiere a los neutrinos de OPERA?
• debido a mala "conexion de un cable"

• vs Berkeley

• Pegawai A akan mempunyai anak X dan Y, maka list anak akan kebawah muncul dari list pegawai
• Jadi ada dua pointer, satu buat hubungan dan satu buat next. Tapi list disatukan

Reviewer #2 (Public review):

To fuse, differentiated muscle cells must rearrange their cytoskeletaon and assemble actin-enriched cytoskeletal structures. These actin foci are proposed to generate mechanical forces necessary to drive close membrane apposition and fusion pore formation.

While the study of these actin-rich structures has been conducted mainly in drosophila, the present manuscript presents clear evidence this mechanism is necessary for the fusion of adult muscle stem cells in vivo, in mice.

However, the authors need to tone down their interpretation of their findings and remember that genetic proof for cytoskeletal actin remodeling to allow muscle fusion in mice has already been provided by different labs (Vasyutina E, et al. 2009 PMID: 19443691; Gruenbaum-Cohen Y, et al., 2012 PMID: 22736793; Hamoud et al., 2014 PMID: 24567399). In the same line of thought, the authors write they "demonstrated a critical function of branched actin-propelled invasive protrusions in skeletal muscle regeneration". I believe this is not a premiere, since Randrianarison-Huetz V, et al., previously reported the existence of finger-like actin-based protrusions at fusion sites in mice myoblasts (PMID: 2926942) and Eigler T, et al., live-recorded said "fusogenic synapse" in mice myoblasts (PMID: 34932950).

Hence, while the data presented here clearly demonstrate that ARP2/3 and SCAR/WAVE complexes are required for differentiating satellite cell fusion into multinucleated myotubes, this is an incremental story, and the authors should put their results in the context of previous literature.

DOI: 10.1186/s13048-024-01578-y

Resource: (MMRRC Cat# 032050-JAX,RRID:MMRRC_032050-JAX)

Curator: @AleksanderDrozdz

SciCrunch record: RRID:MMRRC_032050-JAX

What is this?

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

Traceback (most recent call last): File "/home/ubuntu/dashboard/py/create_release_tables.py", line 54, in format_anno_for_release parsedanno = HypothesisAnnotation(anno) File "/home/ubuntu/dashboard/py/hypothesis.py", line 231, in init self.links = row['document']['link'] TypeError: string indices must be integers

This analysis that compares pooling methods was very informative, but it left me wondering the extent that mean pooling compares to no pooling at all. Is this something y'all considered? It would be interesting to compare the R2 of a more sophisticated transfer learning model that ingests the raw embeddings (like a basic FCN). Though an apples to apples might be hard to create, it would be useful to know the "cost" of mean pooling by observing the extend to which raw embeddings outperform mean pooling (if at all?)

Exploitation des informations de la conférence par les parents d'élèves

La transcription de la conférence d'Anne-Claudine Hollire offre plusieurs points d'intérêt pour les parents d'élèves engagés dans la vie scolaire à différents niveaux.

Voici comment ils peuvent exploiter ces informations :

Niveau Famille (0:00-3:00, 10:00-14:00, 14:00-16:00)

• Comprendre le phénomène du coaching scolaire :

La conférence permet aux parents de saisir les motivations derrière le recours au coaching scolaire, souvent lié à l'angoisse face à la réussite et l'orientation des enfants. * Déconstruire le mythe de l'expert :

La présentation des différents registres d'expertise des coachs (monde de l'entreprise, expérience personnelle, etc.) aide à relativiser la prétendue légitimité de ces intervenants. * Identifier les besoins réels de leur enfant :

La conférence met en lumière les objectifs visés par le coaching (motivation, méthodologie, confiance en soi), permettant aux parents de discerner si leur enfant a réellement besoin d'un tel accompagnement ou si d'autres solutions existent au sein du système scolaire. * Discuter ouvertement de l'orientation :

En comprenant les enjeux affectifs liés à l'orientation, les parents peuvent aborder ce sujet de manière plus sereine avec leur enfant, sans reproduire les tensions décrites dans la conférence.

Niveau Classe et Conseils de Classe (3:00-7:00)

• Sensibiliser aux conséquences des réformes scolaires :

La conférence souligne l'impact des réformes du lycée et de Parcoursup sur l'angoisse des élèves et des parents, alimentant le marché du coaching.

Les parents peuvent relayer ces informations en conseil de classe et plaider pour un accompagnement renforcé au sein de l'établissement.

• S'interroger sur l'utilisation des fonds publics :

L'exemple du dispositif "Ingénieur pour l'école" montre que l'État investit dans des solutions privées d'accompagnement.

Les parents peuvent questionner l'allocation de ces ressources et proposer des alternatives au sein de l'établissement (ateliers de méthodologie, groupes de parole, etc.).

Commissions Éducatives et Conseils de Discipline (10:00-14:00)

• Promouvoir une vision plus inclusive de l'orientation :

La conférence met en avant la tendance du coaching à reproduire les aspirations sociales des familles, notamment vers les métiers de "cadre".

Les parents peuvent s'appuyer sur ces analyses pour encourager une réflexion plus ouverte sur les différentes voies professionnelles possibles, y compris celles moins "valorisées" socialement.

Conseil d'École et Conseil d'Administration (3:00-7:00, 5:00-10:00)

• Plaider pour un meilleur accompagnement à l'orientation :

La conférence met en évidence les limites du système d'orientation actuel, qui pousse les familles à se tourner vers des solutions privées.

Les parents peuvent utiliser ces arguments pour demander un renforcement des moyens alloués aux conseillers d'orientation et aux psychologues scolaires, ainsi que la mise en place de dispositifs d'information et d'accompagnement plus efficaces.

• Interroger le rôle des établissements privés :

La conférence note une proximité entre le discours des coachs et celui de certains établissements privés, notamment sur la valorisation de l'entreprise de soi et du développement personnel.

Les parents peuvent s'appuyer sur ces observations pour initier un dialogue sur les valeurs éducatives promues par les différents types d'établissements.

Liens avec la Municipalité, le Département, l'Académie et la Région (3:00-7:00)

• Dénoncer le soutien de l'État au marché du soutien scolaire :

La conférence souligne les avantages fiscaux accordés aux familles qui recourent au soutien scolaire, y compris certaines formes de coaching.

Les parents peuvent alerter les élus locaux et les représentants de l'Éducation nationale sur les effets pervers de ces dispositifs, qui contribuent à creuser les inégalités d'accès à l'accompagnement scolaire.

• Plaider pour un investissement public dans des solutions alternatives :

Les parents peuvent se mobiliser pour demander aux collectivités territoriales de financer des structures d'accompagnement gratuites et accessibles à tous (maisons des adolescents, centres d'information et d'orientation, etc.).

En conclusion, la conférence d'Anne-Claudine Hollire fournit aux parents d'élèves des clés de compréhension du phénomène du coaching scolaire et de ses implications sur le système éducatif.

En s'appuyant sur ces informations, ils peuvent agir à différents niveaux pour promouvoir un accompagnement à l'orientation plus juste et plus équitaire.

"Questions d'éducation (suite) (6) - Pierre-Michel Menger (2024-2025)"

Thèmes principaux:

Le Concours Kangourou des Mathématiques comme Observatoire Socio-Scolaire:

L'analyse détaillée des données du concours Kangourou des mathématiques révèle des disparités de participation et de performance entre les établissements publics et privés, ainsi qu'en fonction de l'indice de position sociale (IPS) des élèves.

L'implication des établissements privés et des familles favorisées est plus importante, traduisant une "culture d'établissement" tournée vers la performance.

"Les élèves du privé participent bien davantage que ceux du public... dans le quintile supérieur... il y a une implication plus forte des établissements, c'est à la fois peut-être une affaire de culture d'établissement comme je dis toujours ou bien une affaire aussi de ... ressources familiales et d'engagement familial dans le processus éducatif scolaire et périscolaire."

La Quête de la "Bonne École":

Dilemmes des Familles et Arbitrages Public-Privé:

Face à un système éducatif complexe et hétérogène, les familles recherchent la "bonne école" pour leurs enfants.

La performance académique, mesurée par les taux de réussite et les mentions aux examens, reste un critère important, malgré ses limites.

L'enquête de Robert Ballion (1991) met en évidence la prédominance du critère de discipline dans le choix des familles.

"Idéalement, l'établissement efficace serait celui qui réalise l'équilibre le plus parfait entre ces trois types principaux de résultats: la performance académique, l'équité, la formation de la personne aussi bien dans sa dimension individuelle que sociale." - Robert Ballion

Le Rôle de la Discipline et les Effets de Pairs:

La discipline s'avère un facteur déterminant dans la perception de la qualité d'un établissement.

Elle influence non seulement le climat scolaire, mais aussi la performance académique des élèves via les effets de pairs.

Les analyses des enquêtes panel de la DEP (2007 et 2011) confirment l'importance accordée à la discipline par les parents.

"La discipline... est reliée à plus de facteurs que... et plus fortement à d'autres facteurs que la plupart des autres critères... c'est une sorte de point nodal de représentation par les familles de ce qui est attendu."

L'Hypocrisie du Système et la Crise de Vocation:

Le reportage de Zineb Drief (Le Monde, 2017) met en lumière les contradictions et les stratégies des familles de la classe moyenne supérieure, attachées aux valeurs de l'école publique, mais confrontées à ses difficultés, notamment en matière de discipline et de mixité sociale.

L'analyse des données de la base centrale scolarité révèle un taux croissant d'enfants d'enseignants scolarisés dans le privé, illustrant l'ampleur de la crise de confiance envers l'école publique.

"J'ai culpabilisé parce que je tiens à la mixité, mais là ce n'est pas de la mixité, c'est n'importe quoi. On a laissé se former des ghettos. Ce ne sera pas nous qui changerons les choses... À notre tour d'être égoïstes comme beaucoup de parents de la classe moyenne supérieure." - Témoignage d'une mère dans l'article de Zineb Drief

Le Trilemme de la Justice Sociale et l'Avenir de l'École: Le cours conclut sur l'incompatibilité des trois objectifs d'équité procédurale, d'égalité des chances et de liberté individuelle, formant un "trilemme" insoluble.

Les citations de John Maynard Keynes et d'Hannah Arendt illustrent la complexité de la question et l'absence de solution miracle.

"Le problème politique de l'humanité doit concilier trois choses: une meilleure efficacité économique... la justice sociale et la liberté individuelle." - John Maynard Keynes

Points Importants:

L'enseignement public français fait face à des défis majeurs, notamment en matière de discipline, de mixité sociale et d'attractivité des métiers.

Les familles développent des stratégies complexes pour naviguer dans un système hétérogène et trouver la "bonne école" pour leurs enfants.

Le système privé, plus homogène et performant sur certains critères, attire de plus en plus de familles, y compris celles issues du monde de l'éducation.

La crise de vocation des enseignants, particulièrement aiguë dans certaines disciplines et régions, menace la qualité de l'enseignement public.

Il n'existe pas de solution simple au trilemme de la justice sociale, obligeant à des arbitrages difficiles entre équité, performance et liberté.

Conclusion:

Le cours de Pierre-Michel Menger offre une analyse riche et nuancée des enjeux de l'éducation en France, s'appuyant sur des données empiriques et des témoignages éclairants.

Il met en lumière les contradictions, les dilemmes et les défis auxquels le système éducatif est confronté, invitant à une réflexion critique et à la recherche de solutions pragmatiques.

DOI: 10.1016/j.immuni.2024.11.023

Resource: None

Curator: @scibot

SciCrunch record: RRID:AB_2574135

What is this?

法文雞 detailed discussions of A1 grammar

A1 初級文法 Search: 內容 分類 建議順序 初級文法 10 介詞：à / de + 城市和國家 介詞 1 初級文法 11 介詞：上下前後內 介詞 2 初級文法 12 介詞：離...很近 / 很遠 près de, loin de 介詞 3 初級文法 26 表示存在：il y a 表達：空間 4 初級文法 13 疑問形式：qui, que, c'est vs il est 問答 5 初級文法 14 c'est 和 il est 的分別 問答 6 初級文法 17 部分冠詞 - du, de la, des 限定詞 7 初級文法 27 比較級 表達：比較 8 初級文法 28 最高級 表達：比較 9 初級文法 01 第二組 -ir 動詞變位規則 動詞 10 初級文法 02 不規則動詞變位 動詞 11 初級文法 03 動詞 faire & jouer：活動與嗜好 動詞 12 初級文法 04 動詞 aller 與地方 動詞 13 初級文法 05 最近將來時 futur proche：aller + 不定式 動詞 14 初級文法 06 複合過去時 passé composé 動詞 15 初級文法 07 未完成過去時 imparfait 動詞 16 初級文法 08 複合過去時 vs 未完成過去時 動詞 17 初級文法 09 簡單將來時 futur simple 動詞 18 初級文法 25 基本時間標誌介詞：depuis, pendant, il y a... 表達：時間 19 初級文法 15 基本疑問詞 問答 20 初級文法 16 三種疑問句形式：雅語、日常語言、通俗語言 問答 21 初級文法 18 法文直接受詞 COD vs 間接受詞 COI 總論 22 初級文法 19 直接受詞代名詞 pronom COD：le, la, les... 代詞 23 初級文法 20 間接受詞代名詞 pronom COI：lui, leur 代詞 24 初級文法 21 代詞 en 代詞 25 初級文法 22 代詞 y 代詞 26 初級文法 23 重讀人稱代詞 pronom tonique：moi, toi, lui... 代詞 27 初級文法 24 關係代詞 pronom relatif：qui, que 代詞 28 Showing 1 to 28 of 28 entries

法文雞 Same TOC in POS:

動詞 我是誰 動詞 faire 做 (一項活動) 反身動詞 動詞 aller 去 和 venir 來 動詞 pouvoir 可、devoir 要 和 vouloir 想 條件式的 devoir 和 pouvoir 後接不定式的動詞

語式和時態

直陳式

最近將來時 aller + 動詞原形 現在進行時和最近過去時 複合過去時 複合過去時的反身動詞和有兩個助動詞的動詞 未完成過去時 愈過去時 過去時比較：複合過去時、未完成過去時、愈過去時 簡單過去時 最近將來時和簡單將來時 不規則的簡單將來時動詞

命令式

條件式

虛擬式

虛擬式現在時 虛擬式過去時

分詞式

不規則過去分詞 過去分詞的性數配合 被動語態 現在分詞 副動詞 en faisant…

限定詞

定冠詞和不定冠詞 un / une / des、le / la / les 指示形容詞 ce / cet / cette / ces 主有形容詞 mon / ma / mes… 部份冠詞及數量表達 du、beaucoup de… 泛指形容詞和泛指代詞 tout、plusieurs、certains….

代詞

重讀人稱代詞 moi, toi, lui, elle… 非人稱代詞 il 直接和間接代詞 代詞 en 和 y 代詞的位置與順序 關係代詞 qui、que、où 複合關係代詞 中性的關係代詞和強調法 泛指形容詞和泛指代詞 tout、plusieurs、certains…. 主有代詞 指示代詞 celui-ci、celle-là、ceux-ci…

否定

簡單否定 複雜否定 ne…rien、ne…plus、ne…jamais、ne…personne、ne…aucun

提問

以 quel 提問 封閉式問題 開放式問題

表達

C'est & il est Il y a 時間表達：ça fait、il y a…que、pendant、il y a 表達原因 表達結果 表達目的 表達對立 表達讓步 人、事物、動作的比較 數量的比較 引語 假設

其他

國藉 國家與城市 年齡 星期一至日 職業 品質形容詞 談論季節、介詞 日子、月份

大部分譯自《新觀察家》(L'OBS) ，法文原版：傳送門

This Chinese instructor for French grammar is clear and succinct. KEEP!

法文雞的文法筆記

C2 程度法文學習者，幾年前開始在香港教授小班法文。希望網站能提供循序漸進、有系統和容易檢閱的文法教學。有任何問題、想查詢開班或網課事宜等，歡迎發電郵，又或在下方聯絡欄留言給我。

電郵：

coq@lecoqfr.comOpens in your application

A1: 初學階段入門級

〈隱藏 A1 文法〉 1.1. 自我介紹

1.1.1. 我是誰 1.1.2. 國藉 1.1.3. 國家與城市 1.1.4. 年齡 1.1.5. 否定 1.1.6. 以 quel 提問 1.1.7. 封閉式問題 1.1.8. 開放式問題

1.2. 談論活動

1.2.1. 動詞 faire 做 (一項活動) 1.2.2. 星期一至日 1.2.3. 反身動詞 1.2.4. 動詞 aller 去 和 venir 來 1.2.5. 重讀人稱代詞 moi, toi, lui, elle… 1.2.6. 最近將來時 aller + 動詞原形

1.3. 要求和提議

1.3.1. 指示形容詞 ce / cet / cette / ces 1.3.2. 動詞 pouvoir 可、devoir 要 和 vouloir 想

1.4. 描述

1.4.1. C'est & il est 1.4.2. 主有形容詞 mon / ma / mes… 1.4.3. 職業 1.4.4. 品質形容詞 1.4.5. Il y a 1.4.6. 人、事物、動作的比較

1.5. 建議他人

1.5.1. 命令式 1.5.2. 條件式的 devoir 和 pouvoir

1.6. 談論食物

1.6.1. 定冠詞和不定冠詞 un / une / des、le / la / les 1.6.2. 部份冠詞及數量表達 du、beaucoup de…

1.7. 談論天氣

1.7.1. 非人稱代詞 il 1.7.2. 談論季節、介詞

1.8. 談論過去

1.8.1. 複合過去時 1.8.2. 日子、月份

A2: 初學階段初級

〈隱藏 A2 文法〉 2.1. 談論過去

2.1.1. 未完成過去時 2.1.2. 不規則過去分詞 2.1.3. 時間表達：ça fait、il y a…que、pendant、il y a 2.1.4. 複合過去時的反身動詞和有兩個助動詞的動詞

2.2. 談論事物和人

2.2.1. 關係代詞 qui、que、où 2.2.2. 主有代詞 2.2.3. 直接和間接代詞 2.2.4. 數量的比較 2.2.5. 泛指形容詞和泛指代詞 tout、plusieurs、certains… 2.2.6. 代詞 en 和 y 2.2.7. 複雜否定 ne…rien、ne…plus、ne…jamais、ne…personne、ne…aucun

2.3. 談論動作

2.3.1. 現在進行時和最近過去時

2.4. 談論將來

2.4.1. 最近將來時和簡單將來時 2.4.2. 不規則的簡單將來時動詞

B1: 獨立階段中級

〈隱藏 B1 文法〉 3.1. 談論各種事情

3.1.1. 被動語態 3.1.2. 愈過去時 3.1.3. 過去時比較：複合過去時、未完成過去時、愈過去時 3.1.4. 引語 3.1.5. 假設 3.1.6. 過去分詞的性數配合

3.2. 談論動作

3.2.1. 副動詞 en faisant… 3.2.2. 指示代詞 celui-ci、celle-là、ceux-ci…

3.3. 解釋和議論

3.3.1. 表達原因 3.3.2. 表達結果 3.3.3. 表達目的 3.3.4. 表達對立 3.3.5. 複合關係代詞 3.3.6. 代詞的位置與順序 3.3.7. 條件式的運用 3.3.8. 中性的關係代詞和強調法 3.3.9. 表達讓步 3.3.10. 虛擬式的運用

B2: 獨立階段中高級

〈隱藏 B2 文法〉 4.1. 談論過去

4.1.1. 簡單過去時 4.1.2. 虛擬式過去時

4.2. 談論動作

4.2.1. 後接不定式的動詞 4.2.2. 現在分詞

大部分譯自《新觀察家》(L'OBS) ，法文原版：傳送門

French conditional tense: all usages

語式上的用法

語式指說話人對他自己說的話所表達的態度。條件式可用於表達以下東西：

一）禮貌（常見於動詞 pouvoir 和 vouloir）

Je voudrais un café s'il vous plait. 我想要一杯咖啡，謝謝。

Je pourrais parler avec M. Dupontel ? 我可以跟 Dupontel 先生說話嗎？

二）建議（常見用於動詞 devoir）

Tu devrais parler à ton père. 你應該跟你的爸爸談談。

Vous devriez appeler le médecin. 你應該叫醫生。

三）欲望（常見於動詞 vouloir 和 aimer）

J'aimerais devenir pianiste. 我想成為鋼琴家。

Nous voudrions déménager dans le sud de la France. 我們想搬往南法。

四）未經核實的事實

Le ministre rencontrerait son homologue allemand très prochainement. 部長快將會見他的德國同行。

Un accident a eu lieu sur l'autoroute A4. Il y aurait trois blessés. 在 A4 公路發生了意外，有三名傷者。

五）後悔（常見於動詞 vouloir、devoir、aimer 的虛擬式過去時）

J'aurais voulu être un artiste. 我希望我是一個藝術家。

Nous aurions aimé le connaître. 我們本想認識他。

Tu aurais dû lui parler. 你應和他談談的。

假設中的用法

想像一些不存在的東西，及對現在進行假設的時候，我們用 si + 未完成過去時 及 虛擬式現在時。

Si tu avais de l'argent, tu achèterais une nouvelle voiture. (但你沒有錢) 如果你有錢，你會買一輛新車。

S'il faisait beau, nous irions à la mer et nous ferions une balade à vélo. (但天氣不好) 如果天氣很好，我們會去海邊騎單車。

想像一些不存在的東西，及對過去進行假設的時候，我們用 si + 愈過去時 及 虛擬式過去時。

Si tu avais quitté Paris, tu aurais habité dans une belle et grande maison avec un jardin. (但你沒有離開巴黎) 如果你離開了巴黎，你便會住在一間又大又漂亮、有花園的房子了。

Si vos enfants étaient partis en Italie, vous les auriez rejoints. (但你的孩子沒有到意大利去) 如果你的孩子到了意大利去，你便會跟他們重聚了。

more tools: Lute <- LWT (defunct)

Ist b1 immer der Y-Achsenabschnitt egal ob zentriert oder unzentriert?

DOI: 10.1038/s41380-024-02653-y

Resource: RRID:Addgene_55637

Curator: @olekpark

SciCrunch record: RRID:Addgene_55637

What is this?

DOI: 10.1038/s41380-024-02653-y

Resource: RRID:Addgene_105540

Curator: @olekpark

SciCrunch record: RRID:Addgene_105540

What is this?

DOI: 10.1038/s41380-024-02653-y

Resource: RRID:Addgene_65417

Curator: @olekpark

SciCrunch record: RRID:Addgene_65417

What is this?

DOI: 10.1038/s41380-024-02653-y

Resource: RRID:Addgene_44362

Curator: @scibot

SciCrunch record: RRID:Addgene_44362

What is this?

DOI: 10.1038/s41380-024-02653-y

Resource: RRID:Addgene_44361

Curator: @scibot

SciCrunch record: RRID:Addgene_44361

What is this?

DOI: 10.1038/s41380-024-02653-y

Resource: RRID:Addgene_50459

Curator: @scibot

SciCrunch record: RRID:Addgene_50459

What is this?

Author response:

The following is the authors’ response to the original reviews.

The revised manuscript contains new results and additional text. Major revisions:

(1) Additional simulations and analyses of networks with different biophysical parameters and with identical time constants for E and I neurons (Methods, Supplementary Fig. 5).

(2) Additional simulations and analyses of networks with modifications of connectivity parameters to further analyze effects of E/I assemblies on manifold geometry (Supplementary Fig. 6).

(3) Analysis of synaptic current components (Figure 3 D-F; to analyze mechanism of modest amplification in Tuned networks). 

(4) More detailed explanation of pattern completion analysis (Results).

(5) Analysis of classification performance of Scaled networks (Supplementary Fig.8).

(6) Additional analysis (Figure 5D-F) and discussion (particularly section “Computational functions of networks with E/I assemblies”) of functional benefits of continuous representations in networks with E-I assemblies. 

Public Reviews: 

Reviewer #1 (Public Review): 

Summary: 

Meissner-Bernard et al present a biologically constrained model of telencephalic area of adult zebrafish, a homologous area to the piriform cortex, and argue for the role of precisely balanced memory networks in olfactory processing. 

This is interesting as it can add to recent evidence on the presence of functional subnetworks in multiple sensory cortices. It is also important in deviating from traditional accounts of memory systems as attractor networks. Evidence for attractor networks has been found in some systems, like in the head direction circuits in the flies. However, the presence of attractor dynamics in other modalities, like sensory systems, and their role in computation has been more contentious. This work contributes to this active line of research in experimental and computational neuroscience by suggesting that, rather than being represented in attractor networks and persistent activity, olfactory memories might be coded by balanced excitation-inhibitory subnetworks. 

Strengths: 

The main strength of the work is in: (1) direct link to biological parameters and measurements, (2) good controls and quantification of the results, and (3) comparison across multiple models. 

(1) The authors have done a good job of gathering the current experimental information to inform a biological-constrained spiking model of the telencephalic area of adult zebrafish. The results are compared to previous experimental measurements to choose the right regimes of operation. 

(2) Multiple quantification metrics and controls are used to support the main conclusions and to ensure that the key parameters are controlled for - e.g. when comparing across multiple models.  (3) Four specific models (random, scaled I / attractor, and two variant of specific E-I networks - tuned I and tuned E+I) are compared with different metrics, helping to pinpoint which features emerge in which model. 

Weaknesses: 

Major problems with the work are: (1) mechanistic explanation of the results in specific E-I networks, (2) parameter exploration, and (3) the functional significance of the specific E-I model. 

(1) The main problem with the paper is a lack of mechanistic analysis of the models. The models are treated like biological entities and only tested with different assays and metrics to describe their different features (e.g. different geometry of representation in Fig. 4). Given that all the key parameters of the models are known and can be changed (unlike biological networks), it is expected to provide a more analytical account of why specific networks show the reported results. For instance, what is the key mechanism for medium amplification in specific E/I network models (Fig. 3)? How does the specific geometry of representation/manifolds (in Fig. 4) emerge in terms of excitatory-inhibitory interactions, and what are the main mechanisms/parameters? Mechanistic account and analysis of these results are missing in the current version of the paper. 

We agree that further mechanistic insights would be of interest and addressed this issue at different levels:

(1) Biophysical parameters: to determine whether network behavior depends on specific choices of biophysical parameters in E and I neurons we equalized biophysical parameters across neuron types. The main observations are unchanged, suggesting that the observed effects depend primarily on network connectivity (see also response to comment [2]).

(2) Mechanism of modest amplification in E/I assemblies: analyzing the different components of the synaptic currents demonstrate that the modest amplification of activity in Tuned networks results from an “imperfect” balance of recurrent excitation and inhibition within assemblies (see new Figures 3D-F and text p.7). Hence, E/I co-tuning substantially reduces the net amplification in Tuned networks as compared to Scaled networks, thus preventing discrete attractor dynamics and stabilizing network activity, but a modest amplification still occurs, consistent with biological observations.

(3) Representational geometry: to obtain insights into the network mechanisms underlying effects of E/I assemblies on the geometry of population activity we tested the hypothesis that geometrical changes depend, at least in part, on the modest amplification of activity within E/I assemblies (see Supplementary Figure 6). We changed model parameters to either prevent the modest amplification in Tuned networks (increasing I-to-E connectivity within assemblies) or introduce a modest amplification in subsets of neurons by other mechanisms (concentration-dependent increase in the excitability of pseudo-assembly neurons; Scaled I networks with reduced connectivity within assemblies). Manipulations that introduced a modest, input-dependent amplification in neuronal subsets had geometrical effects similar to those observed in Tuned networks, whereas manipulations that prevented a modest amplification abolished these effects (Supplementary Figure 6). Note however that these manipulations generated different firing rate distributions. These results provide a starting point for more detailed analyses of the relationship between network connectivity and representational geometry (see p.12).

In summary, our additional analyses indicate that effects of E/I assemblies on representational geometry depend primarily on network connectivity, rather than specific biophysical parameters, and that the resulting modest amplification of activity within assemblies makes an important contribution. Further analyses may reveal more specific relationships between E/I assemblies and representational geometry, but such analyses are beyond the scope of this study.

(2) The second major issue with the study is a lack of systematic exploration and analysis of the parameter space. Some parameters are biologically constrained, but not all the parameters. For instance, it is not clear what the justification for the choice of synaptic time scales are (with E synaptic time constants being larger than inhibition: tau_syn_i = 10 ms, tau_syn_E = 30 ms). How would the results change if they are varying these - and other unconstrained - parameters? It is important to show how the main results, especially the manifold localisation, would change by doing a systematic exploration of the key parameters and performing some sensitivity analysis. This would also help to see how robust the results are, which parameters are more important and which parameters are less relevant, and to shed light on the key mechanisms.  

We thank the reviewer for raising this point. We chose a relatively slow time constant for excitatory synapses because experimental data indicate that excitatory synaptic currents in Dp and piriform cortex contain a prominent NMDA component. Nevertheless, to assess whether network behavior depends on specific choices of biophysical parameters in E and I neurons, we have performed additional simulations with equal synaptic time constants and equal biophysical parameters for all neurons. Each neuron also received the same number of inputs from each population (see revised Methods). Results were similar to those observed previously (Supplementary Fig.5 and p.9 of main text). We therefore conclude that the main effects observed in Tuned networks cannot be explained by differences in biophysical parameters between E and I neurons but is primarily a consequence of network connectivity.

(3) It is not clear what the main functional advantage of the specific E-I network model is compared to random networks. In terms of activity, they show that specific E-I networks amplify the input more than random networks (Fig. 3). But when it comes to classification, the effect seems to be very small (Fig. 5c). Description of different geometry of representation and manifold localization in specific networks compared to random networks is good, but it is more of an illustration of different activity patterns than proving a functional benefit for the network. The reader is still left with the question of what major functional benefits (in terms of computational/biological processing) should be expected from these networks, if they are to be a good model for olfactory processing and learning. 

One possibility for instance might be that the tasks used here are too easy to reveal the main benefits of the specific models - and more complex tasks would be needed to assess the functional enhancement (e.g. more noisy conditions or more combination of odours). It would be good to show this more clearly - or at least discuss it in relation to computation and function. 

In the previous manuscript, the analysis of potential computational benefits other than pattern classification was limited and the discussion of this issue was condensed into a single itemized paragraph to avoid excessive speculation. Although a thorough analysis of potential computational benefits exceeds the scope of a single paper, we agree with the reviewer that this issue is of interest and therefore added additional analyses and discussion.

In the initial manuscript we analyzed pattern classification primarily to investigate whether Tuned networks can support this function at all, given that they do not exhibit discrete attractor states. We found this to be the case, which we consider a first important result.

Furthermore, we found that precise balance of E/I assemblies can protect networks against catastrophic firing rate instabilities when assemblies are added sequentially, as in continual learning. Results from these simulations are now described and discussed in more detail (see Results p.11 and Discussion p.13).

In the revised manuscript, we now also examine additional potential benefits of Tuned networks and discuss them in more detail (see new Figure 5D-F and text p.11). One hypothesis is that continuous representations provide a distance metric between a given input and relevant (learned) stimuli. To address this hypothesis, we (1) performed regression analysis and (2) trained support vector machines (SVMs) to predict the concentration of a given odor in a mixture based on population activity. In both cases, Tuned E+I networks outperformed Scaled and _rand n_etworks in predicting the concentration of learned odors across a wide range mixtures (Figure 5D-F).  E/I assemblies therefore support the quantification of learned odors within mixtures or, more generally, assessments of how strongly a (potentially complex) input is related to relevant odors stored in memory. Such a metric assessment of stimulus quality is not well supported by discrete attractor networks because inputs are mapped onto discrete network states.

The observation that Tuned networks do not map inputs onto discrete outputs indicates that such networks do not classify inputs as distinct items. Nonetheless, the observed geometrical modifications of continuous representations support the classification of learned inputs or the assessment of metric relationships by hypothetical readout neurons. Geometrical modifications of odor representations may therefore serve as one of multiple steps in multi-layer computations for pattern classification (and/or other computations). In this scenario, the transformation of odor representations in Dp may be seen as related to transformations of representations between different layers in artificial networks, which collectively perform a given task (notwithstanding obvious structural and mechanistic differences between artificial and biological networks). In other words, geometrical transformations of representations in Tuned networks may overrepresent learned (relevant) information at the expense of other information and thereby support further learning processes in other brain areas. An obvious corollary of this scenario is that Dp does not perform odor classification per se based on inputs from the olfactory bulb but reformats representations of odor space based on experience to support computational tasks as part of a larger system. This scenario is now explicitly discussed (p.14).

Reviewer #2 (Public Review): 

Summary: 

The authors conducted a comparative analysis of four networks, varying in the presence of excitatory assemblies and the architecture of inhibitory cell assembly connectivity. They found that co-tuned E-I assemblies provide network stability and a continuous representation of input patterns (on locally constrained manifolds), contrasting with networks with global inhibition that result in attractor networks. 

Strengths: 

The findings presented in this paper are very interesting and cutting-edge. The manuscript effectively conveys the message and presents a creative way to represent high-dimensional inputs and network responses. Particularly, the result regarding the projection of input patterns onto local manifolds and continuous representation of input/memory is very Intriguing and novel. Both computational and experimental neuroscientists would find value in reading the paper. 

Weaknesses: 

that have continuous representations. This could also be shown in Figure 5B, along with the performance of the random and tuned E-I networks. The latter networks have the advantage of providing network stability compared to the Scaled I network, but at the cost of reduced network salience and, therefore, reduced input decodability. The authors may consider designing a decoder to quantify and compare the classification performance of all four networks. 

We have now quantified classification by networks with discrete attractor dynamics (Scaled) along with other networks. However, because the neuronal covariance matrix for such networks is low rank and not invertible, pattern classification cannot be analyzed by QDA as in Figure 5B. We therefore classified patterns from the odor subspace by template matching, assigning test patterns to one of the four classes based on correlations (see Supplementary Figure 8). As expected, Scaled networks performed well, but they did not outperform Tuned networks. Moreover, the performance of Scaled networks, but not Tuned networks, depended on the order in which odors were presented to the network. This hysteresis effect is a direct consequence of persistent attractor states and decreased the general classification performance of Scaled networks (see Supplementary Figure 8 for details). These results confirm the prediction that networks with discrete attractor states can efficiently classify inputs, but also reveal disadvantages arising from attractor dynamics. Moreover, the results indicate that the classification performance of Tuned networks is also high under the given task conditions, which simulate a biologically realistic scenario.

We would also like to emphasize that classification may not be the only task, and perhaps not even a main task, of Dp/piriform cortex or other memory networks with E/I assemblies. Conceivably, other computations could include metric assessments of inputs relative to learned inputs or additional learning-related computations. Please see our response to comment (3) of reviewer 1 for a further discussion of this issue. 

Networks featuring E/I assemblies could potentially represent multistable attractors by exploring the parameter space for their reciprocal connectivity and connectivity with the rest of the network. However, for co-tuned E-I networks, the scope for achieving multistability is relatively constrained compared to networks employing global or lateral inhibition between assemblies. It would be good if the authors mentioned this in the discussion. Also, the fact that reciprocal inhibition increases network stability has been shown before and should be cited in the statements addressing network stability (e.g., some of the citations in the manuscript, including Rost et al. 2018, Lagzi & Fairhall 2022, and Vogels et al. 2011 have shown this).  

We thank the reviewer for this comment. We now explicitly discuss multistability (see p. 12) and refer to additional references in the statements addressing network stability.

Providing raster plots of the pDp network for familiar and novel inputs would help with understanding the claims regarding continuous versus discrete representation of inputs, allowing readers to visualize the activity patterns of the four different networks. (similar to Figure 1B). 

We thank the reviewer for this suggestion. We have added raster plots of responses to both familiar and novel inputs in the revised manuscript (Figure 2D and Supplementary Figure 4A).

Reviewer #3 (Public Review): 

Summary: 

This work investigates the computational consequences of assemblies containing both excitatory and inhibitory neurons (E/I assembly) in a model with parameters constrained by experimental data from the telencephalic area Dp of zebrafish. The authors show how this precise E/I balance shapes the geometry of neuronal dynamics in comparison to unstructured networks and networks with more global inhibitory balance. Specifically, E/I assemblies lead to the activity being locally restricted onto manifolds - a dynamical structure in between high-dimensional representations in unstructured networks and discrete attractors in networks with global inhibitory balance. Furthermore, E/I assemblies lead to smoother representations of mixtures of stimuli while those stimuli can still be reliably classified, and allow for more robust learning of additional stimuli. 

Strengths: 

Since experimental studies do suggest that E/I balance is very precise and E/I assemblies exist, it is important to study the consequences of those connectivity structures on network dynamics. The authors convincingly show that E/I assemblies lead to different geometries of stimulus representation compared to unstructured networks and networks with global inhibition. This finding might open the door for future studies for exploring the functional advantage of these locally defined manifolds, and how other network properties allow to shape those manifolds. 

The authors also make sure that their spiking model is well-constrained by experimental data from the zebrafish pDp. Both spontaneous and odor stimulus triggered spiking activity is within the range of experimental measurements. But the model is also general enough to be potentially applied to findings in other animal models and brain regions. 

Weaknesses: 

I find the point about pattern completion a bit confusing. In Fig. 3 the authors argue that only the Scaled I network can lead to pattern completion for morphed inputs since the output correlations are higher than the input correlations. For me, this sounds less like the network can perform pattern completion but it can nonlinearly increase the output correlations. Furthermore, in Suppl. Fig. 3 the authors show that activating half the assembly does lead to pattern completion in the sense that also non-activated assembly cells become highly active and that this pattern completion can be seen for Scaled I, Tuned E+I, and Tuned I networks. These two results seem a bit contradictory to me and require further clarification, and the authors might want to clarify how exactly they define pattern completion. 

We believe that this comment concerns a semantic misunderstanding and apologize for any lack of clarity. We added a definition of pattern completion in the text: “…the retrieval of the whole memory from noisy or corrupted versions of the learned input.”. Pattern completion may be assessed using different procedures. In computational studies, it is often analyzed by delivering input to a subset of the assembly neurons which store a given memory (partial activation). Under these conditions, we find recruitment of the entire assembly in all structured networks, as demonstrated in Supplementary Figure 3. However, these conditions are unlikely to occur during odor presentation because the majority of neurons do not receive any input.

Another more biologically motivated approach to assess pattern completion is to gradually modify a realistic odor input into a learned input, thereby gradually increasing the overlap between the two inputs. This approach had been used previously in experimental studies (references added to the text p.6). In the presence of assemblies, recurrent connectivity is expected to recruit assembly neurons (and thus retrieve the stored pattern) more efficiently as the learned pattern is approached. This should result in a nonlinear increase in the similarity between the evoked and the learned activity pattern. This signature was prominent in Scaled networks but not in Tuned or rand networks. Obviously, the underlying procedure is different from the partial activation of the assembly described above because input patterns target many neurons (including neurons outside assemblies) and exhibit a biologically realistic distribution of activity. However, this approach has also been referred to as “pattern completion” in the neuroscience literature, which may be the source of semantic confusion here. To clarify the difference between these approaches we have now revised the text and explicitly described each procedure in more detail (see p.6). 

The authors argue that Tuned E+I networks have several advantages over Scaled I networks. While I agree with the authors that in some cases adding this localized E/I balance is beneficial, I believe that a more rigorous comparison between Tuned E+I networks and Scaled I networks is needed: quantification of variance (Fig. 4G) and angle distributions (Fig. 4H) should also be shown for the Scaled I network. Similarly in Fig. 5, what is the Mahalanobis distance for Scaled I networks and how well can the Scaled I network be classified compared to the Tuned E+I network? I suspect that the Scaled I network will actually be better at classifying odors compared to the E+I network. The authors might want to speculate about the benefit of having networks with both sources of inhibition (local and global) and hence being able to switch between locally defined manifolds and discrete attractor states. 

We agree that a more rigorous comparison of Tuned and Scaled networks would be of interest. We have added the variance analysis (Fig 4G) and angle distributions (Fig. 4H) for both Tuned I and Scaled networks. However, the Mahalanobis distances and Quadratic Discriminant Analysis cannot be applied to Scaled networks because their neuronal covariance matrix is low rank and not invertible_. To nevertheless compare these networks, we performed template matching by assigning test patterns to one of the four odor classes based on correlations to template patterns (Supplementary Figure 8; see also response to the first comment of reviewer 2). Interestingly, _Scaled networks performed well at classification but did not outperform Tuned networks, and exhibited disadvantages arising from attractor dynamics (Supplementary Figure 8; see also response to the first comment of reviewer 2). Furthermore, in further analyses we found that continuous representational manifolds support metric assessments of inputs relative to learned odors, which cannot be achieved by discrete representations. These results are now shown in Figure 5D-E and discussed explicitly in the text on p.11 (see also response to comment 3 of reviewer 1).

We preferred not to add a sentence in the Discussion about benefits of networks having both sources of inhibition_,_ as we find this a bit too speculative.

At a few points in the manuscript, the authors use statements without actually providing evidence in terms of a Figure. Often the authors themselves acknowledge this, by adding the term "not shown" to the end of the sentence. I believe it will be helpful to the reader to be provided with figures or panels in support of the statements.  

Thank you for this comment. We have provided additional data figures to support the following statements:

“d_M was again increased upon learning, particularly between learned odors and reference classes representing other odors (Supplementary Figure 9)”

“decreasing amplification in assemblies of Scaled networks changed transformations towards the intermediate behavior, albeit with broader firing rate distributions than in Tuned networks (Supplementary Figure 6 B)”  

Recommendations for the authors:

Reviewer #1 (Recommendations For The Authors): 

Meissner-Bernard et al present a biologically constrained model of telencephalic area of adult zebrafish, a homologous area to the piriform cortex, and argue for the role of precisely balanced memory networks in olfactory processing. 

This is interesting as it can add to recent evidence on the presence of functional subnetworks in multiple sensory cortices. It is also important in deviating from traditional accounts of memory systems as attractor networks. Evidence for attractor networks has been found in some systems, like in the head direction circuits in the flies. However, the presence of attractor dynamics in other modalities, like sensory systems, and their role in computation has been more contentious. This work contributes to this active line of research in experimental and computational neuroscience by suggesting that, rather than being represented in attractor networks and persistent activity, olfactory memories might be coded by balanced excitation-inhibitory subnetworks. 

The paper is generally well-written, the figures are informative and of good quality, and multiple approaches and metrics have been used to test and support the main results of the paper. 

The main strength of the work is in: (1) direct link to biological parameters and measurements, (2) good controls and quantification of the results, and (3) comparison across multiple models. 

(1) The authors have done a good job of gathering the current experimental information to inform a biological-constrained spiking model of the telencephalic area of adult zebrafish. The results are compared to previous experimental measurements to choose the right regimes of operation. 

(2) Multiple quantification metrics and controls are used to support the main conclusions and to ensure that the key parameters are controlled for - e.g. when comparing across multiple models.   (3) Four specific models (random, scaled I / attractor, and two variant of specific E-I networks - tuned I and tuned E+I) are compared with different metrics, helping to pinpoint which features emerge in which model. 

Major problems with the work are: (1) mechanistic explanation of the results in specific E-I networks, (2) parameter exploration, and (3) the functional significance of the specific E-I model. 

(1) The main problem with the paper is a lack of mechanistic analysis of the models. The models are treated like biological entities and only tested with different assays and metrics to describe their different features (e.g. different geometry of representation in Fig. 4). Given that all the key parameters of the models are known and can be changed (unlike biological networks), it is expected to provide a more analytical account of why specific networks show the reported results. For instance, what is the key mechanism for medium amplification in specific E/I network models (Fig. 3)? How does the specific geometry of representation/manifolds (in Fig. 4) emerge in terms of excitatory-inhibitory interactions, and what are the main mechanisms/parameters? Mechanistic account and analysis of these results are missing in the current version of the paper. 

Precise balancing of excitation and inhibition in subnetworks would lead to the cancellation of specific dynamical modes responsible for the amplification of responses (hence, deviating from the attractor dynamics with an unstable specific mode). What is the key difference in the specific E/I networks here (tuned I or/and tuned E+I) which make them stand between random and attractor networks? Excitatory and inhibitory neurons have different parameters in the model (Table 1). Time constants of inhibitory and excitatory synapses are also different (P. 13). Are these parameters causing networks to be effectively more excitation dominated (hence deviating from a random spectrum which would be expected from a precisely balanced E/I network, with exactly the same parameters of E and I neurons)? It is necessary to analyse the network models, describe the key mechanism for their amplification, and pinpoint the key differences between E and I neurons which are crucial for this. 

To address these comments we performed additional simulations and analyses at different levels. Please see our reply to comment (1) of the public review (reviewer 1) for a detailed description. We thank the reviewer for these constructive comments.

(2) The second major issue with the study is a lack of systematic exploration and analysis of the parameter space. Some parameters are biologically constrained, but not all the parameters. For instance, it is not clear what the justification for the choice of synaptic time scales are (with E synaptic time constants being larger than inhibition: tau_syn_i = 10 ms, tau_syn_E = 30 ms). How would the results change if they are varying these - and other unconstrained - parameters? It is important to show how the main results, especially the manifold localisation, would change by doing a systematic exploration of the key parameters and performing some sensitivity analysis. This would also help to see how robust the results are, which parameters are more important and which parameters are less relevant, and to shed light on the key mechanisms.  

We thank the reviewer for this comment. We have now carried out additional simulations with equal time constants for all neurons. Please see our reply to the public review for more details (comment 2 of reviewer 1).

(3) It is not clear what the main functional advantage of the specific E-I network model is compared to random networks. In terms of activity, they show that specific E-I networks amplify the input more than random networks (Fig. 3). But when it comes to classification, the effect seems to be very small (Fig. 5c). Description of different geometry of representation and manifold localization in specific networks compared to random networks is good, but it is more of an illustration of different activity patterns than proving a functional benefit for the network. The reader is still left with the question of what major functional benefits (in terms of computational/biological processing) should be expected from these networks, if they are to be a good model for olfactory processing and learning. 

One possibility for instance might be that the tasks used here are too easy to reveal the main benefits of the specific models - and more complex tasks would be needed to assess the functional enhancement (e.g. more noisy conditions or more combination of odours). It would be good to show this more clearly - or at least discuss it in relation to computation and function.

Please see our reply to the public review (comment 3 of reviewer 1).

Specific comments: 

Abstract: "resulting in continuous representations that reflected both relatedness of inputs and *an individual's experience*" 

It didn't become apparent from the text or the model where the role of "individual's experience" component (or "internal representations" - in the next line) was introduced or shown (apart from a couple of lines in the Discussion) 

We consider the scenario that that assemblies are the outcome of an experience-dependent plasticity process. To clarify this, we have now made a small addition to the text: “Biological memory networks are thought to store information by experience-dependent changes in the synaptic connectivity between assemblies of neurons.”.

P. 2: "The resulting state of "precise" synaptic balance stabilizes firing rates because inhomogeneities or fluctuations in excitation are tracked by correlated inhibition" 

It is not clear what the "inhomogeneities" specifically refers to - they can be temporal, or they can refer to the quenched noise of connectivity, for instance. Please clarify what you mean. 

The statement has been modified to be more precise: “…“precise” synaptic balance stabilizes firing rates because inhomogeneities in excitation across the population or temporal variations in excitation are tracked by correlated inhibition…”.

P. 3 (and Methods): When odour stimulus is simulated in the OB, the activity of a fraction of mitral cells is increased (10% to 15 Hz) - but also a fraction of mitral cells is suppressed (5% to 2 Hz). What is the biological motivation or reference for this? It is not provided. Is it needed for the results? Also, it is not explained how the suppressed 5% are chosen (e.g. randomly, without any relation to the increased cells?). 

We thank the reviewer for this comment. These changes in activity directly reflect experimental observations. We apologize that we forgot to include the references reporting these observations (Friedrich and Laurent, 2001 and 2004); this is now fixed.

In our simulation, OB neurons do not interact with each other, and the suppressed 5% were indeed randomly selected. We changed the text in Methods accordingly to read: “An additional 75 randomly selected mitral cells were inhibited” 

P. 4, L. 1-2: "... sparsely connected integrate-and-fire neurons with conductance-based synapses (connection probability {less than or equal to}5%)." 

Specify the connection probability of specific subtypes (EE, EI, IE, II).  

We now refer to the Methods section, where this information can be found. 

“... conductance-based synapses (connection probability ≤5%, Methods)”  

P. 4, L. 6-7: "Population activity was odor-specific and activity patterns evoked by uncorrelated OB inputs remained uncorrelated in Dp (Figure 1H)" 

What would happen to correlated OB inputs (e.g. as a result of mixture of two overlapping odours) in this baseline state of the network (before memories being introduced to it)? It would be good to know this, as it sheds light on the initial operating regime of the network in terms of E/I balance and decorrelation of inputs.  

This information was present in the original manuscript at (Figure 3) but we improved the writing to further clarify this issue: “ (…) we morphed a novel odor into a learned odor (Figure 3A), or a learned odor into another learned odor (Supplementary Figure 3B), and quantified the similarity between morphed and learned odors by the Pearson correlation of the OB activity patterns (input correlation). We then compared input correlations to the corresponding pattern correlations among E neurons in Dp (output correlation). In rand networks, output correlations increased linearly with input correlations but did not exceed them (Figure 3B and Supplementary Figure 3B)”

P. 4, L. 12-13: "Shuffling spike times of inhibitory neurons resulted in runaway activity with a probability of ~80%, .."   Where is this shown? 

(There are other occasions too in the paper where references to the supporting figures are missing). 

We now provide the statistics: “Shuffling spike times of inhibitory neurons resulted in runaway activity with a probability of 0.79 ± 0.20”

P. 4: "In each network, we created 15 assemblies representing uncorrelated odors. As a consequence, ~30% of E neurons were part of an assembly ..." 

15 x 100 / 4000 = 37.5% - so it's closer to 40% than 30%. Unless there is some overlap? 

Yes: despite odors being uncorrelated and connectivity being random, some neurons (6 % of E neurons) belong to more than one assembly.

P. 4: "When a reached a critical value of ~6, networks became unstable and generated runaway activity (Figure 2B)." 

Can this transition point be calculated or estimated from the network parameters, and linked to the underlying mechanisms causing it? 

We thank the reviewer for this interesting question. The unstability arises when inhibitions fails to counterbalance efficiently the increased recurrent excitation within Dp. The transition point is difficult to estimate, as it can depend on several parameters, including the probability of E to E connections, their strength, assembly size, and others. We have therefore not attempted to estimate it analytically.

P. 4: "Hence, non-specific scaling of inhibition resulted in a divergence of firing rates that exhausted the dynamic range of individual neurons in the population, implying that homeostatic   global inhibition is insufficient to maintain a stable firing rate distribution." 

I don't think this is justified based on the results and figures presented here (Fig. 2E) - the interpretation is a bit strong and biased towards the conclusions the authors want to draw. 

To more clearly illustrate the finding that in Scaled networks, assembly neurons are highly active (close to maximal realistic firing rates) whereas non-assembly neurons are nearly silent we have now added Supplementary Fig. 2B. Moreover, we have toned down the text: “Hence, non-specific scaling of inhibition resulted in a large and biologically unrealistic divergence of firing rates (Supplementary Figure 2B) that nearly exhausted the dynamic range of individual neurons in the population, indicating that homeostatic global inhibition is insufficient to maintain a stable firing rate distribution”

P. 5, third paragraph: Description of Figure 2I, inset is needed, either in the text or caption. 

The inset is now referred to in the text: ”we projected synaptic conductances of each neuron onto a line representing the E/I ratio expected in a balanced network (“balanced axis”) and onto an orthogonal line (“counter-balanced axis”; Figure 2I inset, Methods).”

P. 5, last paragraph: another example of writing about results without showing/referring to the corresponding figures: 

"In rand networks, firing rates increased after stimulus onset and rapidly returned to a low baseline after stimulus offset. Correlations between activity patterns evoked by the same odor at different time points and in different trials were positive but substantially lower than unity, indicating high variability ..." 

And the continuation with similar lack of references on P. 6: 

"Scaled networks responded to learned odors with persistent firing of assembly neurons and high pattern correlations across trials and time, implying attractor dynamics (Hopfield, 1982; Khona and Fiete, 2022), whereas Tuned networks exhibited transient responses and modest pattern correlations similar to rand networks." 

Please go through the Results and fix the references to the corresponding figures on all instances. 

We thank the reviewer for pointing out these overlooked figure references, which are now fixed.

P. 8: "These observations further support the conclusion that E/I assemblies locally constrain neuronal dynamics onto manifolds." 

As discussed in the general major points, mechanistic explanation in terms of how the interaction of E/I dynamics leads to this is missing. 

As discussed in the reply to the public review (comment 3 of reviewer 1), we have now provided more mechanistic analyses of our observations.

P. 9: "Hence, E/I assemblies enhanced the classification of inputs related to learned patterns."   The effect seems to be very small. Also, any explanation for why for low test-target correlation the effect is negative (random doing better than tuned E/I)? 

The size of the effect (plearned – pnovel = 0.074; difference of means; Figure 5C) may appear small in terms of absolute probability, but it is substantial relative to the maximum possible increase (1 – p_novel =  0.133; Figure 5C). The fact that for low test-target correlations the effect is negative is a direct consequence of the positive effect for high test-target correlations and the presence of 2 learned odors in the 4-way forced choice task. 

P. 9: "In Scaled I networks, creating two additional memories resulted in a substantial increase   in firing rates, particularly in response to the learned and related odors"   Where is this shown? Please refer to the figure. 

We thank the reviewer again for pointing this out. We forgot to include a reference to the relevant figure which has now been added in the revised manuscript (Figure 6C).

P. 10: "The resulting Tuned networks reproduced additional experimental observations that were not used as constraints including irregular firing patterns, lower output than input correlations, and the absence of persistent activity" 

It is difficult to present these as "additional experimental observations", as all of them are negative, and can exist in random networks too - hence cannot be used as biological evidence in favour of specific E/I networks when compared to random networks. 

We agree with the reviewer that these additional experimental observations cannot be used as biological evidence favouring Tuned E+I networks over random networks. We here just wanted to point out that additional observations which we did not take into account to fit the model are not invalidating the existence of E-I assemblies in biological networks. As assemblies tend to result in persistent activity in other types of networks, we feel that this observation is worth pointing out.

Methods: 

P. 13: Describe the parameters of Eq. 2 after the equation. 

Done.

P. 13: "The time constants of inhibitory and excitatory synapses were 10 ms and 30 ms, respectively." 

What is the (biological) justification for the choice of these parameters? 

How would varying them affect the main results (e.g. local manifolds)? 

We chose a relatively slow time constant for excitatory synapses because experimental data indicate that excitatory synaptic currents in Dp and piriform cortex contain a prominent NMDA component. We have now also simulated networks with equal time constants for excitatory and inhibitory synapses and equal biophysical parameters for excitatory and inhibitory neurons, which did not affect the main results (see also reply to the public review: comment 2 of reviewer 1).

P. 14: "Care was also taken to ensure that the variation in the number of output connections was low across neurons"   How exactly?

More detailed explanations have now been added in the Methods section: “connections of a presynaptic neuron y to postsynaptic neurons x were randomly deleted when their total number exceeded the average number of output connections by ≥5%, or added when they were lower by ≥5%.“

Reviewer #2 (Recommendations For The Authors): 

Congratulations on the great and interesting work! The results were nicely presented and the idea of continuous encoding on manifolds is very interesting. To improve the quality of the paper, in addition to the major points raised in the public review, here are some more detailed comments for the paper: 

(1) Generally, citations have to improve. Spiking networks with excitatory assemblies and different architectures of inhibitory populations have been studied before, and the claim about improved network stability in co-tuned E-I networks has been made in the following papers that need to be correctly cited: 

• Vogels TP, Sprekeler H, Zenke F, Clopath C, Gerstner W. 2011. Inhibitory Plasticity Balances Excitation and Inhibition in Sensory Pathways and Memory Networks. Science 334:1-7. doi:10.1126/science.1212991 (mentions that emerging precise balance on the synaptic weights can result in the overall network stability) 

• Lagzi F, Bustos MC, Oswald AM, Doiron B. 2021. Assembly formation is stabilized by Parvalbumin neurons and accelerated by Somatostatin neurons. bioRxiv doi: https://doi.org/10.1101/2021.09.06.459211 (among other things, contrasts stability and competition which arises from multistable networks with global inhibition and reciprocal inhibition)   • Rost T, Deger M, Nawrot MP. 2018. Winnerless competition in clustered balanced networks: inhibitory assemblies do the trick. Biol Cybern 112:81-98. doi:10.1007/s00422-017-0737-7 (compares different architectures of inhibition and their effects on network dynamics) 

• Lagzi F, Fairhall A. 2022. Tuned inhibitory firing rate and connection weights as emergent network properties. bioRxiv 2022.04.12.488114. doi:10.1101/2022.04.12.488114 (here, see the eigenvalue and UMAP analysis for a network with global inhibition and E/I assemblies) 

Additionally, there are lots of pioneering work about tracking of excitatory synaptic inputs by inhibitory populations, that are missing in references. Also, experimental work that show existence of cell assemblies in the brain are largely missing. On the other hand, some references that do not fit the focus of the statements have been incorrectly cited. 

The authors may consider referencing the following more pertinent studies on spiking networks to support the statement regarding attractor dynamics in the first paragraph in the Introduction (the current citations of Hopfield and Kohonen are for rate-based networks): 

• Wong, K.-F., & Wang, X.-J. (2006). A recurrent network mechanism of time integration in perceptual decisions. Journal of Neuroscience, 26(4), 1314-1328. https://doi.org/10.1523/JNEUROSCI.3733-05.2006 

• Wang, X.-J. (2008). Decision making in recurrent neuronal circuits. Neuron, 60(2), 215-234. https://doi.org/10.1016/j.neuron.2008.09.034  

• F. Lagzi, & S. Rotter. (2015). Dynamics of competition between subnetworks of spiking neuronal networks in the balanced state. PloS One. 

• Goldman-Rakic, P. S. (1995). Cellular basis of working memory. Neuron, 14(3), 477-485. 

• Rost T, Deger M, Nawrot MP. 2018. Winnerless competition in clustered balanced networks: inhibitory assemblies do the trick. Biol Cybern 112:81-98. doi:10.1007/s00422-017-0737-7. 

• Amit DJ, Tsodyks M (1991) Quantitative study of attractor neural network retrieving at low spike rates: I. substrate-spikes, rates and neuronal gain. Network 2:259-273. 

• Mazzucato, L., Fontanini, A., & La Camera, G. (2015). Dynamics of Multistable States during Ongoing and Evoked Cortical Activity. Journal of Neuroscience, 35(21), 8214-8231. 

We thank the reviewer for the references suggestions. We have carefully reviewed the reference list and made the following changes, which we hope address the reviewer’s concerns:

(1) We adjusted References about network stability in co-tuned E-I networks.

(2) We added the Lagzi & Rotter (2015), Amit et al. (1991), Mazzucato et al. (2015) and GoldmanRakic (1995) papers in the Introduction as studies on attractor dynamics in spiking neural networks. We preferred to omit the two X.J Wang papers, as they describe attractors in decision making rather than memory processes.

(3) We added the Ko et al. 2011 paper as experimental evidence for assemblies in the brain. In our view, there are few experimental studies showing the existence of cell assemblies in the brain, which we distinguish from cell ensembles, group of coactive neurons. 

(4) We also included Hennequin 2018, Brunel 2000, Lagzi et al. 2021 and Eckmann et al. 2024, which we had not cited in the initial manuscript.

(5) We removed the Wiechert et al. 2010 reference as it does not support the statement about geometry-preserving transformation by random networks.

(2) The gist of the paper is about how the architecture of inhibition (reciprocal vs. global in this case) can determine network stability and salient responses (related to multistable attractors and variations) for classification purposes. It would improve the narrative of the paper if this point is raised in the Introduction and Discussion section. Also see a relevant paper that addresses this point here: 

Lagzi F, Bustos MC, Oswald AM, Doiron B. 2021. Assembly formation is stabilized by Parvalbumin neurons and accelerated by Somatostatin neurons. bioRxiv doi: https://doi.org/10.1101/2021.09.06.459211 

Classification has long been proposed to be a function of piriform cortex and autoassociative memory networks in general, and we consider it important. However, the computational function of Dp or piriform cortex is still poorly understood, and we do not focus only on odor classification as a possibility. In fact, continuous representational manifolds also support other functions such as the quantification of distance relationships of an input to previously memorized stimuli, or multi-layer network computations (including classification). In the revised manuscript, we have performed additional analyses to explore these notions in more detail, as explained above (response to public reviews, comment 3 of reviewer 1). Furthermore, we have now expanded the discussion of potential computational functions of Tuned networks and explicitly discuss classification but also other potential functions. 

(3) A plot for the values of the inhibitory conductances in Figure 1 would complete the analysis for that section. 

In Figure 1, we decided to only show the conductances that we use to fit our model, namely the afferent and total synaptic conductances. As the values of the inhibitory conductances can be derived from panel E, we refrained from plotting them separately for the sake of simplicity. 

(4) How did the authors calculate correlations between activity patterns as a function of time in Figure 2E, bottom row? Does the color represent correlation coefficient (which should not be time dependent) or is it a correlation function? This should be explained in the Methods section. 

The color represents the Pearson correlation coefficient between activity patterns within a narrow time window (100 ms). We updated the Figure legend to clarify this: “Mean correlation between activity patterns evoked by a learned odor at different time points during odor presentation. Correlation coefficients were calculated between pairs of activity vectors composed of the mean firing rates of E neurons in 100 ms time bins. Activity vectors were taken from the same or different trials, except for the diagonal, where only patterns from different trials were considered.”

(5) Figure 3 needs more clarification (both in the main text and the figure caption). It is not clear what the axes are exactly, and why the network responses for familiar and novel inputs are different. The gray shaded area in panel B needs more explanation as well.  

We thank the reviewer for the comment. We have improved Figure 3A, the figure caption, as well as the text (see p.6). We hope that the figure is now clearer.

(6) The "scaled I" network, known for representing input patterns in discrete attractors, should exhibit clear separation between network responses in the 2D PC space in the PCA plots. However, Figure 4D and Figure 6D do not reflect this, as all network responses are overlapped. Can the authors explain the overlap in Figure 4D? 

In Figure 4D, activity of Scaled networks is distributed between three subregions in state space that are separated by the first 2 PCs. Two of them indeed correspond to attractor states representing the two learned odors while the third represents inputs that are not associated with these attractor states. To clarify this, please see also the density plot in Figure 4E. The few datapoints between these three subregions are likely outliers generated by the sequential change in inputs, as described in Supplementary Figure 8C.

(7) The reason for writing about the ISN networks is not clear. Co-tuned E-I assemblies do not necessarily have to operate in this regime. Also, the results of the paper do not rely on any of the properties of ISNs, but they are more general. Authors should either show the paradoxical effect associated with ISN (i.e., if increasing input to I neurons decreases their responses) or show ISN properties using stability analysis (See computational research conducted at the Allen Institute, namely Millman et al. 2020, eLife ). Currently, the paper reads as if being in the ISN regime is a necessary requirement, which is not true. Also, the arguments do not connect with the rest of the paper and never show up again. Since we know it is not a requirement, there is no need to have those few sentences in the Results section. Also, the choice of alpha=5.0 is extreme, and therefore, it would help to judge the biological realism if the raster plots for Figs 2-6 are shown.

We have toned down the part on ISN and reduced it to one sentence for readers who might be interested in knowing whether activity is inhibition-stabilized or not. We have also added the reference to the Tsodyks et al. 1997 paper from which we derive our stability analysis. The text now reads “Hence, pDp_sim entered a balanced state during odor stimulation (Figure 1D, E) with recurrent input dominating over afferent input, as observed in pDp (Rupprecht and Friedrich, 2018). Shuffling spike times of inhibitory neurons resulted in runaway activity with a probability of 0.79 ± 0.20, demonstrating that activity was inhibition-stabilized (Sadeh and Clopath, 2020b, Tsodyks et al., 1997).”  

We have now also added the raster plots as suggested by the reviewer (see Figure 2D, Supplementary Figure 1 G, Supplementary Figure 4). We thank the reviewer for this comment.

(8) In the abstract, authors mention "fast pattern classification" and "continual learning," but in the paper, those issues have not been addressed. The study does not include any synaptic plasticity. 

Concerning “continual learning” we agree that we do not simulate the learning process itself. However, Figure 6 show results of a simulation where two additional patterns were stored in a network that already contained assemblies representing other odors. We consider this a crude way of exploring the end result of a “continual learning” process. “Fast pattern classification” is mentioned because activity in balanced networks can follow fluctuating inputs with high temporal resolution, while networks with stable attractor states tend to be slow. This is likely to account for the occurrence of hysteresis effects in Scaled but not Tuned networks as shown in Supplementary

Fig. 8.

(9) In the Introduction, the first sentence in the second paragraph reads: "... when neurons receive strong excitatory and inhibitory synaptic input ...". The word strong should be changed to "weak".

Also, see the pioneering work of Brunel 2000. 

In classical balanced networks, strong excitatory inputs are counterbalanced by strong inhibitory inputs, leading to a fluctuation-driven regime. We have added Brunel 2000.

(10) In the second paragraph of the introduction, the authors refer to studies about structural co-tuning (e.g., where "precise" synaptic balance is mentioned, and Vogels et al. 2011 should be cited there) and functional co-tuning (which is, in fact, different than tracking of excitation by inhibition, but the authors refer to that as co-tuning). It makes it easier to understand which studies talk about structural co-tuning and which ones are about functional co-tuning. The paper by Znamenski 2018, which showed both structural and functional tuning in experiments, is missing here. 

We added the citation to the now published paper by Znamenskyi et al. (2024).  

(11) The third paragraph in the Introduction misses some references that address network dynamics that are shaped by the inhibitory architecture in E/I assemblies in spiking networks, like Rost et al 2018 and Lagzi et al 2021. 

These references have been added.

(12) The last sentence of the fourth paragraph in the Introduction implies that functional co-tuning is due to structural co-tuning, which is not necessarily true. While structural co-tuning results in functional co-tuning, functional co-tuning does not require structural co-tuning because it could arise from shared correlated input or heterogeneity in synaptic connections from E to I cells.  

We generally agree with the reviewer, but we are uncertain which sentence the reviewer refers to.

We assume the reviewer refers to the last sentence of the second (rather than the fourth paragraph), which explicitly mentions the “…structural basis of E/I co-tuning…”. If so, we consider this sentence still correct because the “structural basis” refers not specifically to E/I assemblies, but also includes any other connectivity that may produce co-tuning, including the connectivity underlying the alternative possibilities mentioned by the reviewer (shared correlated input or heterogeneity of synaptic connections).

(13) In order to ensure that the comparison between network dynamics is legit, authors should mention up front that for all networks, the average firing rates for the excitatory cells were kept at 1 Hz, and the background input was identical for all E and I cells across different networks.

We slightly revised the text to make this more clear “We (…) uniformly scaled I-to-E connection weights by a factor of χ until E population firing rates in response to learned odors matched the corresponding firing rates in rand networks, i.e., 1 Hz”

(14) In the last paragraph on page 5, my understanding was that an individual odor could target different cells within an assembly in different trials to generate trial to trail variability. If this is correct, this needs to be mentioned clearly. 

This is not correct, an odor consists of 150 activated mitral cells with defined firing rates. As now mentioned in the Methods, “Spikes were then generated from a Poisson distribution, and this process was repeated to create trial-to-trial variability.”

(15) The last paragraph on page 6 mentions that the four OB activity patterns were uncorrelated but if they were designed as in Figure 4A, dues to the existing overlap between the patterns, they cannot be uncorrelated. 

This appears to be a misunderstanding. We mention in the text (and show in Figure 4B) that the four odors which “… were assigned to the corners of a square…” are uncorrelated.  The intermediate odors are of course not uncorrelated. We slightly modified the corresponding paragraph (now on page 7) to clarify this: “The subspace consisted of a set of OB activity patterns representing four uncorrelated pure odors and mixtures of these pure odors. Pure odors were assigned to the corners of a square and mixtures were generated by selecting active mitral cells from each of the pure odors with probabilities depending on the relative distances from the corners (Figure 4A, Methods).”

(16) The notion of "learned" and "novel" odors may be misleading as there was no plasticity in the network to acquire an input representation. It would be beneficial for the authors to clarify that by "learned," they imply the presence of the corresponding E assembly for the odor in the network, with the input solely impacting that assembly. Conversely, for "novel" inputs, the input does not target a predefined assembly. In Figure 2 and Figure 4, it would be especially helpful to have the spiking raster plots of some sample E and I cells.  

As suggested by the reviewer, we have modified the existing spiking raster plots in Figure 2, such that they include examples of responses to both learned and novel odors. We added spiking raster plots showing responses of I neurons to the same odors in Supplementary Figure 1F, as well as spiking raster plots of E neurons in Supplementary Figure 4A. To clarify the usage of “learned” and “novel”, we have added a sentence in the Results section: “We thus refer to an odor as “learned” when a network contains a corresponding assembly, and as “novel” when no such assembly is present.”.

(17) In the last paragraph of page 8, can the authors explain where the asymmetry comes from? 

As mentioned in the text, the asymmetry comes from the difference in the covariance structure of different classes. To clarify, we have rephrased the sentence defining the Mahalanobis distance: 

“This measure quantifies the distance between the pattern and the class center, taking into account covariation of neuronal activity within the class. In bidirectional comparisons between patterns from different classes, the mean dM may be asymmetric if neural covariance differs between classes.”

(18) The first paragraph of page 9: random networks are not expected to perform pattern classification, but just pattern representation. It would have been better if the authors compared Scaled I network with E/I co-tuned network. Regardless of the expected poorer performance of the E/I co-tuned networks, the result would have been interesting. 

Please see our reply to the public review (reviewer 2).

(19) Second paragraph on page 9, the authors should provide statistical significance test analysis for the statement "... was significantly higher ...". 

We have performed a Wilcoxon signed-rank test, and reported the p-value in the revised manuscript (p < 0.01). 

(20) The last sentence in the first paragraph on page 11 is not clear. What do the authors mean by "linearize input-output functions", and how does it support their claim? 

We have now amended this sentence to clarify what we mean: “…linearize the relationship between the mean input and output firing rates of neuronal populations…”.

(21) In the first sentence of the last paragraph on page 11, the authors mentioned “high variability”, but it is not clear compared with which of the other 3 networks they observed high variability.

Structurally co-tuned E/I networks are expected to diminish network-level variability. 

“High variability” refers to the variability of spike trains, which is now mentioned explicity in the text. We hope this more precise statement clarifies this point.

(22) Methods section, page 14: "firing rates decreased with a time constant of 1, 2 or 4 s". How did they decrease? Was it an implementation algorithm? The time scale of input presentation is 2 s and it overlaps with the decay time constant (particularly with the one with 4 s decrease).  

Firing rates decreased exponentially. We have added this information in the Methods section.

Reviewer #3 (Recommendations For The Authors): 

In the following, I suggest minor corrections to each section which I believe can improve the manuscript. 

- There was no github link to the code in the manuscript. The code should be made available with a link to github in the final manuscript. 

The code can be found here: https://github.com/clairemb90/pDp-model. The link has been added in the Methods section.

Figure 1: 

- Fig. 1A: call it pDp not Dp. Please check if this name is consistent in every figure and the text. 

Thank you for catching this. Now corrected in Figure 1, Figure 2 and in the text.

- The authors write: "Hence, pDpsim entered an inhibition-stabilized balanced state (Sadeh and Clopath, 2020b) during odor stimulation (Figure 1D, E)." and then later "Shuffling spike times of inhibitory neurons resulted in runaway activity with a probability of ~80%, demonstrating that activity was indeed inhibition-stabilized. These results were robust against parameter variations (Methods)." I would suggest moving the second sentence before the first sentence, because the fact that the network is in the ISN regime follows from the shuffled spike timing result. 

Also, I'd suggest showing this as a supplementary figure. 

We thank the reviewer for this comment. We have removed “inhibition-stabilized” in the first sentence as there is no strong evidence of this in Rupprecht and Friedrich, 2018. And removed “indeed” in the second sentence. We also provided more detailed statistics. The text now reads “Hence, pDpsim entered a balanced state during odor stimulation (Figure 1D, E) with recurrent input dominating over afferent input, as observed in pDp (Rupprecht and Friedrich, 2018). Shuffling spike times of inhibitory neurons resulted in runaway activity with a probability of 0.79 ± 0.20, demonstrating that activity was inhibition-stabilized (Sadeh and Clopath, 2020b).”

Figure 2: 

- "... Scaled I networks (Figure 2H." Missing ) 

Corrected.

- The authors write "Unlike in Scaled I networks, mean firing rates evoked by novel odors were indistinguishable from those evoked by learned odors and from mean firing rates in rand networks (Figure 2F)." 

Why is this something you want to see? Isn't it that novel stimuli usually lead to high responses? Eg in the paper Schulz et al., 2021 (eLife) which is also cited by the authors it is shown that novel responses have high onset firing rates. I suggest clarifying this (same in the context of Fig. 3C). 

In Dp and piriform cortex, firing rates evoked by learned odors are not substantially different from firing rates evoked by novel odors. While small differences between responses to learned versus novel odors cannot be excluded, substantial learning-related differences in firing rates, as observed in other brain areas, have not been described in Dp or piriform cortex. We added references in the last paragraph of p.5. Note that the paper by Schulz et al. (2021) models a different type of circuit.  

- Fig. 2B: Indicate in figure caption that this is the case "Scaled I" 

This is not exactly the case “Scaled I”, as the parameter 𝝌𝝌 (increased I to E strength) is set to 1.

- Suppl Fig. 2I: Is E&F ever used in the manuscript? I couldn't find a reference. I suggest removing it if not needed. 

Suppl. Fig 2I E&F is now Suppl Fig.1G&H. We now refer to it in the text: “Activity of networks with E assemblies could not be stabilized around 1 Hz by increasing connectivity from subsets of I neurons receiving dense feed-forward input from activated mitral cells (Supplementary Figure 1GH; Sadeh and Clopath, 2020).”

Figure 3: 

- As mentioned in my comment in the public review section, I find the arguments about pattern completion a little bit confusing. For me it's not clear why an increase of output correlations over input correlations is considered "pattern completion" (this is not to say that I don't find the nonlinear increase of output correlations interesting). For me, to test pattern completion with second-order statistics one would need to do a similar separation as in Suppl Fig. 3, ie measuring the pairwise correlation at cells in the assembly L that get direct input from L OB with cells in the assembly L that do not get direct input from OB. If the pairwise correlations of assembly cells which do not get direct input from OB increase in correlations, I would consider this as pattern completion (similar to the argument that increase in firing rate in cells which are not directly driven by OB are considered a sign of pattern completion). 

Also, for me it now seems like that there are contradictory results, in Fig. 3 only Scaled I can lead to pattern completion while in the context of Suppl. Fig. 3 the authors write "We found that assemblies were recruited by partial inputs in all structured pDpsim networks (Scaled and Tuned) without a significant increase in the overall population activity (Supplementary Figure 3A)."   I suggest clarifying what the authors exactly mean by pattern completion, why the increase of output correlations above input correlations can be considered as pattern completion, and why the results differs when looking at firing rates versus correlations. 

Please see our reply to the public review (reviewer 3).

- I actually would suggest adding Suppl. Fig. 3 to the main figure. It shows a more intuitive form of pattern completion and in the text there is a lot of back and forth between Fig. 3 and Suppl. Fig. 3 

We feel that the additional explanations and panels in Fig.3 should clarify this issue and therefore prefer to keep Supplementary Figure 3 as part of the Supplementary Figures for simplicity.  

- In the whole section "We next explored effects of assemblies ... prevented strong recurrent amplification within E/I assemblies." the authors could provide a link to the respective panel in Fig. 2 after each statement. This would help the reader follow your arguments. 

We thank the reviewer for pointing this out. The references to the appropriate panels have been added. 

- Fig. 3A: I guess the x-axis has been shifted upwards? Should be at zero. 

We have modified the x-axis to make it consistent with panels B and C.  

- Fig. 3B: In the figure caption, the dotted line is described as the novel odor but it is actually the unit line. The dashed lines represent the reference to the novel odor. 

Fixed.

- Fig. 3C: The " is missing for Pseudo-Assembly N

Fixed.

- "...or a learned odor into another learned odor." Have here a ref to the Supplementary Figure 3B.

Added.

Figure 4:   

- "This geometry was largely maintained in the output of rand networks, consistent with the notion that random networks tend to preserve similarity relationships between input patterns (Babadi and Sompolinsky, 2014; Marr, 1969; Schaffer et al., 2018; Wiechert et al., 2010)." I suggest adding here reference to Fig. 4D (left). 

Added.

- Please add a definition of E/I assemblies. How do the authors define E/I assemblies? I think they consider both, Tuned I and Tuned E+I as E/I assemblies? In Suppl. Fig. 2I E it looks like tuned feedforward input is defined as E/I assemblies. 

We thank the reviewer for pointing this out. E/I assemblies are groups of E and I neurons with enhanced connectivity. In other words, in E/I assemblies, connectivity is enhanced not only between subsets of E neurons, but also between these E neurons and a subset of I neurons. This is now clarified in the text: “We first selected the 25 I neurons that received the largest number of connections from the 100 E neurons of an assembly. To generate E/I assemblies, the connectivity between these two sets of neurons was then enhanced by two procedures.”. We removed “E/I assemblies” in Suppl. Fig.2, where the term was not used correctly, and apologize for the confusion.

- Suppl. Fig. 4: Could the authors please define what they mean by "Loadings" 

The loadings indicate the contribution of each neuron to each principal component, see adjusted legend of Suppl. Fig. 4: “G. Loading plot: contribution of neurons to the first two PCs of a rand and a Tuned E+I network (Figure 4D).”

- Fig. 4F: The authors might want to normalize the participation ratio by the number of neurons (see e.g. Dahmen et al., 2023 bioRxiv, "relative PR"), so the PR is bound between 0 and 1 and the dependence on N is removed. 

We thank the reviewer for the suggestion, but we prefer to use the non-normalized PR as we find it more easily interpretable (e.g. number of attractor states in Scaled networks).

- Fig. 4G&H: as mentioned in the public review, I'd add the case of Scaled I to be able to compare it to the Tuned E+I case. 

As already mentioned in the public review, we thank the reviewer for this suggestion, which we have implemented.

- Figure caption Fig. 4H "Similar results were obtained in the full-dimensional space." I suggest showing this as a supplemental panel. 

Since this only adds little information, we have chosen not to include it as a supplemental panel to avoid overloading the paper with figures.

Figure 5: 

- As mentioned in the public review, I suggest that the authors add the Scaled I case to Fig. 5 (it's shown in all figures and also in Fig. 6 again). I guess for Scaled I the separation between L and M will be very good? 

Please see our reply to the public review (reviewer 3).

- Fig. 5A&B: I am a bit confused about which neurons are drawn to calculate the Mahalanobis distance. In Fig. 5A, the schematic indicates that the vector B from which the neurons are drawn is distinct from the distribution Q. For the example of odor L, the distribution Q consists of pure odor L with odors that have little mixtures with the other odors. But the vector v for odor L seems to be drawn only from odors that have slightly higher mixtures (as shown in the schematic in Fig. 5A). Is there a reason to choose the vector v from different odors than the distribution Q? 

The distribution Q and the vector v consist of activity patterns across the same neurons in response to different odors. The reason to choose a different odor for v was to avoid having this test datapoint being included in the distribution Q. We also wanted Q to be the same for all test datapoints. 

What does "drawn from whole population" mean? Does this mean that the vectors are drawn from any neuron in pDp? If yes, then I don't understand how the authors can distinguish between different odors (L,M,O,N) on the y-axis. Or does "whole population" mean that the vector is drawn across all assemblies as shown in the schematic in Fig. 5A and the case "neurons drawn from (pseudo-) assembly" means that the authors choose only one specific assembly? In any case, the description here is a bit confusing, I think it would help the reader to clarify those terms better.  

Yes, “drawn from whole population” means that we randomly draw 80 neurons from the 4000 E neurons in pDp. The y-axis means that we use the activity patterns of these neurons evoked by one of the 4 odors (L, M, N, O) as reference. We have modified the Figure legend to clarify this: “d_M was computed based on the activity patterns of 80 E neurons drawn from the four (pseudo-) assemblies (top) or from the whole population of 4000 E neurons (bottom). Average of 50 draws.”

- Suppl Fig. 5A: In the schematic the distance is called d_E(\bar{Q},\bar{V}) while the colorbar has d_E(\bar{Q},\bar{Q}) with the Qs in different color. The green Q should be a V. 

We thank the reviewer for spotting this mistake, it is now fixed.

- Fig. 5: Could the authors comment on the fact that a random network seems to be very good in classifying patterns on it's own. Maybe in the Discussion? 

The task shown in Figure 5 is a relatively easy one, a forced-choice between four classes which are uncorrelated. In Supplementary Figure 9, we now show classification for correlated classes, which is already much harder.

Figure 6: 

- Is the correlation induced by creating mixtures like in the other Figures? Please clarify how the correlations were induced. 

We clarified this point in the Methods section: “The pixel at each vertex corresponded to one pure odor with 150 activated and 75 inhibited mitral cells (…) and the remaining pixels corresponded to mixtures. In the case of correlated pure odors (Figure 6), adjacent pure odors shared half of their activated and half of their inhibited cells.”. An explicit reference to the Methods section has also been added to the figure legend.

- Fig. 6C (right): why don't we see the clear separation in PC space as shown in Fig. 4? Is this related to the existence of correlations? Please clarify. 

Yes. The assemblies corresponding to the correlated odors X and Y overlap significantly, and therefore responses to these odors cannot be well separated, especially for Scaled networks. We added the overlap quantification in the Results section to make this clear. “These two additional assemblies had on average 16% of neurons in common due to the similarity of the odors.”

- "Furthermore, in this regime of higher pattern similarity, dM was again increased upon learning, particularly between learned odors and reference classes representing other odors (not shown)." Please show this (maybe as a supplemental figure). 

We now show the data in Supplementary Figure 9.

Discussion: 

- The authors write: "We found that transformations became more discrete map-like when amplification within assemblies was increased and precision of synaptic balance was reduced. Likewise, decreasing amplification in assemblies of Scaled networks changed transformations towards the intermediate behavior, albeit with broader firing rate distributions than in Tuned networks (not shown)." 

Where do I see the first point? I guess when I compare in Fig. 4D the case of Scaled I vs Tuned E+I, but the sentence above sounds like the authors showed this in a more step-wise way eg by changing the strength of \alpha or \beta (as defined in Fig. 1). 

Also I think if the authors want to make the point that decreasing amplification in assemblies changes transformation with a different rate distribution in scaled vs tuned networks, the authors should show it (eg adding a supplemental figure). 

The first point is indeed supported by data from different figures. Please note that the revised manuscript now contains further simulations that reinforce this statement, particularly those shown in Supplementary Figure 6, and that this point is now discussed more extensively in the Discussion. We hope that these revisions clarify this general point.

The data showing effects of decreasing amplification in assemblies is now shown in Supplementary Figure 6 (Scaled[adjust])

- I suggest adding the citation Znamenskiy et al., 2024 (Neuron; https://doi.org/10.1016/j.neuron.2023.12.013), which shows that excitatory and inhibitory (PV) neurons with functional similarities are indeed strongly connected in mouse V1, suggesting the existence of E/I assembly structure also in mammals.

Done.

Author response:

The following is the authors’ response to the original reviews.

eLife assessment:

Developing a reliable method to record ancestry and distinguish between human somatic cells presents significant challenges. I fully acknowledge that my current evidence supporting the claim of lineage tracing with fCpG barcodes is inadequate. I agree with Reviewer 1 that fCpG barcodes are essentially a cellular division clock that diverges over time. A division clock could potentially document when cells cease to divide during development, with immediate daughter cells likely exhibiting more similar barcodes than those that are less related. Although it remains uncertain whether the current fCpG barcodes capture useful biological information, refinement of this type of tool could complement other approaches that reconstruct human brain function, development, and aging.

Due to my lack of clarity, the fCpG barcode was perceived to be a new type of cell classifier. However, it is fundamentally different. fCpG sites are selected based on their differences between cells of the same type, while traditional cell classifiers focus on sites with consistent methylation patterns in cells of the same type. Despite these opposing criteria, fCpG barcodes and traditional cell classifiers may align because neuron subtypes often share common progenitors. As a result, cells of the same phenotype are also closely related by ancestry, and ex post facto, have similar fCpG barcodes. fCpG barcodes are complementary to cell type classifiers, and potentially provide insights into aspects such as mitotic ages, diversity within a clade, and migration of immediate daughters---information which is otherwise difficult to obtain. The title has been modified to “Human Brain Ancestral Barcodes” to better reflect the function of the fCpG barcodes. The manuscript is edited to correct errors, and a new Supplement is added to further explain fCpG barcode mechanics and present new supporting data.

Reviewer #1 (Public review):

I thank Reviewer 1 for his constructive comments. Major noted weaknesses were 1) insufficient clarity and brevity of the methodology, 2) inconsistent or erroneous use of neurodevelopmental concepts, and 3) lack of consideration for alternative explanations.

(1) The methodology is now outlined in detailed in a new Supplement, including simulations that indicate that the error rate consistent with the experimental data is about 0.01 changes in methylation per fCpG site per division.

(2) Conceptual and terminology errors noted by the Reviewers are corrected in the manuscript.

(3) I agree completely with the alternative explanation of Reviewer 1 that fCpGs are “a cellular division clock that diverges over 'time'”. Differences between more traditional cell type classifiers and fCpG barcodes are more fully outlined in the new Supplement.  Ancestry recorded by fCpGs and cell type classifiers are confounded because cells of the same phenotype typically have common progenitors---cells within a clade have similar fCpG barcodes because they are closely related. fCpG barcodes can compliment cell type classifiers with additional information such as mitotic ages, ancestry within a clade, and daughter cell migration.

Reviewer #1 (Recommendations for the authors):

(1) A lot of the interpretations suffer from an extremely loose/erroneous use of developmental concepts and a lack of transparency. For instance:

a) The thalamus is not part of the brain stem

Corrected.

b) The pons contains cells other than inhibitory neurons in the data; the same is true for the hippocampus which contains multiple cell types

Corrected to refer to the specific cell types in these regions.

c) The author talks about the rostral-caudal timing a lot which is not really discussed to this degree in the cited references. Thus, it is also unclear how interneurons fit in this model as they are distinguished by a ventral-dorsal difference from excitatory neurons. Also, it is unclear whether the timing is really as distinct as claimed. For instance, inhibitory neurons and excitatory neurons significantly overlap in their birth timing. Finally, conceptually, it does not make sense to go by developmental timing as the author proposes that it is the number of divisions that is relevant. While they are somewhat correlated there are potentially stark differences.

The manuscript attempts to describe what might be broadly expected when barcodes are sampled from different cell types and locations. As a proposed mitotic clock, the fCpG barcode methylation level could time when each neuron ceased division and differentiated. The wide ranges of fCpG barcode methylation of each cell type (Fig 2A) would be consistent with significant overlap between cell types. The manuscript is edited to emphasize overlapping rather than distinct sequential differentiation of the cell types.

d) Neocortical astrocytes and some oligodendrocytes share a lineage, whereas a subset of oligodendrocytes in the cortex shares an origin with interneurons. This could confound results but is never discussed.

The manuscript does not assess glial lineages in detail because neurons were preferentially included in the sampling whereas glial cells were non-systematically excluded. This sampling information is now included in the section “fCpG barcode identification”.

e) Neocortical interneurons should be more closely related in terms of lineage-to-excitatory neurons than other inhibitory neurons of, for instance, the pons. This is not clearly discussed and delineated.

This is not discussed. It may not be possible analyze these details with the current data. The ancestral tree reconstructions indicate that excitatory neurons that appear earlier in development (and are more methylated) are more often more closely related to inhibitory neurons.

f) While there is some spread of excitatory neurons tangentially, there is no tangential migration at the scale of interneurons as (somewhat) suggested/implied here.

The abstract and results have been modified to indicate greater inhibitory than excitatory neuron tangential migration, but that the extent of excitatory neuron tangential migration cannot be determined because of the sparse sampling and that barcodes may be similar by chance.

g) The nature of the NN cells is quite important as cells not derived from the neocortical anlage are unlikely to share a developmental origin (e.g., microglia, endothelial cells). This should be clarified and clearly stated.

The manuscript is modified to indicate that NN cells are microglial and endothelial cells. These cells have different developmental origins, and their data are present in Fig 2A, but are not further used for ancestral analysis.  

(2) The presentation is often somewhat confusing to me and lacks detail. For instance:

a) The methods are extremely short and I was unable to find a reference for a full pipeline, so other researchers can replicate the work and learn how to use the pipeline.

The pipeline including python code is outlined in the new Supplement

b) Often numbers are given as ~XX when the actual number with some indication of confidence or spread would be more appropriate.

Data ranges are often indicated with the violin plots.

c) Many figure legends are exceedingly short and do not provide an appropriate level of detail.

Figure legends have been modified to include more detail

d) Not defining groups in the figure legends or a table is quite unacceptable to me. I do not think that referring to a prior publication (that does not consistently use these groups anyway) is sufficient.

The cell groups are based on the annotations provided with each single cell in the public databases.

e) The used data should be better defined and introduced (number of cells, different subtypes across areas, which cells were excluded; I assume the latter as pons and hippocampus are only mentioned for one type of neuronal cells, see also above).

The data used are present in Supplemental File 2 under the tab “cell summary H01, H02, H04”.

f) Why were different upper bounds used for filtering for H01 and H02, and H04 is not mentioned? Why are inhibitory and excitatory neurons specifically mentioned (Lines 61-66)?

The filtering is used to eliminate, as much as possible, cell type specific methylation, or CpG sites with skewed neuron methylation. The filtering eliminates CpG sites with high or low methylation within each of the three brains, and within the two major neuron subtypes. The goal is to enrich for CpG sites with polymorphic but not cell type specific methylation. This process is ad hoc as success criteria are currently uncertain. The extent of filtering is balanced by the need to retain sufficient numbers of fCpGs to allow comparisons between the neurons.

g) What 'progenitor' does the author refer to? The Zygote? If yes, can the methylation status be tested directly from a zygote? There is no single progenitor for these cells other than the zygote. Does the assumption hold true when taking this into account? See, for instance, PMID 33737485 for some estimation of lineage bottlenecks.

A brain progenitor cell can be defined as the common ancestor of all adult neurons, and is the first cell where each of its immediate daughter cell lineages yield adult neurons. The zygote is a progenitor cell to all adult cells, and barcode methylation at the start of conception, from the oocyte to the ICM, was analyzed in the new Supplement. The proposed brain progenitor cell with a fully methylated barcode was not yet evident even in the ICM.

(3) I am generally not convinced that the fCpGs represent anything but a molecular clock of cell divisions and that many of the similarities are a function of lower division numbers where the state might be more homogenous. This mainly derives from the issues cited above, the lack of convincing evidence to the contrary, and the sparsity of the assessed data.

Agree that the fCpG barcode is a mitotic clock that becomes polymorphic with divisions. As outlined in the new Supplement, ancestry and cell type are confounded because cells of the same type typically have a common progenitor.

a) There appears little consideration or modeling of what the ability to switch back does to the lineage reconstruction.

fCpG methylation flipping is further analyzed and discussed in the new Supplement.

b) None of the data convinced me that the observations cannot be explained by the aforementioned molecular clock and systematic methylation similarities of cell types due to their cell state.

See above

(4) Uncategorized minor issues:

a) The author should explain concepts like 'molecular clock hypothesis' (line 27) or 'radial unit hypothesis' (line 154), as they are somewhat complex and might not be intuitive to readers.

The molecular clock hypothesis is deleted and the radial unit hypothesis is explained in more detail in the manuscript.

b) Line 32: '[...] replication errors are much higher compared to base replication [...]'. I think this is central to the method and should be better explained and referenced. Maybe even through a schematic, as this is a central concept for the entire manuscript.

The fCpG barcode mechanics are better explained in the new Supplement. With simulations, the fCpG flip rate is about 0.01 per division per fCpG.

c) Line 41: 'neonatal'. Does the author mean to say prenatal? Most of the cells discussed are postmitotic before birth.

Corrected to prenatal.

d) Line 96: what does 'flip' mean in this context? Please also see the comment on Figure 2C.

Edited to “chage”

e) Lines 134-135: I am not sure whether the author claims to provide evidence for this question, and I would be careful with claims that this work does resolve the question here.

Have toned down claims as evidence for my analysis is currently inadequate.

f) Lines 192-193: I disagree as the fCpGs can switch back and the current data does not convince me that this is an improvement upon mosaic mutation analysis. In my mind, the main advantage is the re-analysis of existing data and the parallel functional insights that can be obtained.

Lineage analysis is more straightforward with DNA sequencing, but with an error rate of ~10-9 per base per division, one needs to sequence a billion base pairs to distinguish between immediate daughter cells. By contrast, with an inferred error rate of ~10-2 per fCpG per division, much less sequencing (about a million-fold less) is needed to find differences between daughter cells.

g) Lines 208-209: I would be careful with claims of complexity resolution given many of the limitations and inherent systematic similarities, as well as the potential of fCpGs to change back to an ancestral state later in the lineage.

Have modified the manuscript to indicate the analysis would be more challenging due to back changes.

h) There seem to be few figures that assess phenomena across the three brains. Even when they exist there is no attempt to provide any statistical analyses to support the conclusions or permutations to assess outlier status relative to expectations.

The analysis could be more extensive, but with only three brains, any results, like this study itself, would be rightly judged inadequate.

Figure 2B: there appears to be a higher number of '0s' for, for instance, inhibitory neurons compared to excitatory neurons. Is that correct and worth mentioning? The changing axes scales also make it hard to assess.

Inhibitory neurons do appear to have more unmethylated fCpGs compared to excitatory neurons, but in general, most inhibitory fCpGs are methylated with a skew to fully methylated fCpGs, consistent with the barcode starting predominately methylated and inhibitory neurons generally appearing earlier in development relative to excitatory neurons.

j) Figure 2C: I have several issues with this. A minor one is the use of 'Glial' which, I believe, does not appear anywhere else before this, so I am unclear what this curve represents. Generally, however, I am not sure what the y-axis represents, as it is not described in the methods or figure legend. I initially thought it was the cumulative frequency, but I do not think that this squares with the data shown in B. I appreciate the overall idea of having 'earlier'/samples with fewer divisions being shifted to the left, but it is very confusing to me when I try to understand the details of the plot.

This graph is now better described in the legend. “Glial” cells are defined as oligodendrocytes and astrocytes. Other non-neuronal cells (such a microglial cells) have now been removed from the graph.

This graph attempts to illustrate how it may be possible to reconstruct brain development from adult neurons, assuming barcodes are mitotic clocks that become polymorphic with cell division. The X axis is “time”, and the Y axis indicates when different cell types reach their adult levels. The cartoon indicates what is visually present along the X axis during development--- brainstem, then ganglionic eminences with a thin cortex, and finally the mature brain with a robust cortex. Time for the X axis is barcode methylation and starts at 100% and ends at 50% or greater methylation. The fCpG barcode methylation of each cell places it on this timeline and indicates when it ceased dividing and differentiated.

The Y axis indicates the progressive accumulation of the final adult contents of each cell type during this timeline. Early in development, the brain is rudimentary and adult cells are absent. At 90% methylation, only the inhibitory neurons in the pons are present. At 80% methylation, some excitatory neurons are beginning to appear. Inhibitory neurons in the pons have reached their final adult levels and many other inhibitory neuron types are reaching adult levels. By 70% methylation, most inhibitory neurons have reached their adult levels, and more adult excitatory neurons (mainly low cortical neurons, L4-6) and glial cells are beginning to appear. By 60% methylation, inhibitory neurogenesis has largely finished. Adult excitatory neurons and glial cells are more abundant and reach their adult levels by 50% or greater cell barcode methylation levels.

The graph illustrates a rough alignment between mitotic ages inferred by barcode methylation levels and the physical appearances of different neuronal types during development. Many neurons die during development, and this graph, if valid, indicates when neurons that survive to adulthood appear during development.

k) Figure 4Bff: it is confusing to me that the text jumps to these panels after introducing Figure 5. This makes it very hard to read this section of the text.

The Figures appear in the order they are first referred to in the text.

l) Figure 5A: could any of this difference be explained by the shared lineage of excitatory neurons and dorsal neocortical glia?

Not sure

m) Figure 5B: after stating that interneurons have a higher lineage fidelity, the figure legend here states the opposite and I am somewhat confused by this statement.

The legend and text have been clarified. Fig 5A restricts fidelity to within inhibitory cell types. Fig 5B compares between neuron subtypes, and illustrates more apparent inhibitory subtype switching, albeit there are more interneuron subtypes than excitatory subtypes.

n) Figure 5E: generally, the use of tSNE for large pairwise distance analysis is often frowned upon (e.g., PMID 37590228), and I would reconsider this argument.

This analysis was an attempt to illustrate that cells of the same phenotype based on their tSNE metrics can be either closely or more distantly related. Although the tSNE comparisons were restricted to subtypes (and not to the entire tSNE graph), tSNE are not designed for such comparisons. This graph and discussion are deleted. 

Reviewer #2 (Public review):

The manuscript by Shibata proposed a potentially interesting idea that variation in methylcytosine across cells can inform cellular lineage in a way similar to single nucleotide variants (SNVs). The work builds on the hypothesis that the "replication" of methylcytosine, presumably by DNMT1, is inaccurate and produces stochastic methylation variants that are inherited in a cellular lineage. Although this notion can be correct to some extent, it does not account for other mechanisms that modulate methylcytosines, such as active gain of methylation mediated by DNMT3A/B activity and activity demethylation mediated by TET activity. In some cases, it is known that the modulation of methylation is targeted by sequence-specific transcription factors. In other words, inaccurate DNMT1 activity is only one of the many potential ways that can lead to methylation variants, which fundamentally weakens the hypothesis that methylation variants can serve as a reliable lineage marker. With that being said (being skeptical of the fundamental hypothesis), I want to be as open-minded as possible and try to propose some specific analyses that might better convince me that the author is correct. However, I suspect that the concept of methylation-based lineage tracing cannot be validated without some kind of lineage tracing experiment, which has been successfully demonstrated for scRNA-seq profiling but not yet for methylation profiling (one example is Delgado et al., nature. 2022).

I thank Reviewer 2 for the careful evaluation. The validation experiment example (Delgado et al.) introduced sequence barcodes in mice, which is not generally feasible for human studies.

(1) The manuscript reported that fCpG sites are predominantly intergenic. The author should also score the overlap between fCpG sites and putative regulatory elements and report p-values. If fCpG sites commonly overlap with regulatory elements, that would increase the possibility that these sites being actively regulated by enhancer mechanisms other than maintenance methyltransferase activity.

As mentioned for Reviewer 1, fCpGs are filtered to eliminate cell type specific methylation.

(2) The overlap between fCpG and regulatory sequence is a major alternative explanation for many of the observations regarding the effectiveness of using fCpG sites to classify cell types correctly. One would expect the methylation level of thousands of enhancers to be quite effective in distinguishing cell types based on the published single-cell brain methylome works.

As mentioned above, the manuscript did not clearly indicate that the fCpG barcode is not a cell type classifier. The distinctions between fCpG barcodes and cell type classifiers are better explained in the new Supplement.

(3) The methylation level of fCpG sites is higher in hindbrain structures and lower in forebrain regions. This observation was interpreted as the hindbrain being the "root" of the methylation barcodes and, through "progressive demethylation" produced the methylation states in the forebrain. This interpretation does not match what is known about methylation dynamics in mammalian brains, in particular, there is no data supporting the process of "progressive demethylation". In fact, it is known that with the activation of DNMT3A during early postnatal development in mice or humans (Lister et al., 2013. Science), there is a global gain of methylation in both CH and CG contexts. This is part of the broader issue I see in this manuscript, which is that the model might be correct if "inaccurate mC replication" is the only force that drives methylation dynamics. But in reality, active enzymatic processes such as the activation of DNMT3A have a global impact on the methylome, and it is unclear if any signature for "inaccurate mC replication" survives the de novo methylation wave caused by DNMT3A activity.

Reviewer 2 highlights a critical potential flaw in that any ancestral signal recorded by random replication errors could be overwritten by other active methylation processes. I cannot present data that indicates fCpG replication errors are never overwritten, but new data indicate barcode reproducibility and stability with aging.

New data are also present where barcodes are compared between daughter cells (zygote to ICM) in the setting of active and passive demethylation, when germline methylation is erased. This new analysis shows that daughter cells in 2 to 8 cell embryos have more related barcodes than morula or ICM cells. The subsequent active remethylation by a wave of DNMT3A activity may underlie the observation that the barcode appears to start predominately methylated in brain progenitors.

(3) Perhaps one way the author could address comment 3 is to analyze methylome data across several developmental stages in the same brain region, to first establish that the signal of "inaccurate mC replication" is robust and does not get erased during early postnatal development when DNMT3A deposits a large amount of de novo methylation.

See above

(4) The hypothesis that methylation barcodes are homogeneous among progenitor cells and more polymorphic in derived cells is an interesting one. However, in this study, the observation was likely an artifact caused by the more granular cell types in the brain stem, intermediate granularity in inhibitory cells, and highly continuous cell types in cortical excitatory cells. So, in other words, single-cell studies typically classify hindbrain cell types that are more homogenous, and cortical excitatory cells that are much more heterogeneous. The difference in cell type granularity across brain structures is documented in several whole-brain atlas papers such as Yao et al. 2023 Nature part of the BICCN paper package.

As noted above, fCpG barcode polymorphisms and cell type differentiation are confounded because cells of the same phenotype tend to have common progenitors. The fCpG barcode is not a cell type classifier but more a cell division clock that becomes polymorphic with time. Although fCpG barcodes could be more polymorphic in cortical excitatory cells because there are many more types, fCpG barcodes would inherently become more polymorphic in excitatory cells because they appear later in development.

(5) As discussed in comment 2, the author needs to assess whether the successful classification of cell types (brain lineage) using fCpG was, in fact, driven by fCpG sites overlapping with cell-type specific regulatory elements.

Although unclear in the manuscript, the fCpG is not a cell classifier and the barcode is polymorphic between cells of the same type. fCpG barcodes can appear to be cell classifiers because cell types appear at different times during development, and therefore different cell types have characteristic average barcode methylation levels.

(6) In Figure 5E, the author tried to address the question of whether methylation barcodes inform lineage or post-mitotic methylation remodeling. The Y-axis corresponds to distances in tSNE. However, tSNE involves non-linear scaling, and the distances cannot be interpreted as biological distances. PCA distances or other types of distances computed from high-dimensional data would be more appropriate.

The Figure and discussion are deleted (similar comment by Reviewer 1)

Reviewer #3 (Public review):

Summary:

In the manuscript entitled "Human Brain Barcodes", the author sought to use single-cell CpG methylation information to trace cell lineages in the human brain.

Strengths:

Tracing cell lineages in the human brain is important but technically challenging. Lineage tracing with single-cell CpG methylation would be interesting if convincing evidence exists.

Weaknesses:

As the author noted, "DNA methylation patterns are usually copied between cell division, but the replication errors are much higher compared to base replication". This unstable nature of CpG methylation would introduce significant problems in inferring the true cell lineage. The unreliable CpG methylation status also raises the question of what the "Barcodes" refer to in the title and across this study. Barcodes should be stable in principle and not dynamic across cell generations, as defined in Reference#1. It is not convincing that the "dynamic" CpG methylation fits the "barcodes" terminology. This problem is even more concerning in the last section of results, where CpG would fluctuate in post-mitotic cells.

I thank Reviewer 3 for his thoughtful and careful evaluation. I think the “barcode” terminology is appropriate. Dynamic engineered barcodes such as CRISPR/Cas9 mutable barcodes are used in biology to record changes over time. The fCpG barcode appears to start with a single state in a progenitor cell and changes with cell division to become polymorphic in adult cells. Therefore, I think the description of a dynamic fCpG barcode is appropriate.

Reviewer #3 (Recommendations for the authors):

(1) As the author noted, "DNA methylation patterns are usually copied between cell division, but the replication errors are much higher compared to base replication". This unstable nature of CpG methylation would introduce significant problems in inferring the true cell lineage. To establish DNA methylation as a means for lineage tracing, one control experiment would be testing whether the DNA methylation patterns can faithfully track cell lineages for in vitro differentiated & visibly tracked cell lineages. Has this kind of experiment been done in the field?

These types of experiments have not been performed to my knowledge and an appropriate tissue culture model is uncertain. New single cell WGBS data from the zygote to ICM indicate that more immediate daughter cells have more related barcodes even in the setting of active DNA demethylation.

(2) The study includes assumptions that should be backed with solid rationale, supporting evidence, or reference. Here are a couple of examples:

a) the author discarded stable CpG sites with <0.2 or >0.8 average methylation without a clear rationale in H02, and then used <0.3 and >0.7 for a specific sample H01.

The filtering was ad hoc and was used to remove, as much as possible, CpG sites with cell type specific or patient specific methylation. CpG sites with skewed methylation are more likely cell type specific, whereas X chromosome CpG sites with methylation closer to 0.5 in male cells are more likely to be unstable. The ad hoc filtering attempted to remove cell specific CpGs sites while still retaining enough CpG sites to allow comparisons between cells.

b) The author assumed that the early-formed brain stem would resemble progenitors better and have a higher average methylation level than the forebrain. However, this difference in DNA methylation status could reflect developmental timing or cell type-specific gene expression changes.

This observation that brain stem neurons that appear early in development have highly methylated fCpG barcodes in all 3 brains supports the idea that the fCpG barcode starts predominately methylated. Alternative explanations are possible.

(3) The conclusion that excitatory neurons undergo tangential migration is unclear - how far away did the author mean for the tangential direction? Lateral dispersion is known, but it would be striking that the excitatory neurons travel across different brain regions. The question is, how would the author interpret shared or divergent methylation for the same cell type across different brain regions?

As noted with Reviewer 1, this analysis is modified to indicate that evidence of tangential migration is greater for inhibitory than excitatory neurons, but the extent of excitatory neuron migration is uncertain because of sparse sampling, and because fCpG barcodes can be similar by chance.

(4) The sparsity and resolution of the single-cell DNA methylation data. The methylation status is detected in only a small fraction (~500/31,000 = 1.6%) of fCpGs per cell, with only 48 common sites identified between cell pairs. Given that the human genome contains over 28 million CpG sites, it is important to evaluate whether these fCpGs are truly representative. How many of these sites were considered "barcodes"?

fCpG barcodes are distinct from traditional cell type classifiers, and how fCpGs are identified are better outlined in the new Supplement.

(5) While focusing on the X-chromosome may simplify the identification of polymorphic fCpGs, the confidence in determining its methylation status (0 or 1) is questionable when a CpG site is covered by only one read. Did the author consider the read number of detected fCpGs in each cell when calculating methylation levels? Certain CpG sites on autosomes may also have sufficient coverage and high variability across cells, meeting the selection criteria applied to X-chromosome CpGs.

In most cases, a fCpG site was covered by only a single read

(6) The overall writing in the Title, the Main text, Figure legends, and Methods sections are overly simplified, making it difficult to follow. For instance, how did the author perform PWD analysis? How did they handle missing values when constructing lineage trees?

There is not much introduction to lineage tracing in the human brain or the use of DNA methylation to trace cell lineage.

These shortcomings are improved in the manuscript and with the new Supplement. The analysis pipeline including the Python programs are outlined and included as new Supplemental materials. IQ tree can handle the binary fCpG barcode data and skips missing values with its standard settings.

Line 80: it is unclear: "Brain patterns were similar"

Clarified

Line 98: The meaning is unclear here: "Outer excitatory and glial progenitor cells are present" What are these glial progenitor cells and when/how they stop dividing?

The glial cells are the oligodendrocytes and astrocytes. The main take away point is that these glial cells have low barcode methylation, consistent with their appearances later in development.

Line 104: It is unclear if this is a conclusion or assumption -- "A progenitor cell barcode should become increasingly polymorphic with subsequent divisions." The "polymorphic" happens within the progenitors, their progenies, or their progenies at different time points.

The statement is now clarified as an assumption in the manuscript.

Similarly line 134 "Barcodes would record neuronal differentiation and migration." Is this a conclusion from this study or a citation? How is the migration part supported?

The reasoning is better explained in the manuscript.  Migration can be documented if immediate daughter cells with similar barcodes are found in different parts of the adult brain, albeit analysis is confounded by sparse sampling and because barcodes may be similar by chance.

Line 148 and 150: "Nearest neighbor ... neuron pairs" in DNA methylation status would conceivably reflect their cell type-specific gene expression, how did the author distinguish this from cell lineage?

As noted above, because cells with similar phenotypes usually arise from common progenitors, cells within a clade are also usually related. However, the barcodes are still polymorphic within a clade and potentially add complementary information on mitotic ages, ancestry within a clade, and possible cell migration.

Figure 3C: "Cells that emerge early in development" Where are they on the figure?

Hindbrain neurons differentiate early in development and their barcodes are more methylated. The figure has been modified to label some of the values with their neuron types. Also, the older figure mistakenly included data from all 3 brains and now the data are only from brain H01.

Figures 4D and 4E, distinguishing cell subtypes is challenging, as the same color palette is used for both excitatory and inhibitory neurons.

Unfortunate limitations due to complexity and color limitations

Figures 4 and 5, what are these abbreviations?

The abbreviations are presented in Figure 1 and maintained in subsequent figures.

Author response:

Reviewer #1 (Public review):

This manuscript presents an interesting exploration of the potential activation mechanisms of DLK following axonal injury. While the experiments are beautifully conducted and the data are solid, I feel that there is insufficient evidence to fully support the conclusions made by the authors.

In this manuscript, the authors exclusively use the puc-lacZ reporter to determine the activation of DLK. This reporter has been shown to be induced when DLK is activated. However, there is insufficient evidence to confirm that the absence of reporter activation necessarily indicates that DLK is inactive. As with many MAP kinase pathways, the DLK pathway can be locally or globally activated in neurons, and the level of DLK activation may depend on the strength of the stimulation. This reporter might only reflect strong DLK activation and may not be turned on if DLK is weakly activated. The results presented in this manuscript support this interpretation. Strong stimulation, such as axotomy of all synaptic branches, caused robust DLK activation, as indicated by puc-lacZ expression. In contrast, weak stimulation, such as axotomy of some synaptic branches, resulted in weaker DLK activation, which did not induce the puc-lacZ reporter. This suggests that the strength of DLK activation depends on the severity of the injury rather than the presence of intact synapses. Given that this is a central conclusion of the study, it may be worthwhile to confirm this further. Alternatively, the authors may consider refining their conclusion to better align with the evidence presented.

We wish to further clarify a striking aspect of puc-lacZ induction following injury: it is bimodal. It is either induced (in various injuries that remove all synaptic boutons), or not induced, including in injuries that spared only 1-2 remaining boutons. This was particularly evident for injuries that spared the NMJ on muscle 29, which is comprised of only a few boutons. In some instances, only a single bouton was evident on muscle 29. While our injuries varied enormously in the number of branches and boutons that were lost, we did not see a comparable variability in puc-lacZ induction.  In the revision we will include additional images to better demonstrate this observation.

The reviewer (and others) fairly point out that our current study focuses on puc-lacZ as a reporter of Wnd signaling in the cell body. We consider this to be a downstream integration of events in axons that are more challenging to detect. It is striking that this integration appears strongly sensitized to the presence of spared synaptic boutons. Examination of Wnd’s activation in axons and synapses is a goal for our future work.

As noted by the authors, DLK has been implicated in both axon regeneration and degeneration. Following axotomy, DLK activation can lead to the degeneration of distal axons, where synapses are located. This raises an important question: how is DLK activated in distal axons? The authors might consider discussing the significance of this "synapse connection-dependent" DLK activation in the broader context of DLK function and activation mechanisms.

While it has been noted that inhibition of DLK can mildly delay Wallerian degeneration (Miller et al., 2009), this does not appear to be the case for retinal ganglion cell axons following optic nerve crush (Fernandes et al., 2014). It is also not the case for Drosophila motoneurons and NMJ terminals following peripheral nerve injury (Xiong et al., 2012; Xiong and Collins, 2012). Instead, overexpression of Wnd or activation of Wnd by a conditioning injury leads to an opposite phenotype - an increase in resiliency to Wallerian degeneration for axons that have been previously injured (Xiong et al., 2012; Xiong and Collins, 2012). The downstream outcome of Wnd activation is highly dependent on the context; it may be an integration of the outcomes of local Wnd/DLK activation in axons with downstream consequences of nuclear/cell body signaling.  The current study suggests some rules for the cell body signaling, however, how Wnd is regulated at synapses and why it promotes degeneration in some circumstances but not others are important future questions.

For the reviewer’s suggestion, it is interesting to consider DLK’s potential contributions to the loss of NMJ synapses in a mouse model of ALS (Le Pichon et al., 2017; Wlaschin et al., 2023). Our findings suggest that the synaptic terminal is an important locus of DLK regulation, while dysfunction of NMJ terminals is an important feature of the ‘dying back’ hypothesis of disease etiology (Dadon-Nachum et al., 2011; Verma et al., 2022). We propose that the regulation of DLK at synaptic terminals is an important area for future study, and may reveal how DLK might be modulated to curtail disease progression. Of note, DLK inhibitors are in clinical trials (Katz et al., 2022; Le et al., 2023; Siu et al., 2018), but at least some have been paused due to safety concerns (Katz et al., 2022). Further understanding of the mechanisms that regulate DLK are needed to understand whether and how DLK and its downstream signaling can be tuned for therapeutic benefit.

Reviewer #2 (Public review):

Summary:

The authors study a panel of sparsely labeled neuronal lines in Drosophila that each form multiple synapses. Critically, each axonal branch can be injured without affecting the others, allowing the authors to differentiate between injuries that affect all axonal branches versus those that do not, creating spared branches. Axonal injuries are known to cause Wnd (mammalian DLK)-dependent retrograde signals to the cell body, culminating in a transcriptional response. This work identifies a fascinating new phenomenon that this injury response is not all-or-none. If even a single branch remains uninjured, the injury signal is not activated in the cell body. The authors rule out that this could be due to changes in the abundance of Wnd (perhaps if incrementally activated at each injured branch) by Wnd, Hiw's known negative regulator. Thus there is both a yet-undiscovered mechanism to regulate Wnd signaling, and more broadly a mechanism by which the neuron can integrate the degree of injury it has sustained. It will now be important to tease apart the mechanism(s) of this fascinating phenomenon. But even absent a clear mechanism, this is a new biology that will inform the interpretation of injury signaling studies across species.

Strengths:

(1) A conceptually beautiful series of experiments that reveal a fascinating new phenomenon is described, with clear implications (as the authors discuss in their Discussion) for injury signaling in mammals.

(2) Suggests a new mode of Wnd regulation, independent of Hiw.

Weaknesses:

(1) The use of a somatic transcriptional reporter for Wnd activity is powerful, however, the reporter indicates whether the transcriptional response was activated, not whether the injury signal was received. It remains possible that Wnd is still activated in the case of a spared branch, but that this activation is either local within the axons (impossible to determine in the absence of a local reporter) or that the retrograde signal was indeed generated but it was somehow insufficient to activate transcription when it entered the cell body. This is more of a mechanistic detail and should not detract from the overall importance of the study

We agree. The puc-lacZ reporter tells us about signaling in the cell body, but whether and how Wnd is regulated in axons and synaptic branches, which we think occurs upstream of the cell body response, remains to be addressed in future studies.

(2) That the protective effect of a spared branch is independent of Hiw, the known negative regulator of Wnd, is fascinating. But this leaves open a key question: what is the signal?

This is indeed an important future question, and would still be a question even if Hiw were part of the protective mechanism by the spared synaptic branch. Our current hypothesis (outlined in Figure 4) is that regulation of Wnd is tied to the retrograde trafficking of a signaling organelle in axons. The Hiw-independent regulation complements other observations in the literature that multiple pathways regulate Wnd/DLK (Collins et al., 2006; Feoktistov and Herman, 2016; Klinedinst et al., 2013; Li et al., 2017; Russo and DiAntonio, 2019; Valakh et al., 2013). It is logical for this critical stress response pathway to have multiple modes of regulation that may act in parallel to tune and restrain its activation.

Reviewer #3 (Public review):

Summary:

This manuscript seeks to understand how nerve injury-induced signaling to the nucleus is influenced, and it establishes a new location where these principles can be studied. By identifying and mapping specific bifurcated neuronal innervations in the Drosophila larvae, and using laser axotomy to localize the injury, the authors find that sparing a branch of a complex muscular innervation is enough to impair Wallenda-puc (analogous to DLK-JNK-cJun) signaling that is known to promote regeneration. It is only when all connections to the target are disconnected that cJun-transcriptional activation occurs.

Overall, this is a thorough and well-performed investigation of the mechanism of spared-branch influence on axon injury signaling. The findings on control of wnd are important because this is a very widely used injury signaling pathway across species and injury models. The authors present detailed and carefully executed experiments to support their conclusions. Their effort to identify the control mechanism is admirable and will be of aid to the field as they continue to try to understand how to promote better regeneration of axons.

Strengths:

The paper does a very comprehensive job of investigating this phenomenon at multiple locations and through both pinpoint laser injury as well as larger crush models. They identify a non-hiw based restraint mechanism of the wnd-puc signaling axis that presumably originates from the spared terminal. They also present a large list of tests they performed to identify the actual restraint mechanism from the spared branch, which has ruled out many of the most likely explanations. This is an extremely important set of information to report, to guide future investigators in this and other model organisms on mechanisms by which regeneration signaling is controlled (or not).

Weaknesses:

The weakest data presented by this manuscript is the study of the actual amounts of Wallenda protein in the axon. The authors argue that increased Wnd protein is being anterogradely delivered from the soma, but no support for this is given. Whether this change is due to transcription/translation, protein stability, transport, or other means is not investigated in this work. However, because this point is not central to the arguments in the paper, it is only a minor critique.

We agree and are glad that the reviewer considers this a minor critique; this is an area for future study. In Supplemental Figure 1 we present differences in the levels of an ectopically expressed GFP-Wnd-kinase-dead transgene, which is strikingly increased in axons that have received a full but not partial axotomy. We suspect this accumulation occurs downstream of the cell body response because of the timing. We observed the accumulations after 24 hours (Figure S1F) but not at early (1-4 hour) time points following axotomy (data not shown). Further study of the local regulation of Wnd protein and its kinase activity in axons is an important future direction.

As far as the scope of impact: because the conclusions of the paper are focused on a single (albeit well-validated) reporter in different types of motor neurons, it is hard to determine whether the mechanism of spared branch inhibition of regeneration requires wnd-puc (DLK/cJun) signaling in all contexts (for example, sensory axons or interneurons). Is the nerve-muscle connection the rule or the exception in terms of regeneration program activation?

DLK signaling is strongly activated in DRG sensory neurons following peripheral nerve injury (Shin et al., 2012), despite the fact that sensory neurons have bifurcated axons and their projections in the dorsal spinal cord are not directly damaged by injuries to the peripheral nerve. Therefore it is unlikely that protection by a spared synapse is a universal rule for all neuron types. However the molecular mechanisms that underlie this regulation may indeed be shared across different types of neurons but utilized in different ways. For instance, nerve growth factor withdrawal can lead to activation of DLK (Ghosh et al., 2011), however neurotrophins and their receptors are regulated and implemented differently in different cell types. We suspect that the restraint of Wnd signaling by the spared synaptic branch shares a common underlying mechanism with the restraint of DLK signaling by neurotrophin signaling. Further elucidation of the molecular mechanism is an important next step towards addressing this question.

Because changes in puc-lacZ intensity are the major readout, it would be helpful to better explain the significance of the amount of puc-lacZ in the nucleus with respect to the activation of regeneration. Is it known that scaling up the amount of puc-lacZ transcription scales functional responses (regeneration or others)? The alternative would be that only a small amount of puc-lacZ is sufficient to efficiently induce relevant pathways (threshold response).

While induction of puc-lacZ expression correlates with Wnd-mediated phenotypes, including sprouting of injured axons (Xiong et al., 2010), protection from Wallerian degeneration (Xiong et al., 2012; Xiong and Collins, 2012) and synaptic overgrowth (Collins et al., 2006), we have not observed any correlation between the degree of puc-lacZ induction (eg modest, medium or high) and the phenotypic outcomes (sprouting, overgrowth, etc). Rather, there appears to be a striking all-or-none difference in whether puc-lacZ is induced or not induced. There may indeed be a threshold that can be restrained through multiple mechanisms. We posit in figure 4 that restraint may take place in the cell body, where it can be influenced by the spared bifurcation.

References Cited:

Collins CA, Wairkar YP, Johnson SL, DiAntonio A. 2006. Highwire restrains synaptic growth by attenuating a MAP kinase signal. Neuron 51:57–69.

Dadon-Nachum M, Melamed E, Offen D. 2011. The “dying-back” phenomenon of motor neurons in ALS. J Mol Neurosci 43:470–477.

Feoktistov AI, Herman TG. 2016. Wallenda/DLK protein levels are temporally downregulated by Tramtrack69 to allow R7 growth cones to become stationary boutons. Development 143:2983–2993.

Fernandes KA, Harder JM, John SW, Shrager P, Libby RT. 2014. DLK-dependent signaling is important for somal but not axonal degeneration of retinal ganglion cells following axonal injury. Neurobiol Dis 69:108–116.

Ghosh AS, Wang B, Pozniak CD, Chen M, Watts RJ, Lewcock JW. 2011. DLK induces developmental neuronal degeneration via selective regulation of proapoptotic JNK activity. J Cell Biol 194:751–764.

Hao Y, Frey E, Yoon C, Wong H, Nestorovski D, Holzman LB, Giger RJ, DiAntonio A, Collins C. 2016. An evolutionarily conserved mechanism for cAMP elicited axonal regeneration involves direct activation of the dual leucine zipper kinase DLK. Elife 5. doi:10.7554/eLife.14048

Huntwork-Rodriguez S, Wang B, Watkins T, Ghosh AS, Pozniak CD, Bustos D, Newton K, Kirkpatrick DS, Lewcock JW. 2013. JNK-mediated phosphorylation of DLK suppresses its ubiquitination to promote neuronal apoptosis. J Cell Biol 202:747–763.

Katz JS, Rothstein JD, Cudkowicz ME, Genge A, Oskarsson B, Hains AB, Chen C, Galanter J, Burgess BL, Cho W, Kerchner GA, Yeh FL, Ghosh AS, Cheeti S, Brooks L, Honigberg L, Couch JA, Rothenberg ME, Brunstein F, Sharma KR, van den Berg L, Berry JD, Glass JD. 2022. A Phase 1 study of GDC-0134, a dual leucine zipper kinase inhibitor, in ALS. Ann Clin Transl Neurol 9:50–66.

Klinedinst S, Wang X, Xiong X, Haenfler JM, Collins CA. 2013. Independent pathways downstream of the Wnd/DLK MAPKKK regulate synaptic structure, axonal transport, and injury signaling. J Neurosci 33:12764–12778.

Le K, Soth MJ, Cross JB, Liu G, Ray WJ, Ma J, Goodwani SG, Acton PJ, Buggia-Prevot V, Akkermans O, Barker J, Conner ML, Jiang Y, Liu Z, McEwan P, Warner-Schmidt J, Xu A, Zebisch M, Heijnen CJ, Abrahams B, Jones P. 2023. Discovery of IACS-52825, a potent and selective DLK inhibitor for treatment of chemotherapy-induced peripheral neuropathy. J Med Chem 66:9954–9971.

Le Pichon CE, Meilandt WJ, Dominguez S, Solanoy H, Lin H, Ngu H, Gogineni A, Sengupta Ghosh A, Jiang Z, Lee S-H, Maloney J, Gandham VD, Pozniak CD, Wang B, Lee S, Siu M, Patel S, Modrusan Z, Liu X, Rudhard Y, Baca M, Gustafson A, Kaminker J, Carano RAD, Huang EJ, Foreman O, Weimer R, Scearce-Levie K, Lewcock JW. 2017. Loss of dual leucine zipper kinase signaling is protective in animal models of neurodegenerative disease. Sci Transl Med 9. doi:10.1126/scitranslmed.aag0394

Li J, Zhang YV, Asghari Adib E, Stanchev DT, Xiong X, Klinedinst S, Soppina P, Jahn TR, Hume RI, Rasse TM, Collins CA. 2017. Restraint of presynaptic protein levels by Wnd/DLK signaling mediates synaptic defects associated with the kinesin-3 motor Unc-104. Elife 6. doi:10.7554/eLife.24271

Miller BR, Press C, Daniels RW, Sasaki Y, Milbrandt J, DiAntonio A. 2009. A dual leucine kinase-dependent axon self-destruction program promotes Wallerian degeneration. Nat Neurosci 12:387–389.

Nihalani D, Merritt S, Holzman LB. 2000. Identification of structural and functional domains in mixed lineage kinase dual leucine zipper-bearing kinase required for complex formation and stress-activated protein kinase activation. J Biol Chem 275:7273–7279.

Russo A, DiAntonio A. 2019. Wnd/DLK is a critical target of FMRP responsible for neurodevelopmental and behavior defects in the Drosophila model of fragile X syndrome. Cell Rep 28:2581–2593.e5.

Shin JE, Cho Y, Beirowski B, Milbrandt J, Cavalli V, DiAntonio A. 2012. Dual leucine zipper kinase is required for retrograde injury signaling and axonal regeneration. Neuron 74:1015–1022.

Siu M, Sengupta Ghosh A, Lewcock JW. 2018. Dual Leucine Zipper Kinase Inhibitors for the Treatment of Neurodegeneration. J Med Chem 61:8078–8087.

Valakh V, Walker LJ, Skeath JB, DiAntonio A. 2013. Loss of the spectraplakin short stop activates the DLK injury response pathway in Drosophila. J Neurosci 33:17863–17873.

Verma S, Khurana S, Vats A, Sahu B, Ganguly NK, Chakraborti P, Gourie-Devi M, Taneja V. 2022. Neuromuscular junction dysfunction in amyotrophic lateral sclerosis. Mol Neurobiol 59:1502–1527.

Wlaschin JJ, Donahue C, Gluski J, Osborne JF, Ramos LM, Silberberg H, Le Pichon CE. 2023. Promoting regeneration while blocking cell death preserves motor neuron function in a model of ALS. Brain 146:2016–2028.

Xiong X, Collins CA. 2012. A conditioning lesion protects axons from degeneration via the Wallenda/DLK MAP kinase signaling cascade. J Neurosci 32:610–615.

Xiong X, Hao Y, Sun K, Li J, Li X, Mishra B, Soppina P, Wu C, Hume RI, Collins CA. 2012. The Highwire ubiquitin ligase promotes axonal degeneration by tuning levels of Nmnat protein. PLoS Biol 10:e1001440.

Xiong X, Wang X, Ewanek R, Bhat P, Diantonio A, Collins CA. 2010. Protein turnover of the Wallenda/DLK kinase regulates a retrograde response to axonal injury. J Cell Biol 191:211–223.

DOI: 10.1371/journal.pbio.3002941

Resource: RRID:Addgene_12259

Curator: @olekpark

SciCrunch record: RRID:Addgene_12259

What is this?

DOI: 10.1101/2024.12.23.630083

Resource: RRID:IMSR_JAX:000664

Curator: @sjvitug

SciCrunch record: RRID:IMSR_JAX:000664

What is this?

DOI: 10.1007/s10875-024-01846-y

Resource: RRID:Addgene_198735

Curator: @olekpark

SciCrunch record: RRID:Addgene_198735

What is this?

DOI: 10.1007/s10875-024-01846-y

Resource: RRID:Addgene_174038

Curator: @olekpark

SciCrunch record: RRID:Addgene_174038

What is this?

DOI: 10.1007/s10875-024-01846-y

Resource: RRID:Addgene_132777

Curator: @olekpark

SciCrunch record: RRID:Addgene_132777

What is this?

DOI: 10.1007/s00018-024-05506-7

Resource: None

Curator: @olekpark

SciCrunch record: RRID:Addgene_171719

What is this?

Peak PA: * Noche: 10-20% menos que las diurnas * Peak matutino -> Eventual predicción de mayor RCV * HTA bata blanca: 15-20% de las mediciones

Dg diferenciales de encefalopatía hipertensiva: * Isquemia cerebral * ACV hemorrágico o trombótico * Trastorno convulsivo * Efecto de masa * Pseudotumor cerebral * Delirium tremens * Meningitis * Porfiria intermitente aguda * Lesión cerebral traumática o química * Encefalopatía urémica

Receptores adrenérgicos:

a1: ppalmente NA. Ubicados en células postsinápticas del músculo listo -> Vasoconstricción - Intensifican reabsorción renal de Na - Antihipertensivos inhiben receptores a1

a2: ppalmente por catecolaminas. Ubicados en mb presinápticas que sintetizan NA. -> Inhibe liberación de NA - Antihipertensivos inhiben receptores a2

b1: Ubicados en miocardio. Inotrópico y cronotrópico -> Aumenta GC - Estimula liberación de renina renal. - Antihipertensivos inhiben receptores b1

b2: en músculo liso. Estimulada por adrenalina. -> relaja músculo liso y vasodilatación.

Замечание: Здесь стоит уменьшить шаг по оси Y, иначе происходят страшные вещи. breaks_width(50)

Nghiên cứu của bài báo chuyển sự chú ý sang việc tận dụng phản hồi được tạo bởi LLM để làm tăng chất lượng của bản tóm tắt, trong khi đa số các nghiên cứu hiện nay tập trung vào việc sử dụng LLM để đánh giá bản tóm tắt. Cụ thể, mục tiêu là tóm tắt các bản tóm tắt tốt bằng việc khai thác phản hồi của LLM trên 3 tiêu chí chính: Độ trung thực, độ hoàn thiện và độ xúc tích.

For functions whose range is uniform space, the function then thus be defined as uniform space: $$ (f_1 \times f_2) (X) \subseteq U $$ Where \(U\) is entourage. We can apply Moore-Osgood Theorem to such function \( f(x, y) \) if \(f_x \rightrightarrows f\)

Note: This response was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

1. General Statements

We thank the reviewers for their thorough and positive evaluation of the manuscript.

2. Point-by-point description of the revisions

We revised the manuscript following the suggestions of the reviewers to make the article more concise and comprehensible to a wider audience. Specifically, we rearranged Section 5, rewrote the difficult-to-understand sections 5 and 6, and removed unnecessary or overlapping text in Introduction and Discussion. We have also addressed the specific points raised by the reviewers. The responses to individual points are detailed below.

Reviewer 1:

The reviewer did not ask for any changes to the manuscript.

We thank the reviewer for the positive evaluation of the manuscript.

Reviewer 2:

1/ Title: Structure-based mechanism of RyR channel operation by calcium and magnesium ions

The authors may consider using an alternative term instead of "operation".

Thank you for the suggestion. We considered and discussed the term "RyR channel operation" very thoroughly with several colleagues, including native English speakers, and we found it to represent the complex RyR behavior in situ and in experiments most exactly. Alternative terms such as "control" suggest a one-way deterministic action from the ion binding to the protein state, which is not the case. The terms such as "modulation" implicate the presence of a higher RyR state-governing principle, such as phosphorylation, nitrosylation, binding of auxiliary proteins, etc.

2/ Abstract: Please spell out CFF and MWC theorem.

Thank you for the proposal. CFF was changed to caffeine; MWC was changed to Monod-Wyman-Changeaux

3/ Line 87-88: "In striated muscle cells, RyR channels cluster at discrete sites of sarcoplasmic reticulum attached to the sarcolemma where electrical excitation triggers transient calcium release by activation of RyRs."

There is no attachment between sarcoplasmic reticulum and sarcolemma, please rewrite.

We respectfully disagree, since there is strong evidence for the existence of discrete contact sites between the sarcolemma and sarcoplasmic reticulum both at triads of skeletal muscle (Rossi et al., 2019) and at dyads of cardiac muscle (Mackrill, 2022), at which both membranes are firmly attached.

However, to avoid potential misunderstanding, we changed the sentence to "In striated muscle cells, RyR channels cluster at the discrete sites of sarcoplasmic reticulum attached to the sarcolemma in triads or dyads, where electrical excitation triggers transient calcium release by activation of RyRs" (lines 86-87).

4/ Lines 104-107: "Recently, mathematical modeling of the cardiac calcium release site (Iaparov et al., 2022) confirmed that Mg2+ ions could at the same time act as the negative competitor at the calcium activation site and as an inhibitor at the inhibition site. Unfortunately, the structural counterpart of RyR inactivation, an inhibitory binding site for divalent ions, has not been located yet in RyR structures."

Note that the exact structural counterpart exists (Nayak et al., 2022, 2024), where Ca and Mg were found both at the activation and inhibition sites. The paragraph should be updated accordingly.

We respectfully disagree. In the cited works of Nayak et al. (2022; 2024) it was shown that Ca and Mg ions bind firmly at the activation site. Both atoms were also observed at the ACP molecule bound at the ATP binding site. However, they were not observed at the divalent ion-binding inhibition site, which is distinct from the ATP binding site and resides in the loops of the EF-hand region.

However, to clarify the meaning of the disputed sentence, we have changed it to: "Although binding of Ca2+ or Mg2+ to an inhibitory binding site has not been observed yet in RyR structures, a consensus is emerging that the EF-hand loops constitute this site (Gomez et al., 2016; Zheng and Wen, 2020; Nayak et al., 2024; Chirasani et al., 2024 )" (lines 107-109).

5/ Lines 108-110: The activation of RyR by agonists was shown to be accompanied by a conformational change around the Ca2+ binding site that leads to a decrease in the free energy and to a concomitant increase of the Ca2+ binding affinity and a population shift between the closed and open conformations (Dashti et al., 2020).

Please clarify to what state does the "decrease in free energy" refer, to the open or to the closed state?

Thank you for the proposal. The text was changed to: "The activation of RyR by agonists was shown to be accompanied by a conformational change around the Ca2+ binding site that leads to a decrease in the free energy of the open state and concomitantly to an increase of the Ca2+ binding affinity of the activation site. As a result, the occurrence probability of a RyR state/conformation shifts from the closed toward the open (Dashti et al., 2020)" (lines 110-113).

6/ Figure 2: please indicate if distances were measured between the C-alphas or side chains.

Thank you for the proposal. The figure legend was modified to "Distances D1 between the Cα atoms of E4075 and R4736 or equivalent. Right - Distances D2 between the Cα atoms of K4101 and D4730 or equivalent."

7/ Line 353-357: "These data suggest that interactions between the basic arginine residue R4736 and the acidic residues at the start of the initial helix E of the EF1-hand are specific for Ca2+-dependent inactivation in RyR1, whereas the interactions between the lysine K4101 that immediately follows the F helix of EF1 and the middle of the S23 loop (corresponding to D4730 and I4731 in RyR1) may play a part in the inactivation of both RyR1 and RyR2 isoforms.

Sentence is unclear; please rewrite. Overall, the entire section "Spatial interactions between the EF-hand and S23* regions" should be simplified and shortened.

Thank you for the proposal. The text was changed to: "These data suggest that interactions between the basic arginine residue R4736 and the acidic residues E4075 and D4079 are specific for Ca2+-dependent inactivation in RyR1, whereas the interactions between the lysine K4101 and the residues D4730 and I4731 (rRyR1 notation)* may play a part in the inactivation of both RyR1 and RyR2 isoforms." (lines 334-337).

We did not find a way how to make the whole section simpler and shorter at the same time without losing clarity.

8/ Lines 246-249 and Table 1. "all structures corresponding to rRyR1 residues 4063-4196 were<br /> subjected to energy minimization and submitted to the MIB2 server for evaluation of the ion binding score (IBS) of individual amino acid residues and the number of ion binding poses (NIBP) for Ca and Mg ions."

Please elaborate on the "ion binding score" and "number of ion binding poses" concepts and provide reference for the MIB2 server.

Thank you for the proposal. We added the reference for the server (Lu et al., 2022) (line 228) and added the information: "IBS values of individual residues are determined using sequence and structure conservation comparison with 409 and 209 respective templates from the PDB database for Ca2+ and Mg2+ (Lin et al., 2016) and assessing the similarity of the configuration of the residue to its configurations in known structures of its complexes with the given metal (Lu et al., 2012). Ion binding sites are determined by locally aligning the query protein with the metal ion-binding templates and calculating its score as the RMSD-weighted scoring function Z. The site is accepted if it has a scoring function Z>1, and based on the local 3D structure alignment between the query protein and the metal ion-binding template, the metal ion in the template is transformed into the query protein structure (Lin et al., 2016). The larger the IBS value, the higher the tendency of the residue to bind the ion. The larger the NIBP value, the larger the number of such complexes with acceptable structure" (lines 224-234).

9/ Lines 460-466: Nine structural models of RyR were selected, and then these are referred to in the text only with the pdb code. The reviewer understands that it would be difficult to recapitulate all conditions but either a table in the main manuscript file or a minimal description in the text following the pdb code would increase clarity and help readers to follow the content.

Thank you for the proposal. We added a new Table 2 "Model structures used for identifying the allosteric pathways" on line 452 that contains the required information, and inserted a reference to it in the text at line 446 "According to these criteria we selected five RyR1 model structures (Table 2)..."

10/ Line 467: "In the selected structures, we identified residues with high allosteric coupling intensities (ACI) for both the inhibition and activation network and compared them with residues important for ligand binding and gating of RyR (Table 2)."

Please define further the concept of "allosteric coupling intensities". The corresponding methods section appears to focus on the outputs of the OHM server without delving too much on the algorithm or principles followed. Is the allosteric coupling between neighboring residues, or reflect movement of the residues due to ligand binding? Is there a "reference" state or are the comparisons carried out within each allosteric state? This would help to introduce better the sections "The inhibition network" and "The activation network".

Thank you for this suggestion. We have lately realized, considering both the server output and the original work of Wang et al. (2020), that a better term for the variable depicting the role of the residue in the allosteric pathway would be the residue importance RI rather than the ACI. The allosteric pathway is determined on the basis of the network of contacts between pairs of residues in the given structure. The more contacts are present between two residues, the higher is the probability that a perturbation will be propagated from one to the other residue (Eq. 3 of Wang et al. (2020)). An allosteric pathway is then defined as the pathway that transmits the signal the whole way from the allosteric site to the active site.

Based on this we have changed in the manuscript the term "allosteric coupling intensity" to "residue importance" throughout the text and figures of the manuscript. It should be underlined, that this change has no effect whatsoever on presented data and conclusions. We inserted the following formulation in the Results section:

"The term residue importance defines the extent to which the given residue is involved in the propagation of a perturbation from the allosteric site to the active site, i.e., the fraction of simulated perturbations transmitted through this particular residue. The more contacts are present between two residues, the higher is the probability that a perturbation will be propagated from one to the other residue (Wang et al., 2020)." (lines 439-443).

We also inserted the following formulations into the Methods section: "The simulation of the perturbation propagation was performed 10 000 times per structure and pathway to estimate the values of residue importance." (lines 1093-1095), and we expanded the relevant sentence: "Allosteric pathways were traced using the server OHM (https://dokhlab.med.psu.edu/ohm/#/home, (Wang et al., 2020)), in which the allosteric pathway is determined on the basis of the network of contacts between pairs of residues in the given structure." (lines 1082-1084).

11/ Figure 8: The figure would be more meaningful if the pathways were drawn in the context of the 3D structure.

Thank you for the proposal. The pathways described in Fig. 8 are too complex for description in the RyR 3D structure, therefore they were not presented in the original manuscript. However, to follow the reviewer's proposal we have illustrated the pathways observed in the inactivated RyR1 channel (7tdg) and the open RyR2 channel (7u9) in Expanded View Figure EV1 and added the corresponding Expanded View Movie EV1 and EV2. These RyR structures were selected for displaying both the intra- and inter-monomeric inactivation pathways.

12/ Lines 610-612: "The structure of the inactivated RyR2 has not been determined yet; however, it is plausible to suppose that it exists at high concentrations of divalent ions and differs from the inactivated RyR1 structure by the extent of EF-hand - S23* coupling. "

The speculation would be more fit for the discussion section.

Thank you for the proposal; however, the sentence introduces a logical supposition, necessary there for reasoning on the construction of the model. We reformulated the sentence to: "In the absence of a structure of the inactivated RyR2, the model assumes that such a structure exists at high concentrations of divalent ions and differs from the inactivated RyR1 structure by the extent of EF-hand - S23* coupling." (lines 573-575).

13/ Lines 617-619: Closed and primed macrostates could be combined into a single closed macrostate of the model since both are closed and cannot be functionally distinguished at a constant ATP concentration.

The rationale for combining closed with primed does not seem a good idea, especially since the authors also mention that "the primed state is structurally very close to the open state" (lines 925-926). If the COI model is based on the structural findings, in principle it seems that primed should be treated separately.

Thank you for the proposal. The use of both the closed and primed states was crucial for solving the model. As a matter of fact, although the primed and closed states are in part structurally different, functionally they are identical, that is, closed. Consequently, to be distinguished in a functional model we would need to incorporate single-channel data obtained under conditions when the ratio of closed and primed channels was modulated under otherwise identical conditions. Unfortunately, such a set of data, for instance at a varying ATP concentration for a range of cytosolic Ca2+ concentrations, does not exist for either RyR1 or RyR2 channels. Moreover, while there are several RyR1 high-resolution structures in the primed state (such as the 7tzc that we used; 2.45 Å; Melville et al. (2022)), the resolution of the corresponding RyR2 structures (6jg3, 6jh6, 6jhn; 4.5 - 6.1 Å; Chi et al. (2019)) is not sufficient for determination of allosteric pathways. Fortunately, however, the two sets of conditions for RyR2 open probability data that were available in the literature turned out to represent activation of channels either selectively from the closed state (Fig. 10C), or almost selectively from the primed state (Fig. 10A, B). This allowed us to interpret the difference in the allosteric coefficients as a consequence of this fact.

To better clarify the idea, the corresponding text of the Discussion was modified as follows (lines 926-931): "RyR channels can be considered mostly in the primed state under these conditions since the binding of ATP analogs induces the primed structural macrostate in RyRs even in the absence of Ca2+ (Cholak et al., 2023). Fortunately, the two sets of conditions for RyR2 open probability data that were available in the literature turned out to represent activation of channels either selectively from the closed state (Fig. 10C), or selectively from the primed state (Fig. 10A, B).", and "construction of such a model is at present hampered by the lack of open probability data at a sufficiently wide range of experimental conditions and the absence of high-resolution structures of WT RyR2 in the primed state" (lines 934-937).

14/ Line 619. Please define the "COI" acronym. I assume it is closed, open and inactivated but this is not mentioned.

We thank the reviewer for noticing the insufficiency. We expanded the specific sentence as follows: Therefore, we constructed the model of RyR operation, termed the COI (closed-open-inactivated) model, in which we assigned a functional macrostate corresponding to each of the closed, open, and inactivated structural macrostates (Figure 9A)" (line 582).

15/ Figure 9: The diagrams are difficult to follow. Something that could improve it is to differentiate more between open and closed gates, but further elaboration would help the reader.

We thank the reviewer for paying attention to details. The open state was differentiated in Figure 9 (after line 603) by adding a pore opening to the gate.

To elaborate on the gating transitions and to keep the manuscript concise, we added a new Expanded View Figure EV2, which illustrates the relationship between the ion binding within macrostates and the transitions between macrostates.

Nevertheless, for the complexity of the model, which would need a multidimensional presentation, we had to limit the illustration to only the binding of the first ions at the binding sites. We hope that it will help the reader to grasp the principle of the model function more easily.

16/ One comment is that the manuscript is too long; the manuscript exceeds the typical length required by most journals. To enhance its suitability for publication, the content needs to be synthesized and streamlined. The manuscript is written for an audience specialized in the RyR field and may be challenging for outsiders or for readers unfamiliar with structure and/or biophysical models.

We thank the reviewer for opening this problem. The specific contribution to the understanding of RyR operation communicated by this manuscript was achieved by the synergy of approaches coming from different fields of RyR research - the structural, the functional, and the synthetic/systems ones. This needed deep immersion into complex studies performed over several decades to unwrap their complementary contributions. Only then we could synthesize the stepwise advances and integrate the mosaic of partial discoveries into the COI model. When conceptualizing the manuscript we were also considering a two-paper version, one on structural aspects and the other on modeling aspects. We realized that the two papers would need to have a very high overlap at the allosteric mechanism to be understandable in separation and would be difficult to publish in the same journal. We also anticipated a typical side effect that structuralists and modelers would read just their parts and would not appreciate enough the feedback from alternative views - how to design and interpret future structural, functional, and modeling studies.

Compacting the manuscript would be extremely difficult for us. In our view, the dense text would make it even more challenging for readers unfamiliar with some of the numerous approaches used here, as often happens to prominent multidisciplinary journals. Maybe it would be possible with the help of AI, but for now, we prefer to remain authentic.

Nevertheless, we made some effort. To shorten the manuscript, we have removed the paragraph describing the timeline of the search for the RyR inhibition site that was originally on lines 126-151 and replaced it with the paragraph on lines 129-134: "The regulatory domains involved in both, activation and inactivation of RyRs (Figure 1) are located in the C-terminal quarter of the RyR. The Central domain participates in the Ca2+ binding activation site; the C-terminal domain bears several residues of Ca-, ATP- and caffeine-binding activation sites; the U-motif participates at the ATP- and caffeine-binding sites; the EF-hand region contains the putative Ca-binding pair EF1 and EF2; and the S23 loop bears one residue of the caffeine-binding site and two residues interacting with the EF-hand region of a neighboring monomer (Samso, 2017; Hadiatullah et al., 2022)". We also removed the statements about the proposed kinetic mechanism of inactivation by Nayak et al. (2022), originally on lines 175-184. Finally, we removed the discussion of the work of Gomez et al. (2016) originally on lines 882-889, since it fully overlapped with the statements in Results on lines 358-367 (now lines 338-347). We also moved the text of the subsection "Relationship between the COI model and RyR allosteric pathways" (originally lines 670-685) into subsection "Construction of the model of RyR operation", lines 592-603 and 645-662 of the revised version.

17/ Another comment is the limited consideration of two relevant published works. One is by Chirasani et al. (2024), focused on allosteric pathways similar to the ones described here. The other work is by Nayak et al (2024), with cryo-EM structures of RyR1 focused on the interplay with Mg2+ and Ca2+. Overall, the manuscript would be strengthened by incorporating such related results in the literature.

We thank the reviewer for the concerns, but we cannot fully agree. The paper of Chirasani et al. (2024 ) was cited in the manuscript as its online-first version, Chirasani et al. (2023). The manuscript now refers to the printed version proposed by the reviewer. The Chirasani et al. work was discussed on lines 870-881. The paper concentrates on the interaction between the EF-hand region and the S23 segment and its effect on RyR inactivation, which we referenced in the manuscript, but not on the allosteric pathways as mentioned by the reviewer. To broaden the consideration of this important work, we have introduced a more detailed discussion of Chirasani et al. (2024 ) by adding the following text to the manuscript: Lines 881-888: "Based on their structural analysis of the open RyR1 structure 5tal, Chirasani et al. (2024 ) proposed that narrowing the gap between the EF-hand domain and S23 loop, resulting in H-bonding interactions between the EF-hand residue K4101 and the S23 loop residue D4730, and those between the EF-hand residues E4075, Q4076, D4079 and the S23 loop residue R4736, is a consequence of the binding of Ca2+ to the EF-hands. However, our PDBePISA analysis revealed a similar number of interactions between the EF-hand region and the S23 loop not only in open and inactivated but also in primed RyR1 structures (Figure 3). The presence of EF hand-S23 hydrogen bonds in the primed and open RyR1 structures suggests that the proximity of the EF-hand domain and S23 loop is a structural trait distinguishing RyR1 from RyR2, not a consequence of Ca2+ binding to the EF hand.*"

The data and ideas of the illuminating work of Nayak et al. (2024) were discussed and referred to in the manuscript in several places, originally lines 74, 77, 164 (Introduction), 311, 340 (Results), 892-893, and 971 (Discussion). To broaden consideration of this work, we have expanded the discussion of this paper by adding the text shown in bold into the Introduction: "Recent studies reporting RyR structure at a high divalent ion concentration provide only indirect support for the molecular mechanism of Ca2+/Mg2+-dependent inactivation. Wei et al. (2016) and Nayak et al. (2024) observed a change in the conformation of the RyR1 EF-hands in the presence of 100 µM Ca2+ and 10 mM Mg2+, respectively, compared to low-calcium or low-magnesium conditions." (lines 135-138) and in the Discussion (lines 889-891): "The recent RyR1 structure 7umz (Nayak et al., 2024) provided evidence of Mg2+ ion bound in the RyR activation site, thus confirming the functional studies that established competition between Ca2+ and Mg2+ at this activation site (Laver et al., 1997; Zahradnikova et al., 2003; Zahradnikova et al., 2010)."

Reviewer 3:

Minor comment: While I am not an expert in allosteric model construction and therefore cannot fully assess their methodological approach, I observed that the authors fixed a number of parameters to achieve model convergence. A more detailed explanation of the rationale behind these fixed parameters would enhance clarity. Currently, these parameters are not clearly specified in the text and are somewhat obscured by the broader description of all parameters included in the model.

We thank the reviewer very much for this comment, which made us realize that the relevant sections were written in a too technical manner, without sufficient explanation of the ideas behind the derivation and optimization of the model. To clarify the rationale of this process, we have rewritten the subsection "Derivation of the model open probability equation" and the section "Description of RyR operation by the COI model". In the subsection "Derivation of the model open probability equation", we have explained the simplification of the full set of equations (Eqs. 3A-C) into Eqs. 4A-C (lines 642 - 666). In the section "Description of RyR operation by the COI model", we have explained the extent of over-parametrization and the rationale of reducing it by three methods: combining the data into groups with common parameter values; eliminating parameter interdependence by fixation of one parameter at a preset value taken from the literature or postulated a priori; and sharing parameter values between data groups when no significant difference between these values was observed (lines 683-685, 702-710, 719-740).

We hope that these changes make the manuscript more comprehensible.

REFERENCES

Chi, X., D. Gong, K. Ren, G. Zhou, G. Huang, J. Lei, Q. Zhou, and N. Yan. 2019. Molecular basis for allosteric regulation of the type 2 ryanodine receptor channel gating by key modulators. Proceedings of the National Academy of Sciences of the United States of America. 116:25575-25582.

Chirasani, V.R., M. Elferdink, M. Kral, J.S. Carter, S. Heitmann, G. Meissner, and N. Yamaguchi. 2024 Structural and functional interactions between the EF hand domain and S2-S3 loop in the type-1 ryanodine receptor ion channel. The Journal of biological chemistry. 300:105606.

Cholak, S., J.W. Saville, X. Zhu, A.M. Berezuk, K.S. Tuttle, O. Haji-Ghassemi, F.J. Alvarado, F. Van Petegem, and S. Subramaniam. 2023. Allosteric modulation of ryanodine receptor RyR1 by nucleotide derivatives. Structure. 31:790-800 e794.

Dashti, A., G. Mashayekhi, M. Shekhar, D. Ben Hail, S. Salah, P. Schwander, A. des Georges, A. Singharoy, J. Frank, and A. Ourmazd. 2020. Retrieving functional pathways of biomolecules from single-particle snapshots. Nature communications. 11:4734.

Gomez, A.C., T.W. Holford, and N. Yamaguchi. 2016. Malignant hyperthermia-associated mutations in the S2-S3 cytoplasmic loop of type 1 ryanodine receptor calcium channel impair calcium-dependent inactivation. American journal of physiology. 311:C749-C757.

Hadiatullah, H., Z. He, and Z. Yuchi. 2022. Structural Insight Into Ryanodine Receptor Channelopathies. Frontiers in pharmacology. 13:897494.

Laver, D.R., T.M. Baynes, and A.F. Dulhunty. 1997. Magnesium inhibition of ryanodine-receptor calcium channels: Evidence for two independent mechanisms. J.Membrane.Biol. 156:213-229.

Lin, Y.F., C.W. Cheng, C.S. Shih, J.K. Hwang, C.S. Yu, and C.H. Lu. 2016. MIB: Metal Ion-Binding Site Prediction and Docking Server. Journal of chemical information and modeling. 56:2287-2291.

Lu, C.H., C.C. Chen, C.S. Yu, Y.Y. Liu, J.J. Liu, S.T. Wei, and Y.F. Lin. 2022. MIB2: metal ion-binding site prediction and modeling server. Bioinformatics. 38:4428-4429.

Lu, C.H., Y.F. Lin, J.J. Lin, and C.S. Yu. 2012. Prediction of metal ion-binding sites in proteins using the fragment transformation method. PLoS One. 7:e39252.

Mackrill, J.J. 2022. Evolution of the cardiac dyad. Philosophical transactions of the Royal Society of London. Series B, Biological sciences. 377:20210329.

Melville, Z., K. Kim, O.B. Clarke, and A.R. Marks. 2022. High-resolution structure of the membrane-embedded skeletal muscle ryanodine receptor. Structure. 30:172-180 e173.

Nayak, A.R., W. Rangubpit, A.H. Will, Y. Hu, P. Castro-Hartmann, J.J. Lobo, K. Dryden, G.D. Lamb, P. Sompornpisut, and M. Samso. 2024. Interplay between Mg(2+) and Ca(2+) at multiple sites of the ryanodine receptor. Nature communications. 15:4115.

Nayak, A.R., and M. Samso. 2022. Ca(2+) inactivation of the mammalian ryanodine receptor type 1 in a lipidic environment revealed by cryo-EM. eLife. 11.

Rossi, D., A.M. Scarcella, E. Liguori, S. Lorenzini, E. Pierantozzi, C. Kutchukian, V. Jacquemond, M. Messa, P. De Camilli, and V. Sorrentino. 2019. Molecular determinants of homo- and heteromeric interactions of Junctophilin-1 at triads in adult skeletal muscle fibers. Proceedings of the National Academy of Sciences of the United States of America. 116:15716-15724.

Samso, M. 2017. A guide to the 3D structure of the ryanodine receptor type 1 by cryoEM. Protein science : a publication of the Protein Society. 26:52-68.

Wang, J., A. Jain, L.R. McDonald, C. Gambogi, A.L. Lee, and N.V. Dokholyan. 2020. Mapping allosteric communications within individual proteins. Nature communications. 11:3862.

Wei, R., X. Wang, Y. Zhang, S. Mukherjee, L. Zhang, Q. Chen, X. Huang, S. Jing, C. Liu, S. Li, G. Wang, Y. Xu, S. Zhu, A.J. Williams, F. Sun, and C.C. Yin. 2016. Structural insights into Ca(2+)-activated long-range allosteric channel gating of RyR1. Cell research. 26:977-994.

Zahradnikova, A., M. Dura, I. Gyorke, A.L. Escobar, I. Zahradnik, and S. Gyorke. 2003. Regulation of dynamic behavior of cardiac ryanodine receptor by Mg2+ under simulated physiological conditions. American journal of physiology. 285:C1059-1070.

Zahradnikova, A., I. Valent, and I. Zahradnik. 2010. Frequency and release flux of calcium sparks in rat cardiac myocytes: a relation to RYR gating. The Journal of general physiology. 136:101-116.

Zheng, W., and H. Wen. 2020. Investigating dual Ca(2+) modulation of the ryanodine receptor 1 by molecular dynamics simulation. Proteins. 88:1528-1539.

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

Reply to the Reviewers

We sincerely thank the reviewers for their comprehensive and constructive feedback.

Reviewer #1

Major comments:

1. The data and key conclusions of the paper are convincing. However, the reliability of the findings in terms of the new interaction could be improved by not relying solely on proximity ligation approaches (BioID, PLA), but employing a complementary biochemical strategy. The authors state that an immunoprecipitation (IP) was not possible due to a lack of antibodies for IP. This does not seem convincing since in the paper Saito-Diaz et al which they cite commercial antibodies were used to immunoprecipitate APC. Alternatively, the cell line expressing tagged ROBO1 could be used together with endogenous or tagged APC for an biochemical interaction experiment.

Response

We thank the reviewer for this important suggestion. In our initial studies, we attempted co-immunoprecipitation (co-IP) experiments using several different antibodies directed to APC. The signal detected was very low, possibly reflecting relatively low endogenous expression of ROBO1 in COS-7 cells, technical challenges associated with co-IP of APC and ROBO1, which are both large proteins (>200 kDa), and/or transient interactions between the two proteins. As part of the revision plan we will carry out co-IP experiments using HEK293A cells stably expressing full length ROBO1 (5H9 cells).

Regarding the PLA experiment, I was very surprised by the very strong labeling for Clathrin+ROBO1 shown in the representative image. It is hard to believe that this image is representative when the average number of dots in the quantification is about 100. From the image it is also hard to see how it would be possible to quantify individual dots. For this, a zoom would be helpful.

Response

We thank the reviewer for this helpful comment. In the revised manuscript, we have added a magnified panel to Figure 4E.

Clathrin and ROBO1 are likely not even direct interactors but come together by their common interaction with AP2. Therefore, to back this surprisingly strong result up, I would recommend to include one more control such as another rabbit antibody recognizing a protein that does not associate with clathrin or use e.g. the ROBO1 wildtype vs the ROBO1 mutant, that does not bind AP2 and therefore should also not associate with clathrin, for the experiment. Even better, the authors could confirm the PLA results by the mentioned complementary biochemical experiments to bolster the findings by an independent approach.

Response

We thank the reviewer for this suggestion. As recommended, we will use the complementary biochemical approaches suggested, and will perform immunoprecipitation experiments to examine interactions with clathrin in cells that express wildtype ROBO1 vs. cells that express mutant ROBO1 that does not bind AP2. As recommended, we will further perform experiments using control antibody directed to a protein that does not associate with clathrin.

Minor comments:

In general, data and methods are presented in a manner that should make them reproducible by others. Some small things to improve are:

1. In the paragraph on antibodies the used concentrations for the different applications should be provided.

Response

We thank the reviewer for this suggestion and apologize for the omission. In the revised manuscript, we have added a supplementary table to clarify the concentrations of antibodies used for different experimental applications. Please see Table s1.

It should be described how the poly-D-lysine coating was exactly performed.

Response

We thank the reviewer for this comment. In the revised manuscript, we have added the procedure for poly-D-lysine coating in the "Materials and Methods" section. Please see page 7 line 143-144.

The statistical analysis looks adequate. There are just some minor things that should be specified:- Just to make sure: Is it really always SD which is provided and not SEM? Sometimes the error bars look so small that I was wondering about this.

Response

We appreciate the opportunity to clarify that we used SD consistently in the manuscript.

• It should be specified for each experiment which post-hoc test is used or stated that one is always used for the One-Way ANOVA and the other for the Two-Way ANOVA resp. a rationale should be provided why two different post-hoc tests are used.

Response

We have added the post hoc tests used for each assay in the figure legend. The rationale for the different post hoc tests used has also been added in the "Materials and Methods" section as "Two-tailed paired Student's t-test was used for two-group comparisons. One-way ANOVA followed by Tukey's post hoc multiple comparison test was used for multiple-group comparisons with a single independent variable, and two-way ANOVA followed by Sidak's post hoc multiple comparison test was used for multiple-group comparisons with two independent variables". Please see page 12 line 275-279, page 20 line 519-520, page 21 line 524-525, 533-534, 537-538, 540-541, 543-544, page 22 line 551-552, 555-556, 561-562, 576-577, page 23 line 582, 584-585, 587, 589-590, 595, 599, page 24 line 624-625, 629, 635-636, page 25 638-639.

• When using the t-test, it should be stated whether it is paired or unpaired and one- or two-tailed.

Response

Two-tailed paired Student's t-test was used in Fig. 5C. We have added in the "Materials and Methods" section and figure legend in the revised manuscript. Please see page 12 line 275-276, page 23 line 587.

• It should be stated whether it was tested that the data fulfill the requirements for parametric tests (normal distribution).

Response

We have added "The data fulfilled the requirements for normal distribution using the Shapiro-Wilk test" in the "Materials and Methods" section in the revised manuscript. Please see page 12 line 274-275 in the revised manuscript.

Text and figures are mostly clear, apart from some small things:

• I was wondering about figure 1B. If I understand the methods description right, all cells were permeabilized prior to secondary antibody application. Why then is so little fluorescence for Flag visible in the first PBS row at 30 min? That would only make sense for me if the cell was not permeabilized and the protein internalized. So where did the majority of the protein end up after 30 min since you should see the entire population in a permeabilized cell? Could you please comment on this?

Response

We thank the reviewer for this comment. The cells were permeabilized prior to secondary antibody application. Since NSLIT2 binding to ROBO1 can facilitate ADAM10-mediated ROBO1 cleavage to release the extracellular domain of ROBO1 (Coleman et al., 2010), this may have caused little fluorescence for Flag to be visible in the first PBS row at 30 min. In the revised manuscript we have added a comment about the finding described. Please see page 13 line 293-296.

• Fig. 2A the upper left image (0 min PBS) should be very similar to the upper left image in Fig. 1B, shouldn´t it? But it looks quite different to me in terms of surface amount of ROBO1-Flag. Could you please comment on this?

Response

We apologize for the confusing images included in the original version of the manuscript. As noted, the upper left image (0 min PBS) in Fig. 2A should be very similar to the upper left image in Fig. 1B. We have now instead included an image for Fig. 2A that is more representative of the data from the experiments we performed.

• Please explain what the molecular difference between bio-active NSLIT2 and bio-inactive CSLIT2 is. Please provide a rationale why you sometimes use CSLIT2 as negative control and sometimes DD2SLIT2. In Fig. 3G you are using DD2SLIT2. Even though there is no significance reached with the analyzed n, it is very striking that the bars are consistently higher upon DD2SLIT2 application. Can you comment on this effect? Or am I misunderstanding the labeling of the figure?

Response

Bio-active NSLIT2 consists of the N-terminal fragment of SLIT2 and contains the second leucine-rich repeat (LRR) domain (D2), which binds to the first two Ig domains of the ROBO1 receptor (Ig1-2). Bio-inactive CSLIT2 consists of the C-terminal fragment of SLIT2, which does not bind ROBO1. DD2SLIT2 consists of the N-terminal fragment of SLIT2 but lacks D2 LRR domain that is essential for ROBO1 binding. Neither CSLIT2 nor DD2SLIT2 can bind the ROBO1 receptor (Bhosle et al., 2020; Mukovozov et al., 2015; Patel et al., 2012). In Fig. 3G, DD2SLIT2 was used as negative control and did not affect cell spreading, so the bars are consistently higher upon D2SLIT2 application. The use of CSLIT2 or DD2SLIT2 in different experiments was due to the availability of these reagents. In Fig. 3F and 3G, we have made modifications to the X axis to clarify.

• On page 3 it states "...endocytosis of ROBO1...requires...APC": I found this confusing since it is the dissociation of APC that is required for promoting endocytosis. Therefore, it would be good to rephrase this sentence.

Response

We apologize for the confusing language. In the revised manuscript, we have changed "endocytosis of ROBO1 from the cell surface requires the tumor suppressor protein, APC" to "endocytosis of ROBO1 from the cell surface requires the dissociation of the tumor suppressor protein, APC". Please see page 4 line 35-36.

• On page 8 is written "...cells surface ROBO1 [is] removed". Please be more accurate since the acid wash does not remove ROBO1, but only the antibody bound to the extracellular epitope.

Response

We apologize for the confusing language. In the revised manuscript, we have changed "cell surface ROBO1 removed" to "anti-Flag antibody binding ROBO1 removed from the cell surface". Please see page 8 line 153-154.

• On page 8 provide an explanation for the abbreviation HAC.

Response

To enhance clarity, in the revised manuscript we have used the full name "acetic acid" instead of using the abbreviation "HAC". Please see page 8 line 155.

• On page 15 you speak of "mutant AP2". Please be more accurate since there is no mutant AP2 involved, but you are refering to ROBO1 with mutations in its AP2 binding motifs.
• On page 14 you speak of "cells expressing the mutant alleles of AP2". As above, please be more accurate and replace with "cells expressing ROBO1 harboring mutations in both AP2 binding sites".

Response

We thank the reviewer for this suggestion and apologize for the confusion. For the sake of accuracy, we have made the changes as suggested by the reviewer. Please see page 15 line 351 and page 14 line 331-332.

• On page 19 you write: "Using proximity ligation assays, we observed that ROBO1, APC and clathrin interact with one another". I am maybe a bit picky here, but in my eyes with these assays you only show that they are very close together and might be in a complex, but you do not show (direct) interaction in a strict sense. Therefore, I would downtone this a bit.

Response

We thank the reviewer for this important comment. As suggested, in the revised manuscript, we replaced "Using proximity ligation assays, we observed that ROBO1, APC and clathrin interact with one another" with "Using proximity ligation assays, we observed that ROBO1, APC and clathrin are in close proximity to one another". Please see page 18 line 458. We have similarly amended the language throughout the manuscript. Please see page 3 line 11-12, page 4 line 37, page 16 line 391, 394, 396, 398, page 22 line 564.

• In Fig. 5B I would find it easier for the reader if siRNA and control were shown side by side for the different conditions.

Response

In the revised manuscript, we have made the changes suggested by the reviewer to enhance clarity.

• Between the internalization assays and the spreading assays, you switch from HEK293 cells to COS7 cells. Please provide a rationale for this for the reader.

Response

Because the endogenous expression of ROBO1 is relatively low in COS-7 cells, we generated a HEK293A cell line that stably expresses ROBO1, and used these cells to examine subcellular traffic of ROBO1 and explore interactors of ROBO1. We next sought to explore the functional consequences of internalization of ROBO1 and the functional role of APC. As we and others previously showed that SLIT2-ROBO1 signaling inhibits cell spreading (Bhosle et al., 2020; Patel et al., 2012; Tole et al., 2009), we elected to use this measure as a biologic read-out. Because HEK293A cells do not spread as much as COS-7 cells, we instead used COS-7 cells for the spreading assays.

• You provide a table with putative interactors within the paper and as supplementary table. Could you please explain better to the reader what your criteria were for including hits into the "short-list" presented in Table1.

Response

We chose proteins based on two criteria. The first was association with full-length ROBO1, but not with ROBO1 lacking the intracellular domain. The second was association with full-length ROBO1 under basal conditions, but loss of association with full-length ROBO1 after exposure of cells to NSLIT2. In the revised manuscript, we have added the criteria in the manuscript. Please see page 15 line 363-366.

Typos - p. 6: CO2 instead of CO2

• p21 last line: Immunoblotting should not be capitzalized.

• Figure s1 legend: full-lenth is missing a g

Response

We apologize for the oversight. In the revised manuscript, we have corrected these typos. Please see page 6 line 97, page 21 line 535 and page 24 line 611.

Significance

It was already known from Drosophila and for mammalian cells that SLIT2 induces the endocytosis of ROBO1 and that this is necessary for its repulsive function in axon guidance as the authors point out. The key advance of the study is the identification of APC as an interactor of ROBO1 which decreases its endocytosis until it dissociates upon SLIT2 binding to ROBO1. This is an interesting aspect which opens up parallels to the regulation of Wnt signaling by APC as the authors discuss. The significance of this finding would be even greater if it would have been shown that this mechanism actually operates in axon guidance. That not being the case, the authours might want to discuss in more detail if APC has previously been implicated to affect axon guidance.

Researchers working on endocytosis, adhesion, cellular signaling and the development of the nervous system will be interested in these findings.

Response

We thank the reviewer for the positive comments regarding the significance of our findings. As recommended, in the discussion section of the revised manuscript we will discuss in more detail what is known about the role of APC in axon guidance.

Reviewer #2

Major comments:

1. As the authors emphasize the role of NSlit2 in Robo1 internalization throughout their manuscript, I suggest authors include "NSlit" in their title. Something like this "Adenomatous polyposis coli (APC) regulates the NSlit2-induced internalization and signaling of the chemo repellent receptor, hRoundabout (ROBO) 1" or maybe a better title.

Response

As suggested, we have changed the title of the revised manuscript to "Adenomatous polyposis coli (APC) regulates the NSLIT2-induced internalization and signaling of the chemorepellent receptor, Roundabout (ROBO) 1".

In addition to transferrin as the control for their internalization studies, have the authors tested the specificity of NSlit-2-induced internalization with other Robo receptors such as Robo2? Does the APC bind to Robo2 also?

Response

We thank the reviewer for this comment. Due to significant cost constraints, we focused our BioID experiments on identifying proteins that interact with ROBO1. In the revised manuscript, we will expand the discussion to consider the questions raised here by the reviewer.

The N-Slit group at 0' in Figure 1 b and Figure 2a, the Flag-Robo staining looks very different. Is it because the authors did not use ADAM protease inhibitor in Figure 2a that's why they are seeing more internalized Flag-Robo at 0'? It is not very clear either in the Results or the legend.

Response

We apologize for the confusing images. We used ADAM protease inhibitor for all endocytosis assays, as mentioned in the "Materials and Methods" section. The upper left image (0 min PBS) in Fig. 2A should be very similar to the upper left image in Fig. 1B. We have now replaced the image in 2A with one that is more representative of the overall results.

Have the authors tested the Surface Robo1 pool in siAPC cells induced with or without N-Slit2?

Response

We added NSLIT2 to cells as we started endocytosis assay. At the time point of 0 min, the surface ROBO1 pool was not affeacted by NSLIT2.

Does the Robo1 mutated with AP2 binding motifs interact with APC? Have authors performed a Proximity ligation assay with AP2-binding motifs mutated Robo1 and APC?

Response

We thank the reviewer for this suggestion. As recommended, we will perform proximity ligation assays to examine interactions between APC and ROBO1 which lacks AP2-binding motifs.

The resolution of PLA dots in the current version is very low. Authors should include higher magnification pictures for these interactions and also PLA dots channel should be separately represented in addition to the DAPI merged images for better clarity and interpretation.

Response

We thank the reviewer for these suggestions. In the revised manuscript, we have included figures with the recommended modifications to enhance clarity. Please see figure 4A, 4C and 4E.

Do the Slit2 treated cells affect APC mRNA expression? Or does Slit2 only inhibit the interaction between APC and Robo1? Have the authors tested the mRNA expression of APC in slit2-treated and untreated cells?

Response

We thank the reviewer for this question. We will perform the experiments suggested and include the results in the revised manuscript.

The authors have tested the effect of Slit2-induced inhibition of cell spreading under different experimental conditions however it is also important to test the cell migration/proliferation rates under control and siAPC conditions with or without Slit2 treatment.

Response

We thank the reviewer for this comment. In order to test the effect of APC on SLIT2-induced cell migration, a migratory cell type would be required. This would involve introducing a third cell type in addition to the HEK293 and COS-7 cells we have already used, and first validating our key experimental findings in the new cell type. Please see our response to the 10th sub-comment in Minor Comment 4) of Reviewer 1.

Do authors see the inhibition of Robo1 and Cyfip interactions also in the presence of Slit2 by PLA assay?

Response

We thank the reviewer for this interesting question. As this was beyond the scope of the current study, we did not examine whether SLIT2 inhibits interactions between ROBO1 and CYFIP. In the Discussion section of the revised manuscript, we will address this question as a potential line of future investigation.

Studying the endogenous Robo1 and APC interaction by PLA is good but I suggest authors do standard co-IP assays to visualize these interactions since authors have already generated a variety of general epitope- tagged constructs for both Robo1 and APC. These epitope-specific antibodies that are best suitable for IP are easily available with many antibody companies. This is the first study to suggest that the interaction between Robo1 and APC so the strong biochemistry would have a good impact on the findings.

Response

We appreciate this important suggestion. We will perform the recommended studies and include the results in the revised manuscript. Please also see our response to Reviewer 1, Major Comment 1).

Minor comments:

1. I suggest the authors show the single-channel images of Flag-robo (green) in Figure 2B for a clear visualization of internalized Robo in a cell. With DAPI-merged images, it is hard to specifically visualize Robo in these cells.

Response

We assume the reviewer was referring Figure 2A instead of 2B. To enhance clarity, in the revised manuscript we have made the changes suggested by the reviewer.

In Figure 1C, the Y axis should have a clear indication. Instead of "% internalized" it should be mentioned as "% Internalized Robo1".

Response

We thank the reviewer for this suggestion and apologize for the oversight. In the revised manuscript, we have made the suggested change in Figure 1C, 2B, 2D, 2E, 5B and 5D.

I suggest authors to include the simple schematic of the mechanism they are proposing in the manuscript.

Response

We thank the reviewer for the suggestion. To enhance clarity, in the revised manuscript we will include a simple schematic of the mechanism our findings suggest.

The authors should mention the rationale or the function of using the acid wash method for their experimental conditions for a better understanding of the reader.

Response

We thank the reviewer for this suggestion and apologize for the oversight. We performed acid wash experiments to remove the anti-Flag antibody that binds ROBO1 from the cell surface for the endocytosis assay. To increase the clarity, in the "Materials and Methods" section of the revised manuscript we have included the rationale for using acid wash. Please see page 8 line 153-154.

siRNA-mediated knockdown of specific genes should be correctly denoted in the figure. For example, instead of "CLTC", it should be "siCLTC" for easy understanding. The same correction has to be done in all the figures with siRNA data.

Response

We thank the reviewer for this helpful comment and apologize for the oversight. As suggested, we have made the suggested changes throughout the revised manuscript and in Figure 2C, 2D, 5A, 5B, 6A, 6B, 6C, s2C and s2D.

Reference

Bhosle, V.K., Mukherjee, T., Huang, Y.W., Patel, S., Pang, B.W.F., Liu, G.Y., Glogauer, M., Wu, J.Y., Philpott, D.J., Grinstein, S., et al. (2020). SLIT2/ROBO1-signaling inhibits macropinocytosis by opposing cortical cytoskeletal remodeling. Nat Commun 11, 4112.

Coleman, H.A., Labrador, J.P., Chance, R.K., and Bashaw, G.J. (2010). The Adam family metalloprotease Kuzbanian regulates the cleavage of the roundabout receptor to control axon repulsion at the midline. Development (Cambridge, England) 137, 2417-2426.

Mukovozov, I., Huang, Y.W., Zhang, Q., Liu, G.Y., Siu, A., Sokolskyy, Y., Patel, S., Hyduk, S.J., Kutryk, M.J., Cybulsky, M.I., et al. (2015). The Neurorepellent Slit2 Inhibits Postadhesion Stabilization of Monocytes Tethered to Vascular Endothelial Cells. J Immunol 195, 3334-3344.

Patel, S., Huang, Y.W., Reheman, A., Pluthero, F.G., Chaturvedi, S., Mukovozov, I.M., Tole, S., Liu, G.Y., Li, L., Durocher, Y., et al. (2012). The cell motility modulator Slit2 is a potent inhibitor of platelet function. Circulation 126, 1385-1395.

Tole, S., Mukovozov, I.M., Huang, Y.W., Magalhaes, M.A., Yan, M., Crow, M.R., Liu, G.Y., Sun, C.X., Durocher, Y., Glogauer, M., et al. (2009). The axonal repellent, Slit2, inhibits directional migration of circulating neutrophils. Journal of leukocyte biology 86, 1403-1415.

Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

Learn more at Review Commons

Referee #2

Evidence, reproducibility and clarity

Summary:

In this paper, the authors sought to identify the mechanism of Slit2-induced hRobo1 internalization and its signaling. They demonstrated that Slit2-induced hRobo1 internalization is regulated by one of the Robo1 C-terminal binding partners, adenomatous polyposis coli (APC). By using various in vitro experiments, the authors have concluded that APC constitutively interacts with hRobo1, and this interaction is disrupted upon the binding of Slit2 to the extracellular domain of hRobo1. They also showed that the dissociation of interaction between APC and hRobo1 is important for clathrin-mediated endocytosis of hRobo1 and subsequent cell morphology. In conclusion, while this study presents intriguing findings, there are notable experimental concerns. In several instances, the authors fail to sufficiently elucidate the experimental setup or provide specific conditions for certain experiments, which may pose challenges for readers in understanding the methodology thoroughly. Also, the labels for the figures can be more accurate and clearly stated.

Major comments:

1. As the authors emphasize the role of NSlit2 in Robo1 internalization throughout their manuscript, I suggest authors include "NSlit" in their title. Something like this "Adenomatous polyposis coli (APC) regulates the NSlit2- induced internalization and signaling of the chemo repellent receptor, hRoundabout (ROBO) 1" or maybe a better title.
2. In addition to transferrin as the control for their internalization studies, have the authors tested the specificity of NSlit-2-induced internalization with other Robo receptors such as Robo2? Does the APC bind to Robo2 also?
3. The N-Slit group at 0' in Figure 1 b and Figure 2a, the Flag-Robo staining looks very different. Is it because the authors did not use ADAM protease inhibitor in Figure 2a that's why they are seeing more internalized Flag-Robo at 0'? It is not very clear either in the Results or the legend.
4. Have the authors tested the Surface Robo1 pool in siAPC cells induced with or without N-Slit2?
5. Does the Robo1 mutated with AP2 binding motifs interact with APC? Have authors performed a Proximity ligation assay with AP2-binding motifs mutated Robo1 and APC?
6. The resolution of PLA dots in the current version is very low. Authors should include higher magnification pictures for these interactions and also PLA dots channel should be separately represented in addition to the DAPI merged images for better clarity and interpretation.
7. Do the Slit2 treated cells affect APC mRNA expression? Or does Slit2 only inhibit the interaction between APC and Robo1? Have the authors tested the mRNA expression of APC in slit2-treated and untreated cells?
8. The authors have tested the effect of Slit2-induced inhibition of cell spreading under different experimental conditions however it is also important to test the cell migration/proliferation rates under control and siAPC conditions with or without Slit2 treatment.
9. Do authors see the inhibition of Robo1 and Cyfip interactions also in the presence of Slit2 by PLA assay?
10. Studying the endogenous Robo1 and APC interaction by PLA is good but I suggest authors do standard co-IP assays to visualize these interactions since authors have already generated a variety of general epitope- tagged constructs for both Robo1 and APC. These epitope-specific antibodies that are best suitable for IP are easily available with many antibody companies. This is the first study to suggest that the interaction between Robo1 and APC so the strong biochemistry would have a good impact on the findings.

Minor comments:

1. I suggest the authors show the single-channel images of Flag-robo (green) in Figure 2B for a clear visualization of internalized Robo in a cell. With DAPI-merged images, it is hard to specifically visualize Robo in these cells.
2. In Figure 1C, the Y axis should have a clear indication. Instead of "% internalized" it should be mentioned as "% Internalized Robo1".
3. I suggest authors to include the simple schematic of the mechanism they are proposing in the manuscript.
4. The authors should mention the rationale or the function of using the acid wash method for their experimental conditions for a better understanding of the reader.
5. siRNA-mediated knockdown of specific genes should be correctly denoted in the figure. For example, instead of "CLTC", it should be "siCLTC" for easy understanding. The same correction has to be done in all the figures with siRNA data.

Referees cross-commenting

Reviewer 1 comments and suggestions are valid and carry significant weight in improving the manuscript.

Significance

Strengths: The manuscript writing is good and the authors have generated a lot of constructs for a thorough understanding of Robo1 internalization events under different conditions. Studying the differential protein interactions with and without Slit2 with the Robo1-Bir*Flag method is convincing.

Limitations: The representation of figures and their labels, Figure resolution, poor quality, missing important controls and experiments.

Advance: Not very conceptual

Audience: Broad and basic research

My field of expertise: Endocytosis, receptor surface labeling studies, ligand mediated receptor signaling and its effect on axon guidance during embryonic development.

Author response:

The following is the authors’ response to the original reviews.

Public reviews:

Reviewer #1 (public review):

(1) The link between the background in the introduction and the actual study and findings is often tenuous or not clearly explained. A re-working of the intro to better set up and link to the study questions would be beneficial.

We have rewritten the introduction of the manuscript and clearly stated the study questions we were aiming for:

In paragraph 1-we have stated clearly that we need to study why ADC type of cervical cancer is more aggressive. (Line 58 - 77)

In paragraph 2- we have stated clearly that we need to find valuable biomarkers to help diagnose lymph node metastasis, which may compensate the shortage of radiological imaging tools and reduce the rate of misdiagnosis. (Line 78 - 100)

In paragraph 3- we have stated clearly that HPV negative cases is a special group of cervical cancer and we aim to study its cellular features. (Line 101 - 108)

In paragraph 4- we have stated clearly that we need to decode cell-to-cell interaction mode in the tumor immune microenvironment of ADC using scRNA-seq. (Line 109 - 123)

(2) For the sequencing, which kit was used on the Novaseq6000?

For sequencing, we used the Chromium Controller and Chromium Single Cell 3’Reagent Kits (v3 chemistry CG000183) on the Novaseq6000. We feel sorry for lacking this quite important part and have already add the information in Methods section. (Line 196- 197)

(3) Additional details are needed for the analysis pipeline. How were batch effects identified/dealt with, what were the precise functions and settings for each step of the analysis, how was clustering performed and how were clusters validated etc. Currently, all that is given is software and sometimes function names which are entirely inadequate to be able to assess the validity of the analysis pipeline. This could alternatively be answered by providing annotated copies of the scripts used for analysis as a supplement.

We apologize for the inadequacy of descriptions of data analysis process. We have already provided a new part of “data processing” with more details in the Methods section (Line 202 - 221). In addition, we have also provided annotated copies of scripts in the supplementary data as Supplementary Data 1.

(4) For Cell type annotation, please provide the complete list of "selected gene markers" that were used for annotation.

We have already added the list of marker genes for cell type annotation in the revised manuscript as Supplementary Table 3.

(5) No statistics are given for the claims on cell proportion differences throughout the paper (for cell types early, epithelial sub-clusters later, and immune cell subsets further on). This should be a multivariate analysis to account for ADC/SCC, HPV+/- and Early/Late stage.

We feel sorry for lacking statistics when performing analyses of comparisons. In the revision, we have already used statistic approaches to analyze the differences between each set of group comparison. As a result, the corresponding figures have been revised, accordingly.

For examle, Fig. 1F, Fig. 2D, Fig. 4E, Fig. 5D, Fig. 6D had been re-analyzed to compare ADC/SCC；Supplementary Fig. 1A, Supplementary Fig. 2A, Supplementary Fig. 4A, Supplementary Fig. 5A, Supplementary Fig. 6A had been re-analyzed to compare HPV+/HPV-; Supplementary Fig. 1B, Supplementary Fig. 2B, Supplementary Fig. 4B, Supplementary Fig. 5B, Supplementary Fig. 6B had been re-analyzed to compare Early/Late stage. All P values have been listed in the figure legends.

(6) The Y-axis label is missing from the proportion histograms in Figure 2D. In these same panels, the bars change widths on the right side. If these are exclusively in ADC, show it with a 0 bar for SCC, not doubling the width which visually makes them appear more important by taking up more area on the plot.

We feel sorry for impreciseness when presenting histograms of Fig. 2D and we have also revised other figures with similar mistakes, such as Fig. 1F,  Fig. 5D. As for the width of bars, which is due to output style of data processing, we have already corrected all similar mistakes alongside the whole manuscript, for example, Fig. 2D and Supplementary Fig. 2A-B.

(7) Throughout the manuscript, informatic predictions (differentiation potential, malignancy score, stemness, and trajectory) are presented as though they're concrete facts rather than the predictions they are. Strong conclusions are drawn on the basis of these predictions which do not have adequate data to support. These conclusions which touch on essentially all of the major claims made in the manuscript would need functional data to validate, or the claims need to be very substantially softened as they lack concrete support. Indeed, the fact that most of the genes examined that were characteristic of a given cluster did not show the expected expression patterns in IHC highlights the fact that such predictions require validation to be able to draw proper inferences.

Thank you for your insightful comments. As you noted, several conclusions were initially based on bioinformatics predictions. Thus in the revised manuscript, we have rewritten all relevant descriptions in a more softened way, particularly in the paragraph of “epithelial cells” in Results section, as well as the conclusions derived from bioinformatics predictions in other paragraphs throughout the manuscript. We hope our revised descriptions will enhance the precision of our work.

For example, in paragraph “The sub-clusters of epithelial cells in ADC exhibit elevated stem-like features (from Line 353)”, many over-affirmative disriptions had been re-written in Line 353, 362, 371, 375, 379, 383, 390, 392. From Line 395 to 399, the conclusion had been revised as “The observation of cluster Epi_10_CYSTM1 and its possible specificity to ADC makes us question whether or not it may be related to the aggressiveness of ADC” compared to the previous “This observation may partially indicate that high stemness cluster Epi_10_CYSTM1 is essential for ADC to present more aggressive features”. From Line 400 to 408, conclusions from GO analyses had also been rewritten.

In paragraph “ADC-specific epithelial cluster-derived gene SLC26A3 is a potential prognostic marker for lymph node metastasis (from Line 422)”, many conclusions based on predictions had been revises, such as Line 424 - 428, Line 439 - 441, Line 451 - 453, Line 455 - 457, Line 458 - 459, Line 471 - 473, Line 478 - 481, Line 484 - 486, Line 489, etc.

In paragraph “Tumor associated neutrophils (TANs) surrounding ADC tumor area may contribute to the formation of a malignant microenvironment (from Line 536)”, we have changed the descriptions based on bio-infomative predictions, such as Line 560, Line 561, Line 565, Line 566, Line 572, Line 576 - 577, etc.

In paragraph “Crosstalk among tumor cells, Tregs and neutrophils establishes the immunosuppressive TIME in ADC (from Line 601)”, we have already corrected the all the affirmative descriptions, such as Line 604, Line 612, Line 614, Line 626, Line 628 - 629, Line 641, Line 654 – 655, etc.

All the changes have also been listed in Revision Notes in detail.

(8) The cluster Epi_10_CYSTM1 which is the basis for much of the paper is present in a single individual (with a single cell coming from another person), and heavily unconnected from the rest of the epithelial populations. If so much emphasis is placed on it, the existence of this cluster as a true subset of cells requires validation.

We appreciate this suggestion. We agree that the majority of Epi_10_CYSTM1 cells are derived from sample S7. The fact that we have detected this cluster in only one patient may be due to sampling differences and the inherent heterogeneity of tumor specimens. However, the relatively high number of cells in this cluster from one stage III patient suggests its presence in ADC patients and highlights its potential as a diagnostic marker for clinical staging. To further investigate whether this cluster is generally existing in ADC patients, we have identified and selected candidate genes, such as SLC26A3, ORM1, and ORM2, as representative markers of this cluster, which demonstrated high specificity (as shown in Fig. 3B). We then performed IHC staining on a total of 56 tissue samples, and the results showed positive expressions of these markers in the majority of stage IIIC tumor tissues, confirming the existence of this cell cluster (as shown in Supplementary Fig. 3E). In our revised manuscript, we have included an in-depth discussion of this issue in the seventh paragraph of the Discussion section (From Line 801).

(9) Claims based on survival analysis of TCGA for Epi_10_CYSTM1 are based on a non-significant p-value, though there is a slight trend in that direction.

Thank you for your insightful comment. From the data of TCGA survival analysis for Epi_10, we found a not-so-slight trend of difference between groups (with a small P value). As a result, we presented this data and hoped to add more strength to the clinical significance of this cluster. However, this indeed caused controversy because the P value is non-significant. As a result, we have already deleted this data in the revised manuscript.

(10) The claim "The identification of Epi_10_CYSTM1 as the only cell cluster found in patients with stage IIICp raises the possibility that this cluster may be a potential marker to diagnose patients with lymph node metastasis." This is incorrect according to the sample distributions which clearly show cells from the patient who has EPI_10_CYSTM1 in multiple other clusters. This is then used as justification for SLC26A3 which appears to be associated with associated with late stage, however, in the images SLC26A3 appears to be broadly expressed in later tumours rather than restricted to a minor subset as it should be if it were actually related to the EPI_10_CYSTM1 cluster.

We feel thankful for this question. The conclusion that “The identification of Epi_10_CYSTM1 as the only cell cluster found in patients with stage IIICp raises the possibility that this cluster may be a potential marker to diagnose patients with lymph node metastasis” has indeed been written too concrete according to the sample distribution. We feel sorry for this and have already corrected the description into “As one of stage IIIC-specific cell clusters, the cluster of Epi_10_CYSTM1, with its representative marker gene SLC26A3, presents potential diagnostic value to predict lymph node metastasis” from Line 478-481.

However, based on our results, we do think this cluster is a potential diagnostic marker and the hypothesis is right. As for SLC26A3, we have specifically added a new paragraph (from Line 801 - 822) in Discussion section to discuss the rationality and necessity of selecting this gene as our central focus, and the reasons why SLC26A3 should be the representative of cluster Epi_10_CYSTM1. As you noted, SLC26A3 appears to be broadly expressed in later tumors rather than restricted to a minor subset in the images. We apologize for any misunderstanding caused. When presenting the IHC data, we only showed the strongly positive areas of each slide to emphasize the differences. In our revision, we have included whole slide scanning images of the IHC samples, clearly showing that SLC26A3 is restricted to a part of the tumors (Supplementary Fig.9).

(11) The authors claim that cytotoxic T cells express KRT17, and KRT19. This likely represents a mis-clustering of epithelial cells.

We apologize for using data without noticing the contamination of T cells with few epithelial cells. We have re-performed quality control to exclude contamination and re-analyzed all data of T cells. In the reviesed manuscript, we have therefore updated completely new data for T cells in both Fig. 4 and Supplementary Fig. 4.

(12) Multiple claims are made for specific activities based on GO term biological process analysis which while not contradictory to the data, certainly are by no means the only explanation for it, nor directly supported.

Our initial purpose was to use GO analysis as supports for our conclusions. However, we know these are only claims but not evidence, which is also the problem of our writing techniques as in question (7). Therefore, in our revised manuscript, we have already deleted GO data and descriptions in the paragraphs of “T cell (Fig.4)”(from Line 495) and “B/plasma cell (Fig.6)” (from Line 579), because the predictions are quite irrelevant to our conclusions.

However, in the sections of “epithelial cell (Fig.2)” (from Line 352) and “neutrophils (Fig.5)” (from Line 536), we retained the GO data and rewrote the conclusions, because these analyses have provided us with valuable information regarding the role of specific cell clusters in ADC progression. Furthermore, our subsequent analyses, such as CellChat, have further validated the accuracy of the findings from the GO analysis. We do think this logically supports the whole storyline of the study.

Reviewer #2 (public review):

(1) I believe that many of the proposed conclusions are over-interpretations or unwarranted generalizations of the single-cell analysis. These conclusions are often based on populations in the scRNA-seq data that are described as enriched or specific to a given group of samples (eg. ADC). This conclusion is based on the percentage of cells in that population belonging to the given group; for example, a cluster of cells that dominantly come from ADC. The data includes multiple samples for each group, but statistical approaches are never used to demonstrate the reproducibility of these claims.

We feel sorry that many of the conclusions have been written in an over-affirmative way but lack profound supporting evidences. In our revision, we have already optimized the writing techniques and re-written all conclusions or descriptions related to only bio-informatic predictions. Moreover, we have performed statistical re-analyses on all data and rearranged the related figures.

For example, in Line 352, we have changed the sub-title “The sub-clusters of epithelial cells exhibit elevated stem-like features to promote the aggressiveness of ADC” into “The sub-clusters of epithelial cells in ADC exhibit elevated stem-like features”. In this paragraph, many over-affirmative discriptions such as “exclusively”, “significant”, “overwhelmingly”, “remarkably” have been deleted. From Line 486-493, the conclusion of “Moreover, SLC26A3 could be employed as a marker for the Epi_10_CYSTM1 cluster, aiding in the diagnosis of lymph node metastasis to prevent post-surgical upstaging in ADC patients in the future” have been changed into “our results propose that SLC26A3 might be considered as a diagnostic marker to predict lymph node metastasis in ADC patients”. Similar over-affirmative descriptions and conclusions had also been re-written in the other paragraphs, which has been refered to question (7) above.

(2) This leads to problematic conclusions. For example, the "ADC-specific" Epi_10_CYSTM1 cluster, which is a central focus of the paper, only contains cells from one of the 11 ADC samples and represents only a small fraction of the malignant cells from that sample (Sample 7, Figure 2A). Yet, this population is used to derive SLC26A3 as a potential biomarker. SLC26A3 transcripts were only detected in this small population of cells (none of the other ADC samples), which makes me question the specificity of the IHC staining on the validation cohort.

We sincerely feel grateful for this question. This is a quite important question as it is also pointed out by reviewer#1 in question (8) above. In the revised manuscript, we have already optimized our descriptions and have added detailed explanation for the importance of SLC26A3 in the Discussion section  (from Line 802 - 823). We agree that the majority of Epi_10_CYSTM1 cells are derived from sample S7. The fact that we detected this cluster in only one patient may be due to sampling differences and the inherent heterogeneity of tumor specimens. However, the relatively high number of cells in this cluster from one stage III patient suggests its presence in ADC and highlights its potential as a diagnostic marker for staging ADC. To further investigate whether this cluster is generally present in ADC patients, we identified and selected candidate genes, such as SLC26A3, ORM1, and ORM2, as representative markers of this cluster, which demonstrated high specificity (as shown in Fig. 3B). We then performed IHC staining on 56 cases of tissue samples, and the results showed positive expression of these markers in the majority of stage III tumor tissues, confirming the existence of this cell cluster (as shown in Supplementary Fig. 3E). In our revised manuscript, we have included an in-depth discussion of this issue in the seventh paragraph of the Discussion section.

(3) This is compounded by technical aspects of the analysis that hinder interpretation. For example, it is clear that the clustering does not perfectly segregate cell types. In Figures 2B and D, it is evident that C4 and C5 contain mixtures of cell type (eg. half of C4 is EPCAM+/CD3-, the other half EPCAM-/CD3+). These contaminations are carried forward into subclustering and are not addressed. Rather, it is claimed that there is a T cell population that is CD3- and EPCAM+, which does not seem likely.

Thank you for your insightful comment. This important point is also raised by reviewer#1 above. In the revised manuscript, we have reanalyzed our scRNA-seq data and listed the canonical marker genes for cell type annotation. Most importantly, as for T cells and its sub-clustering, we have performed quality control and re-analyzed all data for T cells, with contamination excluded. In the reviesed manuscript, we have added the re-analyzed data for T cells in both Fig. 4 and Supplementary Fig. 4.

Recommendations for the authors:

Reviewer #1 (recommendations for the authors):

The text would substantially benefit from an editorial revision of language usage.

We sincerely feel grateful for this suggestion. In our revision, we have conducted language editing and carefully rewritten our manuscript. The changes have been clearly marked in the tracked version of the revised manuscript.

Reviewer #2 (recommendations for the authors):

(1) Use statistical approaches to claim enrichment/specificity of populations to given groups (ADC, HPV, etc). Analysis packages like Milo for differential abundance testing would be very helpful.

We feel grateful for this suggestion. In our revision, we have performed statistical analyses for all groups of comparison data. Meanwhile, we have rearranged the figures based on these statistical results, for example, Fig. 1F, Fig. 2D, Fig. 4E, Fig. 5D, Fig. 6D, Supplementary Fig. 1A-B, Supplementary Fig. 2A-B, Supplementary Fig. 4A-B, Supplementary Fig. 5A-B, Supplementary Fig. 6A-B.

(2) In the subclustering, consider a round of quality control to ensure that all cells are of the cell type they are claimed to be. Contaminant clusters/cells could be filtered out or reassigned. This could be supplemented with an automated annotation approach using cell-type references.

We feel thankful for this suggestion. As a result, we have provided copies of scripts in the supplementary data to ensure the quality control of cell type annotation.

(3) An explanation for why SLC26A3 is so rare in the scRNA-seq data, but seemingly common in the IHC staining would be helpful. I am concerned about the specificity of the stain.

We apologize for lacking adequate explanation of SLC26A3 and cluster Epi_10_CYSTM1. This is a quite crucial question as it has been listed above in question (8) of reviewer #1 and question (2) of reviewer #2 (public review section). In the revised manuscript, we have added intenstive discussion about this question in the seventh paragraph of Disccusion section (from Line 801 - 822). In fact, because of the heterogeneity among different individuals and different tumor regions even within one sample, Epi_10_CYSTM1 seemed to be derived from only one sample. However, the relatively high number of cells in this cluster from one late-stage (stage IIIC) patient suggests its presence in ADC and highlights its potential as a diagnostic marker for staging ADC. Furthermore, we have identified SLC26A3, ORM1 and ORM2 as specific markers of this cluser and performed IHC staining. With a positive expression of these markers, the existence of this cluster has been indirectly proved (as shown in Fig. 3B).

Author response:

The following is the authors’ response to the current reviews.

The authors agree with the reviewers that future studies are needed to dissect the mechanisms of eIF3 binding to 3'UTRs and their impact on translation, and the impact of this binding on cellular fate.

The following is the authors’ response to the original reviews.

eLife Assessment

This valuable study reveals extensive binding of eukaryotic translation initiation factor 3 (eIF3) to the 3' untranslated regions (UTRs) of efficiently translated mRNAs in human pluripotent stem cell-derived neuronal progenitor cells. The authors provide solid evidence to support their conclusions, although this study may be enhanced by addressing potential biases of techniques employed to study eIF3:mRNA binding and providing additional mechanistic detail. This work will be of significant interest to researchers exploring post-transcriptional regulation of gene expression, including cellular, molecular, and developmental biologists, as well as biochemists.

We thank the reviewers for their positive views of the results we present, along with the constructive feedback regarding the strengths and weaknesses of our manuscript, with which we generally agree. We acknowledge our results will require a deeper exploration of the molecular mechanisms behind eIF3 interactions with 3'-UTR termini and experiments to identify the molecular partners involved. Additionally, given that NPC differentiation toward mature neurons is a process that takes around 3 weeks, we recognize the importance of examining eIF3-mRNA interactions in NPCs that have undergone differentiation over longer periods than the 2-hr time point selected in this study. Finally, considering the molecular complexity of the 13subunit human eIF3, we agree that a direct comparison between Quick-irCLIP and PAR-CLIP will be highly beneficial and will determine whether different UV crosslinking wavelengths report on different eIF3 molecular interactions. Additional comments are given below to the identified weaknesses.

Public Reviews:

Reviewer #1 (Public review):

Summary:

The authors perform irCLIP of neuronal progenitor cells to profile eIF3-RNA interactions upon short-term neuronal differentiation. The data shows that eIF3 mostly interacts with 3'-UTRs - specifically, the poly-A signal. There appears to be a general correlation between eIF3 binding to 3'-UTRs and ribosome occupancy, which might suggest that eIF3 binding promotes protein

Strengths:

The study provides a wealth of new data on eIF3-mRNA interactions and points to the potential new concept that eIF3-mRNA interactions are polyadenylation-dependent and correlate with ribosome occupancy.

Weaknesses:

(1) A main limitation is the correlative nature of the study. Whereas the evidence that eIF3 interacts with 3-UTRs is solid, the biological role of the interactions remains entirely unknown. Similarly, the claim that eIF3 interactions with 3'-UTR termini require polyadenylation but are independent of poly(A) binding proteins lacks support as it solely relies on the absence of observable eIF3 binding to poly-A (-) histone mRNAs and a seeming failure to detect PABP binding to eIF3 by co-immunoprecipitation and Western blotting. In contrast, LC-MS data in Supplementary File 1 show ready co-purification of eIF3 with PABP.

We agree the molecular mechanisms underlying the crosslinking between eIF3 and the end of mRNA 3’-UTRs remains to be determined. We also agree that the lack of interaction seen between eIF3 and PABP in Westerns, even from HEK293T cells, is a puzzle. The low sequence coverage in the LC-MS data gave us pause about making a strong statement that these represent direct eIF3 interactions, given the similar background levels of some ribosomal proteins.

(2) Another question concerns the relevance of the cellular model studied. irCLIP is performed on neuronal progenitor cells subjected to neuronal induction for 2 hours. This short-term induction leads to a very modest - perhaps 10% - and very transient 1-hour-long increase in translation, although this is not carefully quantified. The cellular phenotype also does not appear to change and calling the cells treated with differentiation media for 2 hours "differentiated NPCs" seems a bit misleading. Perhaps unsurprisingly, the minor "burst" of translation coincides with minor effects on eIF3-mRNA interactions most of which seem to be driven by mRNA levels. Based on the ~15-fold increase in ID2 mRNA coinciding with a ~5-fold increase in ribosome occupancy (RPF), ID2 TE actually goes down upon neuronal induction.

We agree that it will be interesting to look at eIF3-mRNA interactions at longer time points after induction of NPC differentiation. However, the pattern of eIF3 crosslinking to the end of 3’-UTRs occurs in both time points reported here, which is likely to be the more general finding in what we present.

(3) The overlap in eIF3-mRNA interactions identified here and in the authors' previous reports is minimal. Some of the discrepancies may be related to the not well-justified approach for filtering data prior to assessing overlap. Still, the fundamentally different binding patterns - eIF3 mostly interacting with 5'-UTRs in the authors' previous report and other studies versus the strong preference for 3'-UTRs shown here - are striking. In the Discussion, it is speculated that the different methods used - PAR-CLIP versus irCLIP - lead to these fundamental differences. Unfortunately, this is not supported by any data, even though it would be very important for the translation field to learn whether different CLIP methodologies assess very different aspects of eIF3-mRNA interactions.

We agree the more interesting aspect of what we observe is the difference in location of eIF3 crosslinking, i.e. the end of 3’-UTRs rather than 5’-UTRs or the pan-mRNA pattern we observed in T cells. The reviewer is right that it will be important in the future to compare PAR-CLIP and Quick-irCLIP side-by-side to begin to unravel the differences we observe with the two approaches.

Reviewer #2 (Public review):

Summary:

The paper documents the role of eIF3 in translational control during neural progenitor cell (NPC) differentiation. eIF3 predominantly binds to the 3' UTR termini of mRNAs during NPC differentiation, adjacent to the poly(A) tails, and is associated with efficiently translated mRNAs, indicating a role for eIF3 in promoting translation.

Strengths:

The manuscript is strong in addressing molecular mechanisms by using a combination of nextgeneration sequencing and crosslinking techniques, thus providing a comprehensive dataset that supports the authors' claims. The manuscript is methodologically sound, with clear experimental designs.

Weaknesses:

(1) The study could benefit from further exploration into the molecular mechanisms by which eIF3 interacts with 3' UTR termini. While the correlation between eIF3 binding and high translation levels is established, the functionality of these interactions needs validation. The authors should consider including experiments that test whether eIF3 binding sites are necessary for increased translation efficiency using reporter constructs.

We agree with the reviewer that the molecular mechanism by which eIF3 interacts with the 3’UTR termini remains unclear, along with its biological significance, i.e. how it contributes to translation levels. We think it could be useful to try reporters in, perhaps, HEK293T cells in the future to probe the mechanism in more detail.

(2) The authors mention that the eIF3 3' UTR termini crosslinking pattern observed in their study was not reported in previous PAR-CLIP studies performed in HEK293T cells (Lee et al., 2015) and Jurkat cells (De Silva et al., 2021). They attribute this difference to the different UV wavelengths used in Quick-irCLIP (254 nm) and PAR-CLIP (365 nm with 4-thiouridine). While the explanation is plausible, it remains a caveat that different UV crosslinking methods may capture different eIF3 modules or binding sites, depending on the chemical propensities of the amino acid-nucleotide crosslinks at each wavelength. Without addressing this caveat in more detail, the authors cannot generalize their findings, and thus, the title of the paper, which suggests a broad role for eIF3, may be misleading. Previous studies have pointed to an enrichment of eIF3 binding at the 5' UTRs, and the divergence in results between studies needs to be more explicitly acknowledged.

We agree with the reviewer that the two methods of crosslinking will require a more detailed head-to-head comparison in the future. However, we do think the title is justified by the fact that we see crosslinking to the termini of 3’-UTRs across thousands of transcripts in each condition. Furthermore, the 3’-UTR crosslinking is enriched on mRNAs with higher ribosome protected fragment counts (RPF) in differentiated cells, Figure 3F.

(3) While the manuscript concludes that eIF3's interaction with 3' UTR termini is independent of poly(A)-binding proteins, transient or indirect interactions should be tested using assays such as PLA (Proximity Ligation Assay), which could provide more insights.

This is a good idea, but would require a substantial effort better suited to a future publication. We think our observations are interesting enough to the field to stimulate future experimentation that we may or may not be most capable of doing in our lab.

Reviewer #3 (Public review):

Summary:

In this manuscript by Mestre-Fos and colleagues, authors have analyzed the involvement of eIF3 binding to mRNA during differentiation of neural progenitor cells (NPC). The authors bring a lot of interesting observations leading to a novel function for eIF3 at the 3'UTR.

During the translational burst that occurs during NPC differentiation, analysis of eIF3-associated mRNA by Quick-irCLIP reveals the unexpected binding of this initiation factor at the 3'UTR of most mRNA. Further analysis of alternative polyadenylation by APAseq highlights the close proximity of the eIF3-crosslinking position and the poly(A) tail. Furthermore, this interaction is not detected in Poly(A)-less transcripts. Using Riboseq, the authors then attempted to correlate eIF3 binding with the translation efficacy of mRNA, which would suggest a common mechanism of translational control in these cells. These observations indicate that eIF3-binding at the 3'UTR of mRNA, near the poly(A) tail, may participate to the closed-loop model of mRNA translation, bridging 5' and 3', and allowing ribosomes recycling. However, authors failed to detect interactions of eIF3, with either PABP or Paip1 or 40S subunit proteins, which is quite unexpected.

Strength:

The well-written manuscript presents an attractive concept regarding the mechanism of eIF3 function at the 3'UTR. Most mRNA in NPC seems to have eIF3 binding at the 3'UTR and only a few at the 5'end where it's commonly thought to bind. In a previous study from the Cate lab, eIF3 was reported to bind to a small region of the 3'UTR of the TCRA and TCRB mRNA, which was responsible for their specific translational stimulation, during T cell activation. Surprisingly in this study, the eIF3 association with mRNA occurs near polyadenylation signals in NPC, independently of cell differentiation status. This compelling evidence suggests a general mechanism of translation control by eIF3 in NPC. This observation brings back the old concept of mRNA circularization with new arguments, independent of PABP and eIF4G interaction. Finally, the discussion adequately describes the potential technical limitations of the present study compared to previous ones by the same group, due to the use of Quick-irCLIP as opposed to the PAR-CLIP/thiouridine.  

Weaknesses:

(1) These data were obtained from an unusual cell type, limiting the generalizability of the model.

We agree that unraveling the mechanism employed by eIF3 at the mRNA 3’-UTR termini might be better studied in a stable cell line rather than in primary cells.

(2) This study lacks a clear explanation for the increased translation associated with NPC differentiation, as eIF3 binding is observed in both differentiated and undifferentiated NPC. For example, I find a kind of inconsistency between changes in Riboseq density (Figure 3B) and changes in protein synthesis (Figure 1D). Thus, the title overstates a modest correlation between eIF3 binding and important changes in protein synthesis.

We thank the reviewer for this question. Riboseq data and RNASeq data are not on absolute scales when comparing across cell conditions. They are normalized internally, so increases in for example RPF in Figure 3B are relative to the bulk RPF in a given condition. By contrast, the changes in protein synthesis measured in Figure 1D is closer to an absolute measure of protein synthesis. 

(3) This is illustrated by the candidate selection that supports this demonstration. Looking at Figure 3B, ID2, and SNAT2 mRNA are not part of the High TE transcripts (in red). In contrast, the increase in mRNA abundance could explain a proportionally increased association with eIF3 as well as with ribosomes. The example of increased protein abundance of these best candidates is overall weak and uncertain.

We agree that using TE as the criterion for defining increased eIF3 association would not be correct. By “highly translated” we only mean to convey the extent of protein synthesis, i.e. increases in ribosome protected fragments (RPF), rather than the translational efficiency.

(4) Despite several attempts (chemical and UV cross-linking) to identify eIF3 partners in NPC such as PABP, PAIP1, or proteins from the 40S, the authors could not provide any evidence for such a mechanism consistent with the closed-loop model. Overall, this rather descriptive study lacks mechanistic insight (eIF3 binding partners).

We agree that it will be important to identify the molecular mechanism used by eIF3 to engage the termini of mRNA 3’-UTRs. Nevertheless, the identification of eIF3 crosslinking to that location in mRNAs is new, and we think will stimulate new experiments in the field.

(5) Finally, the authors suspect a potential impact of technical improvement provided by QuickirCLIP, that could have been addressed rather than discussed.

We agree a side-by-side comparison of eIF3 crosslinks captured by PAR-CLIP versus QuickirCLIP will be an important experiment to do. However, NPCs or other primary cells may not be the best system for the comparison. We think using an established cell line might be more informative, to control for effects such as 4-thiouridine toxicity.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

(1) The Western blot signals for SLC38A2 and ID2 are close to the membrane background and little convincing. Size markers are missing.

We agree these antibodies are not great. They are the best we could find, unfortunately. We have included originals of all western blots and gels as supplementary information. It’s important to note that the Riboseq data for ID2 and SLC38A2 are consistent with the western blots. See Figure 3C and Figure 3–figure supplement 3B.

(2) Figure 1 - Figure Supplement 1 appears to present data from a single experiment. This is far less than ideal considering the minor differences measured.

Thanks for the comment. This is a representative experiment showing the early time course. We have added a second experiment with two different treatments that show the same pattern in the puromycin assay, in Figure 1–figure supplement 1.

(3) Figure 3F: One wonders what this would look like if TE was plotted instead of RPF. Figure 3 - Figure Supplement 4 seems to show something along those lines. However, the data are not mentioned in the main results section are quite unclear. Why are data separated into TE high and low? Doesn't TE high in differentiated cells equal TE low in undifferentiated cells?

This is an interesting question. Note that in Figure 3B, n=6300 genes show no change in TE upon differentiation, compared to a total of n=2127 that show a change in TE, with most of those changes not very large. We have now replotted Figure 3F comparing irCLIP read counts in 3’-UTRs to RPF read counts, which shows a significant positive correlation, regardless of whether we look at undifferentiated or differentiated NPCs (See Figure 3F and a new Figure 3– figure supplement 4A). We also compare irCLIP reads in 3’-UTRs to TE values, which show no correlation (See Figure 3G and Figure 3–figure supplement 4B).

Figure 3-figure supplement 4 was actually a response to a previous round of review (at PLOS Biology) to a rather technical question from a reviewer. We think this figure and associated text should be removed. Instead, we now include supplementary tables with the processed RPF and TE values, for reference (Supplemental files 4-6). We omitted these in the original submission when they should have been included. We also abandoned comparing undifferentiated and differentiated NPCs, and instead look directly at irCLIP reads vs. RPFs or TE, regardless of NPC state, as noted above (Figure 3F, G, and Figure 3–figure supplement 4).

(4) Figure 3C: The data should be plotted on the same y-axis scale. This would make a visual assessment of the differences in mRNA and RFP levels more intuitive.

Thanks for this suggestion. We have rescaled the plots as requested.

Reviewer #2 (Recommendations for the authors):

(1) The quality of the Western blots in several figures is quite poor. Notably, Figure 1C seems to be a composite gel, as each blot appears to come from a different gel. Additionally, in Supplementary Figure 1A, there is only a single data point, yet the authors indicate that this image is representative of multiple assays. The lack of error bars in this figure raises a question vis-a-vis the reproducibility of the experiments.

Thanks for the comments. We now include all the original gels as supplementary information. As noted above, the antibodies for ID2 and SLC38A2 are not great, we agree. And as we noted above, the Riboseq data for ID2 and SLC38A2 are consistent with the western blots.

(2) For the top 500 targets of undifferentiated and differentiated NPCs in the Quick-irCLIP assay, the manuscript does not clarify how many targets are common and how many are unique to each condition. This information is important for understanding the extent of overlap and differentiation-specific interactions of eIF3 with mRNAs. Providing this data would strengthen the interpretation of the results.

There are 449 of the top 500 hits in common between undifferentiated and differentiated NPCs. We have now added this information to the text, to add clarity. 

(3) The manuscript does not provide detailed percentages or numbers regarding the overlap between iCLIP and APA-Seq peaks. Clarifying this overlap, particularly in terms of how many of the APA sites are also targets of eIF3, would bolster the understanding of how these two datasets converge to support the authors' conclusions.

This is a difficult calculation to make, due to the fact that APA-Seq reads are generally much longer than the Quick-irCLIP reads. This is why we focused instead on quantifying the percent of Quick-irCLIP peaks (which are more narrow) overlap with predicted polyadenylation sequences, in Figure 2-figure supplement 1.

Reviewer #3 (Recommendations for the authors):

(1) Perform Quick-irCLIP in HEK293 cells to infer technical limitations and/or to generalize the model. The authors will then compare again eIF3 binding site in Jurkat, HEK293, and NPC.

This is an experiment we plan to do for a future publication, given that we would want to repeat both Quick-irCLIP and PAR-CLIP at the same time.

(2) Select mRNA candidates with high or low TE changes and analyze eIF3 binding and RPF density and protein abundance along NPC differentiation to support the role of eIF3 binding in stimulating translation.

We agree looking at time courses in more depth would be interesting. However, this would require substantial experimentation, which is better suited to a future study. Furthermore, now that we have moved away from comparing undifferentiated NPCs and differentiated NPCs when examining TE and RPF values (Figure 3 and Figure 3–figure supplement 4), we think the results now support a more general mechanism of translation reflected in the irCLIP 3’-UTR vs. RPF correlation, independent of NPC state.

(3) Analyze the interaction of eIF3 with eIF4G and other known partners. This will really provide an improvement to the manuscript. The lack of interaction between eIF3 and the 40S is quite surprising.

We agree more work needs to be done on the mechanistic side. These are experiments we think would be best to carry out in a stable cell line in the future, rather than primary cells.

(4) Perform Oligo-dT pulldown (or cap column if possible) and analyze the relative association of PABP, eIF3, and eIF4F on mRNA in NPC versus HEK293. This will clarify whether this mechanism of mRNA translation is specific to NPC or not.

Thanks for this suggestion. We are uncertain how it would be possible to deconvolute all the possible ways to interpret results from such an experiment. We agree thinking about ways to study the mechanism will keep us occupied for a while.

(5) Citations in the text indicate the first author, whereas the references are numbered! 

Our apologies for this oversight. This was a carryover from previous formatting, and has been fixed.

Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

Learn more at Review Commons

Reply to the reviewers

Reviewer #1 (Evidence, reproducibility____,____ and clarity)

This manuscript by Tsai et al. shows that phage resistance mutations (LPS truncation) confer a cost during interbacterial competition. The authors show that various phage resistant mutants of S. enterica are inhibited by E. cloacae in a contact-dependent manner (on a solid surface but not in liquid). Further experiments showed that this inhibition of S. enterica was mediated by T6SS in E. cloacae. The authors then dissect which parts of the LPS are required for resistance against T6SS attacks and show that a similar resistance is conferred against T6SS of B. thailandensis and C. rodentium. Moreover, the authors show that enzymatic degradation of LPS by a phage enzyme can also increase sensitivity to T6SS (including when such enzymes are on phage particles). Finally, the authors suggest that the change in the thickness of the LPS surface layer could be the reason for changes in T6SS susceptibility. Overall, the manuscript is very well-written. The experiments and controls are explained in sufficient detail and in a logical order. The figures are clear and easy to navigate. The findings are very interesting and important for the T6SS field but also for general understanding how different evolutionary pressures combine and influence each other. I believe that this manuscript will initiate further research in this direction.

• We thank the reviewer for their positive remarks on our manuscript and the valuable suggestions for its improvement. Major comments

The only major point that I would like to raise is that I am not generally convinced that the 2 nm difference in the thickness of LPS is the main reason for the observed differences in T6SS-mediated killing of S. enterica. Based on what we know about T6SS mode of action, we expect that it is potentially pushing effectors by up to several hundreds of nanometers. Therefore, the change in the LPS thickness by a few nanometers (as measured by AFM) seems insufficient to provide enough spacing between the attacker and the prey to significantly decrease T6SS effector delivery. While it is clear that understanding the exact reason for the LPS mediated resistance is beyond the scope of this manuscript, I would suggest that the authors consider the fact that T6SS is known to deliver proteins even to the cytoplasm of target gram-negative cells and discuss the mode of action of the machine in the context of their finding. If the T6SS was drawn to scale in the model figure, it would become apparent that 2 nm change in the distance between two cells has probably no major impact on killing by T6SS and the actual reason for the observed phenotype is likely more complicated than what is proposed.

We appreciate the reviewer's comments and acknowledge that our manuscript leaves open questions regarding the exact mechanisms underlying LPS-mediated resistance. We have now moderated the Discussion in our revised manuscript to reflect the complexity of this phenomenon (Lines 410-423). Although we agree that the nanometer difference in LPS thickness may not fully explain the observed protective phenotype, we believe it remains a plausible contributing factor that is worth considering.

To fully understand how LPS influences T6SS effector delivery, future studies will need to address key mechanistic questions regarding the T6SS injection process. For example, 1) how deeply does the T6SS apparatus penetrate the target Gram-negative cells during injection; 2) what is the magnitude of the injection force generated by the T6SS; and 3) does the structural integrity of the T6SS apparatus remain intact throughout and after contraction? While it is well documented that some T6SS effectors act in the cytosol of target cells, there is evidence to suggest that cytosolic effectors are initially delivered into the periplasm and subsequently translocated into the cytosol for intoxication1,2. Furthermore, although contraction of the T6SS apparatus occurs within milliseconds3,4, this rapid action does not preclude the possibility that the injection force could be influenced by the thickness of the LPS layer. In addition, the stability of T6SS structural or delivered proteins-such as PAAR, VgrG, and Hcp-within the delivery complex might be compromised upon encountering physical barriers such as the LPS layer and the outer membrane of target cells. These potential interactions could affect the efficiency of effector delivery, leading to reduced competitiveness during interbacterial antagonism, as shown in our study.

• We appreciate the reviewer's suggestions and acknowledge that the precise reasons for LPS-mediated resistance likely involve a combination of factors beyond those proposed here. We are actively pursuing these questions as part of an ongoing, long-term effort to better elucidate the mechanisms of T6SS action. Minor comments

Specify which T6SS of B. thailandensis was tested.

• We now cite studies by Schwarz, S., et al., 20105 and LeRoux, M., et al., 20156, from which we used the tssM (BTH_I2954) gene deletion strain abrogating the T6SS-1 of the B. thailandensis E264 (Line 234, Supplementary Table 1). Use a different naming of the two strains used in competition assays than "donor" and "recipient".

• Thank you for this suggestion. In the revised manuscript, we have replaced the terms "donor" and "recipient" with "attacker" and "prey" for clarity. This change has been applied to the text (Lines 441, and 649-667) and to revised Figures 2c-h, Figures 3b, d, g, i, j, Figures 4f, g, Figures 5b, e, g, h, Supplementary Figures 3d-f, and Supplementary Figures 4b-d. Indicate in the material and methods ODs of bacterial mixtures used in the "Bacterial competition assays".

• We apologize for this oversight. The ODs of bacterial mixtures used in the "Bacterial competition assays" have now been specified in the revised Methods section (Line 6____51). Reviewer #1 (Significance)

This manuscript is interesting for researchers who study T6SS, phage predation and other evolutionary pressures shaping bacterial interactions. The work provides new and interesting insights. My expertise in LPS biology is limited.

• We sincerely appreciate the reviewer's interest in and support of our study. Reviewer #____2____ (Evidence, reproducibility____,____ and clarity)

This work investigates the fitness trade-offs in Salmonella enterica resistant to phages. The authors performed co-culture experiments with S. enterica, E. coli, and E. cloacae and found that phage-resistant S. enterica strains displayed reduced fitness in the presence of E. cloacae. Further experiments demonstrated that phage-resistant S. enterica strains were more susceptible to the type VI secretion system (T6SS) of E. cloacae. The authors then examined the role of the O-antigen of lipopolysaccharide (LPS) in T6SS-mediated interbacterial antagonism. By constructing S. enterica mutants with varying O-antigen chain lengths, the authors demonstrated that the O-antigen protects S. enterica from T6SS attack. They then demonstrated that the O-antigen-deficient S. enterica, E. coli, and C. rodentium strains were more susceptible to T6SS attack by E. cloacae. Finally, the authors showed that phage tail spike proteins (TSPs) with endoglycosidase activity could cleave the bacterial O-antigen, thereby increasing susceptibility to T6SS attack.

The study is well-designed and the experiments are well-executed. The findings are significant and have implications for the understanding of microbial community dynamics.

• We thank the reviewer for their positive comments regarding our original submission. Major comments

While the study elegantly demonstrates the link between phage resistance, LPS structure, and T6SS susceptibility, we must remember that these LPS-defective strains are likely at a significant disadvantage in real-world environments without the influence of competing bacteria. Whether it's the gut or external environments, Salmonella needs its LPS for protection against a myriad of host and environmental factors. It seems a bit redundant for T6SS mediated antagonism to select for LPS structures when those structures are essential for bacterial survival outside of this very specific context. It would benefit some discussion about the likelihood of these phage-resistant, LPS-defective strains actually persisting and competing effectively in a more natural setting.

• We thank the reviewer for their insightful comments and appreciate the opportunity to clarify this point. We agree that LPS-defective bacterial strains face significant disadvantages in natural environments, where they must contend with various host and environmental stresses. Consequently, we did not intend to suggest that T6SS-mediated antagonism is the primary driving force in selecting specific LPS structures. Rather, our study highlights an additional role for LPS during interbacterial interactions, complementing its well-established functions. This notion aligns with the hypotheses proposed in prior studies7-9. The reviewer's comments raise an intriguing question about the essentiality of LPS in Gram-negative bacteria under natural conditions. During our revision process, we identified several examples in the literature demonstrating that LPS may not always be indispensable. For instance, LPS-depleted Neisseria meningitidis strains with an early block in lipid A biosynthesis have been shown to remain viable10,11. These strains may possess adaptive advantages under specific circumstances12. Similarly, some pathogenic bacteria produce truncated LPS structures lacking O-antigen or introduce modified LPS to evade host immune responses13. Additionally, evolutionary pressures, such as phage predation, often drive mutations in O-antigen biosynthesis pathways, resulting in alterations to or an absence of O-antigen14. Furthermore, recent studies have also indicated that trade-offs between abiotic and biotic stresses can influence LPS integrity. For instance, LPS-deficient strains may exhibit selective advantages in extreme environments15,16. These findings underscore the context-dependent nature of LPS functionality and its potential dispensability in certain ecological niches.We sincerely appreciate the reviewer's thought-provoking comments. Our current study aims to provide evidence for the role of interbacterial antagonism as an additional factor influencing LPS integrity. However, we did not mean to overstate the contribution of this mechanism. Instead, we only seek to contribute to a broader understanding of the multifaceted functions of LPS in bacterial survival and adaptation. We have modified the Discussion in our revised manuscript to clarify this idea (Lines 453-466). Minor comments

Figure 5 could be more effective is panels b and c are together

• We appreciate this suggestion. We have revised the manuscript accordingly, so panels b and c have been combined in revised Figure 5, __and the respective figure legends have been modified for improved clarity (__Lines 810-823).

69 Authors should define mucoid

• The term "mucoid" has now been defined in the revised manuscript (Lines 69-70).

155 Authors should explain that this result is expected since T6SS acts on solid surface while CDI works in liquid cultures

• Thank you for this comment. Prior studies have demonstrated that while CDI-mediated antibacterial activity is less efficient in liquid environments, it can still occur on both solid surfaces and in liquid cultures, provided the competitors possess the necessary CdiA binding unit, such as BamA17,18. This understanding supports our initial hypothesis that T6SS and/or CDI contribute to the observed protective phenotype in S. enterica phage-resistant variants (Figure 2).

clarify what it is meant by unicellular cultures. Should it be monocultures?

• We apologize for this error and have now replaced "unicellular cultures" with "monocultures" in the revised manuscript (Lines 137, 180, and 258).

618 add to the text how much dead phage was added per bacterial cell

• Apologies for this oversight. The multiplicity of infection (MOI) describing the amount of inactivated phages used to treat bacterial cells has now been included in the revised Methods section (Line 661).

364 references needed for "consistent with predictions for intact LPS structures "

• We thank the reviewer for pointing out this omission. The relevant reference has now been added to the revised manuscript19 (Line 368). Reviewer #____2____ (Significance)

This study offers a new perspective on the interplay between phage resistance and bacterial fitness in the context of microbial communities. While the concept of fitness trade-offs associated with antibiotic resistance is well-established, the authors extend this paradigm to phage resistance. They demonstrate that phage-resistant Salmonella enterica strains exhibit reduced fitness in the presence of Enterobacter cloacae due to increased susceptibility to the type VI secretion system (T6SS). This finding is significant as it highlights the potential for interbacterial antagonism to shape the evolution of phage resistance. The authors further show that the O-antigen of lipopolysaccharide (LPS) plays a crucial role in protecting S. enterica from T6SS attack. This observation provides mechanistic insights into the fitness trade-offs associated with phage resistance.

The study's strength lies in its elegant experimental design and the comprehensive analysis of the interplay between phage resistance, T6SS susceptibility, and O-antigen structure. The authors employ a combination of co-culture experiments, genetic manipulations, and structural analyses to dissect the underlying mechanisms. The findings are robust and have implications for understanding the evolution of bacterial communities in the presence of phages and competing bacterial species.

This research will be of interest to a broad audience, including researchers in microbiology, synthetic biology, and microbial ecology. The findings have implications for understanding the evolution of phage resistance, and the dynamics of microbial communities. The study's insights into the role of the O-antigen in T6SS susceptibility could also inform the design of novel antimicrobial strategies.

My expertise is microbial physiology

• We thank the reviewer for their positive remarks and careful reading of our manuscript. Reviewer #____3____ (Evidence, reproducibility____,____ and clarity)

Tsai et al. describe LPS biosynthesis mutants arising in selection for phage resistance that increase susceptibility to T6SS-mediated interbacterial antagonism. Phage-derived LPS degrading enzymes also contribute to T6SS susceptibility, which may be due to weakening of the physical barrier of LPS. The mechanisms of this fitness trade-off are elucidated with well-executed and presented experiments.

• We are grateful to the reviewer for their kind words and critical reading of the manuscript. Major comments

No major critiques.

Minor comments

Others have described two T6SS in Enterobacter cloacae ATCC 13047 (PMID 33072020). Please clarify which of the two are inactivated by the tssM deletion in this study and either provide compelling evidence that both are inactive or change the text throughout to indicate T6SS-1 or T6SS-2 being inactivated.

• We thank the reviewer for this comment. In our study, we refer to the work by Whitney, J., et al., 201420, from which we used the tssM (ECL_01536) gene deletion strain in which T6SS-1 of the E. cloacae ATCC 13047 is abrogated. Consistent with this detail, we have now clarified in the revised manuscript (Line 155, Supplementary Table 1) that T6SS-1 is inactivated. Moreover, the reference suggested by the reviewer provides additional evidence supporting that T6SS-1, but not T6SS-2, is involved in bacterial competition21, which we also now specify in the revised manuscript. It seems the authors used EHEC EDL933, which has T6SS, in co-culture experiments (Figure 1C). Why do the authors think the S. enterica LPS mutants don't have a competitive disadvantage against EHEC? It seems to run counter to the conclusion that LPS is broadly protective against T6SS.

• We thank the reviewer for raising this point. While it is true that EHEC O157:H7 strain EDL933 possesses a T6SS gene cluster in its genome, a prior study has shown that the T6SS in this strain appears to be inactivated under laboratory conditions, likely due to repression by the global regulator H-NS22. Consistent with these findings, our data indicate that the S. enterica LPS mutants did not exhibit a competitive disadvantage against EHEC EDL933. These results support the conclusion that, under the conditions tested, the truncated LPS in S. enterica does not affect its fitness against EHEC (Figure 1c), likely due to the inactivity of the EHEC T6SS22. It's not clear if the only Felix O1 and P22 phage-resistant transposon hits were in LPS-related genes, or if that pattern was observed in a more complete transposon sequencing dataset and selected for further study. A complete list of the sequence-identified hits, including the non-LPS related variants, would help clarify this and provide a useful resource to the research community.

• We thank the reviewer for the opportunity to clarify this point. For each phage, we initially isolated nine phage-resistant transposon variants, which were subsequently used for co-culture assays and transposon insertion site identification, as described in the original manuscript (Figure 1a __and Supplementary Figure 2a__). We agree with the reviewer that a broader screening approach could reveal non-LPS-related variants and provide a more comprehensive resource for the research community. To address this point, during the manuscript revision period, we followed the same procedure and isolated an additional nine phage-resistant variants for each phage (Supplementary Table 1). Interestingly, from this expanded isolation dataset, the transposon insertions were again found exclusively in LPS-related genes (Author Response Figure 1). We have now included this new dataset in the revised manuscript and believe it strengthens the robustness of our findings. This expanded data has been made available below for further reference. The fact that 8 of the 9 Felix O1 resistant variants all have transposon insertions in waaO should be stated in the results. The initial impression of showing R1-R9 is that 9 disrupted genes are being tested - in this case it's really only two. This is a minor critique because clean deletions by allelic exchange are shown for a more extensive set of genes anyway.

• We thank the reviewer for this comment. As suggested, we have revised the Results section (Lines 126-131) to explicitly state that Felix O1-resistant variants harbor transposon insertions in only two genes (waaO and dagR), which were initially tested in the competition assay (Figure 2). The S. enterica serovar Typhimurium transposon mutagenesis library could benefit from clarification on details. The results section suggests use of a pre-existing "established" transposon library, but the methods and Figure 1 seem to indicate a new library was created based on prior methods. In either case, what is the genome coverage and redundancy of the library? If this is not known or saturation is not reached, the implications of potentially missing phage resistance genes with this approach should be discussed.

• We thank the reviewer for the opportunity to clarify this point. For our study, we created a transposon library following previously established methods23. The library comprises approximately 12,000 variants, as noted in Figure 1a. While doing so provided substantial genome coverage, it did not achieve full saturation. We have now revised the Results section (Lines 93-94, and 115-117) to better describe the potential limitations of this approach, including by stating the possibility that some phage-resistance genes may have been missed during the screening. There is some variation in phenotype among the strains with transposon insertion into the same gene, such as P22 resistant strain R7 which macroscopically agglutinates while the other waaJ insertions R5 and R1 don't. Is this due to polar effects on waaO, or could it be genetic alterations at other sites driven by stringent phage selection?

• We thank the reviewer for this comment. We also suspect that the variation in the macroscopically agglutinative phenotypes among P22-resistant strains, such as strain R7 compared to R5 and R1, may be caused by polar effects on waaO. Additionally, the possibility of genetic alterations at other loci driven by stringent phage selection cannot be excluded. To address this potential variability and ensure consistency, we used clean deletions of each LPS biogenesis gene in all subsequent experiments. This approach eliminates the confounding effects of polar mutations or secondary genetic alterations, thereby providing more robust and interpretable data. Figure S1- The graphs with 12 growth curves are difficult to decipher, and the error bars would suggest maybe there are subtle growth differences among the mutants. Quantifying curve parameter(s) and applying a statistical test may clarify. The CFU counts in panel D seem to be not in log scale. Likewise in Figure S3 panel A, the authors say there are no significant growth defects, but the growth curves are modestly right-shifted for several mutants. This is a point of precision rather than a major critique, because the reversal of competitive growth phenotypes by donor T6SS inactivation indicate the potential minor growth defects aren't playing a major role in competition.

• We thank the reviewer for these suggestions and corrections. We have now revised the manuscript accordingly, including in Supplementary Figures 1 and 3. Quantitative analysis of growth curve parameters and statistical tests have been included below to clarify the observed differences (Author Response Figure 2). The slight right-shift of the growth curves for some mutants, as noted in Supplementary Figure 3, may be attributable to cell aggregation, as shown in Supplementary Figures 2e, f. The growth rate measurements were conducted in a 96-well plate with steady shaking at 200 rpm using a plate reader, which does not fully account for the aggregated cell phenotype. Despite these subtle growth differences, we agree with the reviewer that they do not appear to play a major role in the competitive growth phenotypes, as evidenced by the reversal of phenotypes upon donor T6SS inactivation (Supplementary Figure 3). Figure 3f - The authors say fepE is responsible for very long O-antigen chains, but it is not clear that the delta fepE LPS PAGE differs from wild type, which would fit with the lack of competitive disadvantage against E. cloacae in Figure 3g. The increased VL-modal O-antigen upon fepE overexpression in Figure 3h and increase protection in competition (figure 3i) are convincing. Is there another pathway(s) compensating for fepE deletion?

• We thank the reviewer for this thoughtful comment. We have repeated the experiment independently at least three times and consistently observed a reduction in the VL-modal O-antigen in the ∆fepE strain. To provide additional clarity, we have included supplementary LPS profiles and quantifications below (Author Response Figure 3). We currently do not have evidence from the literature or our experiments to identify an alternative pathway compensating for the deletion of fepE. Nonetheless, we acknowledge this as a possibility and appreciate the reviewer's insight into this topic. Lines 199-200 - I believe the conclusion from wzzB deletion would be that L-modal O-antigen is necessary for protection against T6SS, and not necessarily sufficient.

• We thank the reviewer for pointing out this important distinction. The respective sentence has now been revised in the manuscript (Line 204). Do the environmentally isolated phages As2 and As4 encode TSP homologs?

• We thank the reviewer for this question. We did not identify TSP homologs in the genome of As2 and As4 phages. The genome sequences of As1 to As4 have been uploaded to NCBI's BioProject resource under accession number PRJNA1199570 (Lines 535-544, 741-743). Reviewer #____3____ (Significance)

This manuscript provides a substantial advance in the field's understanding of how phages affect bacterial community interactions. To my knowledge, it is the first to bring together phage and T6SS defense with a strong mechanistic link. It's a conceptual advance in this regard that will stimulate more thought and experimentation on the roles of phage in bacterial communities like gut and environmental microbiomes. The manuscript's strengths include rigorous overall design, clarity of the communication, and depth of mechanistic investigation, all the way down to atomic force microscopy measurements. There are some minor revisions suggested, but these are addressable with minimal/no additional experiments.

As someone with expertise in bacterial secretion systems and interbacterial interactions, I think this work will be of interest to microbiologists generally, and specifically in the fields of phage biology, bacterial secretion systems, and microbiome research. While the phage virology components are straightforward and well described, I think a review from someone with more expertise in this specific area would be beneficial.

• We thank the reviewer for their careful reading of our manuscript and for the suggestions to improve it. References

• Whitney, J.C., Quentin, D., Sawai, S., LeRoux, M., Harding, B.N., Ledvina, H.E., Tran, B.Q., Robinson, H., Goo, Y.A., Goodlett, D.R., et al. (2015). An interbacterial NAD(P)(+) glycohydrolase toxin requires elongation factor Tu for delivery to target cells. Cell 163, 607-619. 10.1016/j.cell.2015.09.027.

• Ali, J., Yu, M., Sung, L.K., Cheung, Y.W., and Lai, E.M. (2023). A glycine zipper motif is required for the translocation of a T6SS toxic effector into target cells. EMBO Rep 24, e56849. 10.15252/embr.202356849.
• LeRoux, M., De Leon, J.A., Kuwada, N.J., Russell, A.B., Pinto-Santini, D., Hood, R.D., Agnello, D.M., Robertson, S.M., Wiggins, P.A., and Mougous, J.D. (2012). Quantitative single-cell characterization of bacterial interactions reveals type VI secretion is a double-edged sword. Proc Natl Acad Sci U S A 109, 19804-19809. 10.1073/pnas.1213963109.
• Basler, M., Pilhofer, M., Henderson, G.P., Jensen, G.J., and Mekalanos, J.J. (2012). Type VI secretion requires a dynamic contractile phage tail-like structure. Nature 483, 182-186. 10.1038/nature10846.
• Schwarz, S., West, T.E., Boyer, F., Chiang, W.C., Carl, M.A., Hood, R.D., Rohmer, L., Tolker-Nielsen, T., Skerrett, S.J., and Mougous, J.D. (2010). Burkholderia type VI secretion systems have distinct roles in eukaryotic and bacterial cell interactions. PLoS Pathog 6, e1001068. 10.1371/journal.ppat.1001068.
• LeRoux, M., Kirkpatrick, R.L., Montauti, E.I., Tran, B.Q., Peterson, S.B., Harding, B.N., Whitney, J.C., Russell, A.B., Traxler, B., Goo, Y.A., et al. (2015). Kin cell lysis is a danger signal that activates antibacterial pathways of Pseudomonas aeruginosa. Elife 4. 10.7554/eLife.05701.
• Hersch, S.J., Manera, K., and Dong, T.G. (2020). Defending against the Type Six Secretion System: beyond Immunity Genes. Cell Rep 33, 108259. 10.1016/j.celrep.2020.108259.
• Unterweger, D., Kitaoka, M., Miyata, S.T., Bachmann, V., Brooks, T.M., Moloney, J., Sosa, O., Silva, D., Duran-Gonzalez, J., Provenzano, D., and Pukatzki, S. (2012). Constitutive type VI secretion system expression gives Vibrio cholerae intra- and interspecific competitive advantages. PLoS One 7, e48320. 10.1371/journal.pone.0048320.
• Toska, J., Ho, B.T., and Mekalanos, J.J. (2018). Exopolysaccharide protects Vibrio cholerae from exogenous attacks by the type 6 secretion system. Proc Natl Acad Sci U S A 115, 7997-8002. 10.1073/pnas.1808469115.
• Steeghs, L., den Hartog, R., den Boer, A., Zomer, B., Roholl, P., and van der Ley, P. (1998). Meningitis bacterium is viable without endotoxin. Nature 392, 449-450. 10.1038/33046.
• Steeghs, L., de Cock, H., Evers, E., Zomer, B., Tommassen, J., and van der Ley, P. (2001). Outer membrane composition of a lipopolysaccharide-deficient Neisseria meningitidis mutant. EMBO J 20, 6937-6945. 10.1093/emboj/20.24.6937.
• Fransen, F., Heckenberg, S.G., Hamstra, H.J., Feller, M., Boog, C.J., van Putten, J.P., van de Beek, D., van der Ende, A., and van der Ley, P. (2009). Naturally occurring lipid A mutants in neisseria meningitidis from patients with invasive meningococcal disease are associated with reduced coagulopathy. PLoS Pathog 5, e1000396. 10.1371/journal.ppat.1000396.
• Maldonado, R.F., Sa-Correia, I., and Valvano, M.A. (2016). Lipopolysaccharide modification in Gram-negative bacteria during chronic infection. FEMS Microbiol Rev 40, 480-493. 10.1093/femsre/fuw007.
• Yu, J., Zhang, H., Ju, Z., Huang, J., Lin, C., Wu, J., Wu, Y., Sun, S., Wang, H., Hao, G., and Zhang, A. (2024). Increased mutations in lipopolysaccharide biosynthetic genes cause time-dependent development of phage resistance in Salmonella. Antimicrob Agents Chemother 68, e0059423. 10.1128/aac.00594-23.
• Burmeister, A.R., Fortier, A., Roush, C., Lessing, A.J., Bender, R.G., Barahman, R., Grant, R., Chan, B.K., and Turner, P.E. (2020). Pleiotropy complicates a trade-off between phage resistance and antibiotic resistance. Proc Natl Acad Sci U S A 117, 11207-11216. 10.1073/pnas.1919888117.
• Carretero-Ledesma, M., Garcia-Quintanilla, M., Martin-Pena, R., Pulido, M.R., Pachon, J., and McConnell, M.J. (2018). Phenotypic changes associated with Colistin resistance due to Lipopolysaccharide loss in Acinetobacter baumannii. Virulence 9, 930-942. 10.1080/21505594.2018.1460187.
• Aoki, S.K., Pamma, R., Hernday, A.D., Bickham, J.E., Braaten, B.A., and Low, D.A. (2005). Contact-dependent inhibition of growth in Escherichia coli. Science 309, 1245-1248. 10.1126/science.1115109.
• Aoki, S.K., Malinverni, J.C., Jacoby, K., Thomas, B., Pamma, R., Trinh, B.N., Remers, S., Webb, J., Braaten, B.A., Silhavy, T.J., and Low, D.A. (2008). Contact-dependent growth inhibition requires the essential outer membrane protein BamA (YaeT) as the receptor and the inner membrane transport protein AcrB. Mol Microbiol 70, 323-340. 10.1111/j.1365-2958.2008.06404.x.
• Gao, Y., Widmalm, G., and Im, W. (2023). Modeling and Simulation of Bacterial Outer Membranes with Lipopolysaccharides and Capsular Polysaccharides. J Chem Inf Model 63, 1592-1601. 10.1021/acs.jcim.3c00072.
• Whitney, J.C., Beck, C.M., Goo, Y.A., Russell, A.B., Harding, B.N., De Leon, J.A., Cunningham, D.A., Tran, B.Q., Low, D.A., Goodlett, D.R., et al. (2014). Genetically distinct pathways guide effector export through the type VI secretion system. Mol Microbiol 92, 529-542. 10.1111/mmi.12571.
• Soria-Bustos, J., Ares, M.A., Gomez-Aldapa, C.A., Gonzalez, Y.M.J.A., Giron, J.A., and De la Cruz, M.A. (2020). Two Type VI Secretion Systems of Enterobacter cloacae Are Required for Bacterial Competition, Cell Adherence, and Intestinal Colonization. Front Microbiol 11, 560488. 10.3389/fmicb.2020.560488.
• Wan, B., Zhang, Q., Ni, J., Li, S., Wen, D., Li, J., Xiao, H., He, P., Ou, H.Y., Tao, J., et al. (2017). Type VI secretion system contributes to Enterohemorrhagic Escherichia coli virulence by secreting catalase against host reactive oxygen species (ROS). PLoS Pathog 13, e1006246. 10.1371/journal.ppat.1006246.
• Mandal, R.K., Jiang, T., and Kwon, Y.M. (2021). Genetic Determinants in Salmonella enterica Serotype Typhimurium Required for Overcoming In Vitro Stressors in the Mimicking Host Environment. Microbiol Spectr 9, e0015521. 10.1128/Spectrum.00155-21.

DOI: 10.1038/s42255-024-01164-y

Resource: RRID:Addgene_12259

Curator: @olekpark

SciCrunch record: RRID:Addgene_12259

What is this?

DOI: 10.1038/s42255-024-01164-y

Resource: RRID:Addgene_12260

Curator: @olekpark

SciCrunch record: RRID:Addgene_12260

What is this?

DOI: 10.1038/s42255-024-01164-y

Resource: RRID:Addgene_79009

Curator: @olekpark

SciCrunch record: RRID:Addgene_79009

What is this?

DOI: 10.1038/s42255-024-01164-y

Resource: RRID:Addgene_86986

Curator: @olekpark

SciCrunch record: RRID:Addgene_86986

What is this?

DOI: 10.1038/s41467-024-55436-y

Resource: Addgene (RRID:SCR_002037)

Curator: @olekpark

SciCrunch record: RRID:SCR_002037

What is this?

DOI: 10.1038/s41522-024-00629-y

Resource: RRID:Addgene_191002

Curator: @scibot

SciCrunch record: RRID:Addgene_191002

What is this?

DOI: 10.1007/s12602-024-10424-y

Resource: (KCLB Cat# 21592, RRID:CVCL_0343)

Curator: @scibot

SciCrunch record: RRID:CVCL_0343

What is this?

DOI: 10.1007/s12602-024-10424-y

Resource: (IZSLER Cat# BS CL 87, RRID:CVCL_0574)

Curator: @scibot

SciCrunch record: RRID:CVCL_0574

What is this?

适合做某事

Author response:

The following is the authors’ response to the original reviews.

Public Reviews:

Reviewer #1 (Public Review):

Summary:

Previous studies have shown that treatment with 17α-estradiol (a stereoisomer of the 17β-estradiol) extends lifespan in male mice but not in females. The current study by Li et al, aimed to identify cell-specific clusters and populations in the hypothalamus of aged male rats treated with 17α-estradiol (treated for 6 months). This study identifies genes and pathways affected by 17α-estradiol in the aged hypothalamus.

Strengths:

Using single-nucleus transcriptomic sequencing (snRNA-seq) on the hypothalamus from aged male rats treated with 17α-estradiol they show that 17α-estradiol significantly attenuated age-related increases in cellular metabolism, stress, and decreased synaptic activity in neurons.

Thanks.

Moreover, sc-analysis identified GnRH as one of the key mediators of 17α-estradiol's effects on energy homeostasis. Furthermore, they show that CRH neurons exhibited a senescent phenotype, suggesting a potential side effect of the 17α-estradiol. These conclusions are supported by supervised clustering by neuropeptides, hormones, and their receptors.

Thanks.

Weaknesses:

However, the study has several limitations that reduce the strength of the key claims in the manuscript. In particular:

(1) The study focused only on males and did not include comparisons with females. However, previous studies have shown that 17α-estradiol extends lifespan in a sex-specific manner in mice, affecting males but not females. Without the comparison with the female data, it's difficult to assess its relevance to the lifespan.

This study was originally designed based on previous findings indicating that lifespan extension is only effective in males, leading to the exclusion of females from the analysis. The primary focus of our research was on the transcriptional changes and serum endocrine alterations induced by 17α-estradiol in aged males compared to untreated aged males. We believe that even in the absence of female subjects, the significant effects of 17α-estradiol on metabolism in the hypothalamus, synapses, and endocrine system remain evident, particularly regarding the expression levels of GnRH and testosterone. Notably, lower overall metabolism, increased synaptic activity, and elevated levels of GnRH and testosterone are strong indicators of health and well-being in males, supporting the validity of our primary conclusions. However, including female controls would enhance the depth of our findings. If female controls were incorporated, we propose redesigning the sample groups to include aged male control, aged female control, aged female treated, aged male treated, as well as young male control, young male treated, young female control, and young female treated. We regret that we cannot provide this data in the short term. Nevertheless, we believe this presents a valuable avenue for future research on this topic. In this study, we emphasize the role of 17α-estradiol in overall metabolism, synaptic function, GnRH, and testosterone in aged males and underscore the importance of supervised clustering of neuropeptide-secreting neurons in the hypothalamus.

(2) It is not known whether 17α-estradiol leads to lifespan extension in male rats similar to male mice. Therefore, it is not possible to conclude that the observed effects in the hypothalamus, are linked to the lifespan extension.

Thanks for the reminding. 17α-estradiol was reported to extend lifespan in male rats similar to male mice (PMID: 33289482). We have added the valuable reference to introduction in the new version.  

(3) The effect of 17α-estradiol on non-neuronal cells such as microglia and astrocytes is not well-described (Figure 1). Previous studies demonstrated that 17α-estradiol reduces microgliosis and astrogliosis in the hypothalamus of aged male mice. Current data suggest that the proportion of oligo, and microglia were increased by the drug treatment, while the proportions of astrocytes were decreased. These data might suggest possible species differences, differences in the treatment regimen, or differences in drug efficiency. This has to be discussed.

We have reviewed reports describing changes in cell numbers following 17α-estradiol treatment in the brain, using the keywords "17α-estradiol," "17alpha-estradiol," and "microglia" or "astrocyte." Only a limited amount of data was obtained. We found one article indicating that 17α-estradiol treatment in Tg (AβPP(swe)/PS1(ΔE9)) model mice resulted in a decreased microglial cell number compared to the placebo (AβPP(swe)/PS1(ΔE9) mice), but this change was not significant when compared to the non-transgenic control (PMID: 21157032). The transgenic AβPP(swe)/PS1(ΔE9) mouse model may differ from our wild-type aging rat model in this context.

Moreover, the calculation of cell numbers was based on visual observation under a microscope across several brain tissue slices. This traditional method often yields controversial results. For example, oligodendrocytes in the corpus callosum, fornix, and spinal cord have been reported to be 20-40% more numerous in males than in females based on microscopic observations (PMID: 16452667). In contrast, another study found no significant difference in the number of oligodendrocytes between sexes when using immunohistochemistry staining (PMID: 18709647). Such discrepancies arising from traditional observational methods are inevitable.

We believe the data presented in this article are reliable because the cell number and cell ratio data were derived from high-throughput cell counting of the entire hypothalamus using single-cell suspension and droplet wrapping (10x Genomics).

(4) A more detailed analysis of glial cell types within the hypothalamus in response to drugs should be provided.

We provided more enrichment analysis data of differentially expressed genes between Y, O, and O.T in microglia and astrocytes in Figure 2—figure supplement 3. In this supplemental data, we found unlike that in neurons, Micro displayed lower levels of synapse-related cellular processes in O.T. compared to O.

(5) The conclusion that CRH neurons are going into senescence is not clearly supported by the data. A more detailed analysis of the hypothalamus such as histological examination to assess cellular senescence markers in CRH neurons, is needed to support this claim.

We also noticed the inappropriate claim and we have changed "senescent phenotype" to "stressed phenotype" and "abnormal phenotype" in abstract and in results.

Reviewer #2 (Public Review):

Summary:

Li et al. investigated the potential anti-ageing role of 17α-Estradiol on the hypothalamus of aged rats. To achieve this, they employed a very sophisticated method for single-cell genomic analysis that allowed them to analyze effects on various groups of neurons and non-neuronal cells. They were able to sub-categorize neurons according to their capacity to produce specific neurotransmitters, receptors, or hormones. They found that 17α-Estradiol treatment led to an improvement in several factors related to metabolism and synaptic transmission by bringing the expression levels of many of the genes of these pathways closer or to the same levels as those of young rats, reversing the ageing effect. Interestingly, among all neuronal groups, the proportion of Oxytocin-expressing neurons seems to be the one most significantly changing after treatment with 17α-Estradiol, suggesting an important role of these neurons in mediating its anti-ageing effects. This was also supported by an increase in circulating levels of oxytocin. It was also found that gene expression of corticotropin-releasing hormone neurons was significantly impacted by 17α-Estradiol even though it was not different between aged and young rats, suggesting that these neurons could be responsible for side effects related to this treatment. This article revealed some potential targets that should be further investigated in future studies regarding the role of 17α-Estradiol treatment in aged males.

Strengths:

(1) Single-nucleus mRNA sequencing is a very powerful method for gene expression analysis and clustering. The supervised clustering of neurons was very helpful in revealing otherwise invisible differences between neuronal groups and helped identify specific neuronal populations as targets.

Thanks.

(2) There is a variety of functions used that allow the differential analysis of a very complex type of data. This led to a better comparison between the different groups on many levels.

Thanks.

(3) There were some physiological parameters measured such as circulating hormone levels that helped the interpretation of the effects of the changes in hypothalamic gene expression.

Thanks.

Weaknesses

(1) One main control group is missing from the study, the young males treated with 17α-Estradiol.

Given that the treatment period lasts six months, which extends beyond the young male rats' age range, we aimed to investigate the perturbation of 17α-Estradiol on the normal aging process. Including data from young males could potentially obscure the treatment's effects in aged males due to age effects, though similar effects between young and aged animals may exist. Long-term treatment of hormone may exert more developmental effects on the young than the old. Consequently, we decided to exclude this group from our initial sample design. We apologize for this omission.

(2) Even though the technical approach is a sophisticated one, analyzing the whole rat hypothalamus instead of specific nuclei or subregions makes the study weaker.

The precise targets of 17α-Estradiol within the hypothalamus remain unresolved. Selecting a specific nucleus for study is challenging. The supervised clustering method described in this manuscript allows us to identify the more sensitive neuron subtypes influenced by 17α-Estradiol and aging across the entire hypothalamus, without the need to isolate specific nuclei in a disturbed hypothalamic environment.

(3) Although the authors claim to have several findings, the data fail to support these claims. You may mean the claim as the senescent phenotype in Crh neuron induced by 17a-estradiol.

Thanks. We have changed the "senescent phenotype" to "stressed phenotype"  or "abnormal phenotype" in the abstract and results to avoid such claim.

(4) The study is about improving ageing but no physiological data from the study demonstrated such a claim with the exception of the testes histology which was not properly analyzed and was not even significantly different between the groups.

The primary objective of this study is to elucidate the effects of 17α-Estradiol on the endocrine system in the aging hypothalamus; exploring anti-aging effects is not the main focus. From the characteristics of the aging hypothalamus, we know that down-regulated GnRH and testosterone levels, along with elevated mTOR signaling, are indicators of aging in these organs (PMID: 37886966, PMID: 37048056, PMID: 22884327). The contrasting signaling networks related to metabolism and synaptic processes significantly differentiate young and aging hypothalami, and 17α-Estradiol helps rebalance these networks, suggesting its potential anti-aging effects.

(5) Overall, the study remains descriptive with no physiological data to demonstrate that any of the effects on hypothalamic gene expression are related to metabolic, synaptic, or other functions.

The study focuses on investigating cellular responses and endocrine changes in the aging hypothalamus induced by 17α-estradiol, utilizing single-nucleus RNA sequencing (snRNA-seq) and a novel data mining methodology to analyze various neuron subtypes. It is important to note that this study does not mainly aim to explore the anti-aging effects. Consequently, we have revised the claim in the abstract from “the effects of 17α-estradiol in anti-aging in neurons” to “the effects of 17α-estradiol on aging neurons.” We observed that the lower overall metabolism and increased expression levels of cellular processes in the synapses align with findings previously reported regarding 17α-estradiol. To address the lack of physiological data and the challenges in measuring multiple endocrine factors due to their volatile nature, we employed several bidirectional Mendelian analyses of various genome-wide association study (GWAS) data related to these serum endocrine factors to identify their mutual causal effects.

Reviewing Editor Comment:

Based on the Public Reviews and Recommendations for Authors, the Reviewers strongly recommend that revisions include an experimental demonstration of the physiological effects of the treatment on ageing in rats as well as the CRH-senescence link. Additional analysis of the glia would greatly strengthen the study, as would inclusion of females and young male controls. The important point was also raised that the work linking 17a-estradiol was performed in mice, and the link with lifespan in rats is not known. Discussion of this point is recommended.

We acknowledge that 17α-estradiol has been reported to extend lifespan in male rats, similar to findings in male mice (PMID: 33289482), and we have noted this in the Introduction. We apologize for not conducting further experiments to validate this point.

Additionally, we have revised the description of the phenotype of senescent CRH neurons to “stressed phenotype” without carrying out further experiments to confirm the senescent phenotype. To provide more clarity on the performance of glial cells during treatment, we have included additional enrichment analysis data of differentially expressed genes among young (Y), old (O), and old treated (O.T) microglia and astrocytes in Figure 2—figure supplement 3. Notably, the behavior of microglia contrasts with that of total neurons concerning synapse-related cellular processes. We apologize for being unable to include female and young controls in this study.

Reviewer #2 (Recommendations For The Authors)

General comments:

(1) The manuscript is very hard to read. Proofreading and editing by software or a professional seems necessary. The words "enhanced", "extensive" etc. are not always used in the right way.

Thanks for the suggestion. We have revised the proofreading and editing. The words "enhanced" and "extensive" were also revised in most sentences.

(2) The numbers of animals and samples are not well explained. Is it 9 rats overall or per group? If there are 8 testes samples per group, should we assume that there were 4 rats per group? The pooling of the hypothalamic how was it done? Were all the hypothalamic from each group pooled together? A small table with the animals per group and the samples would help.

We appreciate your reminder regarding the initial mistake in our manuscript preparation. In the preliminary submission, we reported 9 rats based solely on sequencing data and data mining. The revised version (v1) now includes additional experimental data, with an effective total of 12 animals (4 per group). Unfortunately, we overlooked updating this information in the v1 submission. We have since added detailed information in the Materials and Methods sections: Animals, Treatment and Tissues, and snRNA-seq Data Processing, Batch Effect Correction, and Cell Subset Annotation.

(3) The Clustering is wrong. There are genes in there that do not fall into any of the 3 categories: Neurotransmitters, Receptors, Hormones.

We have changed the description to “Vast majority of these subtypes were clustered by neuropeptides, hormones, and their receptors within all the neurons”.

(4) The coloring of groups in the graphs is inconsistent. It must be more homogeneous to make it easier to identify.

We have changed the colors of groups in Fig. 1D to make the color of cell clusters consistent in Fig. 1A-D.

(5) The groups c1-c4 are not well explained. How did the authors come up with these?

We have added more descriptions of c1-c4 in materials and methods in the new version.

(6) In most cases it's not clear if the authors are talking about cell numbers that express a certain mRNA, the level of expression of a certain mRNA, or both. They need to do a better job using more precise descriptions instead of using general terms such as "signatures", "expression profiles", "affected neurons" etc. It is very hard to understand if the number of neurons is compared between the groups or the gene expression.

We have changed the "signatures" to "gene signatures" to make it more accurate in meaning. The "affected neurons" were also changed to "sensitive neurons". But sorry that we were not able to find better alternatives to the "expression profiles".

(7) Sometimes there are claims made without justification or a reference. For example, the claim about the senescence of CRH neurons due to the upregulation of mitochondrial genes and downregulation of adherence junction genes (lines 326-328) should be supported by a reference or own findings.

The "senescence" here is not appropriate. We have changed it to "stressed phenotype" or "aberrant changes" in abstract and results.

(8) Young males treated with Estradiol as a control group is necessary and it is missing.

Your suggestion is appreciated; however, the treatment duration for aged mice (O.T) was set at 6 months, while the young mice were only 4 months old. This disparity makes it challenging to align treatment timelines for the young animals. The primary aim of this study is to investigate the perturbation of 17α-estradiol on the aging process, and any distinct effects due to age effect observed in young males might complicate our understanding of its role in aged males, though similar endocrine effects may exist in the young animals. Long-term treatment of hormone may exert more developmental effects on the young than the old. Therefore, we made the decision to exclude the young samples in our initial study design. We apologize for any confusion this may have caused.

Specific Comments:

Line 28: "elevated stresses and decreased synaptic activity": Please make this clearer. Can't claim changes in synaptic activity by gene expression.

We have changed it to "the expression level of pathways involved in synapse".

Line 32: "increased Oxytocin": serum Oxytocin.

We have added the “serum”.

Line 52 - 54: Any studies from rats?

Thanks. In rats there is also reported that 17α-estradiol has similar metabolic roles as that in mice (PMID: 33289482) and we have added it to the refences. It’s very useful for this manuscript.

Line 62 - 65: It wasn't investigated thoroughly in this paper so why was it suggested in the introduction?

We have deleted this sentence as being suggested.

Line 70: "synaptic activity" Same as line 28.

We have changed it to "pathways involved in synaptic activity".

Line 79: Why were aged rats caged alone and young by two? Could that introduce hypothalamic gene expression effects?

The young males were bred together in peace. But the aged males will fight and should be kept alone.

Lines 78, 99, 109-110: It is not clear how many animals per group were used and how many samples per group were used separately and/or grouped. Please be more specific.

We have added these information to Materials and methods/Animals, treatment and tissues and Materials and methods/snRNA-seq data processing, batch effect correction, and cell subset annotation.

Line 205: "in O" please add "versus young.".

We have changed accordingly.

Line 207: replace "were" with "was" .

We have alternatively changed the "proportion" to "proportions".

Line 208: replace "that" with "compared to" and after "in O.T." add "compared to?"

We have changed accordingly.

Line 223: "O.T." compared to what? Figure?

We have changed it accordingly.

Line 227: Figure?

We have added (Figure 1E) accordingly.

Line 229: "synaptic activity" Same as line 28.

We have revised it.

Line 235: "synaptic activity" and "neuropeptide secretion" Same as line 28.

We have revised it.

Line 256:" interfered" please revise.

We changed to "exerted".

Line 263: "on the contrary" please revise.

We have changed "on the contrary" to "opposite".

Line 270: "conversed" did you mean "conserved"?

We have changed "conversed" to "inversed".

Line 296-298: Please explain. Why would these be side effects?

It’s hard to explain, therefore, we deleted the words "side effects".

Line 308: "synaptic activity" Same as line 28.

We have changed it to "expression levels of synapse-related cellular processes".

Line 314: "and sex hormone secretion and signaling"Isn't this expected?

Yes, it is expected. We have added it to the sentence "and, as expected, sex hormone secretion and signaling".

Line 325-328: Why is this senescence? Reference?

We have added “potent” to it.

Line 360-361: This doesn't show elevated synaptic activity.

"elevated synaptic activity" was changed to "The elevated expression of synapse-related pathways"

Line 363-364: "Unfortunately" is not a scientific expression and show bias.

We have changed it to "Notably".

Line 376: Similar as above.

Yes, we have change it to "in contrast".

Lines 382-385: This is speculation. Please move to discussion.

Sorry for that. We think the causal effects derived from MR result is evidence. As such, we have not changed it.

Line 389: Please revise "hormone expressing".

We have changed it accordingly.

Line 401: Isn't this effect expected due to feedback inhibition of the biochemical pathway? Please comment.

The binding capability of 17alpha-estradiol to estrogen receptors and its role in transcriptional activation remain core questions surrounded by controversy. Earlier studies suggest that 17alpha-estradiol exhibits at least 200 times less activity than 17beta-estradiol (PMID: 2249627, PMID: 16024755). However, recent data indicate that 17alpha-estradiol shows comparable genomic binding and transcriptional activation through estrogen receptor α (Esr1) to that of 17beta-estradiol (PMID: 33289482). Additionally, there is evidence that 17alpha-estradiol has anti-estrogenic effects in rats (PMID: 16042770). These findings imply possible feedback inhibition via estrogen receptors. Furthermore, 17alpha-estradiol likely differs from 17beta-estradiol due to its unique metabolic consequences and its potential to slow aging in males, an effect not attributed to 17beta-estradiol. For instance, neurons are also targets of 17alpha-estradiol, with Esr1 not being the sole target (PMID: 38776045). Nevertheless, the precise effective targets of 17alpha-estradiol are still unresolved.

Line 409: This conclusion cannot be made because the effect is not statistically significant. Can say "trend" etc.

Thanks for the recommendation. We have added "potential" in front of the conclusion.

Line 426: "suggesting" please revise.

sorry, it’s a verb.

Lines 426-428: This is speculation. Please move to discussion.

The elevated GnRH levels in O.T., observed through EIA analysis, suggest a deduction regarding the direct causal effects of 17alpha-estradiol on various endocrine factors related to feeding, energy homeostasis, reproduction, osmotic regulation, stress response, and neuronal plasticity through MR analysis. Thus, we have not amended our position. We apologize for any confusion.

Lines 431-432: improved compared to what?

The statement have been revised as " The most striking role of 17α-estradiol treatment revealed in this study showed that HPG axis was substantially improved in the levels of serum Gnrh and testosterone".

Line 435: " Estrogen Receptor Antagonists". Please revise.

Thanks for the recommendation. We have changed it to "estrogen receptor antagonists".

Line 438" "Secrete". Please revise.

Sorry, it is "secret".

Lines 439-449: None of this has been demonstrated. Please remove these conclusions.

These are not conclusions but rather intriguing topics for discussion. Given the role of 17alpha-estradiol in promoting testosterone and reducing estradiol levels in males, we believe it is worthwhile to explore the potential application of 17alpha-estradiol in increasing testosterone levels in aged males, particularly those with hypogonadism.

Lines 450-457: No females were included in this study. Why? Also, why is this discussed? It is relevant but doesn't belong in this manuscript since it was not studied here.

Testosterone levels are crucial for male health, while estradiol levels are essential for the health and fertility of females. Previous studies have demonstrated that 17α-estradiol does not contribute to lifespan extension in females. Given the effects of 17α-estradiol on males—specifically, its role in promoting testosterone and reducing estradiol levels—we believe it is important to discuss the potential sex-biased effects of 17α-estradiol, as this could inform future investigations. Therefore, we have chosen not to make changes to this section.

Lines 458-459: This was not demonstrated in this article. Please remove.

We have restricted the claim to "expression level of energy metabolism in hypothalamic neurons".

Line 464: "Promoted lifespan extension" Not demonstrated. Please remove.

At the end of the sentence it was revised as "which may be a contributing factor in promoting lifespan extension".

Line 466: "Showed" No.

The whole sentence was deleted in the new version.

Line 483: "the sex-based effects". Not studied here.

Since the changes in testosterone levels are significant in this dataset and this hormone has a sex-biased nature, we find it worthwhile to suggest this as a topic for future investigation. We have added "which needs further verification in the future" at the end of this sentence.

Author response:

eLife Assessment<br /> This valuable study suggests that Naa10, an N-α-acetyltransferase with known mutations that disrupt neurodevelopment, acetylates Btbd3, which has been implicated in neurite outgrowth and obsessive-compulsive disorder, in a manner that regulates F-actin dynamics to facilitate neurite outgrowth. While the study provides promising insights and biochemical, co-immunoprecipitation, and proteomic data that enhance our understanding of protein N-acetylation in neuronal development, the evidence supporting larger claims is incomplete. Nonetheless, the implications of these findings are noteworthy, particularly regarding neurodevelopmental and psychiatric conditions tied to altered expression of Naa10 or Btbd3.

Thank you very much for recognizing our study, carefully reviewing our work, and providing insightful comments and constructive criticism!

Public Reviews:

Reviewer #1 (Public review):

The manuscript examines the role of Naa10 in cKO animals, in immortalized neurons, and in primary neurons. Given that Naa10 mutations in humans produce defects in nervous system function, the authors used various strategies to try to find a relevant neuronal phenotype and its potential molecular mechanism.

This work contains valuable findings that suggest that the depletion of Naa10 from CA1 neurons in mice exacerbates anxiety-like behaviors. Using neuronal-derived cell lines authors establish a link between N-acetylase activity, Btbd3 binding to CapZb, and F-actin, ultimately impinging on neurite extension. The evidence demonstrating this is in most cases incomplete, since some key controls are missing and clearly described or simply because claims are not supported by the data. The manuscript also contains biochemical, co-immunoprecipitation, and proteomic data that will certainly be of value to our knowledge of the effects of protein N--acetylation in neuronal development and function.

Thanks! It would be appreciated if the Reviewer could point out in the public review which experiment lacks a control group.

Reviewer #2 (Public review):

In this study, the authors sought to elucidate the neural mechanisms underlying the role of Naa10 in neurodevelopmental disruptions with a focus on its role in the hippocampus. The authors use an impressive array of techniques to identify a chain of events that occurs in the signaling pathway starting from Naa10 acetylating Btbd3 to regulation of F-actin dynamics that are fundamental to neurite outgrowth. They provide convincing evidence that Naa10 acetylates Btbd3, that Btbd3 facilitates CapZb binding to F-actin in a Naa10 acetylation-dependent manner, and that this CapZb binding to F-actin is key to neurite outgrowth. Besides establishing this signaling pathway, the authors contribute novel lists of Naa10 and Btbd3 interacting partners, which will be useful for future investigations into other mechanisms of action of Naa10 or Btbd3 through alternative cell signaling pathways.

Thank you very much for recognizing our study!

The evidence presented for an anxiety-like behavioral phenotype as a result of Naa10 dysfunction is mixed and tenuous, and assays for the primary behaviors known to be altered by Naa10 mutations in humans were not tested. As such, behavioral findings and their translational implications should be interpreted with caution.

(1) For the anxiety-like behavioral phenotype, we provided a paragraph titled “Naa10 and stress-induced anxiety” in the Discussion section of the text: “Our investigations revealed that hippocampal CA1-KO of Naa10 did not exhibit significant differences in the open field test (Figure S1K) but led to anxiety-like behavior in mice in the elevated plus maze (EPM) test (Figure 1A). This disparity might be attributed to the specific design of the EPM test, which is tailored to elicit a conflict between an animal's inclination to explore and its fear of open spaces and elevated areas. This distinction implies that Naa10 might play a role in stress responses within the emotional regulation circuitry, particularly in navigating potentially threatening and anxiety-provoking environments.” The open field test offers a less challenging, open environment that primarily promotes exploratory behavior. We agree that additional assays, such as the light-dark box test, would be helpful in clarifying the issue.

(2) We agree that the behavioral findings and their translational implications should be interpreted with caution. The primary neurological behaviors known to be altered by Naa10 mutations in humans include intellectual disability and autism-like syndrome with defective emotional control. These behaviors are influenced by many factors, including defects in the hippocampal CA1. Thus, we tested hippocampal CA1 Naa10-KO mice using the Y-maze, tail suspension test, open field test, and elevated plus maze (EPM). However, only the EPM results were affected, while the other tests showed no significant changes. It should be noted that our study employed a postnatal, CA1-specific Naa10 conditional knockout (cKO) model driven by Camk2a-Cre, which selectively depletes Naa10 from hippocampal CA1 neurons after birth. In contrast, Naa10 mutations in human patients involve global effects and impact multiple brain regions from the embryonic stage, leading to a broader spectrum of phenotypes. The limited disruption in our model likely explains the absence of learning and memory deficits and the incomplete recapitulation of the full range of patient phenotypes. Furthermore, Naa10 knockout may not produce the same effects as Naa10 mutations. Our current study is primarily intended to explore the physiological function of Naa10 in hippocampal function.

(3) We will replace all instances of “anxiety behavior” with “anxiety-like behavior.”

Finally, while not central to the main cell signaling pathway delineated, the characterization of brain region-specific and cell maturity of Naa10 expression patterns was presented in few to single animals and not quantified, and as such should also be interpreted with caution.

We agree that we should provide additional Naa10 immunostaining data from more than three WT and hippocampal CA1 Naa10-KO mouse brains, as well as quantify data such as the silver staining and Light Sheet Fluorescence Microscopy results presented in Figures 1C and 1D, respectively. Nevertheless, the current report presents consistent results across different mice used for various assays. For example, Figures 1B-D, with three different assays, each demonstrate that Naa10-cKO reduces neurite complexity in vivo.

On a broader level, these findings have implications for neurodevelopment and potentially, although not tested here, synaptic plasticity in adulthood, which means this novel pathway may be fundamental for brain health.

Thank you very much again for recognizing our study!

Summarized list of minor concerns

(1) The early claims of the manuscript are supported by very small sample sizes (often 1-3) and/or lack of quantification, particularly in Figures S1 and 1.

We agree that we should provide additional Naa10 immunostaining data from more than three WT and hippocampal CA1 Naa10-KO mouse brains, as well as quantify data such as the silver staining and Light Sheet Fluorescence Microscopy results presented in Figures 1C and 1D, respectively. Nevertheless, the current report presents consistent results across different mice used for various assays. For example, Figures 1B-D, with three different assays, each demonstrate that Naa10-cKO reduces neurite complexity in vivo.

(2) Evidence is insufficient for CA1-specific knockdown of Naa10.

The Camk2a-Cre mice used in this study were derived from Dr. Susumu Tonegawa’s laboratory. According to the referenced paper, this strain restricts Cre/loxP recombination to the forebrain, with particularly high efficiency in the hippocampal CA1. Consistently, our data show that Naa10 was almost completely absent in the CA1 but partially depleted in the DG of the Naa10-cKO mice (Figure S1F in the text). Similar results were observed in a different pair of

(3) The relationship between the behaviors measured, which centered around mood, and Ogden syndrome, was not clear, and likely other behavioral measures would be more translationally relevant for this study. Furthermore, the evidence for an anxiety-like phenotype was mixed.

(1) For the anxiety-like behavioral phenotype, we provided a paragraph titled “Naa10 and stress-induced anxiety” in the Discussion section of the text: “Our investigations revealed that hippocampal CA1-KO of Naa10 did not exhibit significant differences in the open field test (Figure S1K) but led to anxiety-like behavior in mice in the elevated plus maze (EPM) test (Figure 1A). This disparity might be attributed to the specific design of the EPM test, which is tailored to elicit a conflict between an animal's inclination to explore and its fear of open spaces and elevated areas. This distinction implies that Naa10 might play a role in stress responses within the emotional regulation circuitry, particularly in navigating potentially threatening and anxiety-provoking environments.” The open field test offers a less challenging, open environment that primarily promotes exploratory behavior. We agree that additional assays, such as the light-dark box test, would be helpful in clarifying the issue.

(2) We agree that the behavioral findings and their translational implications should be interpreted with caution. The primary neurological behaviors known to be altered by Naa10 mutations in humans include intellectual disability and autism-like syndrome with defective emotional control. These behaviors are influenced by many factors, including defects in the hippocampal CA1. Thus, we tested hippocampal CA1 Naa10-KO mice using the Y-maze, tail suspension test, open field test, and elevated plus maze (EPM). However, only the EPM results were affected, while the other tests showed no significant changes. It should be noted that our study employed a postnatal, CA1-specific Naa10 conditional knockout (cKO) model driven by Camk2a-Cre, which selectively depletes Naa10 from hippocampal CA1 neurons after birth. In contrast, Naa10 mutations in human patients involve global effects and impact multiple brain regions from the embryonic stage, leading to a broader spectrum of phenotypes. The limited disruption in our model likely explains the absence of learning and memory deficits and the incomplete recapitulation of the full range of patient phenotypes. Furthermore, Naa10 knockout may not produce the same effects as Naa10 mutations. Our current study is primarily intended to explore the physiological function of Naa10 in hippocampal function.

(3) We will replace all instances of “anxiety behavior” with “anxiety-like behavior.”

(4) Btbd3 is characterized by the authors as an OCD risk gene, but its status as such is not well supported by the most recent, better-powered genome-wide association studies than the one that originally implicated Btbd3. However, there is evidence that Btbd3 expression, including selectively in the hippocampus, is implicated in OCD-relevant behaviors in mice.

Thanks for clarifying the issue!

(5) The reporting of the statistics lacks sufficient detail for the reader to deduce how experimental replicates were defined.

We believe we have provided sufficient detail for readers to deduce how experimental replicates were defined in each corresponding figure legend. It would be appreciated if the Reviewer could point out which specific figures lack sufficient details.

Author response:

Reviewer #1:

Summary:<br /> In this manuscript, Bisht et al address the hypothesis that protein folding chaperones may be implicated in aggregopathies and in particular Tau aggregation, as a means to identify novel therapeutic routes for these largely neurodegenerative conditions.

The authors conducted a genetic screen in the Drosophila eye, which facilitates the identification of mutations that either enhance or suppress a visible disturbance in the nearly crystalline organization of the compound eye. They screened by RNA interference all 64 known Drosophila chaperones and revealed that mutations in 20 of them exaggerate the Tau-dependent phenotype, while 15 ameliorated it. The enhancer of the degeneration group included 2 subunits of the typically heterohexameric prefoldin complex and other co-translational chaperones.

In a previous paper, we identified 95 Drosophila chaperones (Raut et al., 2017). We request that “all 64 known Drosophila chaperones” be replaced with “64 out of 95 known Drosophila chaperones” to make it factually correct.

Strengths:

Regarding this memory defect upon V377M tau expression. Kosmidis et al (2010) pmid: 20071510, demonstrated that pan-neuronal expression of TauV377M disrupts the organization of the mushroom bodies, the seat of long-term memory in odor/shock and odor/reward conditioning. If the novel memory assay the authors use depends on the adult brain structures, then the memory deficit can be explained in this manner.

If the mushroom bodies are defective upon TauV377M expression does overexpression of Pfdn5 or 6 reverse this deficit? This would argue strongly in favor of the microtubule stabilization explanation.

We agree that the disruptive organization of the mushroom body may cause memory deficits upon hTauV337M expression and that expression of Pfdn5 or Pfdn6 could reverse the deficits. One possible mechanism by which overexpression of Pfdn5/6 could rescue the Tau-induced memory deficits may be due to the stabilization of microtubules in the mushroom bodies.

Proposed revision: We will assess if Tau-induced mushroom body disruption can be rescued with the overexpression of Pfdn5 or Pfdn6.

Weakness:

What is unclear however is how Pfdn5 loss or even overexpression affects the pathological Tau phenotypes. Does Pfdn5 (or 6) interact directly with TauV377M? Colocalization within tissues is a start, but immunoprecipitations would provide additional independent evidence that this is so.

Our data suggests that Pfdn5 stabilizes neuronal microtubules by directly associating with it, and loss of Pfdn5 exacerbates Tau-phenotypes by destabilizing microtubules. However, as the reviewer notes, analysis of direct interaction between Pfdn5 and hTau^V337M might provide further insights into the mechanism of Pfdn5 and Tau-aggregation.

Proposed revision: We will perform colocalization analysis and coimmunoprecipitation to ask if Pfdn5 colocalizes and directly interacts with Tau.

Does Pfdn5 loss exacerbate TauV377M phenotypes because it destabilizes microtubules, which are already at least partially destabilized by Tau expression? Rescue of the phenotypes by overexpression of Pfdn5 agrees with this notion.

However, Cowan et al (2010) pmid: 20617325 demonstrated that wild-type Tau accumulation in larval motor neurons indeed destabilizes microtubules in a Tau phosphorylation-dependent manner. So, is TauV377M hyperphosphorylated in the larvae?? What happens to TauV377M phosphorylation when Pfdn5 is missing and presumably more Tau is soluble and subject to hyperphosphorylation as predicted by the above?

Proposed revisions: We will overexpress Pfdn5 or Pfdn6 with hTau^V337M and ask if microtubule disruption caused by hTau^V337M is rescued. Further, we will analyze the phospho-Tau levels in controls and Pfdn5 mutant background.

Expression of WT human Tau (which is associated with most common Tauopathies other than FTDP-17) as Cowan et al suggest has significant effects on microtubule stability, but such Tau-expressing larvae are largely viable. Will one mutant copy of the Pfdn5 knockout enhance the phenotype of these larvae?? Will it result in lethality? Such data will serve to generalize the effects of Pfdn5 beyond the two FDTP-17 mutations utilized.

Proposed revision: We will incorporate data about the effect of heterozygous mutation of Pfdn5 on the lethality and synaptic phenotypes associated with the hTau^WT and hTau^V337M in the revised manuscript.

Does the loss of Pfdn5 affect TauV377M (and WTTau) levels?? Could the loss of Pfdn5 simply result in increased Tau levels? And conversely, does overexpression of Pfdn5 or 6 reduce Tau levels?? This would explain the enhancement and suppression of TauV377M (and possibly WT Tau) phenotypes. It is an easily addressed, trivial explanation at the observational level, which if true begs for a distinct mechanistic approach.

We thank the reviewer for suggesting an alternate model for the Pfdn5 function. We will perform the Western blot analysis to assess Tau^WT and Tau^V337M levels in the absence of Pfdn5 or animals coexpressing Tau and Pfdn5. We will incorporate these data and conclusions in the revised manuscript.

Finally, the authors argue that TauV377M forms aggregates in the larval brain based on large puncta observed especially upon loss of Pfdn5. This may be so, but protocols are available to validate this molecularly the presence of insoluble Tau aggregates (for example, pmid: 36868851) or soluble Tau oligomers as these apparently differentially affect Tau toxicity. Does Pfdn5 loss exaggerate the toxic oligomers and overexpression promotes the more benign large aggregates??

We will perform the Tau solubility assay in control, in the absence of Pfdn5 or animals coexpressing Tau and Pfdn5. Moreover, we will also ask if the large Tau puncta formed in the absence of Pfdn5 are soluble oligomers or stable aggregates. We have found that the coexpression of Tau and Pfdn5 does not result in the formation of  Tau aggregates. We will incorporate these and other relevant data in the revised manuscript.

Reviewer #2 (Public review):

Bisht et al detail a novel interaction between the chaperone, Prefoldin 5, microtubules, and tau-mediated neurodegeneration, with potential relevance for Alzheimer's disease and other tauopathies. Using Drosophila, the study shows that Pfdn5 is a microtubule-associated protein, which regulates tubulin monomer levels and can stabilize microtubule filaments in the axons of peripheral nerves. The work further suggests that Pfdn5/6 may antagonize Tau aggregation and neurotoxicity. While the overall findings may be of interest to those investigating the axonal and synaptic cytoskeleton, the detailed mechanisms for the observed phenotypes remain unresolved and the translational relevance for tauopathy pathogenesis is yet to be established. Further, a number of key controls and important experiments are missing that are needed to fully interpret the findings.The major weakness relates to the experiments and claims of interactions with Tau-mediated neurodegeneration. In particular, it is unclear whether knockdown of Pfdn5 may cause eye phenotypes independent of Tau. Further, the GMR>tau phenotype appears to have been incorrectly utilized to examine age-dependent, neurodegeneration.

We have consistently found the progression of eye degeneration in the population of animals expressing Tau^V337M, measured as the number of fused ommatidia/total number of ommatidia, with age. A few other studies have also shown age-dependent progressive degeneration in Drosophila retinal axons or lamina (Iijima-Ando et al., 2012; Sakakibara et al., 2018). We appreciate other studies that have proposed hTau-induced eye degeneration as a developmental defect (Malmanche et al., 2017; Sakakibara et al., 2023).

Proposed revision: a) We will analyze the age-dependent neurodegeneration in the adult brain to further support our main conclusion that Pfdn5 ameliorates hTauV337M-induced progressive neurodegeneration.

b) We have used three independent Pfdn5 RNAi lines (the RNAi's target different regions of Pfdn5) – all of which enhance the Tau phenotypes. The knockdown of any of these RNAi lines with GMR-Gal4 does not give detectable eye phenotypes. We will include these data in the revised manuscript.

This manuscript argues that its findings may be relevant to thinking about mechanisms and therapies applicable to tauopathies; however, this is premature given that many questions remain about the interactions from Drosophila, the detailed mechanisms remain unresolved, and absent evidence that tau and Pfdn may similarly interact in the mammalian neuronal context. Therefore, this work would be strongly enhanced by experiments in human or murine neuronal culture or supportive evidence from analyses of human data.

Proteome analysis of Alzheimer's brain tissue shows that the Pfdn5 level is reduced in patients (Askenazi et al., 2023; Tao et al., 2020). Moreover, the Pfdn5 expression level was found to be reduced in the blood samples from AD patients (Ji et al., 2022). Another study further validates the age-dependent reduction of Pfdn5 in the tauopathy transgenic murine model (Kadoyama et al., 2019). Together, these reports highlight a potential link between Pfdn5 levels and tauopathies. We will revise the manuscript to reflect these findings in more detail.

References

Askenazi, M., Kavanagh, T., Pires, G., Ueberheide, B., Wisniewski, T., and Drummond, E. (2023). Compilation of reported protein changes in the brain in Alzheimer's disease. Nat Commun 14, 4466. 10.1038/s41467-023-40208-x.

Iijima-Ando, K., Sekiya, M., Maruko-Otake, A., Ohtake, Y., Suzuki, E., Lu, B., and Iijima, K.M. (2012). Loss of axonal mitochondria promotes tau-mediated neurodegeneration and Alzheimer's disease-related tau phosphorylation via PAR-1. PLoS Genet 8, e1002918. 10.1371/journal.pgen.1002918.

Ji, W., An, K., Wang, C., and Wang, S. (2022). Bioinformatics analysis of diagnostic biomarkers for Alzheimer's disease in peripheral blood based on sex differences and support vector machine algorithm. Hereditas 159, 38. 10.1186/s41065-022-00252-x.

Kadoyama, K., Matsuura, K., Takano, M., Maekura, K., Inoue, Y., and Matsuyama, S. (2019). Changes in the expression of prefoldin subunit 5 depending on synaptic plasticity in the mouse hippocampus. Neurosci Lett 712, 134484. 10.1016/j.neulet.2019.134484.

Malmanche, N., Dourlen, P., Gistelinck, M., Demiautte, F., Link, N., Dupont, C., Vanden Broeck, L., Werkmeister, E., Amouyel, P., Bongiovanni, A., et al. (2017). Developmental Expression of 4-Repeat-Tau Induces Neuronal Aneuploidy in Drosophila Tauopathy Models. Sci Rep 7, 40764. 10.1038/srep40764.

Raut, S., Mallik, B., Parichha, A., Amrutha, V., Sahi, C., and Kumar, V. (2017). RNAi-Mediated Reverse Genetic Screen Identified Drosophila Chaperones Regulating Eye and Neuromuscular Junction Morphology. G3 (Bethesda) 7, 2023-2038. 10.1534/g3.117.041632.

Sakakibara, Y., Sekiya, M., Fujisaki, N., Quan, X., and Iijima, K.M. (2018). Knockdown of wfs1, a fly homolog of Wolfram syndrome 1, in the nervous system increases susceptibility to age- and stress-induced neuronal dysfunction and degeneration in Drosophila. PLoS Genet 14, e1007196. 10.1371/journal.pgen.1007196.

Sakakibara, Y., Yamashiro, R., Chikamatsu, S., Hirota, Y., Tsubokawa, Y., Nishijima, R., Takei, K., Sekiya, M., and Iijima, K.M. (2023). Drosophila Toll-9 is induced by aging and neurodegeneration to modulate stress signaling and its deficiency exacerbates tau-mediated neurodegeneration. iScience 26, 105968. 10.1016/j.isci.2023.105968.

Tao, Y., Han, Y., Yu, L., Wang, Q., Leng, S.X., and Zhang, H. (2020). The Predicted Key Molecules, Functions, and Pathways That Bridge Mild Cognitive Impairment (MCI) and Alzheimer's Disease (AD). Front Neurol 11, 233. 10.3389/fneur.2020.00233.

DOI: 10.1186/s40170-024-00366-y

Resource: Cytoscape (RRID:SCR_003032)

Curator: @scibot

SciCrunch record: RRID:SCR_003032

What is this?

DOI: 10.1186/s40170-024-00366-y

Resource: GraphPad Prism (RRID:SCR_002798)

Curator: @scibot

SciCrunch record: RRID:SCR_002798

What is this?

DOI: 10.1186/s40170-024-00366-y

Resource: Gene Expression Omnibus (GEO) (RRID:SCR_005012)

Curator: @scibot

SciCrunch record: RRID:SCR_005012

What is this?

DOI: 10.1186/s40170-024-00366-y

Resource: (R and D Systems Cat# 4354-MC-050, RRID:AB_1857195)

Curator: @scibot

SciCrunch record: RRID:AB_1857195

What is this?

DOI: 10.1186/s40170-024-00366-y

Resource: Ingenuity Pathway Analysis (RRID:SCR_008653)

Curator: @scibot

SciCrunch record: RRID:SCR_008653

What is this?

DOI: 10.1186/s40170-024-00366-y

Resource: CellProfiler Image Analysis Software (RRID:SCR_007358)

Curator: @scibot

SciCrunch record: RRID:SCR_007358

What is this?

DOI: 10.1186/s40170-024-00366-y

Resource: (Thermo Fisher Scientific Cat# A-11008, RRID:AB_143165)

Curator: @scibot

SciCrunch record: RRID:AB_143165

What is this?

DOI: 10.1186/s40170-024-00366-y

Resource: (Millipore Cat# 05-636, RRID:AB_309864)

Curator: @scibot

SciCrunch record: RRID:AB_309864

What is this?

DOI: 10.1186/s40170-024-00366-y

Resource: ImageJ (RRID:SCR_003070)

Curator: @scibot

SciCrunch record: RRID:SCR_003070

What is this?

DOI: 10.1186/s40170-024-00366-y

Resource: None

Curator: @scibot

SciCrunch record: RRID:CVCL_XA19

What is this?

DOI: 10.1186/s40170-024-00366-y

Resource: (KCLB Cat# 70827, RRID:CVCL_2063)

Curator: @scibot

SciCrunch record: RRID:CVCL_2063

What is this?

Why does \(\log S_i(T)\) have the same randomness as \(\log S_i^{'}(T)\)? Do we need \(Z_i^{'}\) for the second simulated price?

also it might help to write \(x(S_i(T)) \) and \(y(S_i^{'}(T)) \) in the mean estimation for clarity.

THESIS: " This StoryMap will outline these places and attempt to demonstrate X, Y, and/or Z...."?

DOI: 10.1186/s13395-024-00366-y

Resource: (Thermo Fisher Scientific Cat# A-11055, RRID:AB_2534102)

Curator: @scibot

SciCrunch record: RRID:AB_2534102

What is this?

DOI: 10.1186/s13395-024-00366-y

Resource: (Molecular Probes Cat# A-21244, RRID:AB_2535812)

Curator: @scibot

SciCrunch record: RRID:AB_2535812

What is this?

DOI: 10.1186/s13395-024-00366-y

Resource: (SouthernBiotech Cat# 1310-01, RRID:AB_2753206)

Curator: @scibot

SciCrunch record: RRID:AB_2753206

What is this?

DOI: 10.1186/s13395-024-00366-y

Resource: (Millipore Cat# 07-1404, RRID:AB_1587080)

Curator: @scibot

SciCrunch record: RRID:AB_1587080

What is this?

DOI: 10.1186/s13395-024-00366-y

Resource: (Sigma-Aldrich Cat# L9393, RRID:AB_477163)

Curator: @scibot

SciCrunch record: RRID:AB_477163

What is this?