This is a really nice series of investigations and a surprising set of findings. If I understand correctly you suggest the bli-3/H2O2 arm of hsp-6p::GFP activation in the two hit model is an active immune response to the E. faecalis. I'm curious why that response occurs for E. faecalis, but not for similar Gram-positive colonizing bacterium such as S. aureus in figure S2A, which might also be expected to draw increased H2O2 secretion.

You mentioned a screen of 500 bacterial strains that showed 18% activate the mtUPR. Do you know how many of those strains are heme auxotrophs? It would be interesting to see if the heme auxotroph arm ever exists without the bli-3/H2O2 arm, or if they are in some way linked

These are very exciting results. It would be wonderful if an injection-free, high-throughput way to generate extrachromosomal arrays or crispr animals were available for C. elegans and this seems like a very promising development in that direction. I had a few questions about the work if you don’t mind:

1. You mentioned GFP fluorescence wasn’t possible using this route. What plasmids/promoters did you use for the GFP experiment e.g. what tissues were you trying to express in, and what level of microscopy did you use to score it? Did you use plasmids that are resistant to silencing, e.g. contain introns or with codon optimization?

2. What route do you picture the DNA is taking to get to the germline? My best guess would be absorption in the intestine and following the lipids into the eggs, but I’m curious what your thoughts are? Could you put some type of tracer dye with the polyplexes to follow the route?

3. Why did you choose 4C as an incubation temperature, and why do you believe that was more successful than room temperature?

4. It sounds like you incubated the animals in pure M9 with the polyplexes. If that’s the case have you considered trying it with food present to stimulate pharyngeal pumping? If the route is through the intestine, this might stimulate more uptake and more plasmid reaching the germ line.

5. Have you tried doing this in nuclease mutants to reduce how much plasmid is lost? Either nuc-1 for luminal/lysosomal loss or an intracellular nuclease mutant?

Thank you for sharing this interesting work!

I love the name

It would be helpful to include more discussion on the differences in your exposure for the OP50 and P.a. (7 days starting at the L4 larval stage for E. coli strain OP50; 2 days starting at day 5 of adulthood for P. aeruginosa strain PA14). I believe this means the animals are 7 day adults when you start the imaging experiment (in your methods you say day 9 of adulthood?). A timeline graphic would help, as would some discussion of the differences in colony formation rates for these two strains. Basically you could highlight that days after challenge does not correspond to days after exposure, as the OP50 strains were challenged for 7 days, while the P.a strain was challenged for 2. Perhaps change the start point to when the challenge began so they are more directly comparable?

Do you ever see the animals clear or reverse the bacterial colonies as you mentioned in the introduction?

It would be interesting to try a dual-color system here for the OP50 bacteria to determine if the colonization is a random chance event or if some worms are more susceptible to it. If you fed worms on GFP-OP50 and scarlet-OP50 you could see if worms are more likely to be co-colonized than expected by chance.

Great findings on the origins of LDs. Do you expect Pex30 levels to change depending on lipid availability/organismal stress? It would be interesting to see if it's upregulated in response to high PA levels

Do you expect these mutations are strong enough to create a dominant negative effect in a wild-type Pex30 background?

I would love to see some of these images in figure 5.

Could you introduce the significance of this background for this experiment?

It would help to add some quantification here to show the extent of this phenotype

I am having trouble seeing the strength of the phenotype, especially for the TMD2 mutant. Could you add some arrows to highlight WT vs abnormal cortical ER morphology?

How labor intensive is this loading process? Are you able to just open the channel briefly and load a set number of animals into the chip or do you have to carefully monitor the flow of the liquid until you see the correct amount of worms and then turn it off? I'm wondering about this as in regards to making this more high-throughput. If the loading can be standardized, do you think it would be possible to keep the worms in a 96 cell plate, with an automatic sampler and bring worms into the chamber one well at a time for imaging, and then either dispose of them or return them to their well for later imaging? e.g. if you were going through a lot of mutants or drugs or doing a mutagenesis screen.

I think this is a very powerful device with a lot of applications. This is a small issue, but the diagram is a bit confusing the first time you look at it. You show the chip between the waste and reagent, and then an image of the chip on the right side of the red monitoring module box. Is that second box what the camera is seeing from the chip and sending to the control center? It's a bit confusing as it looks like there are two different chips in the system.

It would help to show this data in figure 6 to give more detail on how much body length is impacted.

It would be helpful to discuss how these stock concentrations relate to the concentrations humans are exposed to in their environment. Also, a great addition to the paper would be to do longer term studies of these chemicals' impact on worms. This device is a really powerful way to measure impacts on worm health and behavior over long periods and the introduction to the paper is framed from the perspective of human cumulative exposure to toxins. I'd love to see longer time course assays on how these drugs affect the worms, especially at environmentally relevant concentrations.

These lifespans seem shorter than the average in the literature. Do you think the constrained space in the PDMS chip and on the 96-well plate, is reducing their longevity? Also did you check development rates on the chips and do they go through the larval stages at normal rates?

I'm surprised that the scav-5 and scav-6 paralogs were both able to reduce the large LD phenotype to the same extent as scav-4 (there doesn't appear to be significant difference between the mutants). To me this suggests either they each contribute a third of the BCFA uptake, or that they operate together to internalize BCFAs. The scav-4;scav-6 double mutant suggests the first idea isn't correct as you don't see a stronger effect there. Do you think its possible these transporters are working as a complex? I would be interested to see if you can rescue each of these mutants with scav-4 expression, or if rescue requires all receptors to be present.

It's amazing you were able to find a GOF mutation that exacerbated this phenotype! I think it would help accentuate this if you included a picture of wt animals on the JUb74 food, next to the L462F animals on JUB74 food in Fig 4A to contrast that shift in LD size.

Could you preform the experiment in C. elegans defined media supplemented with BCFAs? That would fully remove other variables from the Microbacterium and confirm this is solely due to BCFAs.