On 2014 Mar 10, Daniel Haft commented:

We noted in this paper that only some genomes with type II secretion systems (T2SS) have rhombosortase and a matching collection of GlyGly-CTERM proteins, but essentially all genomes with rhombosortase have T2SS. This comment discusses proteins from the cholera pathogen that are both T2SS targets and GlyGly-CTERM proteins. In Vibrio cholerae, cholera toxin and at least 20 other proteins depend on T2SS for export, including 16 found by Sikora, et al. (PMID:21385872) in 2011. Three targets found in that study, the related trypsin-like serine proteases VCA0803 (VesA), VC1200 (VesB), and VC1649 (VesC), are three of V. cholerae's six GlyGly-CTERM proteins (the others are VCA0065, VC2621, and VC1485). VesA, in particular, may be highly relevant to pathogenesis, since "...analysis of single, double, and triple protease knock-out strains indicated that VesA is the primary protease responsible for processing the A subunit of cholera toxin during in vitro growth of the V. cholerae strain N16961."

The GlyGly-CTERM domain appears only at a protein's C-terminus, and consists of a signature motif, a highly hydrophobic stretch likely to represent a transmembrane helix, and then a cluster of basic residues. Such a region could serve as a C-terminal transmembrane anchor, but perhaps only temporarily. In Gram-positive bacteria, sortases remove C-terminal regions from with similar structure, but with the signature motif LPXTG. In halophilic archaea, archaeosortases remove a similar region (PGF-CTERM) from the S-layer-forming major cell surface glycoprotein (see PMID:22037399 and PMID:23651326). But while archaeosortases (probably) and sortases both are transpeptidases, rhombosortase may be simply a hydrolase.

The reasons a bacterium would anchor multiple T2SS targets with a GlyGly-CTERM domain, then release them with rhombosortase, are unclear and would be interesting to know. Possibilities include delaying secretion until other steps in protein maturation have completed, apportioning proteins encoded by a single gene to two different destinations, and regulating protein destinations via signals that alter rhombosortase's expression or activity. It would also be interesting to know if every GlyGly-CTERM protein can be a T2SS target.

On 2014 Mar 10, Daniel Haft commented:

We noted in this paper that only some genomes with type II secretion systems (T2SS) have rhombosortase and a matching collection of GlyGly-CTERM proteins, but essentially all genomes with rhombosortase have T2SS. This comment discusses proteins from the cholera pathogen that are both T2SS targets and GlyGly-CTERM proteins. In Vibrio cholerae, cholera toxin and at least 20 other proteins depend on T2SS for export, including 16 found by Sikora, et al. (PMID:21385872) in 2011. Three targets found in that study, the related trypsin-like serine proteases VCA0803 (VesA), VC1200 (VesB), and VC1649 (VesC), are three of V. cholerae's six GlyGly-CTERM proteins (the others are VCA0065, VC2621, and VC1485). VesA, in particular, may be highly relevant to pathogenesis, since "...analysis of single, double, and triple protease knock-out strains indicated that VesA is the primary protease responsible for processing the A subunit of cholera toxin during in vitro growth of the V. cholerae strain N16961."

The GlyGly-CTERM domain appears only at a protein's C-terminus, and consists of a signature motif, a highly hydrophobic stretch likely to represent a transmembrane helix, and then a cluster of basic residues. Such a region could serve as a C-terminal transmembrane anchor, but perhaps only temporarily. In Gram-positive bacteria, sortases remove C-terminal regions from with similar structure, but with the signature motif LPXTG. In halophilic archaea, archaeosortases remove a similar region (PGF-CTERM) from the S-layer-forming major cell surface glycoprotein (see PMID:22037399 and PMID:23651326). But while archaeosortases (probably) and sortases both are transpeptidases, rhombosortase may be simply a hydrolase.

The reasons a bacterium would anchor multiple T2SS targets with a GlyGly-CTERM domain, then release them with rhombosortase, are unclear and would be interesting to know. Possibilities include delaying secretion until other steps in protein maturation have completed, apportioning proteins encoded by a single gene to two different destinations, and regulating protein destinations via signals that alter rhombosortase's expression or activity. It would also be interesting to know if every GlyGly-CTERM protein can be a T2SS target.