SEA-PHAGES | Assignment of gp28 as holin

Link to this post \| posted 24 Mar, 2022 22:44
kcornely	Our team has a question on gp28 (pham 101034) of Langerak. We have a lot of evidence to assign its function as holin (synteny–follows lysins A and B, BLAST hits and HHPred hits), but according to the SEA PHAGES official function list, holins must have multiple transmembrane domains (which totally makes sense). But according to TmHmm and SOSUI, gp28 is soluble. Do we assign it as a holin? If we don't, is it all right to have lysin A and lysin B but not holin? We did notice that this protein is not assigned a function in closely related P1 cluster phages Donovan, Techage, Jebeks and HUHilltop. Thanks in advance for the help!

Link to this post | posted 24 Mar, 2022 22:44

Our team has a question on gp28 (pham 101034) of Langerak. We have a lot of evidence to assign its function as holin (synteny–follows lysins A and B, BLAST hits and HHPred hits), but according to the SEA PHAGES official function list, holins must have multiple transmembrane domains (which totally makes sense). But according to TmHmm and SOSUI, gp28 is soluble. Do we assign it as a holin? If we don't, is it all right to have lysin A and lysin B but not holin? We did notice that this protein is not assigned a function in closely related P1 cluster phages Donovan, Techage, Jebeks and HUHilltop. Thanks in advance for the help!

Link to this post \| posted 25 Mar, 2022 03:17
scaruso	Hello, There is certainly a history of that PHAM being called a holin: https://phagesdb.org/genes/Langerak_CDS_28/ TOPCONS shows it to have two TMDs and most of the internal tools agree, with the ones disagreeing showing one TMD. https://topcons.cbr.su.se/pred/result/rst_3pzb2qao/ HHPred doesn't show the TMDs, as you mention, but has a perfect PFAM hit to the holin: https://toolkit.tuebingen.mpg.de/jobs/Lang. PFAM isn't the best, but it's a full length hit. And it's been called in lots and lots of places: https://blast.ncbi.nlm.nih.gov/Blast.cgi?CMD=Get&RID=3UCM7VF0013. On one of ours, I would feel pretty comfortable with it. I hope that's helpful, Steve

Link to this post | posted 25 Mar, 2022 03:17

scaruso

Hello,

There is certainly a history of that PHAM being called a holin: https://phagesdb.org/genes/Langerak_CDS_28/

TOPCONS shows it to have two TMDs and most of the internal tools agree, with the ones disagreeing showing one TMD. https://topcons.cbr.su.se/pred/result/rst_3pzb2qao/

HHPred doesn't show the TMDs, as you mention, but has a perfect PFAM hit to the holin: https://toolkit.tuebingen.mpg.de/jobs/Lang. PFAM isn't the best, but it's a full length hit. And it's been called in lots and lots of places: https://blast.ncbi.nlm.nih.gov/Blast.cgi?CMD=Get&RID=3UCM7VF0013.

On one of ours, I would feel pretty comfortable with it.

I hope that's helpful,

Steve

Link to this post \| posted 25 Mar, 2022 12:33
debbie	I agree with Steve. This may be one of the only places where I rely on that particluar pFam hit. debbie

Link to this post \| posted 25 Mar, 2022 22:51
kcornely	Thank you, Steve and Debbie! We saw the pfam hit and it was one of the lines of evidence that convinced us that we should assign the function as a holin, but we didn't have convincing data that gp28 is a transmembrane protein. SOSUI classifies it as soluble, and TmmHm shows 0 predicted TMH. So thank you for letting us know that we should also be checking TOPCONS, which convincingly shows two transmembrane domains. In the future, we will check all three programs when we suspect a membrane protein. Thanks again!

Link to this post | posted 25 Mar, 2022 22:51

kcornely

Thank you, Steve and Debbie! We saw the pfam hit and it was one of the lines of evidence that convinced us that we should assign the function as a holin, but we didn't have convincing data that gp28 is a transmembrane protein. SOSUI classifies it as soluble, and TmmHm shows 0 predicted TMH. So thank you for letting us know that we should also be checking TOPCONS, which convincingly shows two transmembrane domains. In the future, we will check all three programs when we suspect a membrane protein. Thanks again!

Link to this post \| posted 15 Dec, 2023 22:40
fbaliraine	We are dealing with the exact situation with Sonah gp 28 (25227-25475 bp; MWTLKFWKDASERAVKSAAQAAILALGGEAFNAWTVDWQTVGGIALGGAALSLLTSLGSDLLPFGTKGTASLAKLDGEGSARZ). This gene is identical to Langerak gp 28. It is downstream of lysin A & B. DeepTMHMM shows 2 transmebrane domains, and we see the same holin hit in HHPred in PFAM & UniProtKB. Using DeepTmHMM, we note that Langerak gp 29, the gene immediately downstream of Langerak gp 28, is a membrane protein, with one transmebrane domain, just like the downstream gp 29 of Sonah. Even syntenically, I want to call it a holin, but there's again this caveat in the Official Functions list, stating that, "to call a holin…at least 2 transmembrane domains found and the gene be adjacent to the endolysins (s), conserved domain hits (4), and the abscence of additional transmembrane domains in the area." In previous P1 annotations where we relied on the less sensitive SOSUI which classified it as soluble, and TmmHm which shows 0 predicted TMH, and the downstream gene not detected as a membrane proten by these less sensitive software, we had simply called NKF. Now, using the more sensitive Using DeepTmHMM, and given that everything else has been met to call this gene a holin, what do we do about the fact that the immediate downstream gene is actually a membrane protein with one transmembrane domain? — in view of the caveate…the abscence of additional transmembrane domains in the area. Edited 18 Dec, 2023 03:02

Link to this post | posted 15 Dec, 2023 22:40

fbaliraine

We are dealing with the exact situation with Sonah gp 28 (25227-25475 bp; MWTLKFWKDASERAVKSAAQAAILALGGEAFNAWTVDWQTVGGIALGGAALSLLTSLGSDLLPFGTKGTASLAKLDGEGSARZ). This gene is identical to Langerak gp 28. It is downstream of lysin A & B. DeepTMHMM shows 2 transmebrane domains, and we see the same holin hit in HHPred in PFAM & UniProtKB. Using DeepTmHMM, we note that Langerak gp 29, the gene immediately downstream of Langerak gp 28, is a membrane protein, with one transmebrane domain, just like the downstream gp 29 of Sonah. Even syntenically, I want to call it a holin, but there's again this caveat in the Official Functions list, stating that, "to call a holin…at least 2 transmembrane domains found and the gene be adjacent to the endolysins (s), conserved domain hits (4), and the abscence of additional transmembrane domains in the area." In previous P1 annotations where we relied on the less sensitive SOSUI which classified it as soluble, and TmmHm which shows 0 predicted TMH, and the downstream gene not detected as a membrane proten by these less sensitive software, we had simply called NKF. Now, using the more sensitive Using DeepTmHMM, and given that everything else has been met to call this gene a holin, what do we do about the fact that the immediate downstream gene is actually a membrane protein with one transmembrane domain? — in view of the caveate…the abscence of additional transmembrane domains in the area.

Edited 18 Dec, 2023 03:02

Link to this post \| posted 18 Dec, 2023 02:22
debbie	Hi Fred, I would call 28 a holin and 29 a membrane protein. debbie

Link to this post \| posted 18 Dec, 2023 06:42
fbaliraine	Thank you Debbie! Fred Edited 18 Dec, 2023 18:57

Link to this post \| posted 11 Apr, 2024 17:42
nietof	Hi everyone We have a putative holin in Alkhayr cluster O gene 66 in PECAAN. Both TmHmm and Sosui show two tmb domains, strong matches in BLAST to other phages in this cluster and others but HHPred doesnt yield robust matches to a holin protein. It has Lysin a and b upstream. I assume it is a Holin but I am not sure why HHPred doesnt find matches.

Link to this post \| posted 11 Apr, 2024 17:42
nietof	Hi everyone We have a putative holin in Alkhayr cluster O gene 66 in PECAAN. Both TmHmm and Sosui show two tmb domains, strong matches in BLAST to other phages in this cluster and others but HHPred doesnt yield robust matches to a holin protein. It has Lysin a and b upstream. I assume it is a Holin but I am not sure why HHPred doesnt find matches.

Link to this post \| posted 12 Apr, 2024 00:24
debbie	Hi Fernando, I think I want to stick with what is written in the Cluster O paper. There are 4 transmembrane proteins in the region downstream of lysin A and lysin B. I would call the 4 genes that hit Deep TMHMM transmembrane proteins - membrane proteins. Deep TMHMM is now the preferred method for transmembrane detection. (See Bioinformatics Guide.) And check out that data on phamerator!!! As to why holins may not be detected in HHPRed, I think that list could include that holins demonstrate a lot of diversity, holin activity may need more than one gene, holins are small proteins and must not be easy to crytallize and study. It is not uncommon to not find a hit to a holin. Best, debbie

Link to this post | posted 12 Apr, 2024 00:24

debbie

Hi Fernando,
I think I want to stick with what is written in the Cluster O paper. There are 4 transmembrane proteins in the region downstream of lysin A and lysin B.

I would call the 4 genes that hit Deep TMHMM transmembrane proteins - membrane proteins.
Deep TMHMM is now the preferred method for transmembrane detection. (See Bioinformatics Guide.)
And check out that data on phamerator!!!

As to why holins may not be detected in HHPRed, I think that list could include that holins demonstrate a lot of diversity, holin activity may need more than one gene, holins are small proteins and must not be easy to crytallize and study.

It is not uncommon to not find a hit to a holin.

Best,
debbie

Recent Activity

Assignment of gp28 as holin