Link to this post | posted 20 Nov, 2018 20:55

scaruso

I have what I believe is a decent candidate for the holin in the A3 subcluster, in Zetzy and would appreciate some feedback. It is immediately downstream of some tRNAs, and upstream of Lysin A and B. Such that, if correct, it would be: Terminase, small subunit Minor tail protein Minor tail protein NKF (probably a minor tail protein) tRNAs (x3) Holin Lysin A Lysin B Terminase, large subunit Portal Capsid maturation Scaffolding Major capsid protein The described synteny shows the Holin between Lysin A and Lysin B, but this seems like a reasonable variation, right in front of Lysin A and B. A link to the HHPred results can be found here: https://toolkit.tuebingen.mpg.de/#/jobs/Zet_3980. I have attached a PDF of the result as well. SOSUI shows that it has a single transmembrane region near the N-terminus: LATIAKLIAQALLPIIAKQIAEE, and it is 255 aa long. The HHPred suggests it is a class VI holin, which has a single transmembrane region near the N-terminus, but it is also in line with the T4 (class III) family, which is closer in size. The gene product: MLATIAKLIAQALLPIIAKQIAEEFGKHVEPLTKALVTAVTEAAATGAERGADKLTDYIPGKLDDQIIDPIVKRGLEIFRDLTRZ Thanks! Steve Edited 21 Nov, 2018 12:15 3.44Mb

Link to this post | posted 21 Nov, 2018 12:15

debbie

Steve, I am usually reluctant to call that hit to the Pfam data. There is no primary evidence provided. I see that no transmembrane domains are found by TmHmm; Sosui calls 1 transmembrane helix. Can you explain that HHPred (at HHPred) reports a probability of 89.9% but is 34.6% in PECAAN? It is in a believable position in the genome (next to the lysins). Are there any other membrane proteins in the genome?

Link to this post | posted 21 Nov, 2018 12:18

scaruso

Not in that area or its environs, there are a lot of identified genes on the left side, but I'll need to get back to you about the rest of the genome. The HHPred search was completed using only the normal DBs suggested by the manual. You might note, though, that it's common to get different results. I check all interesting ones by hand, still. See this example, where the PDB hit from HHPred on the web and on PECAAN had the same match, but very different probabilities (88.55 vs 55.6). https://toolkit.tuebingen.mpg.de/#/jobs/Zet_13556 I'll be working on it over the next week and will update. Steve Edited 22 Nov, 2018 18:05

Link to this post | posted 23 Nov, 2018 15:58

scaruso

There are no genes in the left (forward) part of the genome with transmembrane domains that haven't been assigned functions other than the one above. On the right side after all the forward genes: The first (stop = 24921) gene has a TMD according to PECAAN's TmHmm. It also has a poor HHPred hit to: PF16935.5 ; Hol_Tox ; Putative Holin-like Toxin (Hol-Tox) Probability: 25.57, E-value: 420.0, Score: 17.93, Aligned Cols: 34, Identities: 9%, Similarity: 0.036, PECAAN HHPred is a little different with a prob of 32.3. coverage is only 35.1 The next, oddball forward gene has four TM domains. It has been identified as a DNA binding protein and, curiously as Excise in another phage. I'm stuck on NKF without proof beyond those calls, though. So it could also be looked at. No HHPreds to anything like Holin, though, or anything else. The last has stop=39165, has 1 TMD, but is only 105 bp long. No significant HHPred hits on PECAAN or external. The only issue is, I used the TmHmm program in PECANN to screen all 84 genes. It caught the three above, but I don't know if it would catch every possibility. I don't know what others would be found with SOSUI. Edited 24 Nov, 2018 15:12

Link to this post | posted 24 Nov, 2018 15:22

scaruso

I was thinking Synteny might be compelling, but I ran the gene with the other databases in HHPred, here's what I find: TIGR01673 holin_LLH; phage holin, LL-H family. This model represents a putative phage holin from a number of phage and prophage regions of Gram-positive bacteria. Probability: 91.2   E-value: 2.9   Score: 27.85   Aligned Cols: 83   Identities: 14%   Similarity: 0.192   The second gene's (24921) best hit is to an uncharacterized membrane protein in the host seen by mass spec. NP_217999.1 membrane protein [Mycobacterium tuberculosis H37Rv] Probability: 99.07   E-value: 1.1E-12   Score: 92.87   Aligned Cols: 77   Identities: 30%   Similarity: 0.671 So, seems like a reasonable call, but I can understand if you want to let it go. As for the other gene I mentioned above, I can't find any reason to call () an excise or DNA binding protein other than that it was called that by someone. So I'm going to call it NKF, though if you have info I would be grateful for it. Otherwise, it's all done. You can see the whole thing in PECAAN (or on a DNAM) if you like for reference. Thanks, Steve Edited 25 Nov, 2018 18:07

Link to this post | posted 27 Nov, 2018 17:39
debbie	Steve, I just called a different genome with the same configuration. I think it is appropriate to call this gene the holin.

Link to this post | posted 27 Nov, 2018 17:47
scaruso	Thanks, Debbie. I appreciate you looking at it. Steve