SEA-PHAGES | thiamine-binding protein

Link to this post \| posted 14 May, 2020 21:44
abeyer	Hi there. In annotating Arthrobacter globiformis phage London (cluster AZ), I came across a wonky reverse gene (start 35,904; stop 35,584) that is 100% identical to a putative gene in draft phage Elezi (also AZ); they are the only two currently in this pfam (11727). The predicted length is 321 nts (106 aa's). With NCBI BLAST, CDD,and HHPRED, I got strong hits to a thiamine-binding protein. HHPRED hit: PF01910.17 Thiamine-binding protein, 99.9 probability; 91.5% coverage; target 1-92; query 1-100; 4.7e-25 http://pfam.xfam.org/family/PF01910.17 Top NCBI hit: WP_079552236 Arthrobacter thiamine binding protein, 72.7% identity; 79% alignment; 98.1% coverage; 96 positives; target 1-104; query 1-104; 3.75016e-56. https://www.ncbi.nlm.nih.gov/protein/WP_079552236 (there are several other really good hits; all to Arthrobacter thiamine binding proteins) CDD hit: pfam01910 (same as HHPRED hit); thiamine binding protein crystal structure. 92.4% coverage; target 1-92; query 3-100; 5.50511e-24 It looks like thiamine binding proteins can be 99-125 aa's in length, so the phage protein fits the bill. https://www.ncbi.nlm.nih.gov/Structure/sparcle/archview.html?archid=10487226 There do not appear to be any other conserved domains along with this one. Here is the London phage protein in question: London_47 MIVAFSVAPSGASASGEAPSDASVHEAVAAAVRVVRESGLPNRTSSMFTELEGDWDAVMDVVKRATEAVAPYGSRVSLVLKADIRPGYSGELEGKVERLERAIEKP I'm working off of PECAAN at the moment, but can share the DNA Master file later, if needed. If I'm way off base here, let me know- just thought it looked interesting and possible! Thanks, Andy

Link to this post | posted 14 May, 2020 21:44

Hi there. In annotating Arthrobacter globiformis phage London (cluster AZ), I came across a wonky reverse gene (start 35,904; stop 35,584) that is 100% identical to a putative gene in draft phage Elezi (also AZ); they are the only two currently in this pfam (11727). The predicted length is 321 nts (106 aa's). With NCBI BLAST, CDD,and HHPRED, I got strong hits to a thiamine-binding protein.

HHPRED hit: PF01910.17 Thiamine-binding protein, 99.9 probability; 91.5% coverage; target 1-92; query 1-100; 4.7e-25 http://pfam.xfam.org/family/PF01910.17

Top NCBI hit: WP_079552236 Arthrobacter thiamine binding protein, 72.7% identity; 79% alignment; 98.1% coverage; 96 positives; target 1-104; query 1-104; 3.75016e-56. https://www.ncbi.nlm.nih.gov/protein/WP_079552236 (there are several other really good hits; all to Arthrobacter thiamine binding proteins)

CDD hit: pfam01910 (same as HHPRED hit); thiamine binding protein crystal structure. 92.4% coverage; target 1-92; query 3-100; 5.50511e-24

It looks like thiamine binding proteins can be 99-125 aa's in length, so the phage protein fits the bill. https://www.ncbi.nlm.nih.gov/Structure/sparcle/archview.html?archid=10487226
There do not appear to be any other conserved domains along with this one.

Here is the London phage protein in question: London_47
MIVAFSVAPSGASASGEAPSDASVHEAVAAAVRVVRESGLPNRTSSMFTELEGDWDAVMDVVKRATEAVAPYGSRVSLVLKADIRPGYSGELEGKVERLERAIEKP

I'm working off of PECAAN at the moment, but can share the DNA Master file later, if needed. If I'm way off base here, let me know- just thought it looked interesting and possible!
Thanks,
Andy

Link to this post \| posted 21 May, 2020 00:45
debbie	Andy, I would still call this gene a Hypothetical Protein. Even though has hits to a thiamine binding domain, it is unclear what the function of it is. Considering how many hits it has to other things that have no identified function, it would be prudent to wait to call this anything.

Link to this post \| posted 07 Oct, 2022 17:58
amaya.garciacostas	Hi! It looks like "putative thiamine binding protein" is no longer in the functions list, am I correct? We have quite a few genomes in the AZ cluster with that annotation and I am working on correcting them, thanks!

Link to this post \| posted 07 Oct, 2022 17:58
amaya.garciacostas	Hi! It looks like "putative thiamine binding protein" is no longer in the functions list, am I correct? We have quite a few genomes in the AZ cluster with that annotation and I am working on correcting them, thanks!

Link to this post \| posted 07 Oct, 2022 17:58
amaya.garciacostas	Hi! It looks like "putative thiamine binding protein" is no longer in the functions list, am I correct? We have quite a few genomes in the AZ cluster with that annotation and I am working on correcting them, thanks!

Link to this post \| posted 02 Mar, 2023 23:27
kklyczek	Hi Amaya, Sorry to take so long to reply. We have discussed this offline, but I wanted to post that discussion here. Yes, this protein has very good, full-length HHPred alignments with thiamine-binding domains in a few proteins. Some notes are in these slides. However, a couple of reasons to be cautious about calling it thiamine-binding protein include: - This binding domain is in the ACT domain family, which is found in a lot of different proteins with various functions, binding various ligands (E. coli 3-phosphoglycerate dehydrogenase is the model ACT domain protein). There are equally good alignments in HHPred with several proteins of unknown function. The London gp47 domain contains some but not all of the key residues described for thiamine binding in this paper. Could it be binding something else instead? - Both of the thiamine-binding proteins matched by London gp47 in HHPred are part of a cluster of genes involved in thiamine transport (Bacillus) or oxidative stress responses (Thermotoga). What function would this protein have in the absence of the other proteins in these clusters? Its location in the phage genome is just upstream of the integrase gene, on the opposite strand by itself. Nick Edgington and his students ran Elezi gp47 through AlphaFold and saw very similar folding to the Thermatoga protein TM0486 (see last slide in attachment). Is this sufficient evidence to make the call, or are the other issues still pertinent? Karen Edited 02 Mar, 2023 23:29

Link to this post | posted 02 Mar, 2023 23:27

kklyczek

Hi Amaya,

Sorry to take so long to reply. We have discussed this offline, but I wanted to post that discussion here.

Yes, this protein has very good, full-length HHPred alignments with thiamine-binding domains in a few proteins. Some notes are in these slides. However, a couple of reasons to be cautious about calling it thiamine-binding protein include:

- This binding domain is in the ACT domain family, which is found in a lot of different proteins with various functions, binding various ligands (E. coli 3-phosphoglycerate dehydrogenase is the model ACT domain protein). There are equally good alignments in HHPred with several proteins of unknown function. The London gp47 domain contains some but not all of the key residues described for thiamine binding in this paper. Could it be binding something else instead?

- Both of the thiamine-binding proteins matched by London gp47 in HHPred are part of a cluster of genes involved in thiamine transport (Bacillus) or oxidative stress responses (Thermotoga). What function would this protein have in the absence of the other proteins in these clusters? Its location in the phage genome is just upstream of the integrase gene, on the opposite strand by itself.

Nick Edgington and his students ran Elezi gp47 through AlphaFold and saw very similar folding to the Thermatoga protein TM0486 (see last slide in attachment). Is this sufficient evidence to make the call, or are the other issues still pertinent?

Karen

Edited 02 Mar, 2023 23:29

Recent Activity

thiamine-binding protein