Link to this post | posted 29 Apr, 2025 14:25

nic.vega

Thanks for the papers, this is very helpful! I agree it's *probably* lysis associated based on synteny - the cassette in these AS3 phage seems to be [4 TMD protein][1 TMD protein, spanin-like][endolysin][2 TMD protein, pretty canonical-looking phage holin] at the end of the structural genes & terminating at the back end of the reverse-strand block, not dissimilar to some of the Gordonia cassettes I think. If the function is conserved in other phage, I'd be surprised if it isn't similar here.

Link to this post | posted 28 Apr, 2025 13:04
nic.vega	whoops forgot the .dnam5 33Kb

Link to this post | posted 28 Apr, 2025 13:04

nic.vega

After Tom Bernhardt's talk at the symposium, is there an annotation we can use for lysZ-like proteins? For example, we have Atlantica gp21 (CDS 16204-16494): Canonical structure (per Tom's talk) is a N-terminal TMD + a variable-length helical domain. gp21 has a N-terminal TMD (approx. residues 5-25, annotated as "TMD" rather than "signal" by DeepTMHMMM, confirmed with SOSUI); AlphaFold predicts extended alpha helical domain High-probability, near-full-length hits on HHPred to bacterial FtsB (e.g. 6H9N_B; cell division protein with one TMD; sticks N terminal domain in the membrane & associates other cell division complex stuff on the C terminal domain) and to phage spanins (e.g. P39504). Similar length (97 aa) to these proteins (94-121 aa). Located at the end of the tail assembly genes, immediately before endolysin Amino acid sequence: MSPELLTAILGAGGLAAIVPKLIDGWKAWRSGRAAEEKDKNKGLVDRLAAAEVRLEAEIMWRRANEEYAATLRRLLIEVYGVPADKLPPWPVRRTS

Link to this post | posted 22 Apr, 2025 18:47

nic.vega

Atlantica gp25 (in PECAAN) is showing high-probability full-length hits to prokaryotic Pleckstrin homology (PH)-like domain proteins (e.g. 3HSA_D, Shewanella protein YP_926556.1, type example). Other HHPred hits with similar scores are to the same domain structure in eukaryotic proteins, where the class was originally identified (e.g. 2N23_A, TFIIH p62, canonical example of PH domain). There are dissimilarities in the lengths of some of the disordered regions between structural sub-domains, and identity is low, but conservation of the structure is much better than homology of the sequence. At 104 aa, Atlantica gp25 is similar in length to a known sub-group of bacterial PH domain proteins typified by 3HSA_D but is shorter than a previously identified group of phage-associated PH domain proteins (~180 aa). AlphaFold indicates the characteristic 7-strand beta "sandwich" core + C-terminal alpha helix of this protein class (pic attached). I'm not sure whether PH-like domain meets the criteria to be added as a call. While PH-like domain is a structural class, it is a fairly specific 3' structure, and there is a function associated - canonically these proteins associate with membranes (lipid binding) and assemble other proteins at their location (protein-protein interactions). What do you think? Xu et al. 2010. Bacterial Pleckstrin Homology Domains: A Prokaryotic Origin for the PH Domain, Journal of Molecular Biology 396(1): 31-46. https://doi.org/10.1016/j.jmb.2009.11.006 (https://www.sciencedirect.com/science/article/pii/S0022283609013576) 1.19Mb

Link to this post | posted 16 Apr, 2025 19:23
nic.vega	Thanks Debbie! We thought it might be a frameshift as well, but without wet lab evidence we didn't want to merge the products. We'll annotate both genes with the DNA MTase function & put our observations in the notes.

Link to this post | posted 16 Apr, 2025 17:55

nic.vega

Phage being annotated is Atlantica (AS3); gp's in question (gp42 stop 27,408; gp43 stop 27,967) are forward strand, right side of genome after the reverse block. Both have BLASTP @ PhagesDB hits to gps annotated as DNA methyltransferase. gp42 has pretty solid hits on HHPred and BLAST @NCBI to other phage and bacterial cytosine MTases. However, this gp is short (most targets are ~300 residues, vs 200 in our phage). Correspondingly, most of the expected conserved catalytic/functional domains are found (refs at bottom), but the expected C-terminal domain is not in this product. It is, however, at the C terminal of gp43. Hits to gp43 are to cytosine MTase C term ends and reliably contain the missing domain; the rest of the annotated gp43 (residues 1-178 with our current putative start) has no hits outside SEA-PHAGES, anywhere, for anything. (The start situation/coding potential for gp43 is also a mess. There is a nice, solid, wide peak of coding potential that overlaps 200+ bases into the gp42 ORF.) As of now, my plan is to call DNA methyltransferase function on gp42 and hypothetical on gp43 based on preponderance of domains, but I am wondering whether something is up with this region. https://pmc.ncbi.nlm.nih.gov/articles/PMC317633/pdf/nar00124-0054.pdf https://www.sciencedirect.com/science/article/pii/0022283689904804

Link to this post | posted 13 Mar, 2025 13:50
nic.vega	so cool!

Link to this post | posted 13 Feb, 2025 13:04

nic.vega

The one we found is the same as published attP - specifically, we have a cluster AS phage (Atlantica AS3), and it has the same attP that was published for Galaxy (AS1) in Klyczek et al 2017 "Tales of diversity" (38 bp homology to 3' end of Arthrobacter tRNA-fMet). We just found it, so we haven't done much by way of confirmation, but it does seem to be strongly conserved within cluster. We'll put it in as a note for now, and I'll update if we do find the time to check the lysogen attL/R. Edited 13 Feb, 2025 13:07

Link to this post | posted 12 Feb, 2025 22:14
nic.vega	bump - still not annotating attPs?

Link to this post | posted 20 Dec, 2024 02:56

nic.vega

Vic - I think you're right. We set up (with TrixiePhattel again) standard 1h digests, then a microwave digest and a "standard" 37C digest starting and ending at the same time as the microwave - with handling time, it turned out to be 2 minutes. Overall the banding patterns are very similar except for maybe some differences in undigested band weight, even for PstI which should only have 50% activity in CutSmart. And that's with intentionally overloading the reactions - should be >1 ug DNA/rxn - to really give the enzymes something to do. 134Kb