© 2007-2012 The Authors and Curators of CAZypedia. All rights reserved.
Contributors and readers: You can cite CAZypedia! Here's how.

From CAZypedia

PLEASE NOTE: If you leave a comment here, please sign it by adding four tildes (~~~~) after your post. Thanks!

Click the "+" tab to add a new discussion section... Harry Brumer 11:10, 9 November 2009 (UTC)

Page created 2009-11-09

Table outline created for further modification. Harry Brumer 11:12, 9 November 2009 (UTC)

Three questions from Wim

Adding clan info

1) May I add the clan to the family (so the first row would be called Family / Clan and the first entry would then be GH1 / A)? Because this would be very usefull if I would (later) also introduce family-entries "by similarity" (for which there is as yet no proving complexed structure) just like Heightman&Vasella did in their table, e.g. all families of clan-A are anti by similarity.

• I think this is a good idea, but the clan should be entered as a separate column, so the table can be sorted by Clan (where available). Harry Brumer 14:47, 9 November 2009 (UTC)

• Wim - Great idea. Where clan is unknown, you could write 'unknown'!Spencer Williams 10:17, 10 November 2009 (UTC)

alpha/beta vs. axial/equatorial

2) At the second row (Anomeric specificity), may I add to the entries whether it is axial or equatorial? Because alpha / beta is by IUPAC-definition only for the anomeric versus D/L relationship within the Fisher projection and nothing else, thus alpha / beta is (often but) not always resp axial / equatorial in the sugar's ground state conformation, the prime example being glycosides of sialic/neuraminic acid where their ground state 2C5 is alpha-equatorial.

• I would suggest we stick with alpha/beta for now, since it works in 95+% of the cases (although the limitation is acknowledged). Mike's e/a notation never really achieved widespread use, and the specialist will understand the odd cases. It's best to keep things simple in the beginning. Harry Brumer 14:47, 9 November 2009 (UTC)

• Perhaps another column could be added for a->a, a->e, e->a, e->e. But I agree with Harry - Sinnott's notation is excellent but not often used.Spencer Williams 10:17, 10 November 2009 (UTC)

• I can actually easily add an extra column with Michael's e/a notation, I tried this on a preview and it looks OK (I did not post it). It would be next to "Anomeric specificity" and could be called "Anomeric change". When I'll reach the neuraminidases, then I really think that we'll have to do this, since CAZypedia will certainly get visitors that do not know that alpha/beta is for the Fisher projection only, thus may wrongly assume that their alpha-aglycon is axial in the ground state. By the way, GH-9 is very clearly a syn protonator if I re-check pdb 4tf4 (changed). Added an entry for GH-2 and it looks fine. Wim Nerinckx 21:54, 11 November 2009 (UTC)

exo-anomeric column

The third question - but not for now - is to consider to add an extra row next to the one of the ligand, indicating if it is in or out of the exo-anomeric effect. Once I'll make-post the drawings, it will be very clear that in-anti and out-syn are connected, and indeed you can see this in several complexed structures.

• Let's keep this idea on hold for a bit. Harry Brumer 14:47, 9 November 2009 (UTC)

• I too agree with Harry. Let's try to get a fully consistent table up with well-recognized and accepted data/analysis. Less well-established/accepted theories should stand the test of publication in the open literature before being added to this encyclopedia.Spencer Williams 10:17, 10 November 2009 (UTC)

fold column

I figured out how to insert greek letters. Because I think that a nice completeness-info-addition to the table is to also add the fold for each family. If OK for you both then I do that. Wim Nerinckx 22:38, 10 November 2009 (UTC)

• I think that's a good idea; as far as I'm concerned go ahead. Harry Brumer 07:07, 11 November 2009 (UTC)

Enzyme column

• Wim, can you indicate the specific enzyme for which the various amino acid residues are reported for. For example, there are (at least) two Family 11 'xylanases' from C. fimi : Cex and that correspoinding to GenBank accession number DQ146941. To which one does the numbered residue and X-ray structure apply?Spencer Williams 08:14, 12 November 2009 (UTC)

• Added Cex for the C. fimi GH10 entry (the one on GH11 is from B. circulans). The residue numbering in the table is always as it is in the indicated pdb (it may otherwise become really confusing). Temporarily, may I put a remark-sentence like that on top of the table? This would-could be one of the items in the eventual caption of the table. There are other things that can be put in a caption, e.g. if the indicated pdb is a mutant. Something else: is there a way to make the pdb-code clickable, so that it directly links to its RCSB just like in CAZy? Wim Nerinckx 12:41, 12 November 2009 (UTC)

Integration with other lexicon pages

• I have been thinking of adding a brief section about syn/anti classifications to the 'Classifications' section of the Glycoside hydrolase page, with links to this page. Any comments?

• I think that would be great, but it may be better to wait untill there is at least a first text-and-drawing draft on syn/anti. Wim Nerinckx

Suggestion to move the pdb column

• The question from Spencer is pertinent and now gave me this thought: wouldn't it be better to move the pdb column (instead of at the near-end where it is now) to right before the column with "General acid", and also rename it as "Example PDB ID"? Then it should be at once clear that the residue numbering is for the structure in the example pdb. Wim Nerinckx 18:55, 12 November 2009 (UTC)
• Update: tried it first on a preview and found it much better, so I already did it.Wim Nerinckx 23:21, 12 November 2009 (UTC)

Order of columns

It occurred to me that it might be a bit more logical to re-order the columns as follows:

family, clan, fold, anom. specificity, mechanism, syn/anti, example PDB, enzyme, organism, ligand, general acid, nucleophile/G.B., reference

This groups all the key general mechanistic/structural details in the first 6 columns. Then, the reader is directed to one specific example structure, via the PDB ID, whose details are given. For example, the specific position of the general acid and nucleophile are largely irrelevant, except to someone who actually opens the structure and looks in detail.

I would also suggest the following guideline: The PDB file and reference should be chose to correspond with the first determined structure in the family (listed in the "Family Firsts" section), unless there is different reference in which the syn/anti concept is specifically discussed.

Harry Brumer 08:37, 13 November 2009 (UTC)

• Your proposed column order is great! Will do that.

I however - with respect - disagree on the "family first" (often an empty enzyme) as pdb-example here. I think that the pdb-example here should be one in which the syn or anti position is clear to see. So my order of preference would be 1) Michaelis-type spanning the crucial subsites -1/+1, 2) glycosyl-enzyme intermediate, 3) product if both plus +1 and -1 are occupied, 4) product if several of the minus subsites are occupied including -1. The interested visitor could then easily download the pdb-example (can the pdb-entry in the table be made "clickable" towards the RCSB as in CAZy?), zoom towards the ligand and indicated residues, and see syn or anti for himself. Wim Nerinckx 10:08, 13 November 2009 (UTC)

• I agree regarding the choice of PDB file. I have an idea for easy linking to the PDB that I will try out. Harry Brumer 15:05, 13 November 2009 (UTC)
• I think that the structures of imidazole/tetrazole/lactam oxime-containing enzymes are of special interest as these inhibitors were designed specifically to probe for anti protonation. Could they be added, along with an annotation?Spencer Williams 04:12, 17 November 2009 (UTC)

• Yes the Vasella-inhibitors were very important for the discovery of lateral protonation, and I thought of describing this in the text (later). I would choose only one clear example per family for the table, since otherwise the table would become excessively long - even with only one example it will eventually become about as long as there are families. The best syn/anti examples are pre-TS Michaelis complexes that have an intact glycosidic oxygen (or sulphur) and that are spanning -1/+1, since here not only the glycosidic oxygen (sulphur) is present and is close to the proton donor, but also with these you can see the relationship with the conformation of the glycosidic bond (this is not a speculation as the drawings will show, and it is indicated in the Heightman-Vasella article). My hope is that within about a decade nearly all family-entries would have a beautiful-informative Michaelis. But let me first continue my focus on making a near-complete table. This will take a while: not only I am 3D-rescanning entries before I add them to the table, but in the last years many new complexed structures have appeared, including several Michaelises, so I can replace entries by one with a Michaelis (e.g. at family 1) and can update syn versus anti annotation for many extra families, making the table much more complete and more useful. Wim Nerinckx 10:49, 17 November 2009 (UTC)

Link to PDB data

I've implemented a fairly simple link to PDB data using the template page Template:PDBlink. At the moment, this points to Proteopedia, which I think has a bit nicer look than the RCSB PDB entry pages (and besides, Proteopedia is a wiki!).

Harry Brumer 08:28, 16 November 2009 (UTC)

• Proteopedia now linked throughout in the pdb-column. Very nice! Wim Nerinckx 10:30, 16 November 2009 (UTC)