PTMs

[gocentral]

John G would like to see all the children of 'post-translational protein modification ;GO:0043687' moved up to be direct children of 'protein modification process ; GO:0006464'.

Reasoning: There are 3 stages at which a protein can be modified:

1. before translation (pre-translational)
2. during translation (co-translational)
3. after translation (post-translational)

pre-translationally is the only one that can truly be experimentally determined. We have GO:0019988 for this. Experimentally, it's not possible to distinguish between co- and post-translational modification. The protein ontology has got round this by merging 'co-translational' and 'post-translational' modifications (PRO:000002968).

John G would like a comment added to GO:0043687 to say 'It is not possible to experimentally determine whether a protein modification occurs once protein translation has begun when the protein is on the ribosome, or when protein translation has finished and the protein has left the ribosome.

He would also like to see the following term obsoleted because acetylation does not occur co-translationally. There are no annotations to this term: co-translational protein amino acid acetylation ; GO:0034420

in which case 'post-translational protein amino acid acetylation ; GO:0034421' could also be obsoleted or merged into GO:0006473.

Am happy to do the edits- just want to check there's no objections/comments?

Reported by: rebeccafoulger

Original Ticket: geneontology/ontology-requests/7815

[gocentral]

I have no objection -- in fact, I could swear I remember a bunch of us deciding to do essentially this a while back. I guess we just never got round to actually acting on it ...

m

Original comment by: mah11

[gocentral]

Let me make absolutely clear what I said. When a free protein is modified, then of course, the modification must be post-translational. All C-terminal modifications must be post-translational. For proteins longer than 100 residues, some N-terminal modifications do occur while the down-stream regions are still in the ribosome and before the translation has been completed. Technically these would be co-translational modifications, but there would be no operational definition for a co-translational modifications that could be true in all cases and circumstances for the purposes of an ontology. The protein ontology got around this by defining a post-translational modification as any protein modification occurring after initiation of translation, essentially ignoring "co-translational". What I am suggesting is that the standard definitions of co- and post-translational modifications be provided, but that they not be used for ontological classification purposes, and that we explain in the comments why these terms are not really useful for the ontology.

Original comment by: jsgaravelli

[gocentral]

yes. I think this might be in one of my open requests. I'll try to figure out which one and clode it.

Val

Original comment by: ValWood

[gocentral]

• assigned_to: nobody --> rfoulger

Original comment by: mah11

[gocentral]

John:

>All C-terminal modifications must be post-translational.

does that mean the following terms CAN be given post-translational protein modification ;GO:0043687 as a parent without any TPVs?

C-terminal protein amino acid methylation ; GO:0006481 C-terminal protein lipidation ; GO:0006501 C-terminal protein-tyrosinylation ; GO:0018166

How about protein C-terminal amidation ; GO:0018033 and children?

thanks, Becky

Original comment by: rebeccafoulger

[gocentral]

Becky, Because a protein C-terminal modification requires a free C-terminal, the protein must have been released from the ribosome, and thus is post-translational by either definition of the term. The four examples you listed could be unequivocally called "post-translational". My main concern is something like glycosylation, which everyone regards as post-translational but which can in some cases begin occurring before the protein is released from the ribosome - "co-translational" by the standard definition but "post-translational" by the revisionist protein ontology approach. The ontology needs to say there are classes of things that do not fit neatly into the pre-, co-, post- translational spectrum.

Original comment by: jsgaravelli

[gocentral]

Thanks John- I'll give those four terms the PTM parent when I move the other terms out.

Next Q: is GO:0070930 a valid term?

co-translational protein tagging ; GO:0070930 A protein modification process in which a polypeptide is added to a nascent polypeptide cotranslationally by trans-translation.

current tree: protein modification process ; GO:0006464 --[i]co-translational protein modification ; GO:0043686 --[i]co-translational protein tagging ; GO:0070930

thanks

Original comment by: rebeccafoulger

[gocentral]

GO:007030 was added for this SF item:

https://sourceforge.net/tracker/?func=detail&aid=2856126&group\_id=36855&atid=440764

(it may turn out to be invalid anyway, but at least there's some background) m

Original comment by: mah11

[gocentral]

For request ID:2856126, did Jim Hu give a citation PMID for the process he is talking about? The only trans-peptide splicing I am aware of is post-translational for both substrates, and is "trans-translation" a typo?

Original comment by: jsgaravelli

[gocentral]

John, there is 1 annotation for GO:0070930 for an E.coli transcript, from PMID:8584937. http://amigo.geneontology.org/cgi-bin/amigo/term-assoc.cgi?term=GO:0070930&session\_id=947amigo1286803686

Original comment by: rebeccafoulger

[gocentral]

Jim didn't mention any papers in the SF request but I see that Becky noticed the E. coli annotation supported by PMID:8584937.

"Trans-translation" is not a typo. PMID:18557701 is one of 23 reviews that comes up in a PubMed search.

Original comment by: mah11

[gocentral]

Have done the following:

1/ Moved modification children out from under PTM GO:0043687 2/ Removed 'posttranslational' from the definition of the 90-odd terms that were under GO:0043687. 3/ Added caution comments to: post-translational protein modification ; GO:0043687 co-translational protein modification ; GO:0043686

4/ Have sent out email proposing the obsoletion of co-translational protein amino acid acetylation ; GO:0034420

Kept the following C-terminal modification terms under PTM GO:0043687. Also kept CAAX-box to fit with definition of parent for now: CAAX-box protein modification ; GO:0071587 peptide or protein carboxyl-terminal blocking ; GO:0018410 --[i]C-terminal protein amino acid methylation --[i]C-terminal protein lipidation --[i]C-terminal protein-tyrosinylation --[i]protein C-terminal amidation (and children)

If I don't hear objections from the list, I'll obsolete GO:0034420 and close this SF item. Any new points should get a new SF item please.

thanks

[Note:there is an existing SF item open about C/N-terminal blocking: https://sourceforge.net/tracker/?func=detail&aid=2969457&group\_id=36855&atid=440764]

Original comment by: rebeccafoulger

[gocentral]

I started my graduate work in a lab that demonstrated that proline hydroxylation of collagen occurred during translation by isolated polysomes, and analyzing the nascent chains. I would think a similar experiment could easily be done to determine if nascent chains have acetylation. J Mol Biol. 1971 Jun 28;58(3):831-46. Collagen polysomes: site of hydroxylation of proline residues. Lazarides EL, Lukens LN, Infante AA.

Biochemistry. 1975 Apr 8;14(7):1404-12. Acetylation of nascent polypeptide chains on rat liver polyribosomes in vivo and in vitro. Pestana A, Pitot HC.

See also Gautschi M, Just S, Mun A, Ross S, Rucknagel P, et al. (2003) The yeast N(alpha)-acetyltransferase NatA is quantitatively anchored to the ribosome and interacts with nascent polypeptides. Mol Cell Biol 23: 7403–7414. Find this article online

Original comment by: hdrabkin

[gocentral]

Harold correctly observes that it is possible by well designed and controlled experiments to determine whether a particular modification, in a particular protein of a particular tissue of a particular species occurs as a co- or a post-translational modification. One problem is that such experiments are rarely done, and that it is not wise to assume that the same ontologically defined modification should occur with the same timing in a different species, or tissue, much less a different protein. In the example you mention, proline hydroxylation, it occurs (as I recall) in at least three different locations for eukaryotes, the ER (hence possibly, but not necessarily, co-translationally), as well as the cytoplasm (whence transcription factors migrate to the nucleus although it is not clear whether some hydroxylation may occur in the nucleus) and extracellularly. Both of the latter cases must be post-translational.

The big problem for me is that the term "post-translational modification" is used almost universally in a sense that is very different from the proper timing restricted sense that the words are actually meant to convey. When most people say "post-translational modification" they really mean a natural protein modification. Look at the list of examples they usually add and you will see "like ... glycosylation". Glycosylation in most cases actually begins its first stages as co-translational event, but the bulk of the "glycosylation" modifications are post-translational. Told that a particular protein modification or "PTM" was observed, an annotator looking down a list of possible terms to use would be most likely to pick one that says "post-translational blah-blah" even though there may have been no particular timing significance attached to the report whatsoever.

Original comment by: jsgaravelli

[gocentral]

I see where you are coming from. I have also just now thought of a couple of cases that might be semantically questionable.

co-translational: during translation... so one might look at selenocys incorporation into a protein is co-translational, because it is brought in during translation or "pre-translational: it is on it's special tRNA already before it goes to the A site; similarly, fmet in bacteria and mitochondria: the modification exists prior to the use at the start of protein synthesis. So are those two pre-translational because they come in on a tRNA, or co-translational because the final protein product gets the modification during peptide synthesis. OR, would we restrict PTM to be only those that COULD be added after synthesis, whether they can also be put on at the nascent chain stage during elongation OR might we restrict co-translation to me only those that can only get in during translation, like selenocys and fmet (translformylase. (although I might remember reading somewhere about PTM for selenosys, but I might be remembering incorrectly).

So although I disagree that it is not possible to determine the when, and for acetylation that is can and does occur during chain elongation, I guess the issue is more do we want to remove the ability to be able to annotate to that info if it exists in a paper using a more granular term (which would be an is_a to the "untimed" term.

Original comment by: hdrabkin

[gocentral]

So will the terms co-translational modification and post-translatiopnal modification be obsoleted after the child term moves (I have no objection to this, I just wanted to check if this was the final outcome)

Val

Original comment by: ValWood

[gocentral]

Hi Val,

I believe it's still under debate whether the co-translation and post-translation terms will exist but with caution comments (as they are now) or be obsoleted.

Will anyone want to annotate the timing of a PTM?

John said he had something to add...

Becky

Original comment by: rebeccafoulger

[gocentral]

The issue is not whether anyone will want to annotate the timing of a PTM. It's whether there are distinct gene products involved in co vs post translational modification processes. By analogy to the situtation with protein translocation across membranes, I suspect that there will be such distinct products in those cases where the modification only occurs during or after translation.

Original comment by: jimhu

[gocentral]

The following two terms can stay in the ontology now they have caution comments. Most of their children have been moved up to be direct children of 'protein modification'.

post-translational protein modification ; GO:0043687 co-translational protein modification ; GO:0043686

Therefore, I'm closing this SF item. Thanks, Becky

Original comment by: rebeccafoulger

[gocentral]

• status: open --> closed-works-for-me

Original comment by: rebeccafoulger

[gocentral]

Sorry to go on about this but I still think these terms are confusing, and i'm not sure why we need to keep them.

The confusion is made clear by the fact that there are many IEA mappings to them for terms which are mannosyltransferases amino acid ligases? eh? ubiquitin conjugating enzymes glucosyl transferases and 100's of other random things which would be better anotated to terms in a different branck of the ontology......

(i.e this term is no longer a parent of the terms which it used to be when these annotations were made)

see. http://amigo.geneontology.org/cgi-bin/amigo/term-assoc.cgi?gptype=all&speciesdb=all&taxid=all&evcode=all&term\_assocs=direct&term=GO%3A0043687&session\_id=7489amigo1302535892&action=filter

If we have established that the terms i) aren't useful process groupings and ii) if it is clear the existing annotations should be to more specific terms elsewhere in the ontology,

what do we gain by keeping the terms?

VAl

Original comment by: ValWood

[gocentral]

InterPro have removed all their mappings to GO:0043687, and transferred them to more appropriate GO terms, so the ones in QuickGO are out of date (this is a timing issue).

Like Jim says, they may be useful for cases to annotate cases where the authors show that a modification only occurs post- or only occurs co-translationally.

With exact definitions (and the caution comments), the terms themselves shouldn't be confusing.

Original comment by: rebeccafoulger

[gocentral]

I still find it really confusing as there is nothing in the term name, definition or comment which explicitly excludes the things which I listed below as having existing annotations

mannosyltransferases amino acid ligases? ubiquitin conjugating enzymes glucosyl transferases

so people are inevitable going to use it inappropriately. If this is the first thing people find when they search they are likely to look at the children and assume that the other children are missing.

I don't understand the advantage of having this as a grouping term, for example for
what advantage is there to having post-translational modification as a grouping term for CAAX box modiification if this process term can only occur post-translationall. Why is this needed? why is it important.. Why not just define CAAX box modiification as a "post translational protein modification" and give it a "protein modification process "parent? much simpler....

There are also a large number of TAS annotation (mainly Reactome), but this will still leave many direct annotations which would be better reannotated otherwise will remain confusing......

Original comment by: ValWood

[gocentral]

Hi Val,

The current definition and comment of GO:0043687 are very clear that they should only be used to annotate modification events that occur once the protein has been released from the ribosome:

Definition:The process of covalently altering one or more amino acids in a protein after the protein has been completely translated and released from the ribosome. Comment: This term should only be used to annotate a protein modification process that occurs after the protein has been released from the ribosome, and is therefore strictly post-translational. Modification of a free protein (one not attached to a ribosome) and modification of a C-terminal residue are post-translational processes. Some protein modifications occur while the protein is still in the ribosome but before translation has been completed; these modification processes are considered co-translational and should not be annotated using this term.

They are not so much grouping terms, but there may be distinct gene products involved in co-translational and post-translational modifications, and these terms could be used to annotate such events.

I've contact Reactome to suggest transferring their existing annotations up to the parent term 'protein modification process ; GO:0006464' anyway.

thanks, Becky

Original comment by: rebeccafoulger

[gocentral]

Hi

...OK, a compromise....

At the moment the comment is: This term should only be used to annotate a protein modification process that occurs after the protein has been released from the ribosome, and is therefore strictly post-translational. Modification of a free protein (one not attached to a ribosome) and modification of a C-terminal residue are post-translational processes. Some protein modifications occur while the protein is still in the ribosome but before translation has been completed; these modification processes are considered co-translational and should not be annotated using this term.

can this be more explicit This term should only be used to annotate a protein modification process that *CAN ONLY OCCUR" occurs after the protein has been released from the ribosome, and is therefore strictly post-translational *in every instance**

And add something along the lines of....

Some modification terms which are usually considered to be post-translational (for example phosphorylation, glycosylation, ubiquitination) are not children of this term, because it cannot be assumed that they only occur post-translationally.

(At present the definition or the comment as they stand are still slightly ambiguous and do not restrict this type of modification from being annotated. They will definately continue to accumulate annotations to these processes, and child term requests for these terms to be positioned under here)

Thanks,

Val

Original comment by: ValWood