TMEM147 & NOMO2 annotation to protein destabilization

Antonialock commented 6 years ago

In PMID 20538592 an additional component of the nicalin-NOMO complex is characterized.

From this paper TMEM147 & NOMO2 were annotated to protein destabilization, I don't think this paper supports that annotation?

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2924024/

ValWood commented 6 years ago

@pgaudet should this term even exist? what does it mean (sounds like a phenotype?)

Antonialock commented 6 years ago

I had a look at another human gene that also was annotated to this term. It seemed sensible (something along the lines of protein directly binds target -> target disappears)

pgaudet commented 6 years ago

@ValWood I dont think the term is wrong but in this case perhaps the synonym is more informative ? "negative regulation of protein stability" @thomaspd what do you think ?

ValWood commented 6 years ago

yes would be more consistent with the parent! GO:0031647 regulation of protein stability

Antonialock commented 6 years ago

So there's two things here:

I don't think TMEM147 & NOMO2 should have this annotation
gene products that should have this annotation (like the BHF one) could move to negative regulation of protein stability ?

pgarmiri commented 6 years ago

Thank you for bringing this issue to my attention @Antonialock, @pgaudet , @ValWood.

I had a look at the paper and actually, what the authors see is the opposite (not protein destabilization but rather protein stabilization).

They said that "TMEM147, Nicalin, and NOMO Stabilize Each Other" (figure 3a), but now I think that the purpose of this experiment was to prove that they form a complex. (I have annotated all to GO:0032991 'protein-containing complex' that was missing)

In figure 3d they show that only Nicalin has an effect on the levels of the other two proteins. It is doing so by binding directly to them and 'protecting' them from degradation. ( I have also added the GO:0061635 'regulation of protein complex stability' to Nicalin)

So, I have deleted the annotations for TMEM147 & NOMO2 as @Antonialock suggested. I have changed the annotation for Nicalin to GO:0050821 'protein stabilization', as it has 'positive regulation of protein stability' as an exact synonym. I would be happy to update the annotation to 'positive regulation of protein stability', if you think the current term is not suitable and a new term was to be created.

Let me know if you are happy with the changes.

Penelope

pgaudet commented 6 years ago

Interesting !

@Antonialock do you want to make a ticket to rename the term ?

Thanks, Pascale

Antonialock commented 6 years ago

If you think it's part of a process regulating the availability of these proteins (as opposed to 'just' being needed for stability of complex) then I guess that's fine?

Pascale, I guess that if the parent terms are "regulation" terms then it doesn't hurt to have that in the term name of the child?

cheers both

ValWood commented 6 years ago

Interesting question. I would not normally make these annotations, but now I have a dilemma one of our community curators has used a "telomerase RNA stabilization" term for a protein which is involved in protein complex assembly: http://curation.pombase.org/pombe/curs/b5450581ac9e067c

I don't think it's quite right here, but I can see why people use the terms - because they exist...and they are appropriate to describe the experiment....

However, you could potentially assign anything involved in complex assembly to stability/stabilization, or any chaperone involved in folding, or any negative regulation of degradation, or positive regulation of production, or sequestering....or ..or....or ?)

Is "stability" really a process or do processes like "complex assembly" results in this quality?

Antonialock commented 6 years ago

hmm. I think stability sort of is a process

eg. the destruction complex negatively regulates the stability of beta-catenin through positive regulation of proteasomal-mediated protein catabolism

(you can't see the proteasome in this diagram, but in the absence of a WNT signal, GSK3 phosphorylates beta-cateninm which recruits the proteasome. Once WNT signal binds the receptor this phoshorylation is blocked, and beta-catenin can go into the nucleus and regulate transcription).

so WNT/frizzled/dishellved positively regulates the stability of beta catenin?...

personally I don't think that "being a structurally important part of a complex" per se means regulation....there needs to be a regulatory aspect (maybe that was demonstrated in the paper in question, I don't know, I didn't read it thoroughly)

Antonialock commented 6 years ago

if you know that something is actively targeted for destruction (or not) by a gene product acting on it, then surely it is regulating stability? We just couldn't say that it is proteasomal or whatever other processes that target gene products for destruction (autophagy? what else?)?

pgaudet commented 6 years ago

Right - in our (nextprot/Bioeditor) annotation model we use a special vocabulary called 'protein property (derived from Vario) to capture stability, conformation and these types of observations. The proteins we've annotated (the mismatch repair complex is the best example) is stabilized when the complex is formed, and degraded when they cannot bind the partner, so you cannot really say that the function of A is to stabilize B but this is what you observe.

The other situation where proteins are more or less stable is when they are mutated, so an IMP to 'protein stabilization' is unlikely to reflect a real function.

WRT the catenin example: this is a common case in signaling pathways, we probably need a term like 'regulation of signaling by targeted degradation (as well as sequestering and probably a few other mechanisms I can't think of right now).

Pascale

ValWood commented 6 years ago

if you know that something is actively targeted for destruction (or not) by a gene product acting on it, then surely it is regulating stability? We just couldn't say that it is proteasomal or whatever other processes that target gene products for destruction (autophagy? what else?)?

I agree that it is regulating stability. but isn't this the case for any protein that regulates whether something is targeted for degradation? so regulation of degradation is regulating stability. However, we don't annotate the "mitotic checkpoint complex" which negatively regulates the ApC to prevent securin degradation of "protein stability". I'm saying its somewhat orthogonal, and inconsistently used, which makes it not so useful.....

pgarmiri commented 6 years ago

I can only speak for this case described in the specific paper. You can see the different experiments done in the figure 3 at the link below.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2924024/figure/F3/

In particular they show:

The knockdown of Nicalin, NOMO, or TMEM147 resulted in a strong reduction of the respective binding partners. (figure 3 a)
mRNA levels in HeLa knockdown cells were quantitated by real-time PCR showing a decrease in target gene expression by 75–90%. The mRNA levels of the interaction partners were only marginally changed. (figure 3 b)
Nicalin, NOMO, and TMEM147 immunoreactivities were not significantly changed upon ribosome inhibition. But proteasome inhibition caused an increase at the level of the NOMO and TMEM147 ( figure 3 e)
Induction of Nicalin expression by tetracycline for the indicated times led to a continuous increase in TMEM147 and NOMO levels over time in lysates (four left lanes) and Nicalin immunoprecipitates (four right lanes), demonstrating enhanced complex formation. The increase of NOMO appears to precede the elevation of TMEM147, an indication for a stepwise complex assembly. (figure 3 f)

@pgaudet , I used an IMP because the used KD cell lines but then they study rescue and overexpression lines to demonstrate the reversibility of these effects (figure 3 c,d ). I suppose I could have used IDA based on the overexpression experiment only. I can change it now if it is confusing.

Thanks,

Penelope

pgaudet commented 6 years ago

Hi @pgarmiri

First - Overexpression is not necessarily IDA. In particular in this case it looks like we're looking simply at a rescue experiment, which is more of a control than a test. Good example for the annotation documentation @vanaukenk

Thanks, Pascale

vanaukenk commented 6 years ago

We do address the overexpression issue in the evidence code documentation for both IDA and IMP.

See the overview for each: http://wiki.geneontology.org/index.php/Inferred_from_Direct_Assay_(IDA) http://wiki.geneontology.org/index.php/Inferred_from_Mutant_Phenotype_(IMP)

That said, and with the caveat that I still need to read this paper carefully, I agree with @ValWood and @pgaudet that we should reconsider if/how we use the 'regulation of protein stability' term and its children to make sure we are capturing an evolved regulatory process when we annotate to these terms.

Antonialock commented 6 years ago

From the description of data it sounds like nicalin is just a rate-limiting factor of complex formation.

With less nicalin around, the proteasome degrades the subunits. With more nicalin around, there is a higher chance of complex forming, and the complex doesn't get degraded.

Whatever factor(s) that control nicalin expression levels, is the factor doing the regulation?

pgarmiri commented 6 years ago

With less nicalin around, the proteasome degrades the subunits. With more nicalin around, there is a higher chance of complex forming, and the complex doesn't get degraded.

I would agree with the summary of @Antonialock .

I really don't mind deleting the annotation to 'protein stabilization' but from the definition of the term is not clear that such an annotation is wrong. Perhaps a note explaining the restrictions or some examples (PMIDs) to guide the curators of when they are to use these terms would be useful. That is if the terms are to stay, of course. Thanks, Penelope

ValWood commented 6 years ago

Yes @pgarmiri this is really a use case and documentation issue! I would not worry about this particular annotation per. se. This is really something which needs better curator guidelines for how we should use "protein stability".

pgaudet commented 5 years ago

I have created a spreadsheet with all the annotations: https://docs.google.com/spreadsheets/d/1dp4MtIYTztQ3-H1A9_2gWERSWaIxPYg2njqXDhA_Als/edit#gid=0

Antonialock commented 4 years ago

looks like they are gone from nicalin and nomo which is what the ticket was originally about, so closing, thanks all for input.

pgaudet commented 4 years ago

We still need to provide guidelines.

ValWood commented 4 years ago

I still don't 'get' the stability terms. Stability is just a balance between production and degradation , no?

pgaudet commented 4 years ago

Yes I think so.

Antonialock commented 4 years ago

I still don't 'get' the stability terms. Stability is just a balance between production and degradation , no?

I think I disagree.

Production has nothing to do with stability (but RNA levels are influenced by stability + production)

I think 'mRNA stabilization' must refer to either:

recruiting poly(A) binding proteins to RNAs to negatively regulate RNA catabolism
bind an RNA to prevent catabolism factors binding to negatively regulate RNA catabolism

ValWood commented 4 years ago

So in this case, how does 'stability' differ from "negative regulation of catabolism'

isn't it an 'outcome' rather than a process? "negative regulation of catabolism' -> increase stability

I worry that we say the same thing from 2 different viewpoints.

Antonialock commented 4 years ago

I disagreed with "production" having something to do with "stability".

I don't disagree with 'mRNA stabilization' parentage looking redundant with its parent terms

ValWood commented 4 years ago

OK!

ValWood commented 3 years ago

This ticket is also related: https://github.com/geneontology/go-ontology/issues/20132 discussed multiple possible meanings of stability /stabilization.

ValWood commented 3 years ago

If stability is used for a chaperone, it's an MF...

ValWood commented 2 weeks ago

There are only 91 EXP annotations to this term. I still don't know what this means as a process...

geneontology / go-ontology

TMEM147 & NOMO2 annotation to protein destabilization #20833