search

geneontology / go-ontology

Source ontology files for the Gene Ontology
http://geneontology.org/page/download-ontology
Creative Commons Attribution 4.0 International
220 stars 40 forks source link

NTR: 8-oxo-dATP hydrolase activity #21467

Closed deustp01 closed 3 years ago

deustp01 commented 3 years ago

Please provide as much information as you can:

hdrabkin commented 3 years ago

@pgaudet , why does this need to be discussed?

pgaudet commented 3 years ago

Hi @hdrabkin Again we have this specificity issue - the enzyme P36639 NUDT1 can act on many substrates.

My understanding of @thomaspd 's guidelines for MF is that we would use a more general substrate for those situations, and use 'has input' to specific which is the substrate in a given reaction.

Otherwise we should make sure to add all substrates in the UniProt record, to avoid doing this piecemeal:

KM=8.3 µM for 2-hydroxy-dATP (at 30 degrees Celsius and pH 8.0)1 Publication KM=5.7 µM for 2-hydroxy-dATP (at 30 degrees Celsius and pH 7.2)1 Publication KM=4.3 µM for 2-hydroxy-rATP (at 30 degrees Celsius and pH 8.0)1 Publication KM=13.9 µM for 8-hydroxy-dATP (at 30 degrees Celsius and pH 8.0)1 Publication KM=51.0 µM for 8-hydroxy-rATP (at 30 degrees Celsius and pH 8.0)1 Publication KM=15.2 µM for 8-hydroxy-dGTP (at 30 degrees Celsius and pH 8.0)1 Publication KM=12.8 µM for 8-hydroxy-dGTP (at 30 degrees Celsius and pH 7.2)1 Publication KM=13.2 µM for 8-hydroxy-dGTP (at 22 degrees Celsius and pH 7.5)1 Publication KM=55.0 µM for 8-hydroxy-rGTP (at 30 degrees Celsius and pH 8.0)1 Publication

hdrabkin commented 3 years ago

RHEA:65396 appears to not be there yet. So I can't tell what it is. @amorgat , can you tell me what it will be?

deustp01 commented 3 years ago

I'm fine with adding those substrates. And I agree with @pgaudet that the last substrate listed in the UniProt record, dGTP, should be omitted because the Km for that reaction is so high, though I'm not sure how to square that correct conclusion with the rule "add all substrates in the UniProt record". People who want a 1:1 Rhea:GO MF mapping, though, may not be as happy, especially if a more general substrate can't be specified. Here, that may be hard - a general term like oxidized purine base, will include many molecules on which the enzyme does not act. Is that acceptable?

hdrabkin commented 3 years ago

I'm sure every enzyme ever looked at for a variety of substrates will have an array of Kms but usually that is being done to investigate the enzyme specificity. KM=55.0 µM for 8-hydroxy-rGTP (at 30 degrees Celsius and pH 8.0)1 Publication vs <<< No KM=8.3 µM for 2-hydroxy-dATP (at 30 degrees Celsius and pH 8.0)1 Publication means for rGTP, almost 7x more rGTP is needed for the reaction to proceed at half Vmax. So yeah it would work but that doesn't mean it's significant in vivo.

Also lists two for the same substrate at different pH; not relevant; same reaction, just a little worse at 8.3 vs 7.0 (5.7 vs 8.3); Also looks like rATP better at pH 8 (Km 4.3) but can’t compare it to pH 7 for dATP

hdrabkin commented 3 years ago

What about (d)-8-oxo-RTP + H20 = (d) -8-oxo-RMP + diphosphate + H+ (R meaning A or G). (https://www.qmul.ac.uk/sbcs/iubmb/misc/naseq.html) Might need to also get a new CHEBI (I can find the use of RTP once in Chebi,

deustp01 commented 3 years ago

PMID:11139615, the paper that is causing all these problems has an extensive discussion of the interplay between the various Km's that they measure, which vary by more than a factor of 10, and the likely cellular substrate concentrations. For example, cellular [ATP] is far more abundant than [dATP] so if these concentrations determine the relative concentrations of oxidized bases (a plausible assumption from the mechanisms by which base oxidation occurs, but not actually tested here), the enzyme might be driven to dispose of as many molecules per second of oxidized ATP as of oxidized dATP because the 10-fold greater concentration balances the 10-fold lower reaction efficiency.

pgaudet commented 3 years ago

What about (d)-8-oxo-RTP + H20 = (d) -8-oxo-RMP + diphosphate + H+ (R meaning A or G)

Sure! Would that also cover the terms requested in ##21311? (coming from the same paper?)

Thanks, Pascale

deustp01 commented 3 years ago

Would that also cover the terms requested in ##21311? (coming from the same paper?

No, it wouldn't - #21311 covers reactions involving 2-oxo-RTP And while I'm fussing, wouldn't this also require a new Rhea reaction, and, for sanity, some sort of hierarchy on the Rhea (or ChEBI?) side to indicate that reactions of specific oxidized purine nucleotides are is_a children of the general one. Is that something Rhea (or ChEBI?) will handle (only one instance here but if the decision about generality in MF terms is to be our new rule, the number of cases needing this sort of fix and mapping will be large).

hdrabkin commented 3 years ago

Right; oxo is in different position in 21311 than here (8 oxo). And yes, that's why I was futzing about the CHEBI, because I would need them to request a RHEA corresponding to (d)-8-oxo-RTP + H20 = (d) -8-oxo-RMP + diphosphate + H+ For 21311 I would need (d)-2-oxo-RTP + H20 = (d) -2-oxo-RMP + diphosphate + H+

deustp01 commented 3 years ago

(d)-8-oxo-RTP + H20 = (d) -8-oxo-RMP + diphosphate + H+

More futzing - should that be 8-oxo-(d)RTP + H20 = 8-oxo-(d)RMP + diphosphate + H+ ? and 2-oxo-(d)RTP + H20 = 2-oxo-(d)RMP + diphosphate + H+ ?

deustp01 commented 3 years ago

A non-futzy point to come out of these last few comments is that to make the needed general expressions and to show their relationships to one another, we are going to need a whole DAG to accommodate the combinatorial explosion of diverse possible modifications to the base and to the sugar, and don't forget the ionization states of the phosphate groups (which are also lurking in the background here).

A strategy question is how much we should address this by fitting exact chemistry to MF terms, and how much, as GO-CAMs become the norm, we should annotate the specific molecules on which a gene product acts by making them inputs and outputs of the activity unit that links the gene product to its (general) MF term. But does this really fit well with all the effort Rhea is putting into making reactions with all participants fully specified? @ukemi ?

deustp01 commented 3 years ago

This ticket and #21311 started because I was trying to get Reactome reactions enabled by human NUDT1 (MTH1) annotated with the correct charge states of the small molecules involved in the reactions and fill some gaps in our collection of reactions and @ukemi noted that available GO MF terms did not specify action on oxidized nucleotides. This brought in @amorgat , who provided correct Rhea reactions and also pointed out reactions missing from Reactome in which the NUDT1 gene product enables the hydrolysis of O6-methyl-dGTP (PMID: 30304478 - RHEA:67600 [new]), N6-methyl-dATP (PMID:32144205 - RHEA:67604 [new]), and N6-methyl-ATP (PMID:32144205 - RHEA:67608 [new]) which don't fit any of the patterns for substrates discussed here.

Is it time to retreat to the parent MF term, GO:0047429 "nucleoside-triphosphate diphosphatase activity", the immediate parent of all existing and requested relevant terms and rely on pathway annotation / GO-CAM activity units to do the rest? This indeed looks like an editor-discuss issue.

deustp01 commented 3 years ago

Perhaps some grouping terms would be a workable compromise here. All these new terms would be children of GO:0016818 hydrolase activity, acting on acid anhydrides, in phosphorus-containing anhydrides, would rely on the PMID's cited in the tickets, and would group the individual Rhea reactions cited in those tickets, a violation of the 1 Rhea : 1 GO MF rule, but one required by the GO editorial decision to group molecular entities that the chemists at Rhea and ChEBI have differentiated.

These two handle the four open NTRs, for reactions that convert a modified (d)NTP to (d)nNMP + diphosphate

These continue the pattern, for removal of one phosphate group, so far only observed for oxidized nucleotides

That's enough specifcity to keep me reasonably happy. @ukemi ? @hdrabkin @pgaudet is the 1:1 Rhea:GO violation acceptable?

hdrabkin commented 3 years ago

So for this one are we good to make a term 8-oxo-(d)RTP hydrolase activity

8-oxo-(d)RTP + H20 = 8-oxo-(d)RMP + diphosphate + H+

deustp01 commented 3 years ago

Yes, if that complies with @pgaudet request for generality in these terms, above

pgaudet commented 3 years ago

Sounds good to me. We just need to figure out how to handle RHEA xrefs.

@cmungall are you OK with NarrowMatch?

Thanks, Pascale

hdrabkin commented 3 years ago

+id: GO:0106379 +name: 8-oxo-(d)RTP hydrolase activity

lbreuza commented 3 years ago

Hi, how does that strategy fits with the existing GO terms: 2-hydroxy-ATP hydrolase activity and 2-hydroxy-dATP hydrolase activity for instance? Will GO:0035539 8-oxo-7,8-dihydrodeoxyguanosine triphosphate pyrophosphatase activity (synonym = 8-oxo-dGTPase activity) be a child of 8-oxo-(d)RTP hydrolase activity? Thanks. @ukemi ?

deustp01 commented 3 years ago

Will GO:0035539 8-oxo-7,8-dihydrodeoxyguanosine triphosphate pyrophosphatase activity (synonym = 8-oxo-dGTPase activity) be a child of 8-oxo-(d)RTP hydrolase activity?

It should be a child or, perhaps in line with current reduced-granularity ideas, it should be made obsolete and replaced by its new parent? @pgaudet ? @ukemi ?

hdrabkin commented 3 years ago

My understanding is that these specific terms would NOT be made?. The RHEAs would somehow be made 'narrow' xrefs; don;t know how one would then do logical defs based on the rheas. s.

lbreuza commented 3 years ago

A quick scan of the literature (e.g PMID: 33824325) shows that these different nucleotides (8-oxo-dGTP, 2-hydroxy-ATP, O6-methyl-dGTP, N6-methyl-dATP) might be produced through different pathways and have their own specific physiological significance. As @pgaudet was commenting above, maybe there is something to discuss there. as similar question already came out in #21259. Please note that while @deustp01 queried these specific terms I was also planning to do so as I am updating these proteins in UniProt and I am using protein2GO.