obsolete 'polyketide synthase activity ; GO:0016218'

[gocentral]

Undefined. This is a multimeric complex with many different functional domains. We have the corresponding cellular component terms, and a process term, so I suggest we obsolete this term. It has 37 annotations - mostly ISS.

Reported by: jl242

Original Ticket: geneontology/ontology-requests/7119

[gocentral]

Jane,

Maybe we should ask a biochemist about that. The activity of PKS's is very complicated, as you point out having about 7 different functional domains. I wonder how those genes will be annotated without this term - it seems messy to give 7 different activities. Can you wait a bit before obsoleting this?

Thanks, Pascale

Original comment by: pgaudet

[gocentral]

Hi Pascale - sure, there's no rush. It actually came up because someone asked for a term for nonribosomal peptide synthetase which has the same issue. The way I was thinking we will ultimately model this will be to have the process term with functions as its parts e.g.:

polyketide synthesis ---[p] polyketide synthesis, function A ---[p] polyketide synthesis, function B ---[p] polyketide synthesis, function C

and for the cc term you could do this:

polyketide synthase complex -carries_out_process-> polyketide synthesis

Original comment by: jl242

[gocentral]

also see

https://sourceforge.net/tracker/?func=detail&aid=1866936&group\_id=36855&atid=440764 and http://wiki.geneontology.org/index.php/Polyketide\_synthases

(not that I'm objecting to obsoletion -- just some background on those pages) m

Original comment by: mah11

[gocentral]

Hi Jane,

That might work, but I least in Dicty I think they act as a monomer, so you cannot use the CC -> BP mapping. Would it them be possible to use PRO -> BP mapping? Is that a (future ) possibility? (The distinction between monomers and complexes is 'unfair' when it comes to providing those links, as you loose a lot if you are not a complex!)

Thanks, Pascale

Original comment by: pgaudet

[gocentral]

Hi Pascale - yes, it is a problem where a gp isn't a complex. I guess theoretically we could use PRO...it just starts to looks a bit like an annotation, although PRO is all classes rather than instances I think? I've asked Chris to look at this - he may have some thoughts.

Perhaps we could use this example as a pilot for adding a full set of functions for a process as it's a pretty small case study...

Original comment by: jl242

[gocentral]

Hi Jane, I just chatted with Kristian - you can obsolete.

Original comment by: pgaudet

[gocentral]

Obsoletion request sent out.

Original comment by: girlwithglasses

[gocentral]

• assigned_to: jl242 --> girlwithglasses

Original comment by: girlwithglasses

[gocentral]

Done!

Original comment by: girlwithglasses

[gocentral]

• status: open --> open-fixed

Original comment by: girlwithglasses

[gocentral]

• status: open-fixed --> closed-fixed

Original comment by: girlwithglasses

[marcfeuermann]

I am reopening this ticket because I was wondering if it would be better to reintroduce the MF term "polyketide synthase activity". Indeed, polyketide synthases are composed of different modules with various molecular functions. It is however difficult to annotate each of these functions separately, sometimes we don't know in which order they act. Similarly, Rhea annotates these enzymes as performing only one reaction. Also, I've seen that there are a lot of MF terms related to synthesis of specific polyketides, so why not group them together in a parent term ? There is a ChEBI for polyketide (CHEBI:26188 ). Moreover, in many cases, pathways are characterized via deletion of each enzymes and sometimes intermediates are not all identified, so we know that we have the polyketide synthase activity but the precise product is not known. Finally, with the new GO-CAM visualization tool we could now directly associate this activity with the product (when known).

The situation is exactly the same with non-ribosomal peptide synthetases and a term "non-ribosomal peptide synthetase activity" could also be created.

These MF terms would be useful for the modeling of pathways (and make it more simple). I'm aware that there is no ideal solution and I'm ready to discuss this further. Take care and stay healthy Best regards, Marc.

[pgaudet]

Discussing with @marcfeuermann Polyketide synthases catalyze multistep reactions, iteratively going from basic building blocks (acyl-CoA) and producing a polyketide (chain of alternating ketone (or reduced forms of a ketone) and methylene groups: (-CO-CH2-); cyclic or not).

According to the GO guidelines, we dont create individual reactions for multistep reactions, so I will restore this term.

Thanks, Pascale

[pgaudet]

For the NRPs we will reuse

id: GO:1904091 -name: peptidyl carrier protein activity +name: non-ribosomal peptide synthetase activity

@marcfeuermann will provide definitions for both terms

[marcfeuermann]

Indeed, re-introducing "polyketide synthase activity" and creating "nonribosomal peptide synthetase activity" would not only fit to the guidelines but be really useful for biosynthetic pathways annotation.

A good example is the reaction performed by the highly reducing polyketide synthase bet1 from Neocamarosporium betae (https://www.uniprot.org/uniprotkb/A0A0C6E0I7/entry) It performs the complex multi-step reaction leading to the production of dehydroprobetaenone I (https://www.rhea-db.org/rhea/51348). Rhea to GO mapping would allow the creation of the MF "dehydroprobetaenone I synthase activity" which could be annotated as part_of the "betaenone biosynthesic process" BP (to be created too). In many cases, the exact product of the PKS is not known and we could instead use the "polyketide synthase activity" parent term.

Having a generic MF term for both PKS and NRPS enzymatic activities would also be useful to search for these activities in the databases.

I would propose the following definitions:

Polyketide synthase acvtivity: Catalysis of the multistep reactions that produce polyketides through decarboxylative condensation of carboxylic acids performed by a multimodular enzyme called polyketide synthase (PKS). The fundamental chain-elongation reaction, a C–C bond-forming step, is mediated by a ketosynthase (KS) domain that catalyzes the transfer of the polyketide acyl chain to an active-site cysteine of the KS domain, followed by condensation with a malonyl unit covalently bound to a downstream acyl carrier protein module (malonyl-S-ACP) by a decarboxylative acylation of the malonyl donor unit. Additional essential component of the core PKS chain-elongation apparatus is an associated acyltransferase (AT) domain, which catalyzes the priming of the donor ACP sidearm with the appropriate monomer substrate, usually malonyl-CoA. Supplementing these core chain-elongation domains are variable numbers of auxiliary domains that are responsible for modification of the growing polyketide chain by a small set of iterated reactions including methylation, ketoreduction, dehydration and enoylreduction.

Nonribosomal peptide synthetase activity: Catalysis of the multistep reactions that produce nonribosomal peptides performed by a multimodular enzyme called nonribosomal peptide synthetase (NRPS). The key chain-building reaction, a C–N bond-forming reaction, involves the generation of the characteristic peptide bond by nucleophilic attack of the amino group of an amino-acyl donor unit covalently bound to a downstream peptidyl carrier protein module (amino acyl-S-PCP) on the acyl group of an upstream electrophilic acyl- or peptidyl acyl-S-PCP chain, catalyzed by a condensation (C) domain. Supplementing these core chain-elongation domains are variable numbers of auxiliary domains that are responsible for modification of the growing polypeptide chain by a small set of iterated reactions including epimerization, N-methylation, and heterocyclization.

I think is is worth to describe in details the type of reactions performed by these complex enzymes. Another advantage of having these terms is that users will have access to these details with GO. I don't imagine that each specific "child" reaction (imported from Rhea) will have this precise description. Please feel free to edit if required.

PMID:10631508 is a really nice review for both types of enzymes and I've used this paper to create the definitions.

Finally, many hybrid PKS/NRPS enzymes sharing PKS and NRPS modules have also been described. They could be annotated with both terms.

Thanks a lot. Best regards, Marc.