#10043 targeting vs. localization to nucleus

[Midori Harris]

We at PomBase are wondering if 'protein localization to nucleus' and 'protein targeting to nucleus' are really the same; if they're not, we'd really appreciate having the distinction clarified. Both seem to have standard defs plugged in (I did one of them years ago):

[Term]
id: GO:0034504
name: protein localization to nucleus
namespace: biological_process
def: "A process in which a protein is transported to, or maintained in, a location within the nucleus." [GOC:ecd]
synonym: "protein localisation to nucleus" EXACT [GOC:mah]
synonym: "protein localization in cell nucleus" EXACT []
synonym: "protein localization in nucleus" EXACT [GOC:mah]
is_a: GO:0033365 ! protein localization to organelle

[Term]
id: GO:0044744
name: protein targeting to nucleus
namespace: biological_process
def: "The process of directing proteins towards the nucleus, usually using signals contained within the protein." [GOC:jl]
is_a: GO:0006605 ! protein targeting
created_by: janelomax
creation_date: 2012-11-07T15:45:54Z

Right now the only connections I can easily see are that both have paths to cellular protein localization, and they're both part_of parents of 'protein import into nucleus':

[Term]
id: GO:0006606
name: protein import into nucleus
namespace: biological_process
def: "The directed movement of a protein from the cytoplasm to the nucleus." [GOC:jl]
synonym: "establishment of protein localization to nucleus" EXACT [GOC:mah]
synonym: "protein import into cell nucleus" EXACT []
synonym: "protein nucleus import" EXACT []
synonym: "protein transport from cytoplasm to nucleus" EXACT []
is_a: GO:0006886 {is_inferred="true"} ! intracellular protein transport
is_a: GO:0017038 {is_inferred="true"} ! protein import
is_a: GO:0051170 ! nuclear import
relationship: part_of GO:0034504 ! protein localization to nucleus
relationship: part_of GO:0044744 ! protein targeting to nucleus

This may mean that both protein localization and protein targeting need a more general cleanup, which is a bit scary.

[Tanya Berardini]

I talked to David and here's his take:

targeting - setting up the protein for where it's going
localization - actually moving it or keeping it in one place

I do agree that the definitions are not the greatest. This is hard. I think I might need to kick this up to discussion with the other GO eds.

Thanks,

Tanya

[Midori Harris]

That sounds like a sensible distinction (although I imagine it'll often be hard to tell which to use for actual annotations a lot of the time).

I'd go a bit further and guess that the structure ought to be:

protein targeting to x part_of establishment of protein localization to x (which is in turn already part_of protein localization to x)

and some protein localization to x processes may not have a targeting part well documented, so we probably don't need a targeting term for every x (possibly a source of the present muddle?)

Very happy to put the other GO eds on it!

cheers,
m

[Midori Harris]

We got onto the targeting vs. localization thing again today, talking about the vacuole rather than the nucleus, but same general issue. We thought of another possible distinction: targeting always involves some sort of recognition sequence (e.g. NLS, signal sequence, etc.) on a protein that's getting moved. Not all targeted proteins have to have the targeting sequence, because a protein can be moved around by binding to another protein with the sequence.

Does that hold enough of the time to be one of the distinguishing features?

[Valerie Wood]

targeting - setting up the protein for where it's going

In that case, I don't understand the difference between targeting and transport.
I remember years ago I asked an expert about this, and targeting was used very specifically to refer to proteins which were transported based on signals within their sequence, which served as a signal to identify the proteins destination. It has to be based on information contained in the protein itself.

[Pascale Gaudet]

Hello,

Can you give example of how 'targeting' would be correctly used ? Would ubiquitin be involved in targeting to the proteosome ? What else ? I don't know of any proteins that work to 'target' proteins to certain cellular locations (but perhaps I am just no aware of them).

I am worried that many proteins that have localization signals get annotated to the targeting terms (see for example: MGI Pex7, annotated to GO:0016558 protein import into peroxisome matrix). This seems like a CC annotated as a BP. In this case the annotation reflects a property of the protein's primary structure and not a real biological process.

Pascale

[Valerie Wood]

From my understanding, ubiquitin mediated transport is not targeting because the signal is not intrinsic to protein itself sequence, it's modification dependent.

But I agree, it is a slightly arbitrary classification. I would be happy if the targeting terms were just related synonyms of the transport terms, and we distinguished based on mechanism (I don't think targeting is used so often these days, but others may know better).

[Pascale Gaudet]

Hi Val,

That sounds good to me as well. I'd be in favor or merging those terms.

Pascale

[Tanya Berardini]

Discussed during jamboree today. We will clean this up.

import
--[p] targeting

Definitions will also be clarified.

[Valerie Wood]

How will the definitions be clarified? I'm still not sure what exactly we mean by targeting (and it isn't always part of import , see below)

GO:0072661 protein targeting to plasma membrane
The process of directing proteins towards the plasma membrane; usually uses signals contained within the protein.

This is an odd one because it is used to annotate (mainly) geranylgeranyltransferase s

Although for the proteins which are transported, the signal for geranylgeranyl is in the protein, the regulation of the transport is modification dependent, so it isn't completely
signal dependent. Its hard to use the 'signal within a protein' criteria because its a bit arbitrary. Sometimes the signal sequence is in an accessory protein etc.
In this instance it seems that the targeting described is "regulation of transport to plasma membrane"

[Valerie Wood]

There is also a term

"protein targeting to Golgi'
what does this even mean? How is it different from
GO:0006888 ER to Golgi vesicle-mediated transport
val

[Valerie Wood]

also
CVT pathway
http://www.ebi.ac.uk/QuickGO/GTerm?id=GO:0032258#term=ancchart
is a child of protein targeting, but this seems as though it should be just
"protein transport to vacuole"

[Marc Feuermann]

Hi all,
I think we should indeed make some cleaning in these terms. Most curators do not distinguish between them (even if sometimes there are some subtle differences) and orthologous proteins from different species are annotated differently which gives a lot of headaches to PAINT curators.
However, the CVT pathway is a really specific "protein transport to vacuole" pathway in yeast. The CVT pathway delivers hydrolases to the vacuole and specifically uses the machinery of autophagy. So I would keep it as a separate term (but modify the definition).
Regards,
Marc.

[Valerie Wood]

Hi Mark,
I agree that we should keep CVT pathway as a separate term. However I think it should be described as a type of 'protein transport', rather than 'protein targeting'
Val

[Marc Feuermann]

Hi Val,
Okay, I totally agree. Sorry, I misunderstood the point. I think 'protein transport' and 'protein targeting' are (almost?) the same, so if we obsolete one (or merge them), the problem would be solved.

[Valerie Wood]

Just wanted to point out

GO:0017038 : protein import
OLD: The directed movement of proteins into a cell or organelle. 
NEW: The targeting and directed movement of proteins into a cell or organelle. Not all import involves an initial targeting event.

1.

I spotted this change to import, but no update to this ticket. that for mitochondrial ER and vacuolar protein import, we need the “targeting” terms to be children of the transmembrane transport term
at the moment (for example)
http://www.ebi.ac.uk/QuickGO/GTerm?id=GO:0045039#term=ancchart
“protein targeting to mitochondria”
is a child of
GO:0045039 protein import into mitochondrial inner membrane
it should be the other way around

2.

In addition- note that currently “protein targeting to nucleus” is not a child of “protein import into nucleus”. However I still don’t think there is any difference between these terms. Researchers do not distinguish between import and targeting for nucleocytoplasmic transport as far as I can tell.

One option would be to restrict the use of the defined targeting to “setting up for translocation across a membrane”. This would enable the nucleocytoplasmic term to have a related synonym for targeting, so people find it with a search, but not to be defined as targeting

In fact, this restriction fits with the wikepedia description of targeting which appears to refer to targeting only to ER, mitochondria, peroxisome and chloroplast as part of transmembrane transport:
http://en.wikipedia.org/wiki/Protein_targeting

Localisation to plasma membrane via ‘targeting’ is via targeting to ER so it’s misleading to have the terms for ‘GO:0072661 protein targeting to plasma membrane’ these terms should be localisation and or transport terms as the mechanism is either via i) vesicle mediated transport via secretory system (after ER targeting), ii) diffusion iii) microtubule based so the targeting to the plasma membrane is indirect consequence of ER targeting.

Therefore : protein targeting to plasma membrane
should probably be merged obsoleted (unless there are bacterial exceptions in which case it needs a taxon restriction)
(there are only 31 experimental annotations).

also this one
GO:0072693 protein targeting to prospore membrane
isn’t transport via targeting, there is another open SF ticket about this (this term could be obsoleted as it has only IEA annotations, the pombe and cerevisiae annotations have already been cleaned up)