# HISTORY 27 Mar 2016: Updated by: TOUCHUP-v1.15 18 Mar 2016: Updated by: TOUCHUP-v1.14 # molecular_function 20100723: node_PTN000657188 has function ATPase activity, coupled to transmembrane movement of substances (GO:0042626) 20100723: node_PTN000657462 has function auxin efflux transmembrane transporter activity (GO:0010329) 20100723: Euteleostomi_PTN000657279 has function canalicular bile acid transmembrane transporter activity (GO:0015126) 20100723: Bacteria _PTN000657720 has function lipid-transporting ATPase activity (GO:0034040) # cellular_component 20100726: node_PTN000657188 is found in integral component of membrane (GO:0016021) 20100726: Eukaryota_PTN000657864 is found in mitochondrial inner membrane (GO:0005743) # biological_process 20100728: node_PTN000657188 participates in transmembrane transport (GO:0055085) 20100728: node_PTN000657462 participates in basipetal auxin transport (GO:0010540) 20100728: Eukaryota_PTN000657864 participates in cellular iron ion homeostasis (GO:0006879) 20100728: Euteleostomi_PTN000657279 participates in canalicular bile acid transport (GO:0015722) 20100729: Euteleostomi_PTN000657506 participates in antigen processing and presentation of endogenous peptide antigen via MHC class Ib via ER pathway, TAP-dependent (GO:0002489) 20100729: Euteleostomi_PTN000657506 participates in antigen processing and presentation of exogenous protein antigen via MHC class Ib, TAP-dependent (GO:0002481) 20100729: Euteleostomi_PTN000657506 participates in antigen processing and presentation of endogenous peptide antigen via MHC class I via ER pathway, TAP-dependent (GO:0002485) 20100729: Euteleostomi_PTN000657506 participates in positive regulation of antigen processing and presentation of peptide antigen via MHC class I (GO:0002591) # WARNINGS - THE FOLLOWING HAVE BEEN REMOVED FOR THE REASONS NOTED # NOTES hylogeny This is a large family with 434 members and a number of duplications. There are three paralogous clades off the root as follows: AN568: vertebrate Abcb8, worm haf-6, fly, dicty abcB4, bacterial outgroup AN623: mitochondrial transporters (with bacterial outgroup, but no E. coli) AN1: bacterial outgroup with 2 large eukaryotic clades Basically, this looks like a large set of transporters with diverse substrates. There are no obvious problems with the phylogeny. Full description of structure here: AN0 AN568: vertebrate Abcb8, worm haf-6, fly, dicty abcB4, bacterial outgroup AN623: mitochondrial transporters (with bacterial outgroup, but no E. coli) AN624: vertebrate ABCB6, worm hmt-1, fly, pombe hmt1 (heavy metal tolerance), dicty abcB6, plasmo, chlamy AN676: animal proteins and plant outgroup AN677: vertebrate ABCB7, fly, yeast ATM1, pombe atm1, dicty abcB5 AN726: Arabidopsis ATM1, ATM2, STA1 AN1 AN487: bacterial outgroup, including E. coli mdlA and msbA; msbA is a lipid/drug transporting ATPase AN3: 2 euk clades (with archaeal outgroup) (Note that plasmo, tetrahymena, and leishmania have proteins in each; evidence of duplication) AN311 AN471: plasmo/tetrahymena Q4QA42: leishmania AN477: 2 Arabidopsis ABC's AN312 AN468: 2 dicty ABC's AN313 AN459: yeast MDL1 and MDL2, pombe YNT9 AN314 AN418: Abcb10, human thru fly, mitochondrial? Worm & fly proteins not characterized AN315 AN402: worm haf-2, haf-4, and haf-9 "half transporters" (plus a couple others) AN316-AN317: vertebrate antigen-processing transporters (plus urchin outgroup) AN318: TAP2 (N.B.: fish protein is named differently, abcb3b and -3l1) AN350: TAP1 AN376: ABCB9, uncharacterized and "TAP-like" AN4 AN269: Arabidopsis (and plant) PGP's/ATPases AN264: leishmania ABC transporters and p-glycoproteins AN235: plasmo and TETTH ABC transporters AN5 AN220: dicty abcB2 and 3, plus a clade of entamoeba proteins AN6 AN197: pombe pmd1 and fungal clade AN7 AN143 AN144: 12 paralogous clades of worm pgp's AN190: fly Mdr49 (& 3 paralogs) AN8 AN9: 1 zebrafish & 2 urchin proteins AN14 AN127: bunch of urchin proteins AN15 AN90: Abcb11 human thru Ciona. Bile salt export pump. AN16 AN84: 5 Ciona proteins AN17 AN78, 81: zebrafish and Xenopus outgroups, 1 stray chicken protein, probably not real AN19 AN20: ABCB5 human thru Xenopus AN33 platypus, chicken, and frog outgroups AN36 AN37: Abcb4, human through possum AN52: Abcb1 and -1a, human through possum MF -There are multiple annotations to parents and children of GO:0042626 "ATPase activity, coupled to transmembrane movement of substances." Try to propagate this term as widely as possible, and propagate specific substrates within clades. -GO:0015421 "oligopeptide-transporting ATPase activity" is found on yeast MDL1, but the SGD gene summary paragraph for MDL1 states, "Mdl1p and Mdl2p are similar to each other and to other ABC family members, in particular to human TAP1 and TAP2, which form a heterodimer that transports peptides from the cytoplasm into the endoplasmic reticulum." Therefore, it is justified to propagate 15421 to the clade containing yMDL1 and hTAP1 and hTAP2; AN311 is reasonable. -GO:0034040 "lipid-transporting ATPase activity" is found on E. coli msbA, which transports LPS (see EcoCyc description), so propagate 34040 among bacteria (AN532). msbA also has a "lipid binding" annotation, but I couldn't find any others, so I can't refine the scope. Do not propagate "lipid binding," assuming that it is part of the transport (see ontology questions below). -There are annotations to GO:0042626 in 2 of the 3 major clades, AN1 and AN622. The other clade, AN568, has no MF annotations, but several members described as ATP-binding cassette-containing proteins similar to know transmembrane transporters, so it is justified to propagate this term to all proteins in PTHR24221. Also, this covers (or should, at least) "ATP binding" and "nucleotide binding," so don't propagate those. Also don't propagate GO:0043531 "ADP binding" from the 2 rat TAP's because it seems to be just that this is the product of the ATPase reaction. -GO:0015086 "cadmium ion transmembrane transporter activity" is not defined with ATPase activity, so this is a separate issue. There is an IMP to pombe hmt1, an IDA to worm hmt-1, and a NOT IGI to fly HMT-1 (FBgn0038376) (IGI with pombe hmt1). Propagate to AN624 (but be aware that it will not transfer to fly while the NOT stands). -Propagate GO:0015126 "canalicular bile acid transmembrane transporter activity" within Abcb11 clade, but limit to vertebrates because bile acid is produced in the liver/gall bladder. -Propagate GO:0010329 "auxin efflux transmembrane transporter activity" from 2 Arabidopsis proteins to rest of plant clade (but not to chlamy). -Do not propagate GO:0015238 "drug transmembrane transporter activity" or GO:0008144 "drug binding" since these terms describes an assay, not an existent biological activity and "drug" is one of those nebulous terms. -Propagate GO:0042288 "MHC class I protein binding" from rat TAP1 & TAP2 to AN317 to cover both these clades and a third paralogous vertebrate clade with no annotations. Same for GO:0046980 "tapasin binding," GO:0042605 "peptide antigen binding," GO:0046978 "TAP1 binding," and GO:0046979 "TAP2 binding." CC -Membrane annotations: GO:0016021 "integral to membrane" is found only on worm haf-6 and fly Q9VF20, but these are in 2 of the 3 major clades and "membrane" annotations are found throughout the family. Also, these proteins have multiple membrane-spanning domains and are involved in transmembrane transport. Propagate 16021 to AN0. Try to propagate membrane annotations to specific compartments within clades. -Mitochondrion: There are annotations to GO:0005743 "mitochondrial inner membrane" for mouse Abcb10 and yeast MDL1 and MDL2. The most parsimonious explanation is that this property arose in AN313. (There's nothing in plant or dicty, even in MF or BP, to indicate any mitochondrial function in this clade.) However, this would also propagate the annotation to AN315, and there is no evidence, other than a single HTP "mitochondrion" annotation on mouse Tap1 -- which is otherwise believed to be involved in antigen presentation, which occurs at the cell surface -- to indicate that any descendants of AN315 are found in the mitochondrion. Propagate 5743 to AN313 and block propagation to AN315. -Mitochondrion, again: There are annotations to GO:0005743 "mitochondrial inner membrane" for mouse Abcb7, yeast ATM1, and pombe atm1, plus annotations to GO:0005739 "mitochondrion" for Arabidopsis ATM2 and STA1. There are additional annotations to "mitochondrion" for mouse and rat Abcb6; the mouse paper is HTP, but the rat paper has several lines of supporting evidence, including complementation experiments in yeast atm1 mutants. However, the intervening annotated proteins, such as worm hmt-1 and pombe hmt1, appear to be vacuolar, and the branch length to AN633 is conspicuously long, so it's not justified to propagate "mitochondrial inner membrane all the way to Abcb6. Propagate 5743 to AN676 and "mitochondrion" to AN633. All remaining mitochondrial annotations are HTP; do not propagate. -Vacuole/lysosome: There are several annotations to child terms scattered throughout the family. Propagate narrowly: GO:0005765 "lysosomal membrane" to AN410; GO:0000325 "plant-type vacuole" and GO:0005774 "vacuolar membrane" to AN483; and GO:0000329 "fungal-type vacuole membrane" to AN664 and GO:0005774 "vacuolar membrane" to AN624. -There are IDA's to GO:0000139 "Golgi membrane" and GO:0016324 "apical plasma membrane" for rat Abcb1, 4, and 11, almost all from the same paper. Propagate to AN15 to cover all 3 clades, but not to AN14 to cover urchin as this was observed in liver. (Allow the annotations to propagate to the Ciona proteins in these clades, though.) Propagate GO:0046581 "intercellular canaliculus" from mouse Abcb1, 1a, 4, and 11 to AN15 and note that many of these annotations are consistent as the apical and basolateral membranes are often continuous in epithelial or ductal cells. -Including "intercellular canaliculus" and "apical plasma membrane, there are widespread annotations to GO:0005886 "plasma membrane", including pombe pmt1 and many Arabidopsis proteins out to ATTAP2. Propagate 5886 to AN3, but no farther as archaea don't have plasma membranes. Based on comments from Pombase (SF# 3073344), block propagation into clades already annotated as "mitochondrial": AN418 and 459, children of AN313, above. I'd include AN676 and AN633, too, but they're not descended from AN3 and so don't inherit the "plasma membrane" annotation. -Propagate GO:0009941 "chloroplast envelope" to AN295 and AN480 but not to paralogs. Similarly, propagate GO:0005634 "nucleus" to AN285. -Propagate GO:0042825: "TAP complex" to AN317 as for 42288 under MF, above. BP -Propagate GO:0006879 "cellular iron ion homeostasis" to AN676 based on mouse, yeast, and pombe annotations in that clade. -Propagate GO:0070574 "cadmium ion transmembrane transport" to AN624 to match the MF. -Propagate GO:0015722 canalicular bile acid transport to AN91 to match the MF. -Propagate GO:0007041 "lysosomal transport" to AN410 to match the CC -Do not propagate GO:0015893 "drug transport" or any of its children for the same reason as for 0015238, above, in MF. -Propagate GO:0010540 "basipetal auxin transport" and GO:0010541 "acropetal auxin transport" to AN274 to match the MF. -Propagate GO:0015833 "peptide transport" to AN317 to match other TAP complex annotations in MF and CC. -Propagate GO:0055085 "transmembrane transport" to AN0. -There are annotations to GO:0046686 "response to cadmium ion" on worm, fly, and pombe hmt1. Since the response seems to be sequestration in the vacuole (PMID 19054771) and these are all vacuolar proteins and cadmium transporters, propagate the child term GO:0071585 "detoxification of cadmium ion" to AN624 to match the MF and CC annotations. -Propagate GO:0001916 "positive regulation of T cell mediated cytotoxicity" to AN317 to match TAP complex. Questions for MOD curators General -This is a family of transmembrane transporters. Whenever possible, please try to annotate specific substrates, compartments, and membranes. ST replies August 2, 2010 12:02:17 PM EDT: Added vacuolar membrane for the fly. -Human, mouse, and rat: ABCB6 is mitochondrial, but inherits MF, CC, and BP annotations from AN624 pertaining to vacuolar detoxification of Cd+2. In the absence of contradictory information, these inferences will stand despite the long branch length to the ABCB6 clade. Can you make any annotations that would inform on the function of ABCB6? None currently exist. RH replies, August 6, 2010 8:42:01 AM EDT: I've found a paper from 2008 (PMID:18279659) which states that ABCB6 is not mitochondrial but is in fact associated with the Golgi, so I have annotated to Golgi apparatus and 'NOT' mitochondrion. However, another publication from 2007 (PMID:17661442) demonstrates the localisation of ABCB6 as intergral to mitochondrial outer membrane and plasma membrane - which I have also annotated. cont'd: PMID:17661442 describes the efflux of porphyrins on expression of ABCB6 in mitochondria, so I have annotated to heme transport and efflux transmembrane transporter activity (might be a case for requesting a more specific term 'porphyrin efflux transmembrane transporter activity'?) I have also annotated to heme-transporting ATPase activity and heme binding from PMID:17006453. MF -All organisms with "protein binding" annotations: Is there a more specific child term you can choose? RH replies, August 6, 2010 8:42:01 AM EDT: TAP1 and ABCB9 membrane locations have been upgraded to integral to endoplasmic reticulum membrane -E. coli: The summary for msbA states "MsbA has also been shown to transport N-acetyl-glucosamine." Can you annotate that? (cf. http://biocyc.org/ECOLI/NEW-IMAGE?type=NIL&object=EG10613-MONOMER) -Fly: Please double-check the Hmt-1 GO:0015086 "cadmium ion transmembrane transporter activity" NOT annotation (IGI with pombe hmt1) from PMID 19001374. This paper shows partial complementation of the pombe mutant by the fly protein, but the annotation seems to be based on failure to observe Cd2+ accumulation in the vacuole. The authors conclude that Cd2+ detoxification occurs by "some other mechanism"; is this sufficient to make a NOT annotation without specifying a particular membrane? Please note that pombe and worm have similar complementation experiments annotated in the positive (PMID's 19054771 and 15840570). ST replies August 2, 2010 12:02:17 PM EDT: The partial complementation of the yeast phenotype was captured with the positive 'response to cadmium' annotation. The NOT annotation was made because, contrary to expectation, 'it was concluded that DmHMT-1 does not transport Cd2 and/or CdPC2 and apoPC2 complexes.' It also reflected the sum total of experimental molecular function data for the fly protein and cast some doubt on the role of the pombe protein. However, I acknowledge that the experiment was carried out on a specific membrane so a more specific NOT annotation would be ideal - trouble it that it really isn't practical to make every new specific term and this is a case where I adopted the pragmatic approach of making do with an existing term. I also acknowledge that the experiment carried out in this heterologous system is not ideal. However, I have mixed feelings about removing it because then we'd be left with only positive ISS annotations based on the other species and that wouldn't reflect all the data. I feel there has to be room for capturing conflicting data until there is a definitive answer. - For now, do not propagate cadmium ion transporter activity. CC -All organisms with annotations to membrane-bound organelles: Can you support annotation to the membranes of said organelles? -All organisms with annotations to specific membranes: Can you support annotation to "integral to" said membrane? -Worm: Please verify GO:0030176 "integral to endoplasmic reticulum membrane" for haf-6. Is "integral" justified? Was the identification of the ER strong enough to make the annotation? KVA replies by email, July 30, 2010 5:39:47 PM EDT: These are the statements from the paper on which the annotation is made: We did observe co-compartmentalization using anti-HAF-6 and anti-Calreticulin antibodies, indicating that HAF-6 likely resides in the endoplasmic reticulum and not in mitochondria, as was expected from comparisons of sequence similarity to human half-transporters. Interestingly, HAF-6 localization extended to perinuclear regions in germline tissue, which is consistent with an endoplasmic reticulum (ER) localization (Figure 5G). And from the figure legend: (B) The HAF-6::GFP protein (left) is localized to the same compartment as calreticulin (CRT-1), a primarily ER resident with a C-terminal KDEL sequence. Similar results were obtained in cohybridization experiments using the anti-HAF-6 antibody with anti-CRT-1. HAF-6 is a transmembrane protein, and 100% spectral overlap with CRT-1 is not anticipated; nonetheless, co-compartmentalizaion is observed. Hope that's sufficient for the annotation. BP -Yeast: Does PMID 11251115 justify annotation to GO:0006839 "mitochondrial transport" in addition to the existing "oligopeptide transport" IMP? -Zebrafish: Should the GO:0046686 "response to cadmium ion" IDA for abcb3l1 (Q801W2) from PMID 16799861 actually be an IEP? DH replies by email, August 2, 2010 12:35:02 PM EDT: The paper reports an increase in expression of abcb3l1 in response to Cd exposure. Currently the annotation is to 'response to cadmium ion" by IDA. Perhaps IEP is correct, or, perhaps some groups wouldn't make this annotation at all? I'd like to hear from other curators what they do with this sort of data? Is increased expression in response to a compound enough evidence to annotate that gene product as being involved in the "response" to that compound? I could change to IEP or delete altogether... RB replies, August 2, 2010 2:19:34 PM EDT: There was lot of discussion at the GO camp about using IEP to annotate to response to terms. Please check the discussion page and minutes- http://gocwiki.geneontology.org/index.php/2010_GO-camp_Response_to_terms_issues Minutes from May13th conf.call- http://gocwiki.geneontology.org/index.php/Minutes_from_May_13th_2010_call Some DBs use this data to make BP annotations, but some don't (SGD typically wouldn't make this annotation). But if you decide to use IEP data, please use judgement. DH replies, August 2, 2010 2:47:46 PM EDT: Based on the debate in the links, and the IEP documentation, and the content of the paper, I think I'll keep the annotation and change change the evidence code to IEP in this case. Questions for ontology curators -Does "ATPase activity" has_part "ATP binding"? What about "ADP binding" (since ADP is the product of the ATPase reaction)? Generally speaking, does an enzyme bind both its substrate and product, and should that be reflected in the ontology? -Does GO:0034040 "lipid-transporting ATPase activity" has_part GO:0008289 "lipid binding"? How about is_a? DH at MGI replies, August 2, 2010 8:08:25 AM EDT: This has been a long long....... discussion in the GOC. As part of the entire chemical revamp in GO, our goal is to align all of the chemicals in GO with ChEBI. This will first be done with all of the chemicals in term names. The next step will be to create logical definitions of all GO molecular function terms based on ChEBI as well. Once we do that, we will be able to reason about reactants and products for molecular functions. -There is no definition for GO:0015126 "canalicular bile acid transmembrane transporter activity." DH at MGI replies, August 2, 2010 8:08:25 AM EDT: defined and its process parent -Should GO:0007595: "lactation" be a child of GO:0007589 "body fluid secretion"? DH at MGI replies, August 2, 2010 8:08:25 AM EDT: done -Should GO:0042824 "MHC class I peptide loading complex" be a child of GO:0030176 "integral to endoplasmic reticulum membrane"? 42824 is defined as including TAP, which consists of 2 ABC transporters, which are 6-TM ATP-dependent transmembrane transporters. AD@MGI replies, July 30, 2010 1:44:23 PM EDT: I believe, as you suggest, that GO:0042824 "MHC class I peptide loading complex" should be a child of GO:0030176 "integral to endoplasmic reticulum membrane". I will put in a SF request for this. SF item # 3037241 DH at MGI replies, August 2, 2010 8:08:25 AM EDT: done MSL, revised 2010 Dec 20 PDT, 2011 June 13 (removed cadmium ion transporter activity for AN624) HM, reviewed 2014 May 5 # REFERENCE Annotation inferences using phylogenetic trees The goal of the GO Reference Genome Project, described in PMID 19578431, is to provide accurate, complete and consistent GO annotations for all genes in twelve model organism genomes. To this end, GO curators are annotating evolutionary trees from the PANTHER database with GO terms describing molecular function, biological process and cellular component. GO terms based on experimental data from the scientific literature are used to annotate ancestral genes in the phylogenetic tree by sequence similarity (ISS), and unannotated descendants of these ancestral genes are inferred to have inherited these same GO annotations by descent. The annotations are done using a tool called PAINT (Phylogenetic Annotation and INference Tool).