FITM2

Summary: FITM2 associates with TSPAN18 in the BioPlex interaction screen.

Reason: The exact PMID:28514442 / Q96SJ8 context matches IntAct EBI-21895691: TSPAN18 bait and FITM2 prey in anti-tag coimmunoprecipitation from human HEK293T cells. This supports the association without defining a physiological adaptor or membrane-tether activity. The campaign retains supported generic binding as non-core when no informative functional subtype is established.

Summary: FITM2 interacts with UNC50 in HuRI two-hybrid assays.

Reason: The PMID:32296183 / Q53HI1 context matches prey-pooling, array and validated two-hybrid records EBI-23756614, EBI-24049066 and EBI-24710182 in yeast assay hosts. These are three assay records, not three independent biological functions. They support this incoming interaction but do not transfer the partner function to FITM2 or identify a core binding subtype. Retain as non-core under the explicit generic-binding policy.

Summary: FITM2 interacts with the specified HTATIP2 partner product in HuRI assays.

Reason: The Q9BUP3-3 WITH/FROM entity is HTATIP2 isoform 3, not an alternative FITM2 product. Exact PMID:32296183 records EBI-23729315, EBI-24021946 and EBI-24694273 support two-hybrid association. The recorded partner region 35-276 corresponds to the 242-residue canonical HTATIP2 sequence within its 276-residue isoform 3; its sequence reconstruction matches the official isoform FASTA. This does not establish isoform-exclusive binding or a FITM2 catalytic/adaptor role. Retain the supported generic interaction as non-core.

Summary: BioPlex 3.0 also supports the FITM2-TSPAN18 association.

Reason: The exact PMID:33961781 / Q96SJ8 pair matches EBI-54796289, an anti-tag coimmunoprecipitation association in human HEK293T cells with TSPAN18 bait and FITM2 prey. Preserve this separate publication context. A physical association in a screening system does not identify an informative FITM2 molecular-function subtype, so the explicit campaign policy retains it as non-core.

Summary: FITM2 functions in the endoplasmic reticulum membrane.

Reason: The human Q8N6M3 leaf PTN002508545 lies below PTN000565881, which carries the original positive ER-membrane IBD. Its ungapped family-MSA sequence exactly matches the 262-residue human sequence. ER localization of mouse FIT2 in PMID:18160536 and human FIT2 cellular experiments in PMID:32915949 agree with this ancestral placement; donor count is not used as a confidence measure.

Summary: The combined electronic sources support ER-membrane localization.

Reason: This incoming row combines the mouse P59266 ortholog/Ensembl transfer, InterPro IPR019388 and UniProt SL-0097 mapping. Those sources agree with FITM2 being a multipass ER protein and with the localization/function assays in PMID:18160536 and PMID:32915949. Keep the original combined source context rather than collapsing it into a single donor.

Summary: Manual transfer from mouse Fitm2 supports the ER membrane.

Reason: P59266 is the verified mouse Fitm2/Fit2 ortholog, not FITM1. Mouse membrane fractionation and tagged-protein colocalization support this donor location, and human FIT2 experiments provide independent concordance. The ISS evidence remains a transfer, not a relabeled human localization experiment.

Summary: The original mouse FIT2 study supports this localization transfer.

Reason: PMID:18160536 localizes endogenous mouse FIT2 by membrane fractionation and tests mouse FIT constructs in HEK293 cells; the human host does not make those constructs human. Its brown-adipose fractionation separates ER-associated FIT2 from the droplet fraction. These results support ER membrane as the donor location, consistent with human FITM2 function.

Summary: The Reactome reaction places FITM2 in the ER membrane.

Reason: R-HSA-8857686 describes ER-associated FITM2 binding and partitioning triglycerides into lipid droplets. The location is independently supported even though this same reaction does not settle its more specific membrane-tether activity annotation, reviewed separately.

Summary: FITM2 affects triglyceride pathways in the mouse donor system.

Reason: The Ensembl transfer from mouse P59266 is compatible with reduced triglyceride biosynthetic flux after Fitm2 knockdown in differentiating 3T3-L1 cells. GO:0006641 covers pathways involving triglycerides and does not require FITM2 itself to be the synthesizing enzyme. Retain this broad involvement as non-core: overexpression supports partitioning rather than increased synthesis, and long-term human-knockout synthesis changes in PMID:32915949 were not rescued by FIT2 re-expression.

Summary: The mouse study links FIT2 perturbation to triglyceride flux and storage.

Reason: The full Results of PMID:18160536 distinguish reduced triglyceride biosynthetic flux in knockdown adipocytes from unchanged synthesis and lipolysis in FIT-overexpressing HEK293 cells. This supports contextual pathway involvement, not direct triglyceride synthesis or hydrolysis by FITM2. Preserve the ISS row as non-core; the defining molecular reaction is acyl-CoA diphosphate cleavage.

Summary: Acyl-CoA diphosphatase activity is a supported inherited FITM2 function.

Reason: The positive GO:0010945 IBD at PTN002751040 lies on the exact human target path and is experimentally seeded by human Q8N6M3. A target appearing among the descendant experimental seeds is legitimate, not circular. The official term definition includes acyl-CoA substrates, matching the purified human reaction in PMID:32915949.

Summary: The InterPro mapping agrees with measured acyl-CoA diphosphatase activity.

Reason: IPR019388 maps to GO:0010945, whose official definition explicitly permits acyl-CoA and its derivatives. Purified human FIT2 generates acyl-4-prime-phosphopantetheine and adenosine 3-prime,5-prime-bisphosphate, with loss of activity in catalytic-histidine mutants. This is an experimentally grounded enzyme assignment, not a generic lipid-phosphatase prediction from family membership alone.

Summary: Human FITM2 catalyzes cleavage of the acyl-CoA diphosphate bond.

Reason: PMID:32915949 combines purified human wild-type and H155A/H214A mutant proteins, radiolabeled-substrate assays and product mass spectrometry. The reaction matches GO:0010945, including acyl-CoA substrates. The subsequent microsomal conversion to acyl pantetheine is attributed to another activity and is not assigned to FITM2. Preserve the incoming IMP evidence code despite the paper also containing biochemical assays.

Summary: Triglyceride binding is supported for the mouse FIT2 donor.

Reason: The P59266/Ensembl transfer is consistent with the direct detergent-solubilized FIT2 triolein-binding experiments in PMID:22106267, including saturation and binding/mutant comparisons. Official mouse UniProt donor evidence links this study and the N80A perturbation to reduced binding. The available primary abstract, captions and indexed Results support binding; complete original Methods and all construct species were not recovered, so this remains a transfer rather than a claim of an independently audited human assay.

Summary: Manual ortholog transfer supports FITM2 triglyceride binding.

Reason: The verified mouse P59266 donor carries the direct triglyceride-binding evidence from PMID:22106267. Saturable triolein binding and its relationship to mutant droplet-size phenotypes support the molecular capacity. Keep the ISS context and its species boundary; lipid binding does not by itself demonstrate a physical ER-to-droplet tether.

Summary: FITM2 promotes cellular lipid storage through lipid-droplet organization.

Reason: The InterPro IPR019388/IPR046401 mappings agree with FITM2-dependent droplet assembly and neutral-lipid partitioning in PMID:18160536 and catalytic complementation in PMID:32915949. Lipid storage is the accumulation and maintenance of lipid reserves; this role is broader than catalyzing triglyceride synthesis and does not assign that reaction to FITM2.

Summary: The donor evidence supports diacylglycerol binding.

Reason: The mouse P59266/Ensembl source is consistent with saturable DAG binding and competition with TAG binding in externally accessed indexed Results of PMID:22106267 at PMC3241795, distinct from its abstract-only normal publication cache. This molecular binding capacity is distinct from hydrolysis of DAG or from a membrane-tether assay. Complete original Methods and universal construct provenance were not recovered.

Summary: Manual transfer supports FITM2 diacylglycerol-binding capacity.

Reason: The verified mouse Fitm2 donor and externally accessed indexed DAG-binding/competition Results at PMC3241795 support the existing ISS assertion; the normal PMID:22106267 cache is abstract-only. Retain this access and transfer scope without claiming that DAG is the substrate of the acyl-CoA diphosphatase reaction or that DAG binding alone establishes membrane tethering.

Summary: Combined orthology and InterPro evidence support lipid-droplet organization.

Reason: The P59266/Ensembl and IPR046401 source combination is concordant with mouse droplet-formation assays and human wild-type versus catalytic-mutant complementation. FITM2 contributes through its ER lipid chemistry and associated neutral-lipid handling, rather than being merely a passive droplet cargo.

Summary: The mouse ortholog provides transferable droplet-organization evidence.

Reason: P59266 is the correct mouse Fitm2 ortholog. Its depletion and overexpression alter lipid-droplet accumulation, while human FIT2 catalytic complementation supports the conserved mechanism. The broader organization term is retained alongside the more specific incoming formation term; no redundant NEW assertion is introduced.

Summary: The original mouse FIT2 experiments support lipid-droplet organization.

Reason: PMID:18160536 reports droplet accumulation after mouse FIT2 expression and reduced droplets after knockdown in murine adipocytes. Its fractionation and tracer experiments support partitioning newly synthesized triglyceride into droplets. These are appropriate donor results for the ISS row and are consistent with later human catalytic-rescue evidence.

Summary: Reactome correctly places FITM2 in lipid-particle organization.

Reason: R-HSA-8964572 explicitly includes the FIT proteins in ER-associated neutral-lipid partitioning into droplets. Primary formation and complementation experiments support FITM2 participation. Functions assigned to CIDE proteins or other participants in this broad pathway are not transferred to FITM2.

Summary: FITM2-dependent droplet storage contributes to intracellular triglyceride balance.

Reason: The mouse P59266/Ensembl transfer is consistent with altered cellular triglyceride pools when FIT2 is perturbed and with its supported role in droplet storage. Homeostasis concerns steady-state levels, not a claim of direct triglyceride synthesis. Effects on total triglyceride vary with cell state and perturbation; a universal direction or magnitude is not asserted.

Summary: The mouse study supports a role in intracellular triglyceride homeostasis.

Reason: PMID:18160536 links FIT2-dependent neutral-lipid partitioning and droplet accumulation to cellular triglyceride levels. The annotation describes this storage-related balance, not triglyceride synthase or lipase activity. Later human data support lipid homeostasis while distinguishing secondary long-term transcriptional effects from the catalytic mechanism.

Summary: FITM2 directly breaks down fatty-acyl-CoA by diphosphate cleavage.

Reason: The purified human reaction in PMID:32915949 consumes fatty-acyl-CoA and releases acyl-4-prime-phosphopantetheine plus adenosine 3-prime,5-prime-bisphosphate. FITM2 itself performs this chemical step, supporting direct participation rather than an inference solely from knockout necessity. This is not mitochondrial beta-oxidation or cleavage of the acyl thioester.

Summary: The mouse evidence establishes impaired lipid storage more clearly than participation in adipocyte differentiation.

Reason: In PMID:18160536, PPARgamma, aP2 and adiponectin decreased in the strongest Fitm2 shRNA1 condition, whereas shRNA2/3 retained broadly similar markers despite reduced droplets. The authors' Discussion interprets the primary effect as impaired triglyceride-droplet accumulation rather than differentiation. A secondary consequence of severe storage impairment is plausible but is not proved to explain the shRNA1 marker changes. Transferring participation in fat-cell differentiation therefore overshoots the demonstrated storage role. The later aP2-driven deletion in PMID:24519944 neither establishes nor excludes an earlier differentiation role. This judgment does not use the off-path family IBD as independent evidence against the separate mouse transfer.

Summary: The donor paper's interpretation supports droplet accumulation without establishing a differentiation function.

Reason: The full PMID:18160536 Results preserve a real shRNA-dependent marker distinction, with reduced PPARgamma, aP2 and adiponectin in the strongest shRNA1 condition but broadly retained markers in shRNA2/3. Its Discussion identifies triglyceride-droplet accumulation as the primary affected process, rather than differentiation. Storage impairment could indirectly affect markers, but that causal explanation was not established. The existing ISS assertion is therefore marked over-annotated, while the marker observation and an unresolved early differentiation role are retained. Late aP2-driven deletion in PMID:24519944 cannot establish or exclude that early role; an off-path family IBD is not treated as a separate refutation of this mouse-derived assertion.

Summary: Lipid homeostasis is supported by the conserved FITM2 mechanism.

Reason: PTN000565881 lies on the verified human path and carries the positive lipid-homeostasis IBD. Its descendant seeds include human Q8N6M3, which is legitimate experimental grounding rather than circular transfer. The human acyl-CoA reaction and cellular complementation independently agree with the inherited lipid-homeostasis role.

Summary: FITM2 catalytic activity supports cellular lipid and ER homeostasis.

Reason: PMID:32915949 links human FIT2 loss to altered lipid pools, ER morphology and lipid-droplet formation, with wild-type versus catalytic-mutant complementation. This supports a direct lipid-homeostatic role. Long-term changes in triglyceride synthesis and associated gene expression were not fully rescued, so they are not treated as the defining immediate catalytic output.

Summary: Lipid-droplet formation is a conserved FITM2 role.

Reason: The exact human leaf descends from PTN000565881 and its positive lipid-droplet-formation IBD. The full family topology and original node assertion were checked; no loss along the target path was identified. Human and mouse functional experiments support participation through ER lipid chemistry and partitioning. Other off-path family assertions are not imported.

Summary: The mouse ortholog supports electronic transfer of lipid-droplet formation.

Reason: P59266 is mouse Fitm2, whose gain- and loss-of-function experiments alter droplet formation. Human catalytic-rescue experiments independently support the conserved role. The process is not interpreted as an absolute inability to make any droplets in every tissue after every perturbation.

Summary: Human FITM2 catalytic function is required for normal lipid-droplet formation.

Reason: PMID:32915949 shows that wild-type human FIT2, but not catalytic-histidine mutants, restores the droplet-formation defect in FIT2-deficient cells. The separately established acyl-CoA reaction supplies a molecular mechanism for active participation. Necessity alone is not used to invent an unrelated process or a physical membrane-tether function.

Summary: Manual transfer from mouse Fitm2 supports lipid-droplet formation.

Reason: The verified P59266 donor and its experimental formation/partitioning results support the existing ortholog transfer. Concordant human catalytic complementation strengthens the biological interpretation without changing the incoming ISS source or pretending the mouse assays directly tested human protein.

Summary: The specific ER-to-lipid-droplet tether activity remains unresolved.

Reason: GO:0170007 requires binding that brings an ER membrane and a lipid droplet together to establish a membrane contact site. R-HSA-8857686 describes triglyceride binding and partitioning into droplets; the full PMID:18160536 text and accessible PMID:22106267 binding Results support those activities but do not settle physical membrane bridging. Complete original binding-paper Methods were not recovered, so absence of a visible tether assay is not treated as disproof. Preserve uncertainty instead of removing the TAS row or converting lipid binding into an unverified tether mechanism.