EPM2A

Summary: Laforin regulates and repairs glycogen; the traced donor experiment does not establish glycogen-synthase catalysis.

Reason: The current mouse Epm2a donor Q9WUA5 carries IEP GO:0004373 from PMID:18852261. Its actual Methods (glycogen-synthase assay) and Results/Figure 6 measure glycogen-synthase protein and activity in Epm2a-null tissue fractions; they do not show laforin transferring glucose from UDP-glucose. The same original directly measures laforin-dependent phosphate removal from glycogen (Figure 4). Independent human laforin experiments establish carbohydrate phosphatase activity (PMID:25538239). The source has conflated an enzyme whose abundance/activity was measured with the phosphatase perturbed in the experiment. This is a source-side catalytic-role error, not an orthology failure or a conclusion from the paper title.

Summary: Human laforin has directly measured secondary protein-phosphatase activity in vitro.

Reason: The mouse ortholog and its current phosphoprotein-phosphatase annotation are genuine. Independently, PMID:11001928 Results/Figure 2B and Methods pp.2258–2259 used purified human 331-aa GST-laforin and radiolabeled myelin basic protein, measuring phosphate release after an eight-hour incubation. This is actual protein-substrate evidence, distinct from the separate OMFP assay. PMID:16901901 later reports several negative candidate-protein assays and strong carbohydrate-substrate activity without excluding every protein substrate. Retain the demonstrated in-vitro capacity as secondary, without making a particular physiological protein substrate or all EPM2A products equivalent.

Summary: Human laforin has measured protein serine/threonine phosphatase capacity in vitro.

Reason: The HIRIP5 paper PMID:12915448 reports protein dephosphorylation in its abstract, but its complete operative original remains inaccessible. The actual original PMID:11001928 Methods and Figure 2B use human fetal-brain 331-residue laforin expressed as a bacterial GST fusion and measure phosphate release from myelin basic protein phosphorylated on serine/threonine by PKA or on tyrosine by c-Abl. This is a protein-substrate experiment, distinct from its OMFP assay. The prolonged in-vitro assay supports catalytic capacity, but does not identify a physiological protein substrate or make this the principal glycogen-repair mechanism. Retain the general activity as secondary on this independent positive evidence; the source-specific HIRIP5 substrate experiment is not claimed to have been inspected.

Summary: Human laforin can dephosphorylate a Ser/Thr-phosphorylated protein in vitro.

Reason: ARBA00027653 matches the target domain/taxon conditions; RHEA:20629 and RHEA:47004 describe phosphoserine and phosphothreonine hydrolysis. These mappings specify chemistry but do not alone establish the physiological substrate. PMID:11001928 Methods explicitly generated radiolabeled myelin basic protein with cAMP-dependent protein kinase; Figure 2B measured phosphate release by purified human GST-laforin. This positive assay supports a secondary protein Ser/Thr phosphatase capacity despite subsequent negative GSK3/protein-substrate tests in PMID:16901901 and the stronger physiological glucan-phosphatase evidence. Do not infer that synthetic pNPP/OMFP activity proves protein specificity.

Summary: Protein tyrosine phosphatase activity is supported as an in-vitro secondary activity.

Reason: The original membrane-localization paper PMID:11220751 remains abstract-only in this assessment. The actual original PMID:11001928 Methods and Figure 2B use human fetal-brain 331-residue laforin expressed as a bacterial GST fusion and measure phosphate release from myelin basic protein phosphorylated on serine/threonine by PKA or on tyrosine by c-Abl. This is a protein-substrate experiment, distinct from its OMFP assay. The prolonged in-vitro assay supports catalytic capacity, but does not identify a physiological protein substrate or make this the principal glycogen-repair mechanism. The independent protein assay resolves the general catalytic claim without inferring the missing paper's constructs or extending the result to every splice product.

Summary: Purified human laforin releases phosphate from phosphotyrosine-containing protein substrate.

Reason: The actual original PMID:11001928 Methods and Figure 2B use human fetal-brain 331-residue laforin expressed as a bacterial GST fusion and measure phosphate release from myelin basic protein phosphorylated on serine/threonine by PKA or on tyrosine by c-Abl. This is a protein-substrate experiment, distinct from its OMFP assay. The prolonged in-vitro assay supports catalytic capacity, but does not identify a physiological protein substrate or make this the principal glycogen-repair mechanism. Retain the measured activity outside the primary carbohydrate-phosphatase core; the experiment does not establish myelin basic protein as an endogenous laforin substrate.

Summary: General protein tyrosine phosphatase capacity is supported independently of the inaccessible HIRIP5 assay.

Reason: The complete PMID:12915448 original remains unavailable. The actual original PMID:11001928 Methods and Figure 2B use human fetal-brain 331-residue laforin expressed as a bacterial GST fusion and measure phosphate release from myelin basic protein phosphorylated on serine/threonine by PKA or on tyrosine by c-Abl. This is a protein-substrate experiment, distinct from its OMFP assay. The prolonged in-vitro assay supports catalytic capacity, but does not identify a physiological protein substrate or make this the principal glycogen-repair mechanism. This supports the general molecular function while leaving the particular HIRIP5 dephosphorylation experiment unverified.

Summary: Protein-tyrosine phosphatase capacity is supported in vitro but is not the principal established substrate role.

Reason: The target matches ARBA00027656 and the laforin InterPro family; RHEA:10684 describes protein-phosphotyrosine hydrolysis. The donor PMID:15541350 abstract reports mouse recombinant laforin acting on pNPP, which alone is not protein-substrate proof, and its full original was not recovered. Independent actual human PMID:11001928 Methods used c-Abl-phosphorylated myelin basic protein, with phosphate release in Figure 2B. Retain this secondary capacity while recognizing the later direct glucan evidence and bounded negative protein tests in PMID:16901901; neither the family name nor synthetic-substrate hydrolysis is the sole justification.

Summary: Protein tyrosine phosphatase capacity is retained using an independent protein-substrate assay.

Reason: PMID:11739371 Experimental Procedures and Figure 2A assay pNPP, a synthetic phosphate ester, rather than a phosphotyrosine-containing protein. That assay alone does not establish the protein-substrate specificity. The actual original PMID:11001928 Methods and Figure 2B use human fetal-brain 331-residue laforin expressed as a bacterial GST fusion and measure phosphate release from myelin basic protein phosphorylated on serine/threonine by PKA or on tyrosine by c-Abl. This is a protein-substrate experiment, distinct from its OMFP assay. The prolonged in-vitro assay supports catalytic capacity, but does not identify a physiological protein substrate or make this the principal glycogen-repair mechanism. The positive independent experiment supports the existing broad activity as secondary, without relabeling the pNPP experiment.

Summary: The reported HIRIP5/NFU1 interaction requires inspection of its operative original.

Reason: PMID:12915448 reports yeast-two-hybrid and biochemical association in the abstract, but publisher, author-page and indexed-PDF access did not recover the complete original for adjudicating the HIRIP5 product/domain context. Do not reject the interaction or infer a specific iron-sulfur function from the partner name. Preserve the source assertion pending access to the direct assays.

Summary: Laforin recruits the PTG substrate for malin-dependent ubiquitination.

Reason: The original PMID:14532330 remains abstract-only after publisher and restricted institutional-file access attempts; the independent original PMID:18070875 directly resolves the mechanistic role. PMID:18070875 Methods, Results and Figure 3B show endogenous laforin–PTG co-immunoprecipitation from human HEK293 cells. Figure 3C then reconstitutes PTG ubiquitination with malin, E1/E2 and recombinant laforin: removal of laforin, or use of interaction-impaired malin E280K, prevents PTG modification. Laforin therefore contributes the adaptor linking the ligase to its substrate; malin performs ubiquitin transfer. CHO-IR overexpression effects and catalytic C266S controls are separate from the endogenous human association assay. GO:1990756 specifically describes binding that brings a ubiquitin-like ligase together with its substrate; ubiquitin ligases such as malin fall within this definition. No catalytic activity of malin is assigned to laforin.

Proposed replacements: ubiquitin-like ligase-substrate adaptor activity

Summary: Laforin binds the ubiquitin protein ligase malin.

Reason: PMID:15930137 Methods and Results/Figure 3 demonstrate laforin association with the E3 ligase malin by human-brain-library two-hybrid screening, recombinant laforin plus in-vitro-translated malin binding, and tagged co-immunoprecipitation in HEK293T cells. Its laforin constructs trace to the human 331-aa preparation in PMID:11739371. Binding the ligase does not make laforin an E3 enzyme. Refine the generic source term to the evidenced partner class.

Proposed replacements: ubiquitin protein ligase binding

Summary: Laforin recruits the PTG substrate for malin-dependent ubiquitination.

Reason: PMID:18070875 Methods, Results and Figure 3B show endogenous laforin–PTG co-immunoprecipitation from human HEK293 cells. Figure 3C then reconstitutes PTG ubiquitination with malin, E1/E2 and recombinant laforin: removal of laforin, or use of interaction-impaired malin E280K, prevents PTG modification. Laforin therefore contributes the adaptor linking the ligase to its substrate; malin performs ubiquitin transfer. CHO-IR overexpression effects and catalytic C266S controls are separate from the endogenous human association assay. This directly measured coordinating activity supports the specific GO:1990756 ubiquitin-like ligase-substrate adaptor activity; the paper's broader disease model is not required for this molecular conclusion.

Proposed replacements: ubiquitin-like ligase-substrate adaptor activity

Summary: The malin interaction is a ubiquitin protein ligase interaction.

Reason: PMID:21505799 Methods and Figure 6 use a laforin–malin two-hybrid system with expression controls; wild-type malin binds and disease-associated malin substitutions impair the interaction. The main Methods specify human malin but refer to prior work for laforin construct details. PMID:15930137 Methods and Results/Figure 3 demonstrate laforin association with the E3 ligase malin by human-brain-library two-hybrid screening, recombinant laforin plus in-vitro-translated malin binding, and tagged co-immunoprecipitation in HEK293T cells. Its laforin constructs trace to the human 331-aa preparation in PMID:11739371. Binding the ligase does not make laforin an E3 enzyme. The independent human preparation resolves the partner-class refinement without treating COS-7 host species as construct identity.

Proposed replacements: ubiquitin protein ligase binding

Summary: Laforin associates with the E3 ubiquitin ligase malin in a PRDM8-containing complex.

Reason: PMID:22961547 Methods and Results/Figures 6–7 describe tagged co-expression and three-way co-immunoprecipitation in COS7/HeLa cells: precipitating any one of laforin, malin and PRDM8 recovers the other two. The malin component supports the more informative partner-class term GO:0031625 ubiquitin protein ligase binding. The PRDM8 association and nuclear redistribution remain part of the experimental context; neither a purified binary contact nor transcriptional activity of laforin is inferred. Endogenous PRDM8 also co-precipitates the previously described human C266S myc-laforin transgene from mouse brain/muscle, which is distinct from an endogenous wild-type human assay.

Proposed replacements: ubiquitin protein ligase binding

Summary: Human laforin self-association is specifically homodimerization.

Reason: PMID:25538239 uses purified human 331-aa laforin and analytical ultracentrifugation/small-angle X-ray scattering to support a stable dimer in the tested solution conditions. The inactive isolated DSP crystal and proposed full-length domain arrangement are separate from the solution evidence. Refine generic binding to the already represented homodimerization activity; no oligomer state is imposed on all alternative products.

Proposed replacements: protein homodimerization activity

Summary: Solution measurements establish homodimerization of the tested human laforin preparation.

Reason: PMID:25544560 Results/Figure 5 and solution experiments support a dimer and altered oligomerization of F321S. The main cloning section specifies human residues 1–328 with N-terminal His6, while DXMS separately describes C-terminal His6; the unread supplement may clarify these preparations. Refine self-binding to homodimerization without inventing a new natural product or asserting one construct across all assays.

Proposed replacements: protein homodimerization activity

Summary: The imported BioPlex malin association has independently verified ligase-binding support.

Reason: PMID:33961781 main AP-MS Methods were inspected: human ORFeome baits, C-terminal FLAG-HA tagging, separate293T/HCT116 analyses and gene-level network consolidation do not establish exact natural-product identity or direct binding by themselves. The exact EPM2A–malin supplementary network record was not independently re-extracted. PMID:15930137 Methods and Results/Figure 3 demonstrate laforin association with the E3 ligase malin by human-brain-library two-hybrid screening, recombinant laforin plus in-vitro-translated malin binding, and tagged co-immunoprecipitation in HEK293T cells. Its laforin constructs trace to the human 331-aa preparation in PMID:11739371. Binding the ligase does not make laforin an E3 enzyme. Independent target-specific evidence supports the ubiquitin protein ligase binding refinement while preserving the high-throughput source context.

Proposed replacements: ubiquitin protein ligase binding

Summary: Nuclear localization is a supported context-dependent pool rather than the principal glucan-repair compartment.

Reason: The source original remains inaccessible, so its individual splice-product experiments are not reconstructed from the abstract. Independent PMID:17908927 Methods and Figure 3C show tagged laforin in human HepG2 cytoplasm at baseline and partial nuclear redistribution after glycogenolytic treatment; its construct lineage traces through PMID:15930137. PMID:22961547 Figures 6–7 also document a minor nuclear pool and PRDM8-dependent nuclear accumulation. These positive experiments support gene-level nuclear localization as non-core, without assigning identical targeting to all eight imported products or claiming nuclear residency is constitutive.

Summary: Nuclear localization is a supported context-dependent pool rather than the principal glucan-repair compartment.

Reason: The source original remains inaccessible, so its individual splice-product experiments are not reconstructed from the abstract. Independent PMID:17908927 Methods and Figure 3C show tagged laforin in human HepG2 cytoplasm at baseline and partial nuclear redistribution after glycogenolytic treatment; its construct lineage traces through PMID:15930137. PMID:22961547 Figures 6–7 also document a minor nuclear pool and PRDM8-dependent nuclear accumulation. These positive experiments support gene-level nuclear localization as non-core, without assigning identical targeting to all eight imported products or claiming nuclear residency is constitutive.

Summary: Nuclear localization is a supported context-dependent pool rather than the principal glucan-repair compartment.

Reason: The source original remains inaccessible, so its individual splice-product experiments are not reconstructed from the abstract. Independent PMID:17908927 Methods and Figure 3C show tagged laforin in human HepG2 cytoplasm at baseline and partial nuclear redistribution after glycogenolytic treatment; its construct lineage traces through PMID:15930137. PMID:22961547 Figures 6–7 also document a minor nuclear pool and PRDM8-dependent nuclear accumulation. These positive experiments support gene-level nuclear localization as non-core, without assigning identical targeting to all eight imported products or claiming nuclear residency is constitutive.

Summary: Nuclear localization is a supported context-dependent pool rather than the principal glucan-repair compartment.

Reason: The source original remains inaccessible, so its individual splice-product experiments are not reconstructed from the abstract. Independent PMID:17908927 Methods and Figure 3C show tagged laforin in human HepG2 cytoplasm at baseline and partial nuclear redistribution after glycogenolytic treatment; its construct lineage traces through PMID:15930137. PMID:22961547 Figures 6–7 also document a minor nuclear pool and PRDM8-dependent nuclear accumulation. These positive experiments support gene-level nuclear localization as non-core, without assigning identical targeting to all eight imported products or claiming nuclear residency is constitutive.

Summary: Nuclear localization is supported, with nucleoplasm providing the more specific observed compartment.

Reason: The ARBA rule matches the target and the UniProt nuclear statements have product/condition-specific scope. Independently, current Human Protein Atlas EPM2A subcellular records report supported nucleoplasm localization in SH-SY5Y and U2OS cells with HPA053643/HPA055468; SK-MEL-30 has no staining with HPA055468. PMID:17908927 Figure 3C shows partial nuclear accumulation of tagged laforin in HepG2 cells after glycogenolytic treatment, and PMID:22961547 Figure 6 shows PRDM8-dependent redistribution. Refine the broad nucleus term to the observed nucleoplasm without treating it as constitutive localization of every natural product. HPA record text/XML, not image pixels, was inspected. The combined electronic nucleus mapping is refined with independent target HPA nucleoplasm evidence. The separate historical splice-product and conditional-relocalization sources retain nucleus at their measured scope; HPA does not identify those individual products or resolve every nuclear subcompartment.

Proposed replacements: nucleoplasm

Summary: The HPA target-specific assay supports nucleoplasm localization in the tested cell lines.

Reason: The official EPM2A HPA record identifies nucleoplasm as a supported main location, with HPA053643/HPA055468 staining in SH-SY5Y and U2OS; HPA055468 shows no staining in SK-MEL-30. These are target assay records rather than inference from a generic nuclear protein class. Retain the location as a contextual observation alongside the cytoplasmic glucan role. This assessment reads the actual HPA XML/page annotation, not the underlying image pixels, and does not resolve individual EPM2A products. Independent tagged laforin experiments show condition-dependent nuclear entry (PMID:17908927 Figure 3C) and PRDM8-associated redistribution (PMID:22961547 Figure 6).

Summary: Human 331-aa laforin has directly observed cytoplasmic localization.

Reason: PMID:11001928 uses human 331-aa constructs, including untagged laforin, in HeLa cells and observes punctate cytoplasmic/perinuclear localization with polysome-enriched fractionation. PMID:11739371 independently localizes human 331-aa laforin with cytoplasmic glycogen complexes; PMID:17908927 shows the basal cytoplasmic pool in HepG2 cells. These are positive cellular observations, not an inference from disease necessity. The result supports the broad gene-level location without assigning every alternative product identical targeting. The original Methods, Results and localization captions were read; no integrality in ER membrane is inferred.

Summary: The cytoplasmic location is supported by independent human laforin experiments.

Reason: The complete operative PMID:11220751 original was not available at this checkpoint; its specific construct/variant findings are not claimed as inspected. PMID:11001928 uses human 331-aa constructs, including untagged laforin, in HeLa cells and observes punctate cytoplasmic/perinuclear localization with polysome-enriched fractionation. PMID:11739371 independently localizes human 331-aa laforin with cytoplasmic glycogen complexes; PMID:17908927 shows the basal cytoplasmic pool in HepG2 cells. These are positive cellular observations, not an inference from disease necessity. The result supports the broad gene-level location without assigning every alternative product identical targeting. Independent positive evidence resolves the general compartment while preserving the original citation and its access limits.

Summary: The cytoplasmic location is supported by independent human laforin experiments.

Reason: The complete operative PMID:11883934 original was not available at this checkpoint; its specific construct/variant findings are not claimed as inspected. PMID:11001928 uses human 331-aa constructs, including untagged laforin, in HeLa cells and observes punctate cytoplasmic/perinuclear localization with polysome-enriched fractionation. PMID:11739371 independently localizes human 331-aa laforin with cytoplasmic glycogen complexes; PMID:17908927 shows the basal cytoplasmic pool in HepG2 cells. These are positive cellular observations, not an inference from disease necessity. The result supports the broad gene-level location without assigning every alternative product identical targeting. Independent positive evidence resolves the general compartment while preserving the original citation and its access limits.

Summary: The cytoplasmic location is supported by independent human laforin experiments.

Reason: The complete operative PMID:14532330 original was not available at this checkpoint; its specific construct/variant findings are not claimed as inspected. PMID:11001928 uses human 331-aa constructs, including untagged laforin, in HeLa cells and observes punctate cytoplasmic/perinuclear localization with polysome-enriched fractionation. PMID:11739371 independently localizes human 331-aa laforin with cytoplasmic glycogen complexes; PMID:17908927 shows the basal cytoplasmic pool in HepG2 cells. These are positive cellular observations, not an inference from disease necessity. The result supports the broad gene-level location without assigning every alternative product identical targeting. Independent positive evidence resolves the general compartment while preserving the original citation and its access limits.

Summary: Laforin is localized within neuronal and other cellular cytoplasm.

Reason: PMID:15102711 Results/Figures 2–7 place human C266S myc-laforin expressed in transgenic mice in somatic/dendritic cytoplasm and on polyglucosan bodies. The engineered inactive protein and roughly100-fold overexpression limit extrapolation to endogenous wild-type distribution. PMID:11001928 uses human 331-aa constructs, including untagged laforin, in HeLa cells and observes punctate cytoplasmic/perinuclear localization with polysome-enriched fractionation. PMID:11739371 independently localizes human 331-aa laforin with cytoplasmic glycogen complexes; PMID:17908927 shows the basal cytoplasmic pool in HepG2 cells. These are positive cellular observations, not an inference from disease necessity. The result supports the broad gene-level location without assigning every alternative product identical targeting. Independent human wild-type experiments support retaining the broad cytoplasmic annotation.

Summary: The cytoplasmic location is supported by independent human laforin experiments.

Reason: The complete operative PMID:18311786 original was not available at this checkpoint; its specific construct/variant findings are not claimed as inspected. PMID:11001928 uses human 331-aa constructs, including untagged laforin, in HeLa cells and observes punctate cytoplasmic/perinuclear localization with polysome-enriched fractionation. PMID:11739371 independently localizes human 331-aa laforin with cytoplasmic glycogen complexes; PMID:17908927 shows the basal cytoplasmic pool in HepG2 cells. These are positive cellular observations, not an inference from disease necessity. The result supports the broad gene-level location without assigning every alternative product identical targeting. Independent positive evidence resolves the general compartment while preserving the original citation and its access limits.

Summary: The cytoplasmic location is supported by independent human laforin experiments.

Reason: The complete operative PMID:18617530 original was not available at this checkpoint; its specific construct/variant findings are not claimed as inspected. PMID:11001928 uses human 331-aa constructs, including untagged laforin, in HeLa cells and observes punctate cytoplasmic/perinuclear localization with polysome-enriched fractionation. PMID:11739371 independently localizes human 331-aa laforin with cytoplasmic glycogen complexes; PMID:17908927 shows the basal cytoplasmic pool in HepG2 cells. These are positive cellular observations, not an inference from disease necessity. The result supports the broad gene-level location without assigning every alternative product identical targeting. Independent positive evidence resolves the general compartment while preserving the original citation and its access limits.

Summary: The cytoplasmic location is supported by independent human laforin experiments.

Reason: The complete operative PMID:22036712 original was not available at this checkpoint; its specific construct/variant findings are not claimed as inspected. PMID:11001928 uses human 331-aa constructs, including untagged laforin, in HeLa cells and observes punctate cytoplasmic/perinuclear localization with polysome-enriched fractionation. PMID:11739371 independently localizes human 331-aa laforin with cytoplasmic glycogen complexes; PMID:17908927 shows the basal cytoplasmic pool in HepG2 cells. These are positive cellular observations, not an inference from disease necessity. The result supports the broad gene-level location without assigning every alternative product identical targeting. Independent positive evidence resolves the general compartment while preserving the original citation and its access limits.

Summary: The human laforin carbohydrate-binding module targets cytoplasmic glycogen complexes.

Reason: PMID:11739371 Experimental Procedures and Figures 2/4 use human-muscle331 laforin, cellular fractionation with amylase release and COS1 localization alongside glycogen synthase. CBD mutations impair this distribution. The construct identity comes from the human cDNA, not the COS1 host. Retain cytoplasm as the principal broad cellular location.

Summary: Laforin is predominantly cytoplasmic under the tested basal culture conditions.

Reason: PMID:17908927 Methods and Results/Figure 3C directly show basal cytoplasmic FLAG-laforin in human HepG2 cells and redistribution under glycogenolytic treatment. The laforin constructs cite PMID:15930137; the separately described AGL cDNA is from mouse liver and does not define laforin species. Nuclear movement under stimulation does not contradict a basal cytoplasmic pool.

Summary: Laforin has an established cytoplasmic pool, including glycogen-associated material.

Reason: Human 331-aa laforin is observed in cytoplasmic/perinuclear and polysome-associated preparations in PMID:11001928, Results/Figures 3–8. PMID:11739371 Figure 2B independently shows glycogen-associated cellular laforin released by amylase, and PMID:17908927 Figure 3C shows a basal cytoplasmic pool in human HepG2 cells. These observations support the broad cytoplasm assertion across the experimental and electronic source contexts. Cytosolic localization is also supported and retained in its separate annotations, but does not require replacing every broader cytoplasmic observation. No exclusive distribution or uniform localization of all eight natural products is asserted.

Summary: The inspected source directly supports cytoplasmic polysome association.

Reason: The actual PMID:11001928 Results/Figures 5–8 and Discussion establish association with polysomes in a microsomal fraction: extensive QM-marker overlap, sucrose-gradient co-sedimentation and immunogold labeling over ribosomal units. The Results/caption explicitly report no signal over the associated membranous structures, while direct ER interaction remains a possibility in the Discussion. This is positive cytoplasmic localization, but membrane attachment is not established by the inspected experiments. Current GO membrane-side definitions require an embedded, attached or peripheral membrane association; the historical polysome term is now obsolete. Use the current cytoplasm term rather than claiming the paper establishes an ER lipid-bilayer location. This source-scope refinement is not a universal denial of membrane association in other conditions. This source specifically resolves laforin on polysomes and leaves direct ER-membrane association speculative, allowing a positive cytoplasm refinement. Other ER-membrane assertions have distinct inaccessible splice-product evidence and remain unresolved rather than being rejected globally.

Proposed replacements: cytoplasm

Summary: The membrane-specific localization remains unresolved for this source context.

Reason: The full operative original PMID:11883934 has not been recovered. Independent studies establish cytoplasmic and near-ER/polysome pools, but proximity and microsomal recovery do not by themselves prove attachment to the ER membrane. The experimental assertion is preserved for expert/source inspection; neither a universal membrane-negative claim nor a more specific leaflet location is inferred.

Summary: The membrane-specific localization remains unresolved for this source context.

Reason: The full operative original PMID:14722920 has not been recovered. Independent studies establish cytoplasmic and near-ER/polysome pools, but proximity and microsomal recovery do not by themselves prove attachment to the ER membrane. The experimental assertion is preserved for expert/source inspection; neither a universal membrane-negative claim nor a more specific leaflet location is inferred.

Summary: The membrane-specific localization remains unresolved for this source context.

Reason: The full operative original PMID:18311786 has not been recovered. Independent studies establish cytoplasmic and near-ER/polysome pools, but proximity and microsomal recovery do not by themselves prove attachment to the ER membrane. The experimental assertion is preserved for expert/source inspection; neither a universal membrane-negative claim nor a more specific leaflet location is inferred.

Summary: ER association is reported, but the membrane attachment asserted by this mapping remains unresolved.

Reason: The source is the actual UniProt SL-0097 statement, with isoform-specific cited localization evidence. PMID:11001928 Figure 8 and its operative text localize tagged human laforin on ribosome units rather than the associated membrane. PMID:15102711 Results/Figures 2 and 6 place human C266S transgene signal near ER in mouse tissues, while explicitly leaving associated polyglucosan material unresolved. Proximity and polysome association do not by themselves establish an embedded or attached ER-membrane component. Other underlying isoform/localization references were not fully accessible; therefore retain uncertainty rather than declaring a universal absence of membrane association or overwriting the source.

Summary: A cytosolic laforin pool is supported by independent fractionation experiments.

Reason: The HIRIP5 original PMID:12915448 is inaccessible in full. PMID:11739371 Results/Figure 2B independently fractionates human 331-aa laforin between glycogen–microsomal complexes and cytosol and releases complex-associated laforin into the cytosolic fraction after amylase treatment. This supports the cytosolic pool without treating HIRIP5 localization as evidence for laforin or excluding the polysome/glycogen-associated pools.

Summary: Cytosolic laforin is supported by target experiments and exact database participant provenance.

Reason: ARBA00029243 matches the target; current mouse Q9WUA5 has cytosol evidence including PMID:22669944. That original fractionates mouse brain laforin into free and glycogen-associated pools (Results/Figure 4), and its recombinant mouse constructs in HEK293 cells remain mouse proteins. Independent human PMID:11739371 fractionation and the exact Reactome O95278 cytosol participant corroborate the location. No inference of identical localization or activity across all natural products is made.

Summary: The exact human laforin participant is explicitly located in cytosol.

Reason: Reactome R-HSA-3780995: input 3780991 contains EPM2A dimer 6805697; output 3781004 is ubiquitinated EPM2A. Dimer 6805697 contains EPM2A 3777103, whose referenceEntity 146927 identifies human UniProt O95278 and whose compartment is cytosol (GO:0005829). This GYG2 event is an inference from the muscle/GYG1 counterpart; laforin is a substrate/complex component here, not the ubiquitin ligase. Independent human fractionation in PMID:11739371 corroborates a cytosolic/glycogen-associated pool; the conclusion is not inferred solely from a pathway-level compartment.

Summary: The exact human laforin participant is explicitly located in cytosol.

Reason: Reactome R-HSA-3781001: input dimer 6805697 and output glycogen complex 3780999 contain EPM2A. Dimer 6805697 contains EPM2A 3777103, whose referenceEntity 146927 identifies human UniProt O95278 and whose compartment is cytosol (GO:0005829). This is glycogen-associated binding/assembly context, not evidence that laforin makes glycosidic bonds. Independent human fractionation in PMID:11739371 corroborates a cytosolic/glycogen-associated pool; the conclusion is not inferred solely from a pathway-level compartment.

Summary: The exact human laforin participant is explicitly located in cytosol.

Reason: Reactome R-HSA-3781009: input complex 3781000 contains dimer 6805697; output 3781004 is ubiquitinated EPM2A. Dimer 6805697 contains EPM2A 3777103, whose referenceEntity 146927 identifies human UniProt O95278 and whose compartment is cytosol (GO:0005829). This ubiquitination event distinguishes laforin substrate from the NHLRC1 catalyst. Independent human fractionation in PMID:11739371 corroborates a cytosolic/glycogen-associated pool; the conclusion is not inferred solely from a pathway-level compartment.

Summary: The exact human laforin participant is explicitly located in cytosol.

Reason: Reactome R-HSA-3781011: input complex 3781017 contains dimer 6805697; catalyst 3781002 names activeUnit 3777103. Dimer 6805697 contains EPM2A 3777103, whose referenceEntity 146927 identifies human UniProt O95278 and whose compartment is cytosol (GO:0005829). The GYG2 version is inferred from the GYG1 reaction; the compartment is attached to the actual human participant as well as the event. Independent human fractionation in PMID:11739371 corroborates a cytosolic/glycogen-associated pool; the conclusion is not inferred solely from a pathway-level compartment.

Summary: The exact human laforin participant is explicitly located in cytosol.

Reason: Reactome R-HSA-3781018: input complex 3780999 contains dimer 6805697; catalyst 3781013 names activeUnit 3777103. Dimer 6805697 contains EPM2A 3777103, whose referenceEntity 146927 identifies human UniProt O95278 and whose compartment is cytosol (GO:0005829). The curated reaction explicitly represents laforin-mediated phosphate removal from glycogen. Independent human fractionation in PMID:11739371 corroborates a cytosolic/glycogen-associated pool; the conclusion is not inferred solely from a pathway-level compartment.

Summary: The exact human laforin participant is explicitly located in cytosol.

Reason: Reactome R-HSA-3781021: input dimer 6805697 and output GYG2 glycogen complex 3781017 contain EPM2A. Dimer 6805697 contains EPM2A 3777103, whose referenceEntity 146927 identifies human UniProt O95278 and whose compartment is cytosol (GO:0005829). The GYG2 assembly is inferred from the GYG1 counterpart; the target identity and location are explicit. Independent human fractionation in PMID:11739371 corroborates a cytosolic/glycogen-associated pool; the conclusion is not inferred solely from a pathway-level compartment.

Summary: The exact human laforin participant is explicitly located in cytosol.

Reason: Reactome R-HSA-3791349: normal input complex 3780999 contains dimer 6805697 and EPM2A 3777103. Dimer 6805697 contains EPM2A 3777103, whose referenceEntity 146927 identifies human UniProt O95278 and whose compartment is cytosol (GO:0005829). This is a FailedReaction involving a mutant catalyst. Failure of the represented chemistry does not negate the location of the normal input participant; it does not establish mutant catalytic activity. Independent human fractionation in PMID:11739371 corroborates a cytosolic/glycogen-associated pool; the conclusion is not inferred solely from a pathway-level compartment.

Summary: The exact human laforin participant is explicitly located in cytosol.

Reason: Reactome R-HSA-3797226: input complex 3781000 contains dimer 6805697 and EPM2A 3777103. Dimer 6805697 contains EPM2A 3777103, whose referenceEntity 146927 identifies human UniProt O95278 and whose compartment is cytosol (GO:0005829). This is a FailedReaction with a mutant NHLRC1 catalyst. It supports the target participant compartment, not laforin ubiquitin-ligase activity or successful mutant catalysis. Independent human fractionation in PMID:11739371 corroborates a cytosolic/glycogen-associated pool; the conclusion is not inferred solely from a pathway-level compartment.

Summary: The plasma-membrane assertion requires the original localization experiments.

Reason: PMID:11220751 remains unavailable beyond abstract/metadata after bounded source retrieval. Other inspected laforin originals show cytoplasm, polysomes, glycogen complexes and conditional nuclear distribution; these do not independently establish plasma-membrane attachment. Do not reject an experimental annotation on incomplete evidence or infer a membrane location from downstream effects on glucose transporters.

Summary: The plasma-membrane assertion requires the original localization experiments.

Reason: PMID:11883934 remains unavailable beyond abstract/metadata after bounded source retrieval. Other inspected laforin originals show cytoplasm, polysomes, glycogen complexes and conditional nuclear distribution; these do not independently establish plasma-membrane attachment. Do not reject an experimental annotation on incomplete evidence or infer a membrane location from downstream effects on glucose transporters.

Summary: The isoform-specific plasma-membrane source statement could not be independently adjudicated.

Reason: The official UniProt entry maps SL-0039 to isoform 1 with PMID:11220751 and isoform 2 with PMID:11883934. Those primary localization originals were not recovered at the needed operative scope. The positive cytoplasmic/nuclear and near-ER observations in other papers neither prove nor refute plasma-membrane attachment. Retain uncertainty rather than extending a statement to all products or dismissing the source from its abstract alone.

Summary: Laforin directly participates in glycogen metabolism through phosphate removal.

Reason: PMID:25538239 directly assays amylopectin dephosphorylation and characterizes human 331-aa glucan-binding/catalytic architecture. Independent PMID:25544560 Figure 1C/Methods supply the glycogen-specific bridge: WT laforin releases inorganic phosphate from rabbit-muscle glycogen, unlike C266S and VHR controls. Thus the protein performs chemistry on glycogen, rather than merely being necessary for its normal accumulation. Product/engineered-construct limits remain explicit.

Summary: The broad glycogen metabolic role is supported by direct later biochemistry.

Reason: PMID:9931343 is an early gene-identification/disease source, not independently a direct glycogen enzyme assay. The current role is positively resolved by PMID:25544560 Figure 1C/Methods, which measure phosphate release from rabbit-muscle glycogen by WT laforin relative to catalytic-dead and other-phosphatase controls. Retain the original NAS source and cite the direct experimental support; glycogen dephosphorylation is not glycogen synthesis.

Summary: Glycogen maintenance and regulation are established; glycogen metabolism is the better supported process scope.

Reason: ARBA00043606 matches laforin, and mouse donor glycogen-biosynthesis annotations are present, including PMID:17118331, PMID:18040046, PMID:28536304 and PMID:31042462. Actual originals distinguish direct phosphate removal from glycogen, altered glycogen-synthase readouts in mouse extracts, and phosphatase-independent laforin/malin-dependent control of abnormal chain accumulation. Human laforin also contributes a adaptor for malin-dependent PTG ubiquitination (PMID:18070875 Figure 3C). These are positive roles in glycogen metabolism, not evidence that laforin adds glucose or that every change in glycogen abundance establishes participation in its formation. Use the broader metabolic process to retain both repair and regulatory contributions without forcing a single biosynthetic mechanism.

Proposed replacements: glycogen metabolic process

Summary: The Reactome pathway contains laforin-dependent glycogen maintenance steps, supporting glycogen metabolism.

Reason: Reactome R-HSA-3322077 groups glycogen-chain synthesis with associated binding, laforin-catalyzed dephosphorylation and malin-dependent turnover events. Exact laforin participants/catalyst active units are traceable through R-HSA-3781011 and R-HSA-3781018 to O95278. Laforin directly removes phosphate and contributes to regulation of glycogen-associated proteins; GYS and GBE execute glucose addition/branch formation. Pathway containment does not require assigning laforin their catalytic role. Glycogen metabolic process preserves the positive biochemical and regulatory participation established by human PMID:25538239 and PMID:18070875 without over-specifying formation.

Proposed replacements: glycogen metabolic process

Summary: Protein-substrate chemistry is measured in vitro, but a physiological protein-dephosphorylation role remains unestablished.

Reason: PMID:11001928 Methods pp. 2258–2259 and Figure 2B directly measure phosphate release from PKA- or c-Abl-phosphorylated myelin basic protein using bacterial GST-tagged human 331-residue laforin, including GST and arsenate controls, after an eight-hour incubation. This establishes protein-substrate catalytic capacity. It does not identify an endogenous protein-dephosphorylation pathway. PMID:16901901 Methods and Results/Figures 1–4 instead find no dephosphorylation of the tested GSK3β, GS or PTG substrates and establish selective amylopectin dephosphorylation. These negatives are substrate- and condition-specific, not proof that laforin can never dephosphorylate proteins. The eight-hour artificial-substrate result provides insufficient evidence that protein dephosphorylation is a physiological role of laforin. In the combined evidence, this process assignment is likely over-annotation; the measured molecular-function capacity is retained.

Summary: The HIRIP5-specific protein-dephosphorylation claim remains unresolved at the original experiment level.

Reason: The complete PMID:12915448 HIRIP5 original remains inaccessible. Its abstract positively reports interaction and in-vitro HIRIP5 dephosphorylation, which cannot be discarded from the later negative GSK3β/GS/PTG assays because these are different substrates. PMID:11001928 establishes a prolonged model-protein assay, and PMID:16901901 establishes glucan-directed activity with specific candidate-protein negatives; neither independently verifies or refutes the HIRIP5 experiment or establishes its physiological scope. Retain source-specific uncertainty rather than infer that no protein substrate exists.

Summary: Phosphate-containing compound metabolism is a correct but nonspecific description of laforin catalysis.

Reason: The matched ARBA00028024 rule is supported by purified human laforin releasing phosphate from amylopectin (PMID:25538239, Methods and Figure 4A). Retain this broad phosphate-metabolism annotation as non-core because it does not identify the distinctive glucan-repair activity already represented by carbohydrate phosphatase activity and glycogen metabolism. GO:0016311 dephosphorylation fits the chemistry but its current QuickGO usage value None denotes a restriction against direct annotation, not zero annotation counts; the official term display defines this explicitly. The term is not currently obsolete, and OLS lists it as an obsoletion candidate. The inspected child terms do not provide a suitable glucan-specific process. Therefore no restricted or invented process term is authored as a replacement.

Summary: Carbohydrate phosphatase activity identifies the established physiological substrate class.

Reason: The mouse donor phosphatase assertion is genuine, but its PMID:24430976 original was not recovered beyond metadata/abstract. Independent actual human PMID:25538239 Methods and Results directly establish amylopectin phosphate release by full-length laforin; mouse recombinant laforin also releases phosphate from glycogen in PMID:18040046 Figures 1–2. This supports refining the broad enzyme term to carbohydrate phosphatase without treating pNPP/OMFP or a cell-host species as proof of substrate or protein origin. The current GO:0019203 child sugar-phosphatase and its children do not supply a verified polymer-specific replacement.

Proposed replacements: carbohydrate phosphatase activity

Summary: Laforin directly hydrolyzes phosphate from carbohydrate polymers.

Reason: PMID:16901901 Methods and Results/Figures 1–5 show phosphate release from potato amylopectin by bacterially expressed laforin, with C266S and other active phosphatases as controls. The experiment did not detect glycogen phosphate release under its own conditions; glycogen-specific catalysis is independently established in PMID:25544560. Carbohydrate phosphatase accurately covers the demonstrated polymer chemistry without inferring glucose transfer or an exclusive carbon-position preference.

Summary: Human laforin has directly measured carbohydrate phosphatase activity.

Reason: PMID:25538239 Methods and Results compare malachite-green detection of phosphate released from amylopectin with a separate OMFP synthetic-substrate assay. Human 331-aa full-length laforin is active whereas catalytic C266S/R272A controls and isolated DSP preparations differ. The inactive DSP crystal is not itself an enzyme-activity assay. The glucan-directed reaction is a central molecular function.

Summary: Laforin directly catalyzes carbohydrate dephosphorylation.

Reason: Target matches ARBA00028086 and laforin-like phosphatase InterPro IPR045204; the current mouse donor carries direct PMID:18040046 evidence. That original explicitly uses recombinant mouse laforin for glycogen/amylopectin assays. Independently, human full-length 331-aa laforin in PMID:25538239 Methods and Results releases phosphate from amylopectin, whereas OMFP is assayed separately. This establishes the carbohydrate phosphatase activity without assigning it to untested natural products or equating the isolated inactive DSP crystal construct with full-length enzyme.

Summary: The Reactome catalyst is explicitly human laforin with carbohydrate phosphatase activity.

Reason: R-HSA-3781011 catalystActivity 3781002 has physicalEntity 3781017 and activeUnit 3777103, the EPM2A participant referenced to UniProt O95278 through referenceEntity 146927. Its activity is GO:0019203 and the reaction removes phosphate from glycogen. The GYG2-specific reaction is inferred from the GYG1 counterpart R-HSA-3781018, so it is not a separate direct liver assay. Actual human PMID:25538239 independently establishes the enzyme activity; mouse PMID:18040046 supplies direct glycogen-substrate chemistry.

Summary: The carbohydrate-binding function includes directly demonstrated glycogen recognition.

Reason: PMID:25538239 characterizes carbohydrate-dependent human 331-aa laforin behavior. Independent PMID:11739371 Figure 2C/D uses purified recombinant human 331-aa protein and protein-free glycogen: WT and C266S co-sediment with glycogen, a CBD double mutant does not, and the no-glycogen control excludes simple protein aggregation. Refine to the existing glycogen-binding term, without claiming that glycogen is the only bound glucan.

Proposed replacements: glycogen binding

Summary: Glycogen binding is a directly demonstrated, more informative target function.

Reason: IPR013784 identifies a carbohydrate-binding fold but does not by itself select a ligand. In PMID:11739371, human-muscle-derived laforin was purified and directly cosedimented with protein-free glycogen; carbohydrate-binding-domain mutants disrupted association (Methods and Results/Figure 2). Glycogen binding therefore provides an evidence-backed refinement. This does not exclude experimentally observed starch/polyglucosan recognition or transfer activity to every product.

Proposed replacements: glycogen binding

Summary: The broad polysaccharide-binding mapping can be refined to glycogen binding.

Reason: ARBA00085569 matches laforin and the mouse donor source traces to PMID:15102711. That paper includes human recombinant starch/glycogen binding and human C266S transgene observations in mice; it should not be described as a native mouse-only binding assay. Independently, human laforin binding to protein-free glycogen is directly established in PMID:11739371. Glycogen binding is a supported physiological refinement, while the separate starch-binding annotation retains the in-vitro alternate ligand.

Proposed replacements: glycogen binding

Summary: Dendritic localization was observed in a defined human-laforin transgenic experiment.

Reason: The current mouse Epm2a donor cites PMID:15102711. Its actual Results (generation of transgenic mice) identify human myc-tagged C266S laforin expressed about 100-fold above endogenous levels. Figure 2 and the operative brain Results place the signal in neuronal dendritic cytoplasm. Retain this positive location with the mutant/overexpression and mouse-host boundaries; it is not an endogenous wild-type human neuronal assay or evidence that every EPM2A product is dendritic.

Summary: Protein-substrate chemistry is measured in vitro, but a physiological protein-dephosphorylation role remains unestablished.

Reason: PMID:11739371 directly tests pNPP hydrolysis and glycogen binding/targeting; those experiments do not establish peptidyl-tyrosine dephosphorylation. PMID:11001928 Methods pp. 2258–2259 and Figure 2B directly measure phosphate release from PKA- or c-Abl-phosphorylated myelin basic protein using bacterial GST-tagged human 331-residue laforin, including GST and arsenate controls, after an eight-hour incubation. This establishes protein-substrate catalytic capacity. It does not identify an endogenous protein-dephosphorylation pathway. PMID:16901901 Methods and Results/Figures 1–4 instead find no dephosphorylation of the tested GSK3β, GS or PTG substrates and establish selective amylopectin dephosphorylation. These negatives are substrate- and condition-specific, not proof that laforin can never dephosphorylate proteins. The eight-hour artificial-substrate result provides insufficient evidence that protein dephosphorylation is a physiological role of laforin. In the combined evidence, this process assignment is likely over-annotation; the measured molecular-function capacity is retained.

Summary: Direct human solution measurements resolve self-association as homodimerization.

Reason: The matched ARBA rule and current mouse self-interaction source establish the provenance of the imported assertion. Independent human full-length 331-aa laforin in PMID:25538239 shows dimeric behavior by analytical ultracentrifugation and SAXS (Methods and Results); human laforin solution measurements in PMID:25544560 also support a dimer. Refine identical-protein binding to protein homodimerization, consistently with the direct target experiments, without creating a separate core function or extending dimerization to every natural product. C329S remains dimeric in PMID:25538239, so a universal disulfide mechanism is not inferred.

Proposed replacements: protein homodimerization activity

Summary: Human laforin forms a homodimer in the tested solution conditions.

Reason: PMID:25538239 AUC and SAXS experiments on purified human 331-aa support a stable dimer; the isolated catalytic-domain crystal and the proposed full-length arrangement must be kept separate. The C329S protein remains dimeric in this study, so the result is not evidence that one particular disulfide is universally required. Retain dimerization as part of the characterized catalytic architecture without imposing the state on every splice product.

Summary: The tested human laforin preparation exhibits homodimerization.

Reason: PMID:25544560 solution scattering/ultracentrifugation and Figure 5 support a dimer and disruption of that state by F321S. The main cloning section identifies human 1–328 N-His6, with a separate DXMS C-His6 description; those engineered preparations are not new natural products. Dimerization informs glucan recognition/catalytic organization and does not need a separate disease-level core claim.

Summary: Neuronal soma localization is supported within the documented transgenic context.

Reason: The mouse donor annotation traces to PMID:15102711, whose actual experiment expressed human myc-tagged C266S laforin in mice at high abundance. Brain Results and Figure 2 identify neuronal cell-body/perikaryal signal. This positively supports the compartment in that experiment, while the catalytic mutation, overexpression and species-host context limit extrapolation to endogenous human wild-type protein or other products.

Summary: The rule provenance is verified; its particular negative-regulation claim remains unresolved.

Reason: ARBA00028171 condition set 7 matches the two laforin FunFams and Metazoa taxon, but the returned rule does not identify a source experiment establishing the negative regulation asserted by GO:0045936. Human laforin directly removes phosphate from glucans (PMID:25538239, Methods/Figure 4A), and its malin/PTG scaffold changes PTG abundance and glycogen accumulation (PMID:18070875, Results/Figures 1–3). These findings establish chemistry and regulation without determining whether this specific broad phosphate-metabolism regulatory assertion is correct. The inspected rule does not show that its inference was made by confusing dephosphorylation with inhibition of phosphorylation. Accordingly, no concrete source-side conflation or contradictory target experiment is established that would justify REMOVE or MARK_AS_OVER_ANNOTATED. UNDECIDED records this remaining uncertainty even though relevant primary texts are available.

Summary: The donor phenotype concerns glucan dephosphorylation and maintenance, not phosphate addition.

Reason: Mouse Q9WUA5 carries an IEP carbohydrate-phosphorylation annotation to PMID:18852261. In the actual original, Epm2a deficiency changes glycogen phosphate content, whereas purified active laforin removes phosphate (Figure 4); no laforin-catalyzed phosphate-addition reaction is shown. GO:0046835 specifically describes introduction of phosphate into carbohydrate. Refine to glycogen metabolic process to preserve the established substrate-maintenance role. GO:0016311 dephosphorylation currently has usage None and no inspected glucan-specific child, so it is not authored as a replacement.

Proposed replacements: glycogen metabolic process

Summary: ER-adjacent localization is established, while the exact membrane-face association remains unresolved.

Reason: The combined source has both matched ARBA00027306 condition set 1 and mouse Q9WUA5 donor evidence linked to PMID:15102711. The original uses a human C266S transgene with roughly 100-fold expression in mice: Results/Figure 2F places signal near neuronal ER, while liver Results/Figure 6E,G explicitly leave ER-adjacent polyglucosan association unresolved. Human PMID:11001928 Results/Figures 6–8 places laforin on polysomes; immunogold does not label the associated membranous structure in that preparation. These observations do not establish the embedded, attached or peripheral membrane association required by GO:0098554, but neither do they exclude another membrane-associated pool or resolve the incomplete product-specific original evidence. The ARBA condition match contains no experimental membrane-localization test. Thus the exact assertion remains undecided rather than being declared a demonstrated erroneous transfer merely because some full text was accessible.

Summary: The source establishes cytoplasmic polysome localization rather than a specific membrane leaflet.

Reason: The actual PMID:11001928 Results/Figures 5–8 and Discussion establish association with polysomes in a microsomal fraction: extensive QM-marker overlap, sucrose-gradient co-sedimentation and immunogold labeling over ribosomal units. The Results/caption explicitly report no signal over the associated membranous structures, while direct ER interaction remains a possibility in the Discussion. This is positive cytoplasmic localization, but membrane attachment is not established by the inspected experiments. Current GO membrane-side definitions require an embedded, attached or peripheral membrane association; the historical polysome term is now obsolete. Refine to cytoplasm while retaining the rough-ER-associated polysome observation in the reason. No direct lipid binding or universal exclusion from ER-adjacent structures is claimed.

Proposed replacements: cytoplasm

Summary: The carbohydrate-binding module directly binds glycogen.

Reason: PMID:11739371 Experimental Procedures and Figure 2C/D test purified human 331-aa His-tagged protein with protein-free glycogen. WT and catalytic-dead C266S bind, the W32G/K87A CBD double mutant fails, and the no-glycogen control argues against nonspecific aggregation. The binding result is independent of pNPP hydrolysis and does not establish equal binding by the other imported products.

Summary: The tested human laforin preparation binds glycogen directly.

Reason: PMID:25544560 co-sedimentation and carbohydrate-interaction experiments support glycogen binding, while mutational analysis separates binding from catalytic effects. Main Methods specify an engineered human 1–328 preparation and distinguish its tag description from DXMS. The direct human 331-aa glycogen-binding experiment in PMID:11739371 independently supports the natural long product; no equivalence is assumed for shorter or external products.

Summary: Human laforin binds starch in vitro as an alternate glucan ligand.

Reason: The target CBM20/starch-binding InterPro mapping is consistent with the actual PMID:15102711 Results/Figure 7C: recombinant human GST-laforin cosediments with starch more strongly than glycogen under the compared conditions. This supports starch binding as a biochemical capacity, not human starch synthesis or a physiological intracellular starch substrate. Preserve its distinction from native glycogen/polyglucosan recognition and from untested alternative products.