AK2

Summary: Purified human liver mitochondrial enzyme directly catalyzes adenylate interconversion.

Reason: The original JBC full article was recovered externally from its author-uploaded text. Its Methods assay ATP + AMP to ADP and the reverse reaction with distinct coupled-enzyme systems; Results report purification from normal human liver mitochondria and adenine-nucleotide specificity. The metadata-only local cache does not contain these experiments. This direct chemistry supports the original EXP assertion, independently corroborated by active recombinant human AK2A and AK2B.

Summary: The PAINT kinase assertion agrees with direct human AK2 catalysis.

Reason: Cached PAINT records contain the GO:0004017 IBD at PTN000599576. The human target is among the descendant evidences, which is valid experimental grounding rather than circularity. Direct human enzyme studies provide positive support for the inherited kinase activity. The presence of kinase-loss annotations elsewhere in this family does not contradict those target assays.

Summary: The combined electronic sources identify the measured ATP:AMP phosphotransfer reaction.

Reason: The specific adenylate-kinase signatures, RHEA:12973 and EC:2.7.4.3 agree with ATP + AMP = 2 ADP. Human recombinant assays independently support the target. The UniRule internals were not recovered, so correctness of the target reaction does not imply that every rule condition was audited.

Summary: Both tested human AK2 isoforms are enzymatically active.

Reason: The cached original abstract explicitly identifies recombinant human AK2A and AK2B expressed in E. coli and purified as active enzymes. That statement supports the curated kinase activity. It does not establish the activity of all six sequence isoforms.

Summary: The reported AK2–PMP22 interaction does not specify a useful molecular activity.

Reason: The source is a systematic human interaction study using repeated yeast two-hybrid screens and an integrated interaction network; the immutable UniProt record independently lists this PMP22 pair. Remove the generic protein-binding term under the curation policy, without rejecting the pair or claiming a single unreplicated test. The recovered main article does not establish a specific AK2 mechanism for this partner, and its pair-level supplementary evidence was not independently reanalyzed. DUSP26 activation from a different study is not a replacement for this pair.

Summary: The reported AK2–VIM interaction does not specify a useful molecular activity.

Reason: The source is a systematic human interaction study using repeated yeast two-hybrid screens and an integrated interaction network; the immutable UniProt record independently lists this VIM pair. Remove the generic protein-binding term under the curation policy, without rejecting the pair or claiming a single unreplicated test. The recovered main article does not establish a specific AK2 mechanism for this partner, and its pair-level supplementary evidence was not independently reanalyzed. DUSP26 activation from a different study is not a replacement for this pair.

Summary: The reported AK2–PECAM1 interaction does not specify a useful molecular activity.

Reason: The source is a systematic human interaction study using repeated yeast two-hybrid screens and an integrated interaction network; the immutable UniProt record independently lists this PECAM1 pair. Remove the generic protein-binding term under the curation policy, without rejecting the pair or claiming a single unreplicated test. The recovered main article does not establish a specific AK2 mechanism for this partner, and its pair-level supplementary evidence was not independently reanalyzed. DUSP26 activation from a different study is not a replacement for this pair.

Summary: The reported AK2–PRKACA interaction does not specify a useful molecular activity.

Reason: The source is a systematic human interaction study using repeated yeast two-hybrid screens and an integrated interaction network; the immutable UniProt record independently lists this PRKACA pair. Remove the generic protein-binding term under the curation policy, without rejecting the pair or claiming a single unreplicated test. The recovered main article does not establish a specific AK2 mechanism for this partner, and its pair-level supplementary evidence was not independently reanalyzed. DUSP26 activation from a different study is not a replacement for this pair.

Summary: The reported AK2–TGFBR2 interaction does not specify a useful molecular activity.

Reason: The source is a systematic human interaction study using repeated yeast two-hybrid screens and an integrated interaction network; the immutable UniProt record independently lists this TGFBR2 pair. Remove the generic protein-binding term under the curation policy, without rejecting the pair or claiming a single unreplicated test. The recovered main article does not establish a specific AK2 mechanism for this partner, and its pair-level supplementary evidence was not independently reanalyzed. DUSP26 activation from a different study is not a replacement for this pair.

Summary: The reported AK2–PSEN2 interaction does not specify a useful molecular activity.

Reason: The source is a systematic human interaction study using repeated yeast two-hybrid screens and an integrated interaction network; the immutable UniProt record independently lists this PSEN2 pair. Remove the generic protein-binding term under the curation policy, without rejecting the pair or claiming a single unreplicated test. The recovered main article does not establish a specific AK2 mechanism for this partner, and its pair-level supplementary evidence was not independently reanalyzed. DUSP26 activation from a different study is not a replacement for this pair.

Summary: The reported AK2–SERPINH1 interaction does not specify a useful molecular activity.

Reason: The source is a systematic human interaction study using repeated yeast two-hybrid screens and an integrated interaction network; the immutable UniProt record independently lists this SERPINH1 pair. Remove the generic protein-binding term under the curation policy, without rejecting the pair or claiming a single unreplicated test. The recovered main article does not establish a specific AK2 mechanism for this partner, and its pair-level supplementary evidence was not independently reanalyzed. DUSP26 activation from a different study is not a replacement for this pair.

Summary: The reported AK2–ATXN1 interaction does not specify a useful molecular activity.

Reason: The source is a systematic human interaction study using repeated yeast two-hybrid screens and an integrated interaction network; the immutable UniProt record independently lists this ATXN1 pair. Remove the generic protein-binding term under the curation policy, without rejecting the pair or claiming a single unreplicated test. The recovered main article does not establish a specific AK2 mechanism for this partner, and its pair-level supplementary evidence was not independently reanalyzed. DUSP26 activation from a different study is not a replacement for this pair.

Summary: ATP binding is intrinsic to the adenylate-kinase reaction.

Reason: The adenylate-kinase signatures are consistent with ATP serving as a direct substrate, not merely an indirect energy requirement. The purified human enzyme kinetics and curated reaction independently establish nucleotide binding during catalysis.

Summary: Broad cytoplasmic localization is compatible with mitochondrial AK2.

Reason: GO:0005737 includes cytoplasmic organelles and is not synonymous with cytosol. Its PAINT assertion at PTN000599576 therefore agrees with mitochondrial human AK2. The more precise intermembrane-space annotations describe the same principal pool without making this broad inherited location false. Reported extramitochondrial pools are not required to justify this row.

Summary: The mitochondrial proteomics annotation agrees with established human AK2 localization.

Reason: The cached MitoCoP study examines human mitochondrial proteomes. Its AK2-specific supplementary measurement was not separately recovered, so the original HTP assignment is retained with curator deference and independent human liver mitochondrial biochemistry. This source-level mitochondrion assertion is not converted into an intermembrane-space assay.

Summary: The mitochondrial PAINT assertion matches human AK2.

Reason: The cached GO:0005739 IBD is placed at PTN001916109. Independent human mitochondrial purification and curated localization support this target; full tree and alignment reconstruction were not performed. The broad organelle assertion is retained at the resolution of this source.

Summary: AK2 is assigned to the mitochondrial intermembrane space.

Reason: The UniProt location mapping and UniRule agree with the established AK2 compartment. Reactome explicitly distinguishes its rat-enzyme localization inference from the human catalytic assay. Retain this conserved localization with that provenance; the rule conditions themselves were not independently recovered.

Summary: The reaction is localized to the mitochondrial intermembrane space.

Reason: The curated reaction summary explicitly says that this subcompartment assignment is inferred from the homologous rat enzyme. Its human biochemical citation supports catalysis, not a separate human subcompartment-mapping experiment. The conserved localization agrees with UniProt and is appropriate at this source resolution.

Summary: The reaction is localized to the mitochondrial intermembrane space.

Reason: The curated reaction summary explicitly says that this subcompartment assignment is inferred from the homologous rat enzyme. Its human biochemical citation supports catalysis, not a separate human subcompartment-mapping experiment. The conserved localization agrees with UniProt and is appropriate at this source resolution.

Summary: AK2 directly forms ADP in the ATP-plus-AMP reaction.

Reason: GO:0006172 includes chemical formation of ADP; it does not require de novo synthesis of the adenine or ribose moiety. The recovered PTN000599693 IBD agrees with the direct human reaction. Reverse flux also forms ATP and AMP, without invalidating this forward reaction.

Summary: The AK2 subfamily assignment supports direct ADP formation.

Reason: InterPro:IPR028587 and the UniRule source agree with ATP + AMP = 2 ADP, independently established for human enzyme. The specific chemical step supports this biosynthetic-process term; this is not a claim that AK2 builds the nucleotide backbone. UniRule internals remain unaudited.

Summary: Adenylate interconversion is a direct AK2 pathway step.

Reason: AK2 itself performs reversible phosphate transfer among AMP, ADP and ATP. The Reactome pathway places this reaction within nucleotide interconversion, rather than attributing the entire pathway to a merely required substrate. The focused AK2 reaction records provide target-specific support.

Summary: The broad electronic kinase class can be refined to AMP kinase activity.

Reason: The source class is chemically compatible but less informative than the measured human ATP:AMP phosphotransfer activity. Refine to GO:0004017 using independent human enzymology and the exact RHEA reaction. This does not assert that the broad mapping is false or that unexamined ARBA rule conditions are valid.

Proposed replacements: AMP kinase activity

Summary: The broad electronic kinase class can be refined to AMP kinase activity.

Reason: The source class is chemically compatible but less informative than the measured human ATP:AMP phosphotransfer activity. Refine to GO:0004017 using independent human enzymology and the exact RHEA reaction. This does not assert that the broad mapping is false or that unexamined ARBA rule conditions are valid.

Proposed replacements: AMP kinase activity

Summary: Mouse AK2 localization supports a contextual sperm-flagellum assignment.

Reason: The historical MGI comparative graph traces mouse Ak2 donor Q9WTP6/ENSMUSP00000030583 to IDA evidence in PMID:16790685. Its primary abstract specifically localizes AK2 to the mitochondrial sheath of the sperm midpiece; AK1 has a different flagellar distribution. This is a plausible transfer of a conserved mitochondrial protein to the corresponding human sperm compartment, retained as non-core with explicit mouse provenance. The source is not a direct human sperm assay and does not establish AK2 as an axonemal motor or structural filament.

Summary: AMP is a direct substrate and product of AK2.

Reason: The reversible adenylate-kinase reaction consumes AMP in the ADP-forming direction and produces it in the reverse direction. This directly supports AMP metabolism without adding a de novo purine-backbone synthesis claim. UniRule condition-level provenance is unresolved, but the human chemistry is established.

Summary: ATP is a direct substrate and product of AK2.

Reason: AK2 transfers ATP terminal phosphate to AMP and can catalyze the reverse formation of ATP from ADP. This directly supports ATP metabolism without assigning oxidative phosphorylation or ATP hydrolysis to the enzyme. UniRule internals were not independently read.

Summary: The broad nucleotide-metabolism annotation can be refined to interconversion.

Reason: The direct AK2 step is reversible adenylate interconversion, already represented by the curated pathway row. GO:0015949 captures that work more specifically than generic nucleobase-containing small-molecule metabolism. This is a source-specific refinement, not an additional NEW process. ARBA conditions were not recovered; the target judgment rests on measured chemistry.

Proposed replacements: nucleobase-containing small molecule interconversion

Summary: The source-specific exosomal AK2 detection remains unresolved.

Reason: The paper reports a purified human B-cell exosome proteome and its MHC-II-associated subset. The original main text was inspected externally, but the AK2-specific supplementary peptide entry and its assignment to the total versus co-immunoprecipitated fraction were not recovered. The HDA localization is therefore left unresolved; mitochondrial abundance is not evidence of contamination and no exosome-biogenesis function is inferred.

Summary: Mouse AK2 localization supports a contextual sperm mitochondrial-sheath assignment.

Reason: The historical MGI comparative graph traces mouse Ak2 donor Q9WTP6/ENSMUSP00000030583 to IDA evidence in PMID:16790685. Its primary abstract specifically localizes AK2 to the mitochondrial sheath of the sperm midpiece; AK1 has a different flagellar distribution. This is a plausible transfer of a conserved mitochondrial protein to the corresponding human sperm compartment, retained as non-core with explicit mouse provenance. The source is not a direct human sperm assay and does not establish AK2 as an axonemal motor or structural filament.

Summary: Human AK2 directly binds and activates DUSP26 in a distinct noncanonical activity.

Reason: The original human recombinant-protein experiments show AK2 association with DUSP26 and increased phosphatase activity on pNPP and Plk-phosphorylated FADD. AK2 alone lacks this phosphatase activity, AK3 and FADD controls do not stimulate DUSP26, and kinase-defective AK2 K28E retains activation. Thus AK2 performs the activating step, while DUSP26 performs dephosphorylation. GO:0072542 has the matching binding-plus-activation definition, and local AMBRA1 and CALM1 experimental annotations establish the same regulatory role as legitimate usage. Its parents are enzyme-regulator terms, not kinase or generic binding ancestors of an existing AK2 term. No exact AK2 entry was found in the cached GO-CAM index. This is a bounded activity proposal from tagged recombinant and human-cell experiments; native complex stoichiometry and its contribution to reticular dysgenesis remain unresolved. No proliferation, apoptosis or differentiation BP is proposed from necessity evidence. The study places this regulatory interaction in an extramitochondrial nuclear-enriched pool. HeLa fractionation detects nuclear AK2, endogenous AK2-DUSP26 interaction is enriched there, and forced mitochondrial targeting fails to reduce p-FADD. This source-specific context is distinct from the intermembrane-space catalytic pool.