# AGK notes

## 2026-06-03 - Proteostasis PN review

- Deep research provenance: `just deep-research-falcon human AGK --fallback perplexity-lite` was attempted for this review. Falcon timed out after 600 seconds, then the configured `perplexity-lite` fallback failed with an API quota error. No provider deep-research file was produced, so this review uses cached publications, UniProt, GOA, and PN project reports directly.

- AGK has two separable core activities. It is a mitochondrial lipid kinase that phosphorylates monoacylglycerol and diacylglycerol to produce lysophosphatidic acid and phosphatidic acid [PMID:15939762 "AGK, that phosphorylates monoacylglycerol and diacylglycerol to form LPA and PA, respectively"]. It is also a kinase-independent TIM22 complex subunit required for carrier-protein import into the mitochondrial inner membrane [PMID:28712726 "AGK functions in a kinase-independent manner to maintain the integrity of the TIM22 complex, where it facilitates the import and assembly of mitochondrial carrier proteins"].

- The direct human lipid-kinase paper does not support sphingosine or ceramide kinase activity as a robust AGK function. It reports significant phosphorylation only for monoacylglycerols and diacylglycerols, not ceramide or sphingosine [PMID:15939762 "Significant phosphorylated products were only detected with monoacylglycerols and diacylglycerols as substrates, but not with any other lipid tested, including ceramide and sphingosine"]. A second ceramide-kinase paper also cautions that no evidence was found for ceramide phosphorylation by the multiple lipid kinase [PMID:16269826 "No evidence for phosphorylation of ceramide by the recently described multiple lipid kinase was found"].

- The strongest PN-relevant role is the TIM22 import-complex role, not a generic protein-transport assertion. The PN projection reports `AGK -> GO:0042721 TIM22 mitochondrial import inner membrane insertion complex` as already in GOA exactly, and separately proposes broad `GO:0015031 protein transport` from the parent mitochondrial protein-transport class. Because AGK already has the narrower TIM22 complex and `protein insertion into mitochondrial inner membrane` annotations, adding generic `protein transport` would reduce specificity rather than improve the review [file:projects/PROTEOSTASIS/reports/pn_projection/pn_projected_annotations.tsv "AGK"; "GO:0015031"; "new_to_goa"; "GO:0042721"; "already_in_goa_exact"].

- Several Reactome-derived cytosol rows in the seeded GOA are attached to BRAF/RAF fusion events rather than native AGK biology. They should be removed for AGK because the native protein is experimentally supported as mitochondrial inner-membrane/intermembrane-space associated, and the Reactome event titles in the seeded review refer to BRAF or RAF fusion mutant dimers rather than AGK function.

- Generic `protein binding` rows are treated as over-annotations. The SMIM26 interaction is biologically interesting in renal-cell carcinoma context, but it is not a core molecular function for AGK and should not displace the better-supported lipid kinase and TIM22 complex roles [PMID:37009826 "SMIM26, but not LINC00493, suppresses ccRCC growth and metastatic lung colonization by interacting with acylglycerol kinase (AGK)"].

## Falcon deep research findings (2026-06-07)

A Falcon deep-research run completed for AGK (the earlier 2026-06-03 attempt had timed out). The report largely CONFIRMS the existing review; below is a synthesis of KEY findings, flagged CONFIRMS / NEW / PROVISIONAL with provenance. PMIDs resolved via PubMed (DOI->PMID) where possible; Jiang 2020 (Blood) and Vukotic 2017 (Mol Cell) PMIDs could not be reliably resolved in this run and are cited by DOI only (not used to alter annotations).

- NEW (substrate class / pathway): The TIM22 complex (with AGK) imports sideroflexins (SFXN proteins) as a novel TIM22 substrate class, beyond the canonical SLC25 carriers; AGK loss perturbs SFXN biogenesis and reduces serine-independent proliferation, linking AGK deficiency to impaired mitochondrial one-carbon metabolism [PMID:33476211 Jackson 2021 "we observed down-regulation of SFXN proteins... and show that they represent a novel class of TIM22 complex substrate"; "dysregulation of one-carbon metabolism is a molecular feature in the biology of Sengers syndrome"]. This extends the import role beyond the SLC25/ANT carriers in the existing review.

- NEW (pathway context): A 2023 expert review of mitochondrial phospholipid metabolism positions AGK-mediated DAG->PA phosphorylation in the inner mitochondrial membrane as an intramitochondrial PA-generating route, with PA feeding downstream phospholipid synthesis including cardiolipin [PMID:37655851 Joshi 2023, "Mitochondrial phospholipid metabolism in health and disease"]. This situates AGK's lipid-kinase product (PA) upstream of cardiolipin biogenesis; the existing review captures glycerolipid metabolism but not this downstream phospholipid-homeostasis link. PROVISIONAL as a direct AGK->cardiolipin causal claim (review-level synthesis, not direct AGK enzymology).

- CONFIRMS (mechanism): Vukotic et al. 2017 (Mol Cell; DOI:10.1016/j.molcel.2017.06.013 — the same study underlying the already-cited PMID:28712724/28712726 cluster) showed the G126E mutation in the conserved GDG catalytic motif abolishes lipid phosphorylation, and that kinase-dead AGK still restores TIM22 complex integrity and ANT1 import — directly supporting the separable kinase-independent import role already in the review. Kinase activity is instead required for cristae maintenance and apoptotic resistance [Vukotic 2017, DOI:10.1016/j.molcel.2017.06.013]. Reinforces existing ACCEPT calls on TIM22/import and the kinase-independent reason text.

- CONFIRMS (disease genetics): Mayr et al. 2012 established biallelic loss-of-function AGK as causal for Sengers syndrome (congenital cataracts, hypertrophic cardiomyopathy, skeletal myopathy, lactic acidosis), with AGK loss reducing the adenine nucleotide translocator in the inner membrane in muscle [PMID:22284826 "Lack of the mitochondrial protein acylglycerol kinase causes Sengers syndrome"; "loss of AGK led to a decrease of the adenine nucleotide translocator in the inner mitochondrial membrane in muscle"]. This is the foundational genetics paper already referenced as an additional_reference_id but not previously in the references list; now added.

- NEW / PROVISIONAL (non-mitochondrial context): An AGK-JAK2 interaction has been proposed to promote proplatelet formation / thrombopoiesis, with cell-permeable JAK2 JH2-domain peptides increasing AGK-JAK2 binding (preclinical concept) [Jiang 2020, Blood 136:119, DOI:10.1182/blood.2019003851]. PMID unresolved in this run; kept in notes only, not used to add annotations. This is outside the core mitochondrial lipid-kinase/TIM22 roles and may reflect a cytosolic/signaling moonlighting context.

- CONFIRMS (localization/topology): Protease-protection and fractionation place AGK exclusively in mitochondria, exposed to the intermembrane space and peripherally associated with the inner membrane without predicted transmembrane segments — consistent with the existing IMM/IMS annotations and the MODIFY of the Reactome outer-membrane row.