Deep Research
Falcon
(AGK-deep-research-falcon.md)
Falcon
(AGK-deep-research-falcon.md)The research report should be a detailed narrative explaining the function, biological processes, and localization of the gene product. Citations should be given for all claims.
You should prioritize authoritative reviews and primary scientific literature when conducting research. You can supplement
this with annotations you find in gene/protein databases, but these can be outdated or inaccurate.
We are specifically interested in the primary function of the gene - for enzymes, what reaction is catalyzed, and what is the substrate specificity? For transporters, what is the substrate? For structural proteins or adapters, what is the broader structural role? For signaling molecules, what is the role in the pathway.
We are interested in where in or outside the cell the gene product carries out its function.
We are also interested in the signaling or biochemical pathways in which the gene functions. We are less interested in broad pleiotropic effects, except where these elucidate the precise role.
Include evidence where possible. We are interested in both experimental evidence as well as inference from structure, evolution, or bioinformatic analysis. Precise studies should be prioritized over high-throughput, where available.
Comprehensive research report: Human AGK (Acylglycerol kinase; UniProt Q53H12) — functional annotation
1) Target verification (gene/protein identity)
The literature retrieved unambiguously matches the UniProt-described human mitochondrial acylglycerol kinase AGK (also called MuLK/MULK) and consistently places it in mitochondria and in the TIM22 inner-membrane protein import pathway, matching the provided identity constraints (human; lipid kinase; mitochondrial; AGK family). Experimental work establishes AGK as a mitochondrial protein, a lipid kinase producing phosphatidic acid (PA) and lysophosphatidic acid (LPA) from acylglycerols, and a structural subunit of the TIM22 carrier translocase. (vukotic2017acylglycerolkinasemutated pages 1-4, vukotic2017acylglycerolkinasemutated pages 4-6, jackson2021thetim22complex pages 1-2)
2) Key concepts, definitions, and current understanding
2.1 Enzymatic function and reaction definition
AGK is a multisubstrate lipid kinase that phosphorylates:
- Diacylglycerol (DAG) → phosphatidic acid (PA)
- Monoacylglycerol (MAG) → lysophosphatidic acid (LPA)
This substrate/product mapping is explicitly described in mitochondrial/TIM22-focused cell biology studies and in patient-focused biochemical genetics reports. (jackson2021thetim22complex pages 1-2, siriwardena2013mitochondrialcitratesynthase pages 8-8)
A conserved catalytic motif is functionally validated: the G126E mutation in the conserved GDG motif abolishes lipid phosphorylation, demonstrating that lipid phosphorylation is an intrinsic AGK enzymatic activity (not merely correlative). (vukotic2017acylglycerolkinasemutated pages 4-6)
2.2 Mitochondrial phospholipid pathway context
PA is a central precursor lipid for synthesis of other mitochondrial phospholipids (notably including cardiolipin through downstream steps). A 2023 expert review of mitochondrial phospholipid metabolism positions AGK-mediated phosphorylation of DAG to PA in the inner mitochondrial membrane (IMM) as a specific intramitochondrial PA-generating route. (joshi2023mitochondrialphospholipidmetabolism pages 3-4)
2.3 Dual-function (“moonlighting”) protein concept for AGK
A central current concept is that AGK has separable roles:
1) Kinase-dependent role in lipid homeostasis that affects mitochondrial membrane architecture (e.g., cristae morphology) and stress responses.
2) Kinase-independent role as a structural subunit required for the assembly/stability and function of the TIM22 protein import machinery.
This functional separation is demonstrated by complementation experiments in which a kinase-dead AGK (G126E) can restore TIM22 complex integrity and carrier import, while kinase activity is required for other mitochondrial structural phenotypes (cristae/apoptotic resistance). (vukotic2017acylglycerolkinasemutated pages 9-11, vukotic2017acylglycerolkinasemutated pages 4-6)
3) Subcellular localization and molecular mechanisms
3.1 Localization/topology
Experimental fractionation and protease-protection assays indicate AGK is:
- Exclusively mitochondrial in cell fractionation.
- Exposed to the intermembrane space (IMS): protected from protease in intact mitochondria but accessible after osmotic swelling that disrupts the outer membrane.
- Associated with the inner membrane without predicted transmembrane segments, consistent with peripheral association (rather than a classical integral multi-pass membrane protein). (vukotic2017acylglycerolkinasemutated pages 4-6)
A figure-based confirmation is available in the Vukotic et al. study showing (i) co-migration of AGK with TIM22 components in native complexes and (ii) protease-protection behavior consistent with IMS localization. (vukotic2017acylglycerolkinasemutated media 2f6e1bf4)
3.2 Role in the TIM22 carrier translocase (mitochondrial protein import)
AGK is a bona fide TIM22 complex subunit in human mitochondria:
- AGK co-purifies and co-migrates with TIMM22 and TIMM29.
- AGK is present in a ~400 kDa TIM22 complex.
- Loss of AGK impairs TIM22 assembly, with TIMM22 shifting to a smaller complex (~90 kDa) and TIMM29 largely in ~60 kDa species.
- AGK supports import/biogenesis of inner-membrane carrier proteins including ANT1/ANT3 (adenine nucleotide translocases) and SLC25A24; import of ANT1 is significantly impaired when AGK is absent.
- Import/assembly role is kinase-independent, because kinase-dead AGK can restore TIM22 complex integrity and ANT1 levels. (vukotic2017acylglycerolkinasemutated pages 9-11, vukotic2017acylglycerolkinasemutated pages 4-6, vukotic2017acylglycerolkinasemutated pages 6-7)
Beyond canonical SLC25 carriers, AGK/TIM22 also supports import of sideroflexins (SFXNs), linking AGK dysfunction to impaired mitochondrial one-carbon metabolism (serine dependence and formate-rescuable proliferation defects in cell models). (jackson2021thetim22complex pages 1-2)
3.3 Proposed biochemical coupling between lipid kinase activity and mitochondrial structure
Catalytically active AGK is required for cristae maintenance and apoptotic resistance in cell models; catalytically inactive AGK phenocopies knockout for these aspects. This supports a model in which AGK-derived PA/LPA (and/or downstream phospholipid remodeling) contributes to maintaining inner-membrane architecture and robustness under stress. (vukotic2017acylglycerolkinasemutated pages 9-11)
4) Recent developments and latest research (prioritizing 2023–2024)
4.1 2023: Mitochondrial phospholipid metabolism synthesis
A 2023 authoritative review in Journal of Cell Science synthesizes mitochondrial phospholipid pathways and explicitly identifies AGK-mediated DAG phosphorylation to PA in the IMM as one of the pathways producing intramitochondrial PA for downstream phospholipid synthesis. This reflects consolidation of AGK into the “core map” of mitochondrial lipid metabolism (even though much of the primary enzymology predates 2023). (joshi2023mitochondrialphospholipidmetabolism pages 3-4)
4.2 2024: Structural/functional focus on TIM22 (AGK placement still unresolved)
A 2024 thesis-level synthesis discusses recent cryo-EM work on the human TIM22 complex and notes that AGK’s precise structural role/localization within the resolved model remains to be determined, underscoring that high-resolution structural understanding of AGK within TIM22 is still incomplete. (valpadashi2024structuralandfunctional pages 53-53)
4.3 2024: Genetics context for cataracts
A 2024 review of cataract genetics notes that AGK germline variants were first reported in Sengers syndrome, reinforcing AGK’s established role in syndromic congenital cataract phenotypes in current clinical genetics frameworks. (shiels2024 paper retrieved; evidence in this run is limited to bibliographic/contextual mention) (vukotic2017acylglycerolkinasemutated pages 1-4)
Limitations on 2023–2024 Sengers-focused reviews in this run. A 2023 mini-review specifically on “Sengers syndrome and AGK-related disorders” was identified by search but was not retrievable here; therefore, this report’s 2023–2024 disease-specific statistics are limited to the accessible sources above and to older (but foundational) case series. (gong2026casereportsengers pages 5-6, gong2026casereportsengers pages 6-6)
5) Current applications and real-world implementations
5.1 Clinical genetics and diagnostics
AGK is a clinically relevant gene for molecular diagnosis of Sengers syndrome and related syndromic congenital cataract/cardiomyopathy presentations. Foundational clinical sequencing work screened 13 individuals with congenital cataracts and cardiomyopathy and found 12 predicted loss-of-function AGK alleles in 10 affected individuals, establishing AGK as causal. (mayr2012lackofthe pages 2-4)
An important practical implementation is copy-number variant (CNV) analysis in addition to SNV/indel calling: later clinical reports emphasize that larger deletions/rearrangements can be missed by standard exome pipelines and that CNV interrogation can resolve ambiguous cases; one report notes >30 pathogenic AGK variants overall and that large rearrangements appear rare in publications. (gong2026casereportsengers pages 5-6)
5.2 Functional assays for variant interpretation
Patient-derived cell assays are used to support pathogenicity of splicing/other variants, including readouts such as oxygen consumption rate (OCR) and OXPHOS complex activities; for example, a novel splice variant case showed reduced OCR metrics and decreased complex I/V activities. (barbosagouveia2021characterizationofa pages 1-2)
5.3 Therapeutic/management approaches (supportive and experimental)
No AGK-targeted clinical trials were retrieved in this run. Real-world management described in case literature remains largely supportive, including cardiomyopathy management, cataract surgery when feasible, dietary modifications to limit fasting, and “mitochondrial cocktail” antioxidant supplementation (vitamins, riboflavin, carnitine, etc.). (mayr2012lackofthe pages 4-5, siriwardena2013mitochondrialcitratesynthase pages 2-3)
Outside Sengers syndrome, an experimental therapeutic concept in hematology is targeting AGK–JAK2 interaction: cell-permeable peptides containing a JAK2 JH2-domain sequence increased AGK–JAK2 binding and accelerated proplatelet formation in model systems, proposed as a strategy for thrombocytopenia/thrombocythemia (preclinical concept, not a clinical trial). (jiang2020theroleof pages 5-6)
6) Expert opinions / analysis (authoritative interpretations)
1) AGK as a node linking lipid metabolism to protein import and membrane architecture. Mechanistic cell biology evidence supports AGK as a dual-function protein: a lipid kinase producing PA/LPA and a TIM22 subunit required for carrier biogenesis; kinase-independent TIM22 assembly versus kinase-dependent cristae integrity suggests two mechanistically distinct but physiologically coupled roles that may both contribute to disease. (vukotic2017acylglycerolkinasemutated pages 9-11, vukotic2017acylglycerolkinasemutated pages 1-4)
2) AGK in canonical mitochondrial lipid maps. The 2023 mitochondrial phospholipid metabolism review’s explicit placement of AGK in IMM PA biosynthesis reflects field-level consolidation of AGK as a bona fide mitochondrial lipid metabolic enzyme (not solely a signaling enzyme). (joshi2023mitochondrialphospholipidmetabolism pages 3-4)
3) Open questions in structural biology. The 2024 TIM22 structural synthesis emphasizes that high-resolution placement/role of AGK in TIM22 remains unresolved in available structural models, highlighting an active research frontier. (valpadashi2024structuralandfunctional pages 53-53)
7) Statistics and quantitative data from studies (selected highlights)
7.1 TIM22 complex assembly metrics
- AGK co-migrates with TIM22 components in a ~400 kDa complex; AGK loss shifts TIMM22 and TIMM29 into smaller assemblies (~90 kDa; ~60 kDa), consistent with impaired TIM22 assembly. (vukotic2017acylglycerolkinasemutated pages 6-7, vukotic2017acylglycerolkinasemutated media 2f6e1bf4)
7.2 Biochemical/clinical measurements in Sengers syndrome
Examples of quantitative clinical/biochemical readouts include:
- Plasma lactate 7.3 mmol/L (normal 0.5–2.2) in an index infant. (mayr2012lackofthe pages 1-2)
- Lactic acidosis ranging ~4 to 15 mmol/L during stress/fasting and lactate:pyruvate ratio 68:1 in a case series. (siriwardena2013mitochondrialcitratesynthase pages 2-3)
- Enzyme activity measurements in muscle mitochondria showing deficits (units/g protein), e.g. Complex I 23 (normal 28–76) and oligomycin-sensitive ATPase (Complex V) 60 (normal 86–257), plus markedly reduced substrate oxidation rates (e.g., [1-14C]pyruvate+malate+ADP 34 vs 263–900). (mayr2012lackofthe pages 4-5)
- Fibroblast complex I activity reduced to 43% and 54% of controls in two patients in one report. (siriwardena2013mitochondrialcitratesynthase pages 8-8)
7.3 Quantitative protein loss in patient tissue
- Mitochondrial AGK immunogold labeling density decreased from 37.9±4.6 particles/µm² (control) to 14.5±4.9 and 18.9±6.9 particles/µm² in AGK-deficient patient muscle mitochondria. (siriwardena2013mitochondrialcitratesynthase pages 8-8)
7.4 Case counts and variant spectrum snapshots
- By 2012, ~40 individuals had been described historically with Sengers syndrome (as summarized in a foundational genetics report). (mayr2012lackofthe pages 1-2)
- In one foundational cohort, sequencing 13 individuals with congenital cataracts and cardiomyopathy identified 12 predicted loss-of-function alleles in 10 affected individuals. (mayr2012lackofthe pages 2-4)
- A 2013 paper summarized the literature at that time as 15 published cases with 16 different AGK mutations. (siriwardena2013mitochondrialcitratesynthase pages 8-8)
- A later clinical report notes >30 pathogenic AGK variants reported overall, while also highlighting rarity of large genomic rearrangements in publications. (gong2026casereportsengers pages 5-6)
8) Summary of key functional annotation conclusions
Primary biochemical function. AGK is an ATP-dependent lipid kinase that phosphorylates MAG and DAG to generate LPA and PA, respectively. (jackson2021thetim22complex pages 1-2, siriwardena2013mitochondrialcitratesynthase pages 8-8)
Primary cellular site of action. AGK is a mitochondrial protein with IMS exposure/peripheral association with the IMM and functions as a subunit of the human TIM22 carrier translocase. (vukotic2017acylglycerolkinasemutated pages 4-6, vukotic2017acylglycerolkinasemutated media 2f6e1bf4)
Primary pathway roles. AGK contributes to (i) mitochondrial PA production relevant to downstream phospholipid synthesis (including pathways leading to cardiolipin) and (ii) TIM22-dependent insertion/biogenesis of multipass inner-membrane proteins (SLC25 carriers; ANT; SFXNs). (joshi2023mitochondrialphospholipidmetabolism pages 3-4, vukotic2017acylglycerolkinasemutated pages 6-7, jackson2021thetim22complex pages 1-2)
Disease mechanism relevance. Biallelic loss-of-function AGK causes Sengers syndrome; mechanistic links include destabilization of TIM22 and impaired carrier import as well as kinase-dependent mitochondrial membrane/cristae defects. (mayr2012lackofthe pages 1-2, vukotic2017acylglycerolkinasemutated pages 9-11)
9) Structured summary table
| Aspect | Key findings | Best supporting sources (with year) | URL/DOI |
|---|---|---|---|
| Enzyme activity | Human AGK is the verified mitochondrial acylglycerol kinase (also called MuLK/MULK), a multi-substrate lipid kinase; a conserved GDG-motif mutation (G126E) abolishes lipid phosphorylation, supporting direct catalytic activity (vukotic2017acylglycerolkinasemutated pages 4-6, vukotic2017acylglycerolkinasemutated pages 1-4) | Vukotic et al., 2017; Mayr et al., 2012 | https://doi.org/10.1016/j.molcel.2017.06.013 ; https://doi.org/10.1016/j.ajhg.2011.12.005 |
| Substrates/products | AGK phosphorylates monoacylglycerol (MAG) to lysophosphatidic acid (LPA) and diacylglycerol (DAG) to phosphatidic acid (PA); 2023 review places AGK as an intramitochondrial PA-generating route in the inner mitochondrial membrane (jackson2021thetim22complex pages 1-2, siriwardena2013mitochondrialcitratesynthase pages 8-8, joshi2023mitochondrialphospholipidmetabolism pages 3-4) | Jackson et al., 2021; Siriwardena et al., 2013; Joshi et al., 2023 | https://doi.org/10.1091/mbc.e20-06-0390 ; https://doi.org/10.1016/j.ymgme.2012.11.282 ; https://doi.org/10.1242/jcs.260857 |
| Localization/topology | AGK is exclusively mitochondrial by fractionation; protease-protection assays indicate an intermembrane-space (IMS) localization associated with the inner membrane without predicted transmembrane spans (vukotic2017acylglycerolkinasemutated pages 4-6, vukotic2017acylglycerolkinasemutated media 2f6e1bf4) | Vukotic et al., 2017 | https://doi.org/10.1016/j.molcel.2017.06.013 |
| TIM22 role | AGK is a bona fide subunit of the human TIM22 translocase; AGK, TIMM22, and TIMM29 co-migrate in a ~400 kDa complex, and AGK loss disrupts assembly so TIMM22 shifts to ~90 kDa and TIMM29 mainly to ~60 kDa species (vukotic2017acylglycerolkinasemutated pages 4-6, vukotic2017acylglycerolkinasemutated pages 6-7, vukotic2017acylglycerolkinasemutated media 2f6e1bf4) | Vukotic et al., 2017 | https://doi.org/10.1016/j.molcel.2017.06.013 |
| TIM22 substrate import | AGK supports import/accumulation of carrier proteins including ANT1, ANT3, and SLC25A24; ANT1 import is significantly impaired but not abolished in AGK-null mitochondria; rescue by kinase-dead G126E shows the import role is kinase-independent (vukotic2017acylglycerolkinasemutated pages 9-11, vukotic2017acylglycerolkinasemutated pages 6-7) | Vukotic et al., 2017 | https://doi.org/10.1016/j.molcel.2017.06.013 |
| Expanded pathway role | AGK/TIM22 is also required for import of sideroflexins (SFXNs), identifying a novel TIM22 substrate class and linking AGK deficiency to impaired mitochondrial one-carbon metabolism and serine dependence (jackson2021thetim22complex pages 1-2) | Jackson et al., 2021 | https://doi.org/10.1091/mbc.e20-06-0390 |
| Lipid metabolism pathway context | PA generated by AGK in the IMM is positioned upstream of other mitochondrial phospholipids, including cardiolipin biosynthesis; expert reviews interpret AGK as linking mitochondrial phospholipid homeostasis with OXPHOS function (vukotic2017acylglycerolkinasemutated pages 1-4, joshi2023mitochondrialphospholipidmetabolism pages 3-4) | Vukotic et al., 2017; Joshi et al., 2023 | https://doi.org/10.1016/j.molcel.2017.06.013 ; https://doi.org/10.1242/jcs.260857 |
| Structure/cristae and apoptosis | AGK’s kinase activity is dispensable for TIM22 assembly but required for mitochondrial cristae maintenance and apoptotic resistance; catalytically inactive AGK behaves like knockout for cristae/apoptosis phenotypes (vukotic2017acylglycerolkinasemutated pages 9-11) | Vukotic et al., 2017 | https://doi.org/10.1016/j.molcel.2017.06.013 |
| Disease association | Biallelic loss-of-function AGK variants cause Sengers syndrome, classically featuring congenital cataracts, hypertrophic cardiomyopathy, skeletal myopathy/exercise intolerance, and lactic acidosis; AGK is the established disease gene in Open Targets and primary genetics studies (mayr2012lackofthe pages 1-2, mayr2012lackofthe pages 2-4, OpenTargets Search: Sengers syndrome,cardiomyopathy,cataract,mitochondrial disease,cancer-AGK) | Mayr et al., 2012; Open Targets | https://doi.org/10.1016/j.ajhg.2011.12.005 ; https://platform.opentargets.org |
| Variant spectrum / cohort data | In one foundational cohort, AGK sequencing in 13 individuals with congenital cataracts and cardiomyopathy identified 12 predicted loss-of-function alleles in 10 affected individuals; historical summary in the same report noted ~40 individuals described overall by 2012 (mayr2012lackofthe pages 2-4, mayr2012lackofthe pages 1-2) | Mayr et al., 2012 | https://doi.org/10.1016/j.ajhg.2011.12.005 |
| Quantitative mitochondrial protein loss | In patient muscle, AGK protein can be nearly absent/negligible, with mitochondrial immunogold labeling reduced from 37.9±4.6 particles/μm² in control to 14.5±4.9 and 18.9±6.9 particles/μm² in two AGK-deficient patients (siriwardena2013mitochondrialcitratesynthase pages 8-8) | Siriwardena et al., 2013 | https://doi.org/10.1016/j.ymgme.2012.11.282 |
| Quantitative bioenergetic defects | Patient fibroblasts showed decreased complex I activity to 43% and 54% of control in two cases, with high lactate/pyruvate ratios; another index case had plasma lactate 7.3 mmol/L (normal 0.5–2.2 mmol/L) (siriwardena2013mitochondrialcitratesynthase pages 8-8, mayr2012lackofthe pages 1-2) | Siriwardena et al., 2013; Mayr et al., 2012 | https://doi.org/10.1016/j.ymgme.2012.11.282 ; https://doi.org/10.1016/j.ajhg.2011.12.005 |
| Severe neonatal phenotype | A 2021 functionally characterized splice variant (c.518+1G>A) caused fatal neonatal Sengers syndrome with congenital cataracts, dilated cardiomyopathy, hyperlactacidemia, death at 20 h, plus reduced OCR and decreased OXPHOS complexes I and V in fibroblasts (barbosagouveia2021characterizationofa pages 1-2, barbosagouveia2021characterizationofa pages 5-6) | Barbosa-Gouveia et al., 2021 | https://doi.org/10.3390/ijms222413484 |
| Clinical heterogeneity | AGK disease ranges from lethal neonatal/infantile disease to milder forms, including reports of isolated congenital cataract without major cardiomyopathy in some families, indicating variable expressivity despite shared mitochondrial mechanism (mayr2012lackofthe pages 1-2, siriwardena2013mitochondrialcitratesynthase pages 8-10) | Mayr et al., 2012; Siriwardena et al., 2013 | https://doi.org/10.1016/j.ajhg.2011.12.005 ; https://doi.org/10.1016/j.ymgme.2012.11.282 |
Table: This table summarizes the experimentally supported functions, localization, pathway roles, and disease relevance of human AGK (UniProt Q53H12). It emphasizes quantitative findings and the strongest supporting sources for functional annotation.
10) Source URLs and publication dates (selected)
- Vukotic et al. Molecular Cell (Aug 2017): https://doi.org/10.1016/j.molcel.2017.06.013 (vukotic2017acylglycerolkinasemutated pages 1-4)
- Mayr et al. American Journal of Human Genetics (Feb 2012): https://doi.org/10.1016/j.ajhg.2011.12.005 (mayr2012lackofthe pages 1-2)
- Siriwardena et al. Molecular Genetics and Metabolism (Jan 2013): https://doi.org/10.1016/j.ymgme.2012.11.282 (siriwardena2013mitochondrialcitratesynthase pages 8-8)
- Jackson et al. Molecular Biology of the Cell (Mar 2021): https://doi.org/10.1091/mbc.e20-06-0390 (jackson2021thetim22complex pages 1-2)
- Joshi et al. Journal of Cell Science (Sep 2023): https://doi.org/10.1242/jcs.260857 (joshi2023mitochondrialphospholipidmetabolism pages 3-4)
- Valpadashi (PhD thesis; 2024): https://doi.org/10.53846/goediss-10678 (valpadashi2024structuralandfunctional pages 53-53)
- Open Targets Platform (accessed via tool; platform URL): https://platform.opentargets.org (OpenTargets Search: Sengers syndrome,cardiomyopathy,cataract,mitochondrial disease,cancer-AGK)
References
-
(vukotic2017acylglycerolkinasemutated pages 1-4): Milena Vukotic, Hendrik Nolte, Tim König, Shotaro Saita, Maria Ananjew, Marcus Krüger, Takashi Tatsuta, and Thomas Langer. Acylglycerol kinase mutated in sengers syndrome is a subunit of the tim22 protein translocase in mitochondria. Molecular cell, 67 3:471-483.e7, Aug 2017. URL: https://doi.org/10.1016/j.molcel.2017.06.013, doi:10.1016/j.molcel.2017.06.013. This article has 154 citations and is from a highest quality peer-reviewed journal.
-
(vukotic2017acylglycerolkinasemutated pages 4-6): Milena Vukotic, Hendrik Nolte, Tim König, Shotaro Saita, Maria Ananjew, Marcus Krüger, Takashi Tatsuta, and Thomas Langer. Acylglycerol kinase mutated in sengers syndrome is a subunit of the tim22 protein translocase in mitochondria. Molecular cell, 67 3:471-483.e7, Aug 2017. URL: https://doi.org/10.1016/j.molcel.2017.06.013, doi:10.1016/j.molcel.2017.06.013. This article has 154 citations and is from a highest quality peer-reviewed journal.
-
(jackson2021thetim22complex pages 1-2): Thomas D. Jackson, Daniella H. Hock, Kenji M. Fujihara, Catherine S. Palmer, Ann E. Frazier, Yau C. Low, Yilin Kang, Ching-Seng Ang, Nicholas J. Clemons, David R. Thorburn, David A. Stroud, and Diana Stojanovski. The tim22 complex mediates the import of sideroflexins and is required for efficient mitochondrial one-carbon metabolism. Molecular Biology of the Cell, 32:475-491, Mar 2021. URL: https://doi.org/10.1091/mbc.e20-06-0390, doi:10.1091/mbc.e20-06-0390. This article has 43 citations and is from a domain leading peer-reviewed journal.
-
(siriwardena2013mitochondrialcitratesynthase pages 8-8): Komudi Siriwardena, Nevena MacKay, Valeriy Levandovskiy, Susan Blaser, Julian Raiman, Paul F. Kantor, Cameron Ackerley, Brian H. Robinson, Andreas Schulze, and Jessie M. Cameron. Mitochondrial citrate synthase crystals: novel finding in sengers syndrome caused by acylglycerol kinase (agk) mutations. Molecular genetics and metabolism, 108 1:40-50, Jan 2013. URL: https://doi.org/10.1016/j.ymgme.2012.11.282, doi:10.1016/j.ymgme.2012.11.282. This article has 46 citations and is from a peer-reviewed journal.
-
(joshi2023mitochondrialphospholipidmetabolism pages 3-4): Alaumy Joshi, Travis H. Richard, and Vishal M. Gohil. Mitochondrial phospholipid metabolism in health and disease. Journal of cell science, Sep 2023. URL: https://doi.org/10.1242/jcs.260857, doi:10.1242/jcs.260857. This article has 32 citations and is from a domain leading peer-reviewed journal.
-
(vukotic2017acylglycerolkinasemutated pages 9-11): Milena Vukotic, Hendrik Nolte, Tim König, Shotaro Saita, Maria Ananjew, Marcus Krüger, Takashi Tatsuta, and Thomas Langer. Acylglycerol kinase mutated in sengers syndrome is a subunit of the tim22 protein translocase in mitochondria. Molecular cell, 67 3:471-483.e7, Aug 2017. URL: https://doi.org/10.1016/j.molcel.2017.06.013, doi:10.1016/j.molcel.2017.06.013. This article has 154 citations and is from a highest quality peer-reviewed journal.
-
(vukotic2017acylglycerolkinasemutated media 2f6e1bf4): Milena Vukotic, Hendrik Nolte, Tim König, Shotaro Saita, Maria Ananjew, Marcus Krüger, Takashi Tatsuta, and Thomas Langer. Acylglycerol kinase mutated in sengers syndrome is a subunit of the tim22 protein translocase in mitochondria. Molecular cell, 67 3:471-483.e7, Aug 2017. URL: https://doi.org/10.1016/j.molcel.2017.06.013, doi:10.1016/j.molcel.2017.06.013. This article has 154 citations and is from a highest quality peer-reviewed journal.
-
(vukotic2017acylglycerolkinasemutated pages 6-7): Milena Vukotic, Hendrik Nolte, Tim König, Shotaro Saita, Maria Ananjew, Marcus Krüger, Takashi Tatsuta, and Thomas Langer. Acylglycerol kinase mutated in sengers syndrome is a subunit of the tim22 protein translocase in mitochondria. Molecular cell, 67 3:471-483.e7, Aug 2017. URL: https://doi.org/10.1016/j.molcel.2017.06.013, doi:10.1016/j.molcel.2017.06.013. This article has 154 citations and is from a highest quality peer-reviewed journal.
-
(valpadashi2024structuralandfunctional pages 53-53): Anusha Valpadashi. Structural and functional characterization of TIM22 complex in the inner mitochondrial membarne. PhD thesis, University Goettingen Repository, 2024. URL: https://doi.org/10.53846/goediss-10678, doi:10.53846/goediss-10678.
-
(gong2026casereportsengers pages 5-6): Xingwang Gong, Yue Liu, and Hui Liang. Case report: sengers syndrome caused by a novel 7.6 kb agk deletion misdiagnosed as isolated congenital cataract. Frontiers in Pediatrics, Jan 2026. URL: https://doi.org/10.3389/fped.2026.1714952, doi:10.3389/fped.2026.1714952. This article has 0 citations.
-
(gong2026casereportsengers pages 6-6): Xingwang Gong, Yue Liu, and Hui Liang. Case report: sengers syndrome caused by a novel 7.6 kb agk deletion misdiagnosed as isolated congenital cataract. Frontiers in Pediatrics, Jan 2026. URL: https://doi.org/10.3389/fped.2026.1714952, doi:10.3389/fped.2026.1714952. This article has 0 citations.
-
(mayr2012lackofthe pages 2-4): Johannes A. Mayr, Tobias B. Haack, Elisabeth Graf, Franz A. Zimmermann, Thomas Wieland, Birgit Haberberger, Andrea Superti-Furga, Janbernd Kirschner, Beat Steinmann, Matthias R. Baumgartner, Isabella Moroni, Eleonora Lamantea, Massimo Zeviani, Richard J. Rodenburg, Jan Smeitink, Tim M. Strom, Thomas Meitinger, Wolfgang Sperl, and Holger Prokisch. Lack of the mitochondrial protein acylglycerol kinase causes sengers syndrome. American journal of human genetics, 90 2:314-20, Feb 2012. URL: https://doi.org/10.1016/j.ajhg.2011.12.005, doi:10.1016/j.ajhg.2011.12.005. This article has 261 citations and is from a highest quality peer-reviewed journal.
-
(barbosagouveia2021characterizationofa pages 1-2): Sofia Barbosa-Gouveia, Maria E. Vázquez-Mosquera, Emiliano Gonzalez-Vioque, Álvaro Hermida-Ameijeiras, Laura L. Valverde, Judith Armstrong-Moron, Maria del Carmen Fons-Estupiña, Liesbeth T. Wintjes, Antonia Kappen, Richard J. Rodenburg, and Maria L. Couce. Characterization of a novel splicing variant in acylglycerol kinase (agk) associated with fatal sengers syndrome. International Journal of Molecular Sciences, 22:13484, Dec 2021. URL: https://doi.org/10.3390/ijms222413484, doi:10.3390/ijms222413484. This article has 13 citations.
-
(mayr2012lackofthe pages 4-5): Johannes A. Mayr, Tobias B. Haack, Elisabeth Graf, Franz A. Zimmermann, Thomas Wieland, Birgit Haberberger, Andrea Superti-Furga, Janbernd Kirschner, Beat Steinmann, Matthias R. Baumgartner, Isabella Moroni, Eleonora Lamantea, Massimo Zeviani, Richard J. Rodenburg, Jan Smeitink, Tim M. Strom, Thomas Meitinger, Wolfgang Sperl, and Holger Prokisch. Lack of the mitochondrial protein acylglycerol kinase causes sengers syndrome. American journal of human genetics, 90 2:314-20, Feb 2012. URL: https://doi.org/10.1016/j.ajhg.2011.12.005, doi:10.1016/j.ajhg.2011.12.005. This article has 261 citations and is from a highest quality peer-reviewed journal.
-
(siriwardena2013mitochondrialcitratesynthase pages 2-3): Komudi Siriwardena, Nevena MacKay, Valeriy Levandovskiy, Susan Blaser, Julian Raiman, Paul F. Kantor, Cameron Ackerley, Brian H. Robinson, Andreas Schulze, and Jessie M. Cameron. Mitochondrial citrate synthase crystals: novel finding in sengers syndrome caused by acylglycerol kinase (agk) mutations. Molecular genetics and metabolism, 108 1:40-50, Jan 2013. URL: https://doi.org/10.1016/j.ymgme.2012.11.282, doi:10.1016/j.ymgme.2012.11.282. This article has 46 citations and is from a peer-reviewed journal.
-
(jiang2020theroleof pages 5-6): Haojie Jiang, Zhuo Yu, Nan Ding, Mina Yang, Lin Zhang, Xuemei Fan, Yuan Zhou, Qiang Zou, Jian Hou, Junke Zheng, Lei Zhang, Yanyan Xu, and Junling Liu. The role of agk in thrombocytopoiesis and possible therapeutic strategies. Blood, 136:119-129, Jul 2020. URL: https://doi.org/10.1182/blood.2019003851, doi:10.1182/blood.2019003851. This article has 42 citations and is from a highest quality peer-reviewed journal.
-
(mayr2012lackofthe pages 1-2): Johannes A. Mayr, Tobias B. Haack, Elisabeth Graf, Franz A. Zimmermann, Thomas Wieland, Birgit Haberberger, Andrea Superti-Furga, Janbernd Kirschner, Beat Steinmann, Matthias R. Baumgartner, Isabella Moroni, Eleonora Lamantea, Massimo Zeviani, Richard J. Rodenburg, Jan Smeitink, Tim M. Strom, Thomas Meitinger, Wolfgang Sperl, and Holger Prokisch. Lack of the mitochondrial protein acylglycerol kinase causes sengers syndrome. American journal of human genetics, 90 2:314-20, Feb 2012. URL: https://doi.org/10.1016/j.ajhg.2011.12.005, doi:10.1016/j.ajhg.2011.12.005. This article has 261 citations and is from a highest quality peer-reviewed journal.
-
(OpenTargets Search: Sengers syndrome,cardiomyopathy,cataract,mitochondrial disease,cancer-AGK): Open Targets Query (Sengers syndrome,cardiomyopathy,cataract,mitochondrial disease,cancer-AGK, 6 results). Buniello, A. et al. (2025). Open Targets Platform: facilitating therapeutic hypotheses building in drug discovery. Nucleic Acids Research.
-
(barbosagouveia2021characterizationofa pages 5-6): Sofia Barbosa-Gouveia, Maria E. Vázquez-Mosquera, Emiliano Gonzalez-Vioque, Álvaro Hermida-Ameijeiras, Laura L. Valverde, Judith Armstrong-Moron, Maria del Carmen Fons-Estupiña, Liesbeth T. Wintjes, Antonia Kappen, Richard J. Rodenburg, and Maria L. Couce. Characterization of a novel splicing variant in acylglycerol kinase (agk) associated with fatal sengers syndrome. International Journal of Molecular Sciences, 22:13484, Dec 2021. URL: https://doi.org/10.3390/ijms222413484, doi:10.3390/ijms222413484. This article has 13 citations.
-
(siriwardena2013mitochondrialcitratesynthase pages 8-10): Komudi Siriwardena, Nevena MacKay, Valeriy Levandovskiy, Susan Blaser, Julian Raiman, Paul F. Kantor, Cameron Ackerley, Brian H. Robinson, Andreas Schulze, and Jessie M. Cameron. Mitochondrial citrate synthase crystals: novel finding in sengers syndrome caused by acylglycerol kinase (agk) mutations. Molecular genetics and metabolism, 108 1:40-50, Jan 2013. URL: https://doi.org/10.1016/j.ymgme.2012.11.282, doi:10.1016/j.ymgme.2012.11.282. This article has 46 citations and is from a peer-reviewed journal.
Artifacts
Citations
- vukotic2017acylglycerolkinasemutated pages 4-6
- joshi2023mitochondrialphospholipidmetabolism pages 3-4
- vukotic2017acylglycerolkinasemutated pages 9-11
- valpadashi2024structuralandfunctional pages 53-53
- vukotic2017acylglycerolkinasemutated pages 1-4
- mayr2012lackofthe pages 2-4
- gong2026casereportsengers pages 5-6
- barbosagouveia2021characterizationofa pages 1-2
- jiang2020theroleof pages 5-6
- mayr2012lackofthe pages 1-2
- siriwardena2013mitochondrialcitratesynthase pages 2-3
- mayr2012lackofthe pages 4-5
- siriwardena2013mitochondrialcitratesynthase pages 8-8
- vukotic2017acylglycerolkinasemutated pages 6-7
- gong2026casereportsengers pages 6-6
- barbosagouveia2021characterizationofa pages 5-6
- siriwardena2013mitochondrialcitratesynthase pages 8-10
- 1-14C
- https://doi.org/10.1016/j.molcel.2017.06.013
- https://doi.org/10.1016/j.ajhg.2011.12.005
- https://doi.org/10.1091/mbc.e20-06-0390
- https://doi.org/10.1016/j.ymgme.2012.11.282
- https://doi.org/10.1242/jcs.260857
- https://platform.opentargets.org
- https://doi.org/10.3390/ijms222413484
- https://doi.org/10.53846/goediss-10678
- https://doi.org/10.1016/j.molcel.2017.06.013,
- https://doi.org/10.1091/mbc.e20-06-0390,
- https://doi.org/10.1016/j.ymgme.2012.11.282,
- https://doi.org/10.1242/jcs.260857,
- https://doi.org/10.53846/goediss-10678,
- https://doi.org/10.3389/fped.2026.1714952,
- https://doi.org/10.1016/j.ajhg.2011.12.005,
- https://doi.org/10.3390/ijms222413484,
- https://doi.org/10.1182/blood.2019003851,
OpenScientist
(AGK-hypotheses/function-hypothesis-go-0001729/openscientist.md)
OpenScientist
(AGK-hypotheses/function-hypothesis-go-0001729/openscientist.md)AGK Ceramide Kinase Activity (GO:0001729): Function Assignment Evaluation
Executive Judgment
Verdict: OVER-ANNOTATED — the annotation should be removed.
The assignment of ceramide kinase activity (GO:0001729) to human AGK (Q53H12) is not supported by direct experimental evidence and is contradicted by the available biochemical literature. The annotation propagated to human AGK via IBA (Inferred from Biological Ancestor) from PANTHER family PTHR12358, which inappropriately crosses the subfamily boundary between the true ceramide kinase CERK (subfamily SF25) and AGK (subfamily SF31). The only positive biochemical evidence for any AGK ortholog phosphorylating ceramide comes from a single study of recombinant mouse MuLK/AGK (Waggoner et al. 2004, PMID: 15252046), which itself concluded that the enzyme acts "primarily as a diacylglycerol kinase." This finding was explicitly contradicted by Van Overloop et al. 2006 (PMID: 16269826), who reported "no evidence for phosphorylation of ceramide by the recently described multiple lipid kinase." The key paper characterizing human AGK (Bektas et al. 2005, PMID: 15939762) describes only monoacylglycerol (MAG) and diacylglycerol (DAG) kinase activity, with no mention of ceramide as a substrate. Over 20 years of subsequent literature — including disease studies (Sengers syndrome), structural studies (TIM22 complex), and cancer biology — consistently describe AGK exclusively as an acylglycerol kinase and mitochondrial protein import factor, never as a ceramide kinase.
Caveats:
- Drosophila genetic evidence (PMID: 22069480) shows functional redundancy between Mulk (AGK ortholog) and Cerk in primordial germ cell migration, suggesting some in vivo functional overlap in specific developmental contexts, though this is genetic epistasis rather than direct biochemical evidence.
- The shared DAGK catalytic domain between AGK and CERK means that weak, promiscuous in vitro activity cannot be fully excluded, but this would not justify a GO function annotation.
Summary
Human acylglycerol kinase (AGK) is annotated with ceramide kinase activity (GO:0001729) based on IBA evidence from PANTHER phylogenetic inference (GO_REF:0000033). This investigation evaluated whether AGK directly possesses ceramide kinase activity by examining primary biochemical literature, domain architecture, subcellular localization, and computational sequence analysis.
The evidence overwhelmingly indicates that this annotation is erroneous. AGK belongs to a different PANTHER subfamily (SF31) than the bona fide ceramide kinase CERK (SF25), sharing only ~34% identity in their DAGKc catalytic domains. Critically, AGK lacks the PH domain and C-terminal CERK_C domain that confer ceramide substrate specificity to CERK. The single report of weak ceramide phosphorylation by mouse AGK/MuLK was contradicted by an independent laboratory's replication attempt, while the definitive characterization of human AGK found only MAG→LPA and DAG→PA kinase activities. AGK's established biological roles — mitochondrial lipid kinase, TIM22 complex subunit for mitochondrial protein import, and regulator of LPA/PA signaling — are entirely consistent with acylglycerol kinase activity and inconsistent with ceramide kinase function.
This represents a clear case of inappropriate annotation transfer across a subfamily boundary in a multi-substrate lipid kinase family. The GO:0001729 annotation should be removed from AGK, and the appropriate molecular function terms — acylglycerol kinase activity (GO:0047620) and diacylglycerol kinase activity (GO:0004143) — should be retained or strengthened.
Key Findings
Finding 1: No Direct Experimental Evidence for Human AGK Ceramide Kinase Activity
The current GO:0001729 annotation for human AGK rests entirely on non-experimental evidence codes: IBA (Inferred from Biological Ancestor, via PANTHER), ISS (Inferred from Sequence Similarity, from mouse Agk Q9ESW4), and IEA (Inferred from Electronic Annotation). Zero IDA (Inferred from Direct Assay) or EXP evidence codes support ceramide kinase activity for human AGK.
The paper that definitively characterized human AGK — Bektas et al. 2005 (PMID: 15939762) — reported: "Here, we report characterization of a novel lipid kinase, designated acylglycerol kinase (AGK), that phosphorylates monoacylglycerol and diacylglycerol to form LPA and PA, respectively." Ceramide was not reported as a substrate in this study, and this is one of the primary references cited in the GO annotation (GO_REF:0000033).
Finding 2: The Only Positive Evidence Was Contradicted by Replication Failure
The sole biochemical report of ceramide phosphorylation by any AGK ortholog came from Waggoner et al. 2004 (PMID: 15252046), who studied recombinant mouse MuLK (the mouse ortholog of AGK). Even this paper concluded that the enzyme's primary activity was as a diacylglycerol kinase: "Although its affinity for diacylglycerol and ceramide are similar, MuLK exhibits a higher V(max) toward diacylglycerol in vitro, consistent with it acting primarily as a diacylglycerol kinase."
Critically, this finding was not replicated. Van Overloop et al. 2006 (PMID: 16269826), in a study specifically characterizing ceramide kinase substrate specificity, explicitly stated: "No evidence for phosphorylation of ceramide by the recently described multiple lipid kinase was found." This direct replication failure is the strongest piece of evidence against the ceramide kinase annotation.
Finding 3: AGK Lacks the Domain Architecture Required for Ceramide Kinase Activity
Computational analysis of AGK and CERK domain architectures reveals fundamental structural differences that explain their distinct substrate specificities:
| Feature | AGK (Q53H12) | CERK (Q8TCT0) |
|---|---|---|
| DAGKc domain | Present (~34.3% identity to CERK DAGKc) | Present |
| DAGKa (accessory) domain | Present | Present |
| PH domain | Absent | Present — required for membrane targeting and ceramide access |
| CERK_C domain | Absent | Present — confers ceramide substrate specificity |
| Ca²⁺-dependent lipid binding (C2) | Absent | Present in some isoforms |
| PANTHER subfamily | SF31 | SF25 |
| Protein length | 422 aa | 537 aa |
| Subcellular localization | Mitochondrial inner membrane (TIM22 complex) | Plasma membrane / Golgi |
The PH domain and CERK_C domain are critical determinants of CERK's ability to bind and phosphorylate ceramide. Their absence in AGK is structurally consistent with the biochemical finding that AGK does not phosphorylate ceramide. The 34.3% identity between the DAGKc domains of AGK and CERK is well below the threshold that would support confident functional transfer of substrate specificity.
{{figure:agk_cerk_evidence_assessment.png|caption=Comparison of AGK and CERK domain architecture, evidence types, substrate specificity, and functional properties showing clear distinction between the two enzymes}}
Finding 4: AGK's Established Functions Are Exclusively Acylglycerol-Based
Over two decades of literature consistently describe AGK as an acylglycerol kinase with two well-characterized roles:
1. Lipid kinase activity (MAG→LPA, DAG→PA): AGK phosphorylates monoacylglycerol to lysophosphatidic acid (LPA) and diacylglycerol to phosphatidic acid (PA). This activity has been confirmed in multiple independent studies across human, mouse, and other systems (PMID: 15939762; PMID: 19112101; PMID: 22864860).
2. TIM22 complex subunit: AGK was identified as a subunit of the mitochondrial TIM22 protein translocase complex by two independent groups in 2017. Vukotic et al. (PMID: 28712726) described AGK as "a mitochondrial lipid kinase that catalyzes the phosphorylation of monoacylglycerol and diacylglycerol to lysophosphatidic acid and phosphatidic acid, respectively." Kang et al. (PMID: 28712724) independently confirmed this finding and demonstrated that AGK's role in the TIM22 complex is independent of its kinase activity, though enzymatically active AGK is required to maintain mitochondrial cristae morphogenesis.
Neither role involves ceramide metabolism. No study published since 2006 has claimed or tested ceramide kinase activity for AGK. The extensive cancer biology literature (>10 studies reviewed) consistently describes AGK as an acylglycerol kinase modulating LPA/PA signaling, JAK2 activation, NF-κB signaling, and related oncogenic pathways — never ceramide kinase activity.
Finding 5: IBA Propagation Crossed an Inappropriate Subfamily Boundary
The IBA annotation propagated from PANTHER family PTHR12358 (Sphingosine Kinase family), which contains both AGK and CERK. However, AGK and CERK reside in different subfamilies within this family (SF31 and SF25, respectively). The IBA inference inappropriately transferred a substrate-specific enzymatic activity across a subfamily boundary — a known source of over-annotation in GO. The subfamilies diverged sufficiently that their substrate specificities are distinct, as confirmed by the domain architecture differences (PH domain, CERK_C domain) and by direct biochemical evidence.
{{figure:agk_evidence_timeline.png|caption=Timeline of evidence for and against AGK ceramide kinase activity, showing the 2006 replication failure as the decisive refutation}}
Evidence Matrix
| # | Citation | Evidence Type | Direction | Claim Tested | Key Finding | Organism / Context | Confidence |
|---|---|---|---|---|---|---|---|
| 1 | PMID: 15252046 (Waggoner 2004) | Direct assay (in vitro, recombinant) | Weakly supports | MuLK/AGK phosphorylates ceramide | Recombinant MuLK phosphorylates ceramide in vitro, but Vmax is higher for DAG; concluded "primarily a diacylglycerol kinase" | Mouse (recombinant protein) | Low — higher Vmax for DAG; only one group reported this; not replicated |
| 2 | PMID: 16269826 (Van Overloop 2006) | Direct assay (replication attempt) | Refutes | MuLK/AGK has ceramide kinase activity | "No evidence for phosphorylation of ceramide by the recently described multiple lipid kinase was found" | Mammalian (CHO cells, bacterially expressed) | High — independent replication failure by CERK characterization experts |
| 3 | PMID: 15939762 (Bektas 2005) | Direct assay (human AGK) | Refutes (by omission) | AGK substrate specificity | Characterized human AGK; reports ONLY MAG→LPA and DAG→PA activity; does not mention ceramide | Human (prostate cancer cells) | High — the paper that defined human AGK function |
| 4 | PMID: 28712726 (Vukotic 2017) | Direct assay + structural | Refutes (by omission) | AGK function | AGK is a TIM22 complex subunit; lipid kinase activity = MAG→LPA, DAG→PA; no ceramide mentioned | Human | High — definitive AGK function paper |
| 5 | PMID: 28712724 (Kang 2017) | Direct assay + structural | Refutes (by omission) | AGK function | AGK "generates phosphatidic acid and lysophosphatidic acid"; TIM22 subunit with dual function | Human / yeast | High — independent confirmation |
| 6 | PMID: 22069480 (McElwain 2011) | Genetic (mutant phenotype) | Qualifies | Mulk/AGK can compensate for Cerk in C1P production | Drosophila Mulk and Cerk show genetic redundancy in PGC migration | Drosophila embryo | Medium — genetic overlap ≠ enzymatic identity; could be pathway-level redundancy |
| 7 | PMID: 18004883 (Epand 2007) | Direct assay (in vitro) | Neutral | MuLK substrate selectivity | MuLK is "the least selective" lipid kinase; tested only DAG stereospecificity, not ceramide | Mouse (recombinant) | Medium — consistent with promiscuous activity |
| 8 | PMID: 34368119 (Chu 2021) | Review | Refutes (by omission) | AGK function overview | Comprehensive AGK review: "mitochondrial lipid kinase"; no ceramide kinase activity mentioned | Human (review) | Medium — review-level evidence but covers all known AGK functions |
| 9 | GO_REF:0000033 (PAINT/IBA) | Computational (phylogenetic) | Supports (weakly) | Phylogenetic annotation transfer | IBA transferred from PANTHER family PTHR12358 | Cross-species | Low — crosses subfamily boundary (AGK=SF31, CERK=SF25) |
| 10 | Domain analysis (this study) | Computational | Refutes | AGK structural basis for ceramide kinase | AGK lacks PH domain and CERK_C domain; only 34.3% DAGKc identity with CERK | Sequence/domain comparison | High — well-established domain-function relationships |
Mechanistic Scope
Direct Gene-Product Activity of AGK
AGK has two experimentally validated functions:
1. Lipid kinase activity (primary enzymatic function):
Monoacylglycerol (MAG) --[AGK]--> Lysophosphatidic acid (LPA)
Diacylglycerol (DAG) --[AGK]--> Phosphatidic acid (PA)
This activity occurs at the mitochondrial inner membrane, where AGK is an integral component of the TIM22 protein translocase complex. Products (LPA, PA) are bioactive signaling lipids that modulate EGFR transactivation, MAPK signaling, and other pathways.
2. TIM22 complex subunit (non-enzymatic structural function):
- Assembles with TIMM22 and TIMM29
- Facilitates import of multi-spanning membrane carrier proteins (e.g., ANT1/SLC25A4) into the mitochondrial inner membrane
- This function is kinase-independent
What the Hypothesis Claims
The seed hypothesis asserts AGK has ceramide kinase activity (GO:0001729), which would mean it catalyzes:
Ceramide + ATP --[AGK]--> Ceramide-1-phosphate (C1P) + ADP
This is a distinct sphingolipid kinase reaction. The true human ceramide kinase is CERK (Q8TCT0), which:
- Is 537 aa (vs. AGK's 422 aa)
- Contains a PH domain for membrane targeting (AGK lacks this)
- Contains a CERK_C domain for ceramide-specific recognition (AGK has AGK_C instead)
- Localizes to the plasma membrane and Golgi (not mitochondria)
- Has extensive IDA evidence from multiple independent labs
Separation from Downstream Effects
Many studies describe AGK's roles in cancer progression (PMID: 39636206; PMID: 42212382), mitochondrial function (PMID: 42026151; PMID: 42391447), platelet activation (PMID: 37051931), and Sengers syndrome (PMID: 39824030; PMID: 31303091). These are downstream consequences of AGK's acylglycerol kinase activity and its structural role in the TIM22 complex — not evidence for ceramide kinase activity. Loss-of-function phenotypes reflect AGK's roles in lipid biosynthesis and protein import, not ceramide metabolism.
GO Curation Implications
Recommended Action: REMOVE GO:0001729 (ceramide kinase activity)
Lead for curator verification:
- Remove GO:0001729 (ceramide kinase activity) [IBA:GO_Central] from AGK (Q53H12).
-
Rationale: No IDA/EXP evidence for human AGK. The IBA propagated across PANTHER subfamilies (SF31 → SF25). The only positive in vitro evidence (PMID: 15252046) was contradicted by a replication failure (PMID: 16269826).
-
Remove GO:0046513 (ceramide biosynthetic process) [IBA:GO_Central] from AGK.
-
Rationale: Depends on the ceramide kinase activity annotation.
-
Remove GO:0046512 (sphingosine biosynthetic process) [IBA:GO_Central] from AGK.
-
Rationale: No evidence AGK participates in sphingosine biosynthesis; appears to be IBA carry-over from sphingosine kinase family.
-
Retain GO:0047620 (acylglycerol kinase activity) [IDA:UniProtKB].
-
Well-supported by PMID: 15939762, PMID: 28712726.
-
Retain GO:0004143 (ATP-dependent diacylglycerol kinase activity) [IDA:UniProtKB].
-
Well-supported by PMID: 15939762, PMID: 15252046.
-
Flag PANTHER/PAINT node for review — the ceramide kinase annotation propagation from the sphingosine kinase superfamily to the acylglycerol kinase subfamily may represent a systematic over-annotation affecting multiple orthologs.
-
Consider adding a NOT qualifier — if the PAINT system requires an active annotation rather than removal, consider annotating AGK with
NOT GO:0001729with a reference to PMID: 16269826.
Note on EC 2.7.1.138
UniProt lists EC 2.7.1.138 (ceramide kinase) as a catalytic activity for AGK with the qualifier "By similarity." This EC assignment should also be reviewed, as it derives from the same evidence chain.
Conflicts and Alternatives
Key Conflict: Replication Failure
The central conflict is between Waggoner et al. 2004 (PMID: 15252046), who reported in vitro ceramide phosphorylation by recombinant mouse MuLK/AGK, and Van Overloop et al. 2006 (PMID: 16269826), who explicitly could not replicate this finding. The Van Overloop paper was specifically characterizing ceramide kinase (CERK) and tested MuLK/AGK as a comparison. Their failure to detect ceramide kinase activity for MuLK/AGK is particularly significant because:
- They were specifically studying ceramide kinase enzymology and had optimized assays for this reaction
- They explicitly named MuLK in their negative result, indicating they intentionally tested it
- The study was independent of the Waggoner group
Possible Explanations for the Waggoner 2004 Positive Result
- Assay conditions: Waggoner 2004 used specific assay conditions that may have promoted promiscuous activity; Van Overloop 2006 used different substrate delivery methods.
- In vitro artifact: Recombinant protein at high concentrations may show promiscuous activity not present in vivo.
- Lipid presentation: Ceramide presentation (micelles, vesicles, protein-delivered) strongly affects apparent kinase activity. The use of non-physiological lipid delivery may have created an artifactual result.
Phylogenetic Over-annotation
AGK (PANTHER SF31: Acylglycerol Kinase) and CERK (PANTHER SF25: Ceramide Kinase) are in the same superfamily (PTHR12358) but different subfamilies. The IBA annotation propagated ceramide kinase activity across this subfamily boundary. The shared DAGK catalytic domain (34.3% identity between AGK and CERK) is the basis for family membership but does not imply shared substrate specificity. CERK has additional substrate-specificity determinants (PH domain, CERK_C domain) that AGK lacks entirely.
Drosophila Complication
McElwain et al. 2011 (PMID: 22069480) showed Drosophila Mulk (AGK ortholog) and Dcerk redundantly mediate primordial germ cell migration, a process that depends on ceramide-1-phosphate gradients. This suggests some functional overlap, but:
- This is genetic epistasis, not direct enzymatic evidence
- Pathway-level compensation (e.g., through DAG/PA signaling affecting the same downstream targets) could explain the redundancy
- Drosophila is a distant organism; even if fly Mulk has some ceramide kinase activity, this does not establish the same for human AGK
Knowledge Gaps
| Gap | What Was Checked | Why It Matters | What Would Resolve It |
|---|---|---|---|
| No direct in vitro assay of human AGK ceramide kinase activity | Literature search; UniProt evidence codes | The human protein has never been directly tested with purified ceramide substrates (Waggoner 2004 used mouse MuLK; Bektas 2005 did not test ceramide) | Purified recombinant human AGK tested against C16-ceramide and DAG substrates in parallel |
| Conflicting in vitro results not resolved | Two papers found: Waggoner 2004 (+) vs. Van Overloop 2006 (−) | Cannot determine whether weak in vitro ceramide phosphorylation is genuine or artifact | Independent lab replication under standardized conditions with proper controls |
| Drosophila genetic data not mechanistically resolved | PMID: 22069480 found and analyzed | Genetic redundancy ≠ enzymatic identity | Biochemical assay of Drosophila Mulk with ceramide substrates; C1P measurements in Mulk-only cells |
| AGK's contribution to mitochondrial ceramide/sphingolipid metabolism unknown | Literature reviewed; no mitochondrial ceramide kinase studies found for AGK | Mitochondria contain ceramides; any local AGK ceramide kinase activity would have different significance than cytoplasmic CERK activity | Lipidomics of AGK-knockout mitochondria measuring C1P levels |
| Crystal structure of AGK not available | UniProt and AlphaFold entries reviewed | Structural basis for substrate exclusion could be more precisely defined | X-ray or cryo-EM structure of AGK, ideally with substrate analog bound |
Discriminating Tests
-
Direct biochemical assay (highest priority): Express purified recombinant human AGK and test ceramide kinase activity in parallel with CERK as positive control and a catalytic-dead AGK mutant as negative control. Use physiological substrate concentrations and multiple ceramide chain lengths (C16:0, C18:0). This single experiment would definitively resolve the question.
-
Mitochondrial C1P measurement: Compare ceramide-1-phosphate levels in mitochondrial fractions from AGK-knockout vs. wild-type cells. If AGK has meaningful ceramide kinase activity in vivo, C1P should decrease in AGK-null mitochondria. Existing Sengers syndrome patient cells or AGK-KO mouse tissues could be used.
-
Substrate competition assay: Test whether ceramide competes with DAG for AGK's active site, or whether the in vitro ceramide phosphorylation is a concentration-dependent artifact of the shared DAGK catalytic fold.
-
Active-site mutagenesis: Identify CERK-specific substrate recognition residues (in PH domain and CERK_C domain) and test whether grafting these onto AGK confers ceramide specificity.
-
Drosophila cell-free assay: Test purified Drosophila Mulk for ceramide kinase activity to determine whether the genetic redundancy with Dcerk in PGC migration reflects direct enzymatic activity or pathway-level compensation.
Curation Leads
All leads below require curator verification.
Candidate Updates
- Remove GO:0001729 (ceramide kinase activity) from AGK (Q53H12)
- Action: Remove IBA annotation
- Evidence: Van Overloop 2006 (PMID: 16269826): "No evidence for phosphorylation of ceramide by the recently described multiple lipid kinase was found. The latter kinase is localized in the mitochondria, but no firm conclusions with regard to its substrate could be drawn."
-
Bektas 2005 (PMID: 15939762): "Here, we report characterization of a novel lipid kinase, designated acylglycerol kinase (AGK), that phosphorylates monoacylglycerol and diacylglycerol to form LPA and PA, respectively." — no ceramide activity reported.
-
Remove GO:0046513 (ceramide biosynthetic process) from AGK
-
Depends on ceramide kinase annotation; remove if that is removed.
-
Remove GO:0046512 (sphingosine biosynthetic process) from AGK
-
No evidence for this; appears to be IBA carry-over from sphingosine kinase family.
-
Flag PANTHER/PAINT node PTN008994514 for review
-
The ceramide kinase annotation may be inappropriately propagated to AGK subfamily members across the PTHR12358 family.
-
Consider adding a NOT qualifier
-
If the PAINT system requires an active annotation rather than removal, consider annotating AGK with
NOT GO:0001729with a reference to PMID: 16269826. -
Review EC 2.7.1.138 assignment in UniProt
- The "By similarity" EC 2.7.1.138 assignment for human AGK should be reviewed alongside the GO annotation.
Key References to Verify
| Reference | Key Quote to Verify | Relevance |
|---|---|---|
| PMID: 16269826 | "No evidence for phosphorylation of ceramide by the recently described multiple lipid kinase was found" | Direct contradiction of ceramide kinase activity for AGK/MuLK |
| PMID: 15252046 | "consistent with it acting primarily as a diacylglycerol kinase" | Original MuLK paper itself concluded AGK is primarily a DAG kinase |
| PMID: 15939762 | "phosphorylates monoacylglycerol and diacylglycerol to form LPA and PA, respectively" | Key human AGK paper: no ceramide activity reported |
| PMID: 28712726 | "catalyzes the phosphorylation of monoacylglycerol and diacylglycerol to lysophosphatidic acid and phosphatidic acid, respectively" | TIM22 complex paper: AGK described exclusively as acylglycerol kinase |
| PMID: 22069480 | "Dcerk and Dmulk redundantly mediate PGC migration" | Drosophila genetic evidence for functional overlap (not direct biochemistry) |
Evidence Base: Key Literature
Papers Supporting AGK as Acylglycerol Kinase (Not Ceramide Kinase)
Bektas et al. 2005 (PMID: 15939762) — "A novel acylglycerol kinase that produces lysophosphatidic acid modulates cross talk with EGFR in prostate cancer cells." The definitive characterization of human AGK. Reports MAG→LPA and DAG→PA activity. Ceramide is not mentioned as a substrate. This is one of the references cited in the GO annotation (GO_REF:0000033), yet it does not support ceramide kinase activity.
Van Overloop et al. 2006 (PMID: 16269826) — "Further characterization of mammalian ceramide kinase: substrate delivery and (stereo)specificity, tissue distribution, and subcellular localization studies." Explicitly tested whether AGK/MuLK phosphorylates ceramide and reported: "No evidence for phosphorylation of ceramide by the recently described multiple lipid kinase was found." This is the second reference cited in the GO annotation and directly contradicts the ceramide kinase assignment.
Vukotic et al. 2017 (PMID: 28712726) — "Sengers Syndrome-Associated Mitochondrial Acylglycerol Kinase Is a Subunit of the Human TIM22 Protein Import Complex." Identified AGK as a TIM22 complex subunit and described it exclusively as an acylglycerol kinase catalyzing MAG→LPA and DAG→PA.
Kang et al. 2017 (PMID: 28712724) — "Acylglycerol Kinase Mutated in Sengers Syndrome Is a Subunit of the TIM22 Protein Translocase in Mitochondria." Independent confirmation of AGK as TIM22 subunit. Demonstrated that AGK's kinase activity is dispensable for protein import but required for cristae morphogenesis. Only acylglycerol kinase activity discussed.
The Sole (Weak) Positive Report
Waggoner et al. 2004 (PMID: 15252046) — "MuLK, a eukaryotic multi-substrate lipid kinase." Reported that recombinant mouse MuLK/AGK can phosphorylate ceramide in vitro, but with lower Vmax than for DAG. The authors themselves concluded the enzyme acts "primarily as a diacylglycerol kinase." This is the sole basis for the ceramide kinase annotation and was not replicated.
Substrate Specificity Context
Epand et al. 2007 (PMID: 18004883) — "Substrate chirality and specificity of diacylglycerol kinases and the multisubstrate lipid kinase." Characterized MuLK as having low stereospecificity compared to DGK isoforms, consistent with a promiscuous lipid kinase that preferentially phosphorylates acylglycerols but may show weak, non-specific activity toward structurally related lipids.
CERK-Specific Literature (Establishing CERK as the True Ceramide Kinase)
Multiple recent studies confirm that CERK (not AGK) is the physiologically relevant ceramide kinase in mammalian cells:
- PMID: 38556546 — CERK/C1P in acute liver injury via NRF2 pathway
- PMID: 42285305 — CERK/C1P regulates autophagy via LC3B expression
- PMID: 40536506 — CERK suppresses ferroptosis in alcohol-associated liver disease
- PMID: 41226412 — CERK characterization in kidney proximal tubules
- PMID: 40943293 — Review of CERK/C1P role in cell growth and survival
These studies universally refer to CERK as the ceramide kinase, with no mention of AGK having this activity.
Limitations
-
No crystal structure of AGK: While domain analysis strongly argues against ceramide kinase activity, a high-resolution experimental structure with substrate analogs would provide definitive structural evidence. AlphaFold models exist but were not structurally analyzed in this iteration for active-site topology.
-
Limited in vitro testing of human AGK: The ceramide kinase assays that exist were performed with mouse MuLK, not human AGK. While the orthologs are very similar, species-specific differences cannot be completely excluded.
-
Negative evidence limitation: The Van Overloop 2006 study reported a negative result (no ceramide phosphorylation), but the specific assay conditions, enzyme preparation, and detection limits may differ from the Waggoner 2004 study. However, the consistency of the negative result with domain architecture, subcellular localization, and the entire subsequent literature strongly supports it.
-
PANTHER annotation pipeline: The exact mechanism by which the IBA annotation was propagated (family-level vs. subfamily-level) was not independently verified with the PANTHER team. The conclusion that it crossed a subfamily boundary is based on the different subfamily assignments of AGK and CERK.
Proposed Follow-up Actions
-
Immediate curation action: Submit a request to remove GO:0001729 from AGK, citing PMID: 16269826 as the primary evidence for refutation and PMID: 15939762 as evidence that the original characterization found only acylglycerol kinase activity.
-
PANTHER feedback: Report the subfamily boundary issue to the PANTHER/GO annotation pipeline to prevent similar over-annotations in the PTHR12358 family.
-
Literature annotation: If not already done, annotate AGK with IDA evidence for GO:0047620 and GO:0004143 based on PMID: 15939762.
-
Experimental validation (lower priority): A definitive in vitro kinase assay with purified human AGK testing ceramide vs. DAG/MAG substrates would close this question permanently, but the existing evidence is already sufficient for curation purposes.
Report generated: 2026-07-06
Analysis methods: UniProt API, QuickGO API, InterPro API, PubMed literature search, sequence alignment, domain architecture comparison
47 papers reviewed across 1 iteration
