Final Report: CG4860 Short-Chain Acyl-CoA Dehydrogenase Activity (GO:0016937) Hypothesis Evaluation

Executive Judgment

Verdict: Over-annotated

The GO:0016937 (short-chain fatty acyl-CoA dehydrogenase activity) annotation for Drosophila melanogaster CG4860 (Q9VGC2) is over-specific and should be generalized. While CG4860 possesses clear ACAD-family sequence features — including 57.1% identity to human ACADS, 100% conserved catalytic glutamate, and all three canonical Pfam domains — every line of evidence assigning SCAD-specific activity is computationally derived (IBA, IEA, ARBA). No direct enzymatic assay has ever been performed on CG4860. Critically, the sole experimental study directly testing this prediction (PMID: 40519079) demonstrates that only the paralog Arc42 (Q9VDT1) — not CG4860 — recapitulates the metabolic signature of SCAD deficiency in CRISPR loss-of-function mutants. Furthermore, a Leu→Thr substitution at a key substrate-specificity position in the CG4860 binding pocket, and the FlyBase ISS annotation referencing ACAD11 (a long/very-long-chain ACAD) rather than ACADS, introduce additional uncertainty about substrate specificity. The broader parent term GO:0003995 (acyl-CoA dehydrogenase activity) is well-supported and should be retained; GO:0016937 should be flagged as over-specific pending direct enzymatic characterization of CG4860's substrate preference.

Most important caveats: (1) No direct enzymatic assay of recombinant CG4860 protein has been published; the in vivo data show CG4860 does not serve as the primary SCAD but do not exclude residual or overlapping activity. (2) The acylcarnitine profile of CG4860 mutants is described in PMID: 40519079 but the abstract does not specify what profile CG4860 LOF does produce, leaving the alternative substrate preference uncharacterized. (3) GO:0016937 is broader than the common biochemical term "SCAD" — it encompasses any short-chain acyl-CoA dehydrogenase including branched-chain substrates (IVD: GO:0008470, SBCAD: GO:0003853 are children of GO:0016937). CG4860 could still correctly carry GO:0016937 if it acts on any short-chain (<6C) acyl-CoA, even non-canonical substrates.

Summary

CG4860 (UniProt: Q9VGC2) is one of two Drosophila melanogaster genes predicted to be orthologs of human ACADS, the short-chain acyl-CoA dehydrogenase. It currently carries GO:0016937 (short-chain fatty acyl-CoA dehydrogenase activity) via IBA evidence from PANTHER phylogenetic inference (GO_REF:0000033). This investigation evaluated whether that specific functional assignment is justified.

Through sequence analysis, literature review, annotation provenance tracing, and substrate-binding site comparison, we find that CG4860 is unambiguously a member of the acyl-CoA dehydrogenase family but cannot be confidently assigned SCAD-specific activity. The paralog Arc42 (Q9VDT1) has substantially higher sequence identity to human ACADS (71.1% vs 57.1%), and is the only Drosophila gene whose loss-of-function mirrors SCAD deficiency metabolically. CG4860 shows binding-pocket substitutions that may alter substrate specificity, and the only non-PAINT annotation pathway (FlyBase ISS) references ACAD11, a structurally and functionally distant long-chain ACAD. All SCAD-specific annotations on CG4860 trace back to computational pipelines with no experimental corroboration.

We recommend that curators retain the general acyl-CoA dehydrogenase activity term (GO:0003995) for CG4860, flag GO:0016937 as over-specific, and prioritize direct substrate-specificity assays to resolve CG4860's true chain-length preference.

Key Findings

Finding 1: CG4860 Has SCAD-Like Sequence Features but Lacks Functional Validation

Pairwise alignment using BLOSUM62 scoring reveals that CG4860 shares 57.1% identity and 75.5% similarity with human ACADS (P16219), confirming membership in the SCAD-type ACAD subfamily. All three canonical Pfam domains are present (Acyl-CoA_dh_N, Acyl-CoA_dh_M, Acyl-CoA_dh_1), and the catalytic glutamate within the EIYEGTSEIQ motif is 100% conserved. The ALSEPGNGSDAGAA motif, characteristic of SCAD-family enzymes, is also 100% identical between CG4860 and human ACADS.

However, CG4860 has never been subjected to direct enzymatic characterization. No published study has measured its substrate specificity, kinetic parameters, or dehydrogenase activity on any acyl-CoA substrate. The single experimental study that examines CG4860 function in vivo (PMID: 40519079) found that CG4860 CRISPR loss-of-function mutants do not produce the acylcarnitine profile expected for SCAD deficiency:

"We find that while Arc42 and CG4860 are both predicted orthologs of human ACADS, only Arc42 loss of function mirrors the acylcarnitine profile of ACADS loss of function." — PMID: 40519079

This is the most important piece of evidence in this evaluation. Sequence similarity predicts SCAD activity, but functional testing contradicts the prediction for CG4860 while confirming it for Arc42.

{{figure:evidence_comparison.png|caption=Comparison of CG4860 and Arc42 as candidate SCAD orthologs. Sequence identity heatmap and functional evidence summary showing Arc42 as the primary functional SCAD in Drosophila.}}

Finding 2: Arc42 Is the Primary Functional SCAD Ortholog in Drosophila

Arc42 (Q9VDT1) has substantially higher sequence identity to human ACADS than CG4860 does (71.1% vs 57.1%, with 84.9% similarity). More importantly, CRISPR-Cas9 loss-of-function analysis in PMID: 40519079 demonstrates that only Arc42 mutants recapitulate the acylcarnitine signature of human SCAD deficiency. Both CG4860 and Arc42 carry GO:0016937 IBA annotations, but the functional validation supports only Arc42.

This finding is significant for curation because it establishes that the PANTHER phylogenetic inference, while correctly identifying both genes as SCAD-related, does not distinguish between a gene that performs SCAD activity in vivo (Arc42) and a paralog that may have diverged in substrate specificity (CG4860).

Finding 3: PANTHER Subfamily PTHR43884:SF12 Encompasses Both SCAD and IVD

The PANTHER subfamily PTHR43884:SF12, which provides the basis for the IBA annotation, is labeled "ISOVALERYL-COA DEHYDROGENASE, MITOCHONDRIAL-RELATED" but actually contains both SCAD-type and IVD-type members: Q9VGC2 (CG4860), P16219 (human ACADS), P26440 (human IVD), Q9VSL9 (Drosophila IVD), and Q9VDT1 (Arc42). This broad grouping means the subfamily assignment alone does not distinguish SCAD from IVD specificity.

Our pairwise identity analysis confirmed that CG4860 is SCAD-type rather than IVD-type (57.1% to HsACADS vs 36.4% to HsIVD), so the PANTHER grouping is not a source of misannotation per se — but it does illustrate that the subfamily is broader than its name implies, and curators should not rely on the subfamily label alone.

Finding 4: Substrate-Binding Pocket Substitutions Distinguish CG4860 from SCAD

Detailed comparison of the substrate-binding cavity residues revealed a critical Leu→Thr substitution at the RIGIA+8 position in CG4860. This position lines the substrate-binding pocket and influences chain-length preference:

• CG4860: Thr (polar, small side chain)
• Human ACADS: Leu (hydrophobic, bulky)
• Arc42: Leu (hydrophobic, conserved with ACADS)

Additional divergences include:
- RIGIA+5: CG4860 has Ala (unique) vs Ser (HsACADS) / Gly (Arc42)
- Upstream pocket: CG4860 has SLDCG vs TLDMG (HsACADS) / TLDAG (Arc42)

The Leu→Thr change at a binding-pocket position is structurally significant. The smaller, polar threonine could accommodate longer or differently branched substrates, potentially shifting CG4860's substrate preference away from strict short-chain specificity. This structural evidence independently supports the hypothesis that CG4860 may not be a canonical SCAD.

Finding 5: FlyBase ISS Annotation References ACAD11, Not ACADS

The FlyBase ISS annotation for CG4860 (GO:0003995, acyl-CoA dehydrogenase activity, PMID: 22758915) uses ACAD11 (UniProtKB:Q709F0) as its "with/from" reference. ACAD11 is a human long-chain/very-long-chain ACAD (780 amino acids, optimal substrate C22:0 docosanoyl-CoA) with IDA evidence for GO:0004466 (long-chain-acyl-CoA dehydrogenase activity), GO:0070991, and GO:0017099. ACAD11 belongs to a different PANTHER family (PTHR48083) than CG4860 (PTHR43884) and lacks the SCAD-specific conserved motifs.

The source paper (PMID: 22758915) is a peroxisomal proteome inventory:

"We have analyzed the proteome of Drosophila to identify the proteins involved in peroxisomal biogenesis and homeostasis as well as metabolic enzymes that function within the organelle." — PMID: 22758915

This means the FlyBase ISS annotation actually connects CG4860 to peroxisomal metabolism via similarity to a long-chain ACAD — a functionally different context from mitochondrial short-chain fatty acid oxidation. Importantly, FlyBase correctly assigned only the general term GO:0003995, not the specific GO:0016937, suggesting curatorial caution even before the recent experimental evidence.

{{figure:go_hierarchy_evidence.png|caption=GO term hierarchy and evidence flow diagram showing all annotation paths to CG4860's GO:0016937 assignment. All paths are computational; the sole experimental evidence challenges the assignment.}}

Finding 6: All SCAD-Specific Annotations Are Computationally Derived

A complete provenance trace of CG4860's GO:0016937 annotation reveals that every contributing evidence line is computational:

1. GO:0016937 IBA — PANTHER phylogenetic inference (GO_REF:0000033), with/from P16219 (HsACADS) + Q3ZBF6 (bovine ACADS)
2. GO:0016937 IEA — EC:1.3.8.1 mapping (GO_REF:0000003)
3. EC:1.3.8.1 — assigned by ARBA automatic rule (ECO:0000256, ARBA00012046)
4. GO:0003995 ISS — FlyBase, with/from Q709F0 (HsACAD11, a long-chain ACAD)

No evidence code is experimental (IDA, IMP, IEP, or IGI). The only experimental paper directly examining CG4860 function (PMID: 40519079) challenges rather than supports the SCAD assignment.

{{figure:go_decision_summary.png|caption=GO annotation decision summary showing the complete provenance chain, verdict, and curation recommendations for CG4860 GO:0016937. All annotations are computationally derived, and the sole experimental study challenges the SCAD-specific assignment.}}

Evidence Matrix

GO Curation Implications

Current Annotation Status

• GO:0016937 (short-chain fatty acyl-CoA dehydrogenase activity) — IBA (GO_Central) + IEA (UniProt) — Under review
• GO:0003995 (acyl-CoA dehydrogenase activity) — ISS (FlyBase, with/from Q709F0/ACAD11) + IEA (UniProt) — Well supported
• GO:0006635 (fatty acid beta-oxidation) — ISS (FlyBase) — Reasonable but not specifically validated
• GO:0050660 (FAD binding) — IEA (InterPro) — Well supported by domain analysis
• GO:0016627 (oxidoreductase activity, acting on CH-CH) — IEA (InterPro) — Well supported

GO Hierarchy Context (Critical for Interpretation)

GO:0016937 (short-chain fatty acyl-CoA dehydrogenase activity) is broader than the common biochemical term "SCAD." In the GO hierarchy:
- GO:0016937 is a child of GO:0003995 (acyl-CoA dehydrogenase activity)
- GO:0016937 is the parent of GO:0008470 (IVD/3-methylbutanoyl-CoA DH activity) and GO:0003853 (SBCAD activity)
- GO:0016937 encompasses activity on any short-chain acyl-CoA substrate (<6 carbons), including branched-chain substrates like isovaleryl-CoA and 2-methylbutyryl-CoA

This means that if CG4860 acts on any short-chain acyl-CoA (straight or branched), GO:0016937 would be technically correct. The annotation would only be definitively wrong if CG4860 acts exclusively on medium- or long-chain substrates. However, the lack of positive evidence for any short-chain activity, combined with the negative LOF data, makes the annotation unsupported.

Annotation Provenance (All Computational — No Experimental Evidence)

All SCAD-specific annotations on CG4860 are computationally derived:
- GO:0016937 IBA: PAINT phylogenetic inference (GO_REF:0000033), with/from P16219 (HsACADS)
- GO:0016937 IEA: EC:1.3.8.1 mapping (GO_REF:0000003); EC itself assigned by ARBA automatic rule (ECO:0000256)
- GO:0003995 ISS: FlyBase (PMID: 22758915), with/from Q709F0 (HsACAD11, a long-chain ACAD)
- EC:1.3.8.1: ARBA automatic rule (ARBA00012046) — not from experimental enzyme assay

The sole experimental paper mentioning CG4860 (PMID: 40519079) provides counter-evidence against SCAD function.

Recommended Curation Actions (Leads Requiring Curator Verification)

Lead 1 (High priority): Generalize GO:0016937 to GO:0003995

Consider removing or NOT-qualifying GO:0016937 (short-chain fatty acyl-CoA dehydrogenase activity) for CG4860; retain GO:0003995 (acyl-CoA dehydrogenase activity). The IBA-level annotation is challenged by experimental evidence (PMID: 40519079) showing CG4860 does not serve the SCAD function in vivo. If retained, the annotation should be flagged as potentially over-specific and not elevated above IBA evidence without direct enzymatic data.

Lead 2 (Moderate priority): Review the FlyBase ISS reference protein

Verify whether the FlyBase ISS annotation (GO:0003995, with/from ACAD11 Q709F0) is appropriate given that ACAD11 is phylogenetically distant and functionally distinct from the SCAD subfamily. ACAD11 is in PANTHER family PTHR48083, while CG4860 is in PTHR43884.

Lead 3 (Moderate priority): Flag GO:0046359 (butyrate catabolic process) for review

If CG4860 does not function as a SCAD, it likely does not participate in butyrate catabolism. This IBA annotation follows from the GO:0016937 assignment and should be reviewed concurrently.

Lead 4 (Lower priority): Strengthen Arc42 SCAD annotation

Consider adding experimental evidence-supported annotation for Arc42 (Q9VDT1) GO:0016937, citing PMID: 40519079. Arc42 LOF mirrors human ACADS deficiency acylcarnitine profile — this constitutes mutant phenotype evidence (IMP) supporting the SCAD function.

Lead 5 (Lower priority): Review EC:1.3.8.1 assignment

EC:1.3.8.1 was assigned by ARBA automatic rule, not experiment. This feeds the IEA GO:0016937 annotation. If the specific SCAD GO term is flagged, the EC assignment should also be reconsidered to prevent re-propagation of over-specific terms.

GO Term Decision Table

Mechanistic Scope

Direct Gene Product Activity

CG4860 encodes a 415-amino acid protein with all three canonical acyl-CoA dehydrogenase domains (N-terminal IPR013786, middle IPR006091, C-terminal IPR009075), FAD as cofactor, and a perfectly conserved catalytic glutamate. The protein is predicted to catalyze the alpha,beta-dehydrogenation of an acyl-CoA thioester substrate, transferring electrons to electron transfer flavoprotein (ETF). This is a direct enzymatic activity — the general ACAD function is well-supported.

Uncertainty in Substrate Specificity

What remains uncertain is the chain-length and branching preference of CG4860's substrate. Three lines of evidence suggest it may differ from canonical short-chain (C4-C6) specificity:

1. Binding-pocket divergence: The Leu→Thr substitution at RIGIA+8 alters the hydrophobic character of the substrate-binding cavity, potentially accommodating longer or branched substrates
2. No SCAD-deficient phenotype on LOF: CG4860 CRISPR mutants do not accumulate the short-chain acylcarnitines expected in SCAD deficiency
3. FlyBase cross-reference to ACAD11: The ISS annotation references a very-long-chain ACAD, suggesting computational methods detect broader ACAD-family similarity

Separation from Downstream Phenotypes

It is important to distinguish the molecular function question (what substrate does CG4860 dehydrogenate?) from downstream biological consequences. The loss-of-function phenotype in PMID: 40519079 is informative because it uses a direct metabolic readout (acylcarnitine profiling) rather than a distal developmental or behavioral phenotype. The absence of a SCAD-deficient acylcarnitine signature in CG4860 mutants is therefore strong evidence against SCAD-specific activity, not merely a failure to observe a downstream consequence.

ACAD Family Substrate Hierarchy in Drosophila:

  C2-C6 (short-chain) ──── SCAD ──── Arc42 (confirmed, PMID:40519079)
                                      CG4860 (?? — NOT confirmed as SCAD)
  C6-C12 (medium-chain) ── MCAD ──── CG3902/Mcad (Drosophila)
  C12-C20 (long/very-long) LCAD/VLCAD ── distinct ACAD families
  Branched-chain ────────── IVD/IBDH/SBCADH ── CG4860 may overlap here?

  All share: FAD cofactor, ETF electron transfer, ACAD fold
  Differ in: substrate binding pocket → chain-length selectivity

Mechanistic Model / Interpretation

The evidence supports a model of paralog functional divergence following gene duplication in the Drosophila lineage. CG4860 and Arc42 both descended from an ancestral SCAD-type ACAD but have diverged in substrate specificity. Arc42 retained the canonical SCAD function, as demonstrated by its higher sequence identity to human ACADS (71.1%), conservation of substrate-pocket residues (Leu at RIGIA+8), and functional validation through CRISPR LOF metabolomics.

CG4860, while retaining the overall ACAD architecture and catalytic machinery, has accumulated binding-pocket substitutions — particularly Leu→Thr at RIGIA+8 — that may shift its substrate preference. The direction of this shift is unknown but could include:

1. Branched-chain short acyl-CoAs (e.g., isobutyryl-CoA, 2-methylbutyryl-CoA) — consistent with the PANTHER subfamily's inclusion of IVD members and with documented cross-reactivity between BCAA ACAD pathways (PMID: 31133529)
2. Medium-chain acyl-CoAs — the smaller Thr residue could enlarge the pocket to accommodate longer substrates
3. A unique or partially overlapping substrate range — CG4860 may retain low-level SCAD activity but primarily function on alternative substrates

The FlyBase ISS connection to ACAD11 (a peroxisomal long-chain ACAD) via PMID: 22758915 raises the additional possibility of subcellular compartment divergence — if CG4860 localizes to peroxisomes rather than mitochondria, its functional context would be fundamentally different from mitochondrial SCAD. However, this remains speculative without localization data.

Evidence Base

Primary Literature

PMID: 40519079 — Characterizing fatty acid oxidation genes in Drosophila

This is the most important paper for this evaluation. Using CRISPR-Cas9 to generate loss-of-function alleles for both Arc42 and CG4860, the authors performed acylcarnitine profiling and found that only Arc42 LOF recapitulates the metabolic signature of human SCAD deficiency. This directly challenges the GO:0016937 annotation on CG4860. The verified snippet: "We find that while Arc42 and CG4860 are both predicted orthologs of human ACADS, only Arc42 loss of function mirrors the acylcarnitine profile of ACADS loss of function."

PMID: 41898520 — Structure and Substrate Specificity of Human Short-Chain Acyl-CoA Dehydrogenase and Insights into Pathogenicity of Disease-Associated Mutations

Provides structural context for understanding which binding-pocket residues determine short-chain specificity in SCAD. The substrate-binding cavity residues identified in human ACADS served as the basis for our comparison showing CG4860's divergence at key positions.

PMID: 22758915 — An inventory of peroxisomal proteins and pathways in Drosophila melanogaster

The source paper for the FlyBase ISS annotation connecting CG4860 to ACAD11. This study cataloged peroxisomal proteins in Drosophila and identified CG4860 as similar to ACAD11, a long-chain peroxisomal ACAD — a very different functional context from mitochondrial SCAD activity. The verified snippet: "We have analyzed the proteome of Drosophila to identify the proteins involved in peroxisomal biogenesis and homeostasis as well as metabolic enzymes that function within the organelle."

PMID: 31133529 — Metabolic analysis reveals evidence for branched chain amino acid catabolism crosstalk and the potential for improved treatment of organic acidurias

Demonstrates that ACAD substrate specificity is not absolute in mammalian systems. IBDH and SBCADH show cross-pathway activity in branched-chain amino acid catabolism, raising the possibility that CG4860 could function in BCAA catabolism rather than (or in addition to) fatty acid oxidation.

PMID: 29563254 — Phosphorylation of MCAD selectively rescues PINK1 deficiencies in behavior and metabolism

Establishes that specific ACAD family members have distinct, non-redundant roles in Drosophila mitochondrial metabolism, supporting the principle that ACAD paralogs should not be assumed to share identical substrate specificity.

PMID: 17304052 — Development of a newborn screening follow-up algorithm for the diagnosis of isobutyryl-CoA dehydrogenase deficiency

Provides context on the overlap between SCAD and IBDH deficiency biomarkers, particularly C4-acylcarnitine, which represents both butyrylcarnitine and isobutyrylcarnitine and is relevant for interpreting acylcarnitine profiling results.

Conflicts and Alternatives

Central Conflict: Sequence vs Function

The central conflict is between sequence-based annotation (supporting SCAD activity) and in vivo functional evidence (failing to support SCAD as the primary role):

• Sequence says: CG4860 is a SCAD-subfamily member (57.1% identity to human ACADS, conserved catalytic glutamate, SCAD-specific motifs)
• Genetics says: CG4860 LOF does not mirror SCAD deficiency; Arc42 LOF does

Alternative Interpretations

1. Functional redundancy: CG4860 could have SCAD activity but be expressed in different tissues or at lower levels than Arc42. Loss of CG4860 alone may not produce a detectable phenotype if Arc42 compensates. Bgee expression data show partially overlapping but distinct expression: CG4860 is highest in Malpighian tubule (+ 56 tissues), while Arc42 is highest in enteroendocrine cells (+ 77 tissues).

2. Subfunctionalization after duplication: CG4860 and Arc42 may have arisen from a gene duplication and partitioned the ancestral function. CG4860 may act on slightly different substrates (e.g., C4-branched or C5 substrates).

3. Neofunctionalization: CG4860 may have evolved a new substrate preference. The Leu→Thr change and other pocket substitutions could shift specificity toward branched-chain, odd-chain, or other acyl-CoA substrates.

4. Paralog over-annotation risk: The PAINT/IBA annotation system annotates based on phylogenetic position. When two paralogs both fall in the SCAD clade, both receive GO:0016937. This is a known limitation of phylogenetic annotation: it cannot distinguish which paralog retains the ancestral function after duplication and divergence.

5. ISS reference discrepancy: FlyBase's ISS annotation for CG4860 uses ACAD11 (Q709F0) as the reference — a long-chain/very-long-chain ACAD in a different PANTHER family (PTHR48083 vs PTHR43884). This likely reflects broad ACAD-family similarity used in the peroxisome proteome inventory, not a substrate-specificity-level assignment. Notably, FlyBase correctly assigned only the general term GO:0003995, not GO:0016937.

Limitations and Knowledge Gaps

Gap 1: No Direct Enzymatic Assay on CG4860

What was checked: Literature searches (PubMed), UniProt, FlyBase, QuickGO evidence codes. No published enzymatic assay found.

Why it matters: The fundamental question — what substrates does CG4860 dehydrogenate? — cannot be answered without measuring enzyme kinetics on purified CG4860 protein with a panel of acyl-CoA substrates.

What would resolve it: Recombinant expression, purification, and kinetic characterization (Km, Vmax, kcat) with butyryl-CoA (C4), hexanoyl-CoA (C6), octanoyl-CoA (C8), isobutyryl-CoA, 2-methylbutyryl-CoA, and isovaleryl-CoA.

Gap 2: CG4860 LOF Metabolic Profile Not Fully Characterized

What was checked: Abstract and citation information from PMID: 40519079. The abstract states CG4860 LOF does not mirror SCAD deficiency but does not specify what metabolic signature CG4860 mutants do produce.

Why it matters: The mutant profile could reveal the actual substrate preference of CG4860.

What would resolve it: Full-text review of PMID: 40519079, particularly supplementary acylcarnitine data for CG4860 mutants.

Gap 3: CG4860 Subcellular Localization Unknown

What was checked: FlyBase ISS references peroxisomal study (PMID: 22758915); ACAD family members are typically mitochondrial; N-terminal targeting peptide predicted for mitochondria.

Why it matters: If CG4860 localizes to peroxisomes rather than mitochondria, its substrate specificity and biological role would differ substantially from mitochondrial SCAD.

What would resolve it: Fluorescent protein tagging or immunolocalization of CG4860 in Drosophila tissues.

Gap 4: Structural Prediction of Altered Substrate Preference

What was checked: Sequence-level binding-pocket residue comparison against human ACADS crystal structure positions.

Why it matters: AlphaFold model comparison with substrate docking could predict whether the Leu→Thr substitution meaningfully alters cavity volume and substrate accommodation.

What would resolve it: AlphaFold structure comparison (AF-Q9VGC2-F1) with human ACADS crystal structure, computational docking with different acyl-CoA substrates.

Gap 5: Double Mutant (CG4860 + Arc42) Phenotype Unknown

What was checked: Only single-mutant data available from PMID: 40519079.

Why it matters: If CG4860 has partially redundant SCAD activity, it would only manifest when Arc42 is also absent.

What would resolve it: CG4860/Arc42 double-mutant acylcarnitine profiling.

Proposed Follow-up Experiments / Discriminating Tests

Test 1: Recombinant CG4860 Enzyme Kinetics (Highest Priority — Definitive)

Express and purify CG4860 protein, measure Km and kcat for a substrate panel: butyryl-CoA (C4), hexanoyl-CoA (C6), octanoyl-CoA (C8), isobutyryl-CoA, 2-methylbutyryl-CoA, isovaleryl-CoA, and palmitoyl-CoA (C16). Compare kinetic parameters to those of Arc42 and human ACADS. This is the gold-standard test that would definitively resolve the GO:0016937 annotation question.

Test 2: Branched-Chain Acylcarnitine Profiling of CG4860 Mutants

Analyze CG4860 CRISPR mutant flies (from PMID: 40519079) for branched-chain acylcarnitines (C4-isobutyryl, C5-isovaleryl, C5-2-methylbutyryl). Accumulation of any of these would suggest CG4860 functions in BCAA catabolism rather than fatty acid beta-oxidation.

Test 3: CG4860/Arc42 Double Mutant Analysis

Generate double mutants and profile acylcarnitines. If C4:0-acylcarnitine accumulates beyond Arc42-single-mutant levels, CG4860 has partial SCAD activity. If not, CG4860 is functionally distinct from SCAD.

Test 4: CG4860 Subcellular Localization

Tag CG4860 with GFP or mCherry at the endogenous locus and co-stain with mitochondrial (MitoTracker) and peroxisomal (anti-SKL/PTS1) markers. Peroxisomal localization would fundamentally redefine the annotation framework.

Test 5: Structural Comparison via AlphaFold

Compare CG4860 AlphaFold model (AF-Q9VGC2-F1) to human ACADS crystal structure with substrate docked. Measure binding-pocket volume differences caused by the Leu→Thr substitution. This computational analysis could predict altered substrate preference without wet-lab experiments.

Curation Leads

Lead 1: Generalize GO:0016937 to GO:0003995 (High Priority)

• Action: Remove or NOT-qualify GO:0016937 (short-chain fatty acyl-CoA dehydrogenase activity) for CG4860; retain GO:0003995 (acyl-CoA dehydrogenase activity)
• Rationale: All SCAD-specific evidence is computational; sole experimental study contradicts SCAD-specific function; general ACAD activity well-supported by domain architecture
• Candidate reference to verify: PMID: 40519079 — snippet: "We find that while Arc42 and CG4860 are both predicted orthologs of human ACADS, only Arc42 loss of function mirrors the acylcarnitine profile of ACADS loss of function."

Lead 2: Review FlyBase ISS Reference Protein (Moderate Priority)

• Action: Verify whether the FlyBase ISS annotation (GO:0003995, with/from ACAD11 Q709F0) is appropriate given ACAD11's distant phylogenetic and functional relationship
• Reference to verify: PMID: 22758915 — check whether CG4860 was identified as peroxisomal in this proteomics study

Lead 3: Flag GO:0046359 (Butyrate Catabolic Process) (Moderate Priority)

• Action: Consider removing butyrate catabolism annotation for CG4860
• Rationale: If CG4860 does not function as a SCAD, it likely does not participate in butyrate catabolism. This IBA annotation follows from the GO:0016937 assignment.

Lead 4: Strengthen Arc42 SCAD Annotation (Lower Priority)

• Action: Consider adding IMP-level annotation for Arc42 (Q9VDT1) GO:0016937, citing PMID: 40519079
• Rationale: Arc42 LOF mirrors human ACADS deficiency acylcarnitine profile — mutant phenotype evidence supporting SCAD function

Lead 5: Review EC:1.3.8.1 Assignment (Lower Priority)

• Action: EC:1.3.8.1 was assigned by ARBA automatic rule, not experiment. If the SCAD GO term is flagged, the EC assignment should also be reconsidered to prevent re-propagation
• Rationale: The EC number drives automatic GO annotation through IEA mapping

Lead 6: Consider NOT Annotation for GO:0016937

• Action: If full data from PMID: 40519079 robustly shows no SCAD-deficient phenotype in CG4860 mutants, consider a NOT annotation for GO:0016937 with IMP evidence
• Caveat: Appropriate only if authors specifically tested and excluded C4:0-acylcarnitine accumulation; partial redundancy with Arc42 cannot be excluded without double-mutant data

Lead 7: Candidate Alternative GO Terms

If future experiments reveal CG4860's actual substrate preference, consider:
- GO:0003995 — acyl-CoA dehydrogenase activity (general; currently recommended)
- GO:0008470 — isovaleryl-CoA dehydrogenase activity (if isovaleryl-CoA is preferred)
- GO:0003853 — 2-methylacyl-CoA dehydrogenase activity (if branched-chain short substrates)
- A new child term of GO:0003995 if CG4860 has a unique substrate profile

Investigation Timeline

Artifacts

• OpenScientist final report
• OpenScientist final report
• OpenScientist evidence comparison
  
• OpenScientist go decision summary
  
• OpenScientist go hierarchy evidence
  
• OpenScientist plot 1
  
• OpenScientist plot 2
  
• OpenScientist plot 3