COQ4 is required for the oxidative decarboxylation of the C1 carbon of Coenzyme Q in eukaryotic cells

Abstract

Coenzyme Q (CoQ) is a redox lipid that fulfills critical functions in cellular bioenergetics and homeostasis. CoQ is synthesized by a multi-step pathway that involves several COQ proteins. Two steps of the eukaryotic pathway, the decarboxylation and hydroxylation of position C1, have remained uncharacterized. Here, we provide evidence that these two reactions occur in a single oxidative decarboxylation step catalyzed by COQ4. We demonstrate that COQ4 complements an Escherichia coli strain deficient for C1 decarboxylation and hydroxylation and that COQ4 displays oxidative decarboxylation activity in the non-CoQ producer Corynebacterium glutamicum. Overall, our results substantiate that COQ4 contributes to CoQ biosynthesis, not only via its previously proposed structural role, but also via oxidative decarboxylation of CoQ precursors. These findings fill a major gap in the knowledge of eukaryotic CoQ biosynthesis, and shed new light on the pathophysiology of human primary CoQ deficiency due to COQ4 mutations.

Keywords: COQ4; Coenzyme Q; Coenzyme Q biosynthesis; Coenzyme Q deficiency; Oxidative decarboxylation; mitochondria; respiratory chain.

Figure 1.. COQ4-deficient human cells accumulate ₁₀P-HB.

A) The biosynthetic pathway of CoQ in eukaryotes. Proteins shared by S. cerevisiae and humans are in green, and those specific to S. cerevisiae or humans are in orange and blue, respectively. 4-hydroxybenzoic acid (4-HB) is the aromatic precursor of CoQ and the numbering of the carbon atoms used throughout this study is shown for polyprenyl-4-HB (nP-HB). Isopentenyl pyrophosphate (IPP) and farnesyl pyrophosphate (FPP) are building blocks for the synthesis of the polyprenyl pyrophosphate tail (PPP) which is added onto 4-HB by Coq2. The polyprenyl chain (n=6 for S. cerevisiae, n=10 for humans) is depicted by R on all intermediates derived from nP-HB. The uncharacterized steps for C1-decarboxylation and C1-hydroxylation are indicated with red question marks. An alternative order for the early steps is depicted by dashed arrows. The biosynthetic intermediates are represented in reduced form but they may also exist in their oxidized form as depicted for polyprenyl-1,4-benzoquinone (PBQn). B-C) Superimposition of electrochomatograms (B) and chromatograms of single ion monitoring (m/z 836.6 in positive mode) of ₁₀P-HB in lipid extracts from WT, ΔCOQ4 and ΔCOQ6 HEK293 cells (0.2 mg protein), with CoQ₈ used as standard. D) Targeted mass spectrometry quantification of CoQ₁₀ (left) and ₁₀P-HB from human primary fibroblast samples normalized to a CoQ₈ internal standard. The data are representative of biological triplicates (B-D).

Figure 2.. The D164A mutation impairs COQ4 function and metal binding, but does not compromise stability or interaction with complex Q.

A) Zn²⁺/protein molar ratio of wild type and mutant COQ4. Proteins (3 µM) were treated with 10 mM EDTA (Apo) or with 100 µM ZnCl₂ (Zn), followed by extensive dialysis against 25 mM Tris-HCl, pH 8.0, 100 mM NaCl. Zinc ion concentration was determined by atomic absorption spectrometry. **=p<0.05. B) Far UV CD spectra of human recombinant COQ4 wild type and of the D164A mutant. Proteins (0,4 mg/ml) were treated with 10 mM EDTA (Apo) or with 100 µM ZnCl₂ (Zn), followed by extensive dialysis against 25 mM Tris-HCl, pH 8.0, 100 mM NaCl. C) Targeted LC-MS-MS measurements of CoQ₁₀ and ₁₀P-HB in WT or ΔCOQ4 HEK293 expressing a GFP control empty vector, hCOQ4 WT, or hCOQ4 D164A, shown as mean ± SD (n=3). **p < 0.05, two-sided Student’s t-test. D) Affinity enrichment mass spectrometry-based interaction profile of hCOQ4-Flag WT (left) and hCOQ4-Flag D164A (right) versus GFP control in WT HEK293 cells. Results representative of biological triplicate. Significant interactors, as determined by a two sample t-test with FDR < 0.05, are shown in dark gray, and COQ proteins are highlighted in red.

Figure 3.. COQ4 complements defects in C1-decarboxylation and C1-hydroxylation in E. coli.

A,D) Plate-growth assay of serial dilutions of ∆ubiD (A) or ∆ubiD ∆ubiH (D) cells containing empty vectors (vec) or vectors expressing yeast Coq4 (yCOQ4), human COQ4 (hCOQ4) or the human COQ4 D164A allele (D164A). B,E) CoQ₈ quantification by HPLC-ECD-MS of ∆ubiD (B) or ∆ubiD ∆ubiH (E) cells containing vectors described in A, mean ± SD. ****=p<0.0001, ns = non-significant by unpaired t-test. CoQ₈ levels for WT cells are similar in both experiments. C) CoQ₈ quantification by targeted LC-MS-MS of ΔubiD cells containing vectors expressing yCOQ4 or other S. cerevisiae CoQ biosynthetic proteins. **=p<0.05, ns = non-significant. F) Labelling of CoQ₈ (percentage above the bars) from the ¹³C₇-4HB precursor, mean ± SD (n = 3 to 4).

Figure 4.. COQ4 synthesizes polyprenyl-1,4-benzoquinone in Corynebacterium glutamicum.

A) Endogenous MK_n(H₂) biosynthesis pathway in C. glutamicum (highlighted in beige) and artificial polyprenyl compounds obtained via expression of heterologous genes. Strain 405 expresses ubiA and ddsA, and produces _8–11P-HB. Additional expression of E. coli ubiD and ubiX or S. cerevisiae COQ4 (yCOQ4) yields, _8–11PP or PBQ_8–11, respectively. B) Electrochromatograms obtained by HPLC-ECD-MS analysis of lipid extracts from strain 405 containing the empty plasmid pEC-XT99A (vec) or the pEC-XT99A-yCOQ4 plasmid. The polyprenyl-1,4-benzoquinones (PBQ_8–11) synthesized specifically in the presence of yCOQ4 are labelled based on their identification by MS (see fig 4C). C-E) Single ion monitoring analyses of nonaprenyl and decaprenyl forms of PBQ_n (C), nP-HB (D), and nPP (E) in lipid extracts of strain 405 expressing yCOQ4 (pEC-XT99A-yCOQ4), UbiX and UbiD (pEC-XT99A-ubiDX), or a control plasmid (pEC-XT99A, vec). Results representative of biological triplicates (B-E).