A novel role for kynurenine 3-monooxygenase in mitochondrial dynamics

Abstract

The enzyme kynurenine 3-monooxygenase (KMO) operates at a critical branch-point in the kynurenine pathway (KP), the major route of tryptophan metabolism. As the KP has been implicated in the pathogenesis of several human diseases, KMO and other enzymes that control metabolic flux through the pathway are potential therapeutic targets for these disorders. While KMO is localized to the outer mitochondrial membrane in eukaryotic organisms, no mitochondrial role for KMO has been described. In this study, KMO deficient Drosophila melanogaster were investigated for mitochondrial phenotypes in vitro and in vivo. We find that a loss of function allele or RNAi knockdown of the Drosophila KMO ortholog (cinnabar) causes a range of morphological and functional alterations to mitochondria, which are independent of changes to levels of KP metabolites. Notably, cinnabar genetically interacts with the Parkinson's disease associated genes Pink1 and parkin, as well as the mitochondrial fission gene Drp1, implicating KMO in mitochondrial dynamics and mitophagy, mechanisms which govern the maintenance of a healthy mitochondrial network. Overexpression of human KMO in mammalian cells finds that KMO plays a role in the post-translational regulation of DRP1. These findings reveal a novel mitochondrial role for KMO, independent from its enzymatic role in the kynurenine pathway.

Fig 1. The kynurenine pathway (KP) in mammals and Drosophila.

The KP is the major route of tryptophan (TRP) degradation in mammals, with >95% of the essential amino acid degraded through the pathway. TRP is metabolised by IDO1/2 or TDO2 to produce L-kynurenine (L-KYN). L-KYN is metabolised through two distinct branches of the pathway; by kynurenine aminotransferases (KATs) to produce kynurenine acid (KYNA) or by KMO to produce 3-hydroxykynurenine (3-HK), 3-hydroxyanthranillic acid (3-HANA), quinolinic acid (QUIN) and ultimately nicotinamide adenosine dinucleotide (NAD+). Genes encoding enzymes in Drosophila are in red text. Flies lack homologues for 3-hydroxyanthranilic acid (3-HAO) and quinolinic acid phosphoribosyltransferase (QPRT), so the pathway is uncoupled from NAD⁺ synthesis in flies. 3-HK is converted into the brown ommochrome pigment xanthommatin by phenoxazinone synthetase (cardinal) in Drosophila.

Fig 2. Mitochondria are elongated due to cinnabar reduction in vitro and in vivo.

(A) Mitochondrial morphology is altered due to cinnabar (vii-ix) and parkin (iv-vi) dsRNA treatment in Drosophila S2 cells, compared to non-targeting control dsRNA cells (i-iii) (scale bars = 5 μm). (B) Transmission electron microscopy images from eye tissue of cn³ flies. (scale bar = 1 μm). (C) Aspect ratio (major axis length / minor axis length). (D) Feret’s diameter (distance between two most distal points). (E) Form factor (degree of branching, calculated as the inverse of circularity) (median ± 95% CI, Mann-Whitney U test, *** P < 0.001, **** P < 0.0001, n = 3, > 350 total mitochondria analyzed per group,). (F) Mitochondrial area coverage (percentage of total area analyzed covered by mitochondria. Mean ± SD, Welch’s t test, * P < 0.05, n = 3).

Fig 3. Changes in TCA cycle, OXPHOS, locomotor ability and lifespan in cn³ flies.

(A) Respiratory capacity is reduced in cn³ mutants flies. (mean ± SEM; paired t test, Holm-Sidak post hoc, ** P < 0.01, * P < 0.05, n = 8). (B) 3-HK supplementation does not rescue cn³ respirometry phenotypes (mean ± SEM; paired t test, Holm-Sidak post hoc, ns = not significant, n = 9) (C) Citrate synthase activity is increased cn³ amorphs even when supplemented with 3-HK (mean ± SEM; two-way ANOVA, Tukey post hoc, ** P < 0.01, *** P < 0.001, ns = not significant, n = 6). (D) Climbing ability is reduced in cn³ mutants (mean ± SEM; student t-test, *** P < 0.001, n = 6). (E) Longevity is recued in cn³ but enhanced in v^36f mutants (n = 100, Mantel-Cox test, * P < 0.05, **** P < 0.0001).

Fig 4. Homozygous cn LOF causes synthetic lethality in Pink1^B9 and park²⁵ flies.

(A) cn³/CyO in a w^- wild-type background were self-crossed. The expected Mendelian ratio of cn³ to cn³/CyO flies is 1:2. CyO is homozygous lethal. % expected progeny was calculated from the total number of progeny. (B) Crosses were set up to generate Pink1^B9 flies carrying the cn³ allele. The ratio of Pink1^B9 to FM6 flies is 1:1. % expected progeny was calculated as the proportion of Pink1^B9 male progeny in relation to FM6 (χ² test, 1 d.f., *** P < 0.001). (C) Expected progeny was calculated in proportion to the number of corresponding park²⁵/TM6B flies that eclosed (χ² test, 1 d.f., *** P < 0.001). (D) Penetrance male defective thorax phenotype of Day 1 Pink1^B9 or park²⁵ males combined with homozygous cn³. Penetrance in the Pink1^B9; cn³ and cn³; park²⁵ genotypes was compared to the park²⁵ stock. χ ² test, 1 d.f., * P < 0.0125, *** P < 0.00025. (E) Effect of KMO inhibitor Ro 61–8048 on defective thorax phenotype in park²⁵ and Pink1^B9 (F) flies (χ ² test, 1 d.f., * P < 0.0167, *** P < 0.0003) (G) Effect of KYNA supplementation on defective thorax phenotype in park²⁵ and Pink1^B9 (H) flies.

Fig 5. cn or hKMO overexpression rescues climbing ability in Pink1^B9 and park²⁵ flies while not affecting Marf levels.

(A) Overexpression of UAS-cn was driven by GAL4^Act5C_. Expression of cn was assessed by qPCR of cDNA from whole flies (mean ± SD, randomisation test, *** P < 0.0001). (B) Overexpression of UAS-hKMO was driven by GAL4^Act5C_. Expression of hKMO was assessed by SDS-PAGE and immunoblotting of 20 μg whole fly protein extract and ran alongside protein from HEK 293T cells transiently transfected with pcDNA3.1(KMO). (C) GAL4^Act5C –driven overexpression of UAS-cn or UAS-hKMO is sufficient to rescue the eye colour phenotype in cn³ mutant flies. (D) 3-HK levels in heads from control flies, cn³, or cn³ overexpressing cn or hKMO constructs (mean ± SD, one way ANOVA, Tukey post hoc, * P < 0.05, ** P < 0.001, *** P < 0.0001, n = 5). (E, F) cn or hKMO expression rescues Pink1 and parkin climbing phenotypes (mean ± SEM; two-way ANOVA, Tukey post hoc, *** P < 0.001, **** P < 0.0001, ns = not significant, 10 flies per n, n = 8–10). (G) Marf levels are not affected by cn overexpression. (H) β-actin normalised Marf levels are increased in Pink1^B9 flies compared to FM6 (Pink1⁺) but unchanged by cn overexpression (Mean ± SD, two-way ANOVA, Tukey post hoc, * P < 0.05, ** P < 0.01, ns = not significant, n = 5).

Fig 6. cn genetically interacts with Drp1.

(A) Overexpression of Drp1 via the FLAG-Drp1 transgene improves climbing ability in cn³ flies. The FLAG-Drp1 transgene, under control of the endogenous Drp1 promoter, was introduced into cn³ flies. Climbing ability was assessed in flies aged 7–35 days (mean ± SEM, Two-way ANOVA, Tukey post hoc, n = 5–10). (B) Overexpression of Drp1 via the FLAG-Drp1 transgene improves climbing ability in cn RNAi flies (mean ± SEM, Two way ANOVA, Tukey post hoc, * P <0.05; ** P < 0.01; *** P <0.001; **** P <0.0001. * = P value RNAi control vs cn^RNAi; GAL4^Act5C/+. † = P value RNAi control; FLAG-Drp1 vs cn^RNAi; GAL4^Act5C/FLAG-Drp1. o = P value RNAi control vs RNAi control; FLAG-Drp1. o = P value cn^RNAi; GAL4^Act5C vs cn^RNAi; GAL4^Act5C/FLAG-Drp1. n = 8–10 (10 flies per n). (C) Longitudinal sections from cn³; FLAG-Drp1 flies. (D) Form factor of traced mitochondria (median ± 95% CI, Mann-Whitney test, **** P < 0.0001, n = 4, >500 mitochondria analysed per group). (E) Feret’s diameter of traced mitochondria (median ± 95% CI, Mann-Whitney test, **** P < 0.0001, n = 4, >500 mitochondria analysed per group). (F) Aspect ratio of traced mitochondria (median ± 95% CI, Mann-Whitney test, **** P < 0.0001, n = 4, >500 mitochondria analysed per group). (G) % area coverage of traced mitochondria (mean ± SD, t test, *** P < 0.05, n = 4).

Fig 7. Mitochondrial DRP1 Ser637 phosphorylation is reduced by KMO overexpression, resulting in an increase in mitochondrial fission.

(A) Immunoblotting against DRP1 (total, pSer616 and pSer637), KMO, GAPDH and VDAC1 in cytosolic and mitochondrial fractions of HEK 293T cell transfected with pcDNA3.1 (empty of KMO), treated with DMSO or CCCP (20 μM) for 2 hrs. (B) Total cytosolic DRP1 normalised to GAPDH levels by densitometry of immunoblot bands (i). Cytosolic DRP1 pSer616 (ii) and pSer637 (iii) normalised to total cytosolic DRP1. Total mitochondrial DRP1 normalised to VDAC1 (iv). Mitochondrial DRP1 pSer616 (v) and pSer637 (vi) normalised to total mitochondrial DRP1 (mean ± SD, two-way ANOVA, Tukey post hoc, *** P < 0.001, n = 3). (C) MitoTracker Red FM staining in HEK 293T cells transfected with pcDNA3.1 vector [empty (i) or KMO (ii)] (scale bar = 10 μm). (D) Aspect ratio of mitochondria from HEK 293T cells transfected with control pcDNA3.1 or pcDNA3.1 (KMO) vectors (Mann Whitney U Test, **** P < 0.0001). (E) Form Factor of mitochondria from HEK 293T cells transfected with control pcDNA3.1 or pcDNA3.1 (KMO) vectors (Mann Whitney U Test, **** P < 0.0001).

Fig 8. KMO regulates mitochondrial dynamics by modulating DRP1 phosphorylation.

PINK1/PRKN mediated mitophagy is tightly linked to mitochondrial dynamics, via PINK1 phosphorylation of AKAP, leading to dephosphorylation of Ser637 of DRP1. Overexpression of KMO has a similar effect on DRP1 phosphorylation levels, indicating a mechanism for KMO interaction with both mitochondrial dynamics and mitophagy mechanisms.