Vitamin K2 Is a Mitochondrial Electron Carrier That Rescues Pink1 Deficiency

See allHide authors and affiliations

Science  08 Jun 2012:
Vol. 336, Issue 6086, pp. 1306-1310
DOI: 10.1126/science.1218632

Keeping Mitochondria in the Pink

Pink1 is a mitochondrial kinase, and loss of Pink1 function in flies and mice results in the accumulation of inefficient mitochondria. In a screen for modifiers of the Parkinson-associated gene, pink1, Vos et al. (p. 1306, published online 10 May; see the Perspective by Bhalerao and Clandinin) identified the fruit fly homolog of UBIAD1, “Heix.” UBIAD1 was localized in mitochondria and was able to convert vitamin K1 into vitamin K2/menaquinone (MK-n, n the number of prenylgroups). In bacteria, vitamin K2/MK-n acts as an electron carrier in the membrane and, similarly, in Drosophila, mitochondrial vitamin K2 appeared to act as an electron carrier to facilitate adenosine triphosphate production. Fruit flies that lack heix showed severe mitochondrial defects that could be rescued by administering vitamin K2.


Human UBIAD1 localizes to mitochondria and converts vitamin K1 to vitamin K2. Vitamin K2 is best known as a cofactor in blood coagulation, but in bacteria it is a membrane-bound electron carrier. Whether vitamin K2 exerts a similar carrier function in eukaryotic cells is unknown. We identified Drosophila UBIAD1/Heix as a modifier of pink1, a gene mutated in Parkinson’s disease that affects mitochondrial function. We found that vitamin K2 was necessary and sufficient to transfer electrons in Drosophila mitochondria. Heix mutants showed severe mitochondrial defects that were rescued by vitamin K2, and, similar to ubiquinone, vitamin K2 transferred electrons in Drosophila mitochondria, resulting in more efficient adenosine triphosphate (ATP) production. Thus, mitochondrial dysfunction was rescued by vitamin K2 that serves as a mitochondrial electron carrier, helping to maintain normal ATP production.

Parkinson’s disease (PD) is a common neurodegenerative disorder, and genetic causes of the disease allow us to elucidate the molecular pathways involved (1, 2). Mutations in pink1, encoding an evolutionarily conserved mitochondrial kinase, cause PD in humans and mitochondrial defects in model organisms (36). To understand Pink1 function in vivo, we performed a genetic modifier screen in Drosophila. Because PD affects the nervous system we screened 193 chemically induced recessive lethal mutants that were selected for defects in neuro-communication (79). We tested dominant modification of pink1B9 null mutant flight defects (fig. S1A). Although none of the chemically induced mutants showed dominant flight defects when crossed to a wild-type pink1RV allele, 24 mutants suppressed and 32 enhanced the pink1B9 flight defect, such that pink1B9 flies failed to fly (fig. S1A).

To reveal the mechanism by which the modifiers affected Pink1, we mapped one of the strongest enhancers that, in combination with pink1B9, results in enhanced lethality to heixuedian (heix). We named this allele heix2 and identified several additional heix alleles (fig. S1, B to E) (10). To test whether loss of heix specifically exacerbated pink1 phenotypes, we assessed flight, adenosine triphosphate (ATP) levels, and neuronal mitochondrial membrane potential (Ψm) (10). Heterozygosity for heix combined with pink1RV did not display defects, whereas it strongly enhanced defects in pink1B9 null mutants that are rescued by overexpressing Heix (Fig. 1, A to C, and fig. S2). Conversely, overexpression of Heix rescued pink1B9-associated defects and those in parkin1/Δ21, which also encodes a PD-associated gene affecting mitochondria (Fig. 1, D and E) (4, 5). Although Heix overexpression does not fully rescue flight, suggesting Pink1 acts in different pathways, the effect is similar in magnitude to previously identified suppressors (4, 5, 11). Thus, Heix is a dosage-sensitive modifier of pink1.

Fig. 1

Identification of heix as modifier of pink1B9. Enhancement of pink1B9 phenotypes by heterozygosity for heix (pink) compared with pink1RV controls heterozygous for heix (black) and rescue with UAS-heix (k11403/+;da>heix); flight (n > 50) (A), ATP levels (n = 10 assays) (B), and Ψm measurements using the potentiometric dye JC-1. Red aggregate/green monomeric fluorescence ratio measured at third instar neuromuscular boutons; aggregates accumulate at negative Ψm (n = 20 synapses) (C). Suppression of pink1B9 and park1/Δ21 phenotypes by overexpression of Heix (da>heix); flight (n > 50) (D), ATP levels (n = 10 assays) (E), and JC-1 red/green fluorescence ratio (n = 20 synapses) (F). Data are percentage [(A) and (D)] and mean normalized to control [(B), (C), (E), and (F)]. Error bars indicate SEM. Analysis of variance (ANOVA)/Dunnett: *P < 0.05; **P < 0.01; ***P < 0.001.

Heix is evolutionary well conserved (fig. S3A), and its bacterial and human homologs, MenA and UBIAD1, harbor a prenyltransferase domain involved in synthesis of vitamin K2/menaquinone (MK-n, n indicating the number of prenyl groups) (fig. S3A) (12, 13). In a dendrogram, the three proteins cluster closely, suggesting the enzymes are functionally conserved from bacteria to human (fig. S3B). Feeding heix mutants during development with MK-4 but not with menadione, a synthetic precursor that is prenylated to form vitamin K2, significantly improved their survival rate (Fig. 2, A and B) and other cellular defects (below), whereas ubiquinone (Q-10) feeding did not (Fig. 2A), indicating specificity (10). Thus, similar to UBIAD1 and MenA, Heix produces vitamin K2 in Drosophila. We thus tested whether supplementing pink1B9 with vitamin K2 alleviated its phenotypes. Because Escherichia coli produces vitamin K2 (MK-8) (14), we supplemented minimal medium with living E. coli and then placed 1-day-old adult control and pink1B9 mutant flies on this medium. A significant improvement in flight and ATP levels was observed in pink1B9 (Fig. 3, A to C). In contrast, pink1B9 flies on medium with menA mutant E. coli that barely produce vitamin K2 (fig. S4A) (12) were not rescued (Fig. 3, A to C), whereas pink1B9 on medium with menA mutant E. coli complemented with a wild-type menA gene (fig. S4A) showed a significant improvement in their flight and ATP defects (Fig. 3, A to C). Thus, vitamin K2 produced in E. coli alleviates pink1-related defects.

Fig. 2

Heix is involved in vitamin K2 production. (A) Improved survival of first instar heix mutants placed on MK-4, on ubiquinone (Q10), or on menadione (black) shown as the difference in the percentage of control larvae that survive on control medium (n = 10 experiments, each with 10 larvae); that is, about 20% more heix mutant larvae survive with vitamin K2 compared with heix mutants on control medium. Error bars, SEM. ANOVA Dunnett: *P < 0.05; **P < 0.01. ns, not significant. (B) Schematic of the biochemical conversion of menadione to vitamin K2.

Fig. 3

Vitamin K2 rescues pink1B9 mutant phenotypes. Flying ability of pink1RV (A) and pink1B9 (B) and ATP levels in pink1RV and pink1B9 flies (C) that were placed on control medium (black), on medium with E. coli that produce vitamin K2 (MK-8, yellow), or on medium with menA mutant E. coli that do not produce vitamin K2 (green) or with menA mutant E. coli transformed with wild-type menA that produce vitamin K2 (blue) (fig. S4). Flying ability (D and E), ATP levels (F), Ψm at neuromuscular boutons determined by using JC-1 red/green ratio (G), and mitochondrial morphology in third instar larval muscles (H to J) in control pink1RV and in pink1B9 on MK-4 containing medium (gray) or control medium (black) [(D) to (I)] or upon overexpression of Heix (da>heix) (J). The amount of rounded mitochondria is quantified (H). Arrows are normal mitochondrial structure in (I) and (J). Error bars, SEM. n > 50 flies, and data are percentages in (A), (B), (D), and (E); 10 assays in (C) and (F); 20 synapses or muscles in (G) to (J); and mean normalized to control in (C) and (F) to (H). ANOVA/Dunnett: *P < 0.05; **P < 0.01; ***P < 0.001.

Next, we placed 1-day-old adult pink1B9 mutants and controls on MK-4–supplemented or on control medium. Control flies flew normally and did not show altered ATP levels or Ψm after feeding with various concentrations of MK-4 (100 to 1000 μM) (Fig. 3, D, F, and G, and fig. S4, B and C). Pink1B9 on MK-4 showed a dose- and time-dependent improvement of flight, increased ATP amount, and a more negative Ψm (Fig. 3, E to G, and fig. S4, B and C). Similarly, park1/Δ21 mutants showed improved flight, increased ATP levels, and a more negative Ψm when placed on MK-4 compared with park1/Δ21 mutants on control medium (fig. S5, A to D). Thus, MK-4 acutely rescued PD-gene-associated mitochondrial and systemic defects in flies.

Larval and adult pink1B9 and parkin1/Δ21 muscles show enlarged, clumped mitochondria, a defect that is modified by mitochondrial remodeling (15). This mitochondrial morphology defect was partially rescued by placing pink1B9 or park1/Δ21 on MK-4 (Fig. 3, H and I, and figs. S6 and S5, E to I) (10) and by overexpressing Heix (Fig. 3J and fig. S5J), indicating that this rescue of mitochondrial morphology was because of a heix-dependent vitamin K2 deficit. We believe that this effect of vitamin K2 is the result of improved mitochondrial function and not because of an important role for vitamin K2 in mitochondrial remodeling, because feeding MK-4 to drp11/2 or pink1B9;drp1/+ mutants that are defective in mitochondrial fission did not rescue their mitochondrial morphological defects (fig. S7, A and B). Conversely, mitochondrial morphological defects observed in animals expressing RNA interference (RNAi) to electron transport chain (ETC) complex I subunits were alleviated when placed on MK-4 (fig. S7C). Thus, morphological defects that arise because of functional defects in mitochondria can be rescued by vitamin K2.

In bacteria, vitamin K2 transfers electrons in the ETC, establishing a proton motive force across the membrane (16). Could vitamin K2 contribute to electron transport in the ETC in mitochondria? Heix-His expressed in fly cells was present in mitochondria (Fig. 4, A and B) (10), similar to the human ortholog UBIAD1 (17), and vitamin K2 has previously been detected in mitochondrial fractions (18, 19). We tested the consequence of loss of heix on mitochondrial function (10) and found that, compared with controls, heix mutants showed a less negative Ψm and lower ATP levels that could be significantly rescued by rearing heix mutants on MK-4 or by reexpression of Heix (Fig. 4C and fig. S8, A and B). Similarly, severe Ψm defects in pink1B9;heix/+ were rescued almost to control levels when placed on MK-4 or reexpression of Heix (fig. S8, C and D), and intact mitochondria purified from heix mutants showed a significantly reduced capacity to produce ATP in vitro (Fig. 4D). Thus, MK-4 in heix mutants is necessary to maintain a negative Ψm and to produce ATP.

Fig. 4

Vitamin K2 acts as a mitochondrial electron carrier. Immunolabeling of S2 Drosophila cells (A) and Western blot of fractionations (B) transfected with Heix-His and labeled with anti-His and anti–ATP synthase β (A). (C) ATP levels of heix1/Df and heix/Df;da>heix (n = 8). (D) Time-dependent ATP production in mitochondria isolated from heix1/Df mutant larvae (n = 8). (E) Measurements of complex II activity [normalized to citrate synthase activity (CS)] using a dose-response of MK-4 as electron carrier (n = 4). (F) Measurements of complex II activity using ubiquinone (0.13 mM; Q-10, light gray) or MK-4 (0.35 mM, gray) as electron carrier (n = 4). (G) Time-dependent ATP production in mitochondria isolated from pink1B9 mutant flies. Reactions are either not supplemented with MK-4 (black) or supplemented with 0.15 mM (dark gray) or 0.35 mM MK-4 (light gray) (n = 8). (H) Adenosine diphosphate–stimulated complex I–driven respiration rate (oxygen consumption) in mitochondria isolated from pink1RV controls and pink1B9 mutants placed for 72 hours on MK-4 medium or on control medium. Mitochondria from vitamin K2-fed mutant flies consume oxygen faster. Data are mean normalized to controls in (C) and (H) or mean in (D) to (G). Error bars, SEM. ANOVA/Dunnett: *P < 0.05, **P < 0.01, ***P < 0.001.

To determine whether vitamin K2 is sufficient to facilitate electron transport in mitochondria, we prepared mitochondrial fractions and measured reduction of an artificial electron acceptor, 2,6-dichlorophenolindophenol (DCPIP), downstream of complex II/succinate dehydrogenase (EC In this reaction, succinate is the electron donor and Q-10 the electron carrier (fig. S9) (20). When we added MK-4 rather than Q-10, we found that it was also effective at reducing DCPIP, and increasing concentrations of MK-4 resulted in more efficient electron transport (Fig. 4E) (10). Compared with MK-4, Q-10 was more effective at reducing DCPIP (Fig. 4F); however, when we placed 1-day-old adult pink1B9 animals on Q-10 medium, we observed a systemic and mitochondrial rescue that did not exceed that obtained when mutants were reared on MK-4 (fig. S10 and Fig. 3). Direct application of MK-4 to mitochondria purified from pink1B9 significantly facilitated ATP production (Fig. 4G). Furthermore, time-dependent oxygen consumption of intact mitochondria prepared from pink1B9 mutants was increased when the flies were first placed on MK-4 (Fig. 4H) (10). Thus, vitamin K2 is sufficient for electron transport downstream of a eukaryotic ETC complex, resulting in improved mitochondrial oxygen consumption and energy production, particularly in pink1B9 mutants.

Our data predict that vitamin K2 may also alleviate the defects in other conditions that impair mitochondrial function. We thus placed flies that express RNAi to different complex I components, sbo mutants that produce less ubiquinone, and rotenone-treated animals on control medium or on MK-4 medium and assessed Ψm and ATP levels. MK-4 rescued the mitochondrial defects in all of these flies (fig. S11, A and B). These results thus support a role for vitamin K2 in the transport of electrons in eukaryotic mitochondria to produce ATP, similar to its role in prokaryotic membranes (16), suggesting that vitamin K2 serves a conserved function in prokaryotes and mitochondria. Because mitochondria are involved in aging (21), we reared aged flies on MK-4 or overexpressed Heix but did not observe a significant rescue in mobility (fig. S11C). We surmise that long-lasting compensatory changes may be more important in these situations.

Human UBIAD1 mutations may also affect mitochondrial function. Electron microscopic analyses of corneal samples from Schnyder’s crystalline corneal dystrophy patients who harbor mutations in the UBIAD1 gene (2224) indicate cystic swelling of mitochondria, but the nature for this defect is unknown (25). Heix/UBIAD1 produces vitamin K2, and in our studies vitamin K2 rescued mitochondrial defects in numerous conditions that affect mitochondrial function. Vitamin K2 was even effective at improving systemic locomotion defects in fully developed adult pink1 and parkin mutant flies. Vitamin K2 did not affect mitochondrial remodeling directly, but, by increasing ETC efficiency, it contributed to the proton motif force that facilitates ATP production, similar to ubiquinone (26). Vitamin K2 may thus constitute a promising compound to treat mitochondrial pathology, also in PD patients suffering from Pink1 or Parkin deficiency.

Supplementary Materials

Materials and Methods

Figs. S1 to S11

References (2744)

References and Notes

  1. Materials and methods are available as supplementary materials on Science Online.
  2. Acknowledgments: We thank H. Bellen, J. Park, J. Chung, the Bloomington Drosophila stock center, and the Yale E. coli stock center. Support was provided by an Agentschap voor Innovatie door Wetenschap en Technologie (IWT)–Vlaanderen to M.V., a European Research Council Starting Grant (no. 260678), Fonds Voor Wetenschappelijk Onderzoek (G074709-G095511-G094011), the Research Fund KU Leuven, a Methusalem grant, KU Leuven, the Hercules and Francqui foundations, and VIB. The data reported in this paper are tabulated in the main paper and in the supplementary materials. B.D.S. is a paid consultant for Janssen Pharmaceutica (Johnson and Johnson) with regard to Alzheimer’s disease and Parkinson’s disease.
View Abstract

Stay Connected to Science

Navigate This Article