Protein lipoylation in mitochondria requires Fe-S cluster assembly factors NFU4 and NFU5

Abstract

Plants have evolutionarily conserved NifU (NFU)-domain proteins that are targeted to plastids or mitochondria. "Plastid-type" NFU1, NFU2, and NFU3 in Arabidopsis (Arabidopsis thaliana) play a role in iron-sulfur (Fe-S) cluster assembly in this organelle, whereas the type-II NFU4 and NFU5 proteins have not been subjected to mutant studies in any plant species to determine their biological role. Here, we confirmed that NFU4 and NFU5 are targeted to the mitochondria. The proteins were constitutively produced in all parts of the plant, suggesting a housekeeping function. Double nfu4 nfu5 knockout mutants were embryonic lethal, and depletion of NFU4 and NFU5 proteins led to growth arrest of young seedlings. Biochemical analyses revealed that NFU4 and NFU5 are required for lipoylation of the H proteins of the glycine decarboxylase complex and the E2 subunits of other mitochondrial dehydrogenases, with little impact on Fe-S cluster-containing respiratory complexes or aconitase. Consequently, the Gly-to-Ser ratio was increased in mutant seedlings and early growth improved with elevated CO2 treatment. In addition, pyruvate, 2-oxoglutarate, and branched-chain amino acids accumulated in nfu4 nfu5 mutants, further supporting defects in the other three mitochondrial lipoate-dependent enzyme complexes. NFU4 and NFU5 interacted with mitochondrial lipoyl synthase (LIP1) in yeast 2-hybrid and bimolecular fluorescence complementation assays. These data indicate that NFU4 and NFU5 have a more specific function than previously thought, most likely providing Fe-S clusters to lipoyl synthase.

Figure 1

Genetic analysis of Arabidopsis mutants in NFU4 and NFU5. A, Gene models of NFU4 and NFU5 and the positions of T-DNA insertions. Black bars represent exons, gray bars are the 5′ and 3′ untranslated regions of the transcript. Triangles represent T-DNA insertions, their orientation is marked with an arrow to indicate the outward facing left border primer. The position of the T-DNA relative to the ATG start codon is indicated by the number of the nucleotide next to the left-border sequence. B, Transcript levels of NFU4 and NFU5 in leaf tissue of WT (Col-0, Col-4, or Ler) and the indicated T-DNA insertion lines, determined by RT-qPCR (graphs) or standard RT-PCR (right). For RT-qPCR, values are the average of three biological samples ± se. C, Protein blot analysis of NFU4 and NFU5 in mitochondria isolated from seedlings. Blots were labeled with antibodies against NFU4. Ponceau S stain was used to confirm equal loading and transfer. D, Decrease in NFU5 protein as a consequence of the nfu5-2 allele, quantified in the nfu4-2 mutant background. See Supplemental Figure S2 for more details of the quantification.

Figure 2

NFU4 and NFU5 proteins are abundant in all plant organs. A, Protein blot analysis of NFU4 and NFU5 in different organs of a 6-week-old Arabidopsis plant (Col-0), 20 µg protein per lane, labelled with NFU4 antibodies. Coomassie Blue staining of the gel after transfer was used as loading control. lvs, leaves. B, Specific affinity of the polyclonal antibodies raised against NFU4 and NFU5. Luminescence signals of known amounts of recombinant proteins were compared with signals in purified mitochondria from WT leaves and from cell culture of nfu4-2 and nfu5-1 mutants. Each antiserum cross-reacts with the other isoform (90% amino acid identity), but has a stronger affinity for the protein it was raised against.

Figure 3

Phenotypes of nfu4 and nfu5 single and double mutants. A, Growth phenotype of 4-week-old plants of the indicated genotype. Scale bar: 1 cm. B, Images of open siliques with immature seeds in WT (Col-0) and the indicated mutant lines. Black arrowheads point at aborted seeds. Scale bars: 0.5 mm. C, Frequency of normal and aborted embryos in nfu4 nfu5-1/+ plants. ***P < 0.0001 for 1:3 segregation ratio (χ² test). D, An aborted and healthy embryo from the silique of a nfu4-2 nfu5-1/+ plant. Plant tissue was cleared with Hoyer’s solution and imaged with DIC microscopy. Scale bars: 50 µm.

Figure 4

Analysis of nfu4 nfu5 hemizygous and double mutants. A, Protein blot analysis using protein extracts of mitochondria isolated from callus of WT (Col-0), hemizygous, and nfu4-2 nfu5-2 double mutants as indicated. Antibodies against the following proteins were used: NFU4 and NFU5; the Fe–S scaffold protein ISU1, glutaredoxin GRXS15, complex I assembly factor INDH, aconitase (ACO), the H protein subunit of the GDC, E1α subunit of PDH, and the translocase of the outer membrane TOM40. B, Growth phenotype of 4-week-old WT and nfu4-2 nfu5 plants. Scale bar: 0.5 cm. C, Blue-Native PAGE of mitochondrial complexes I, III, and V stained with Coomassie Blue (left panel) and by NADH/NBT activity staining for complex I (right panel) in the indicated plant lines.

Figure 5

Seedlings depleted of NFU4 and NFU5 proteins have a pleiotropic phenotype. A, Scheme for obtaining a mutant line expressing NFU4 under the control of the ABI3 promoter in a nfu4 nfu5 knockout background. The ABI3 promoter is active during embryogenesis but switched off after seed germination. The observed segregation numbers in T2 seedling from three independent lines were: 76 chlorotic/small and 260 WT appearance (total n = 336). B, Representative images of a WT seedling and the two classes of segregants, mutant (m) and WT-like (wtl), grown for 21 d on 0.5 MS medium in 8-h light/16-h dark cycles. Scale bars: 0.5 cm. C, PCR genotyping results of mutant (m) and WT-like (wtl) seedlings from three independent lines, showing the absence or presence of a functional NFU5 sequence using primers AM84 and AM85. D, Protein blot analysis of NFU4 and NFU5 in plants lines as in (C).

Figure 6

The nfu4 nfu5 double mutant shows decreased protein lipoylation affecting lipoyl-dependent metabolism. A, Representative images of nfu4 nfu5 double mutant segregants (m) and sibling WT-like (wtl) seedlings, grown on 0.5 MS agar plates under ambient CO₂ and 1% CO₂ in the greenhouse (8/16-h light/dark cycles, variable temperature, ∼90% humidity). Scale bars: 0.5 cm. Additional images in Supplemental Figure S7. B, Concentrations of selected free amino acids and organic acids in 3-week-old seedlings of the indicated genotypes. ACO, aconitase. Values represent the average of three to six biological replicates ± sd. *P < 0.05, **P < 0.01, ***P < 0.001 (Student’s t test, pairwise comparison to WT). See Supplemental Tables S2, S3 for complete data sets. C, Protein blot analysis for lipoyl cofactor (top) and H protein isoforms of GDC (bottom), in WT, nfu4 nfu5 (m), and wtl segregants. See also Supplemental Figure S8.

Figure 7

NFU4 and NFU5 are not required for the Fe–S enzymes aconitase and complex II. A, Respiration in intact seedlings measured using a liquid-phase oxygen electrode. WT (Col-0); m, nfu4-2 nfu5-1 mutant expressing ABI3prom:NFU4; wtl, WT-like nfu4-2 NFU5 segregants. Values represent the mean oxygen consumption per g fresh weight ± sd (n = 3). ***P < 0.001 (Student’s t test). B, Aconitase activity in total cell extracts in seedlings grown under ambient and 1% CO₂ for 3 weeks. Values represent the mean ± sd (n = 3–4). *P < 0.05 (Student’s t test). C, Complex II activity measured as electron transfer from succinate to ubiquinol (SQR) using the electron acceptor 2,6-dichloroindophenol (DCIP) in enriched mitochondrial fractions. The complex II inhibitor TTFA was added at a concentration of 0.1 mM, and only the TTFA-sensitive activity is given here. Values represent the mean SQR activity in two independent small-scale mitochondrial preparations of mutant and WT-like seedlings. Protein levels of NFU4 and NFU5 were confirmed by protein blot analysis (right panel).

Figure 8

NFU4 and NFU5 proteins interact with LIP1. A, Yeast 2-hybrid analysis to test direct interaction between NFU4/NFU5 and mitochondrial lipoyl synthase (LIP1), biotin synthase (BIO2), and the main mitochondrial aconitase (ACO2). AD-, Gal4 AD; BD-, DNA BD, both at the N-terminal position. Images were taken after 5 d and are representative of at least three independent transformations. B, BiFC to test interaction between NFU4/NFU5 and LIP1. The coding sequences were placed upstream of the N-terminal or C-terminal region of YFP, and the plasmids transformed into Arabidopsis protoplasts. Results are representative of at least two independent transfection experiments and ≥20 fluorescent cells per transformation event. Images are provided with (Figure  8, B) and without (Supplemental Figure S7) maximal Z-stack intensity projections. Scale bars: 10 μm.