Cysteine desulfurase is regulated by phosphorylation of Nfs1 in yeast mitochondria

Abstract

The cysteine desulfurase Nfs1/Isd11 uses the amino acid cysteine as the substrate and its activity is absolutely required for contributing persulfide sulfur to the essential process of iron-sulfur (Fe-S) cluster assembly in mitochondria. Here we describe a novel regulatory process involving phosphorylation of Nfs1 in mitochondria. Phosphorylation enhanced cysteine desulfurase activity, while dephosphorylation decreased its activity. Nfs1 phosphopeptides were identified, and the corresponding phosphosite mutants showed impaired persulfide formation. Nfs1 pull down from mitochondria recovered an associated kinase activity, and Yck2, a kinase present in the pull down, was able to phosphorylate Nfs1 in vitro and stimulate cysteine desulfurase activity. Yck2 exhibited an eclipsed distribution in the mitochondrial matrix, although other cellular localizations have been previously described. Mitochondria lacking the Yck2 protein kinase (∆yck2) showed less phosphorylating activity for Nfs1. Compared with wild-type mitochondria, ∆yck2 mitochondria revealed slower persulfide formation on Nfs1 consistent with a role of Yck2 in regulating mitochondrial cysteine desulfurase activity. We propose that Nfs1 phosphorylation may provide a means of rapid adaptation to increased metabolic demand for sulfur and Fe-S clusters within mitochondria.

Keywords: Iron–sulfur; Kinase; Mitochondria; Persulfide; Phosphorylation.

Fig. 1. Mitochondria phosphorylate Nfs1

(A) Nfs1 phosphorylation in [γ-³²P] ATP kinase assay. Nfs1/Isd11 alone (lane 1), lysate alone (lane 2) or various amounts of Nfs1/Isd11 plus lysate (lanes 3–5), were tested in a kinase assay. For specificity control, 75 pmol BSA was used as a substrate (lane 6). The metal dependency of the activity was tested by addition of 20 mM EDTA to one sample (lane 7). Yeast mitochondrial lysate (7.5 μg), substrates Nfs1/Isd11 or BSA, and [γ-³²P] ATP were incubated for 15 min at 30 degrees followed by recovery of proteins by TCA precipitation. The proteins were separated by SDS-PAGE, subjected to Coomassie blue staining (left panel) and autoradiography (right panel). (B) Nfs1 phosphorylation depends on the amount of added mitochondrial lysate. Recombinant Nfs1/Isd11 (10 pmol) was subjected to kinase assays in the presence of different amounts of mitochondrial lysates 15 μg (lane 1), 1.5 μg (lane 2), 0.75 μg (lane 3), 0.15 μg (lane 4). (C) Inactive Nfs1 is an equivalently good substrate for phosphorylation. Nfs1/Isd11 (lanes 2 and 3), Nfs1_C421A/Isd11 (lanes 4 and 5) or Nfs1 alone (lanes 6 and 7) were added to WT mitochondrial lysate in the presence of [γ-³²P] ATP. ³²P autoradiogram (upper panel) and Coomassie blue staining (lower panel) are shown. (D) Human protein complex hNFS1/hISD11 is a good substrate for phosphorylation. Recombinant hNFS1_His6/hISD11 was expressed and purified from E. coli, and (10 pmol) was subjected to kinase assay in the presence or absence of 7.5 μg of yeast WT mitochondrial lysate (YPH499). The autoradiogram is shown in the upper panel and protein stain in the lower panel. The mobility of recombinant hNFS1 is marked.

Fig. 2. Identification of Nfs1 phosphorylation sites by mass spectrometry

(A) Nfs1 phosphopeptides identified by LC-MS/MS in three independent experiments. Experiment 1. Nfs1_His6/Isd11 expressed in E. coli was purified, digested with trypsin, and analyzed by mass spectrometry for phosphopeptides. Experiment 2. Nfs1_His6 inserted into yeast mitochondria was purified, similarly processed and analyzed for phosphopeptides. Experiment 3. Recombinant Nfs1_His6/Isd11 was purified from E. coli, exposed to ATP and a kinase active fraction from mitochondria. Phosphophopeptides are shown with black boxes enclosing conserved phosphorylation sites including S₃₃₄ and T₃₃₆ and the confirmed phosphorylation site on T₁₉₅. Clear boxes indicate potential phosphorylation sites. (B) Alignment of Nfs1 phosphorylation sites. Yeast Nfs1 (P25374) was aligned with human NFS1 (Q9Y697) and E. coli IscS (P0A6B7) using Clustal Omega, conserved phosphorylation sites (T₁₉₅, S₃₃₄, T₃₃₆) are highlighted by black boxes. (C) Monomer structure of yeast Nfs1 modeled on the crystal structure available from the human homolog (PDB: 5USR) and using the I-TASSER modeling program. The PLP co-factor is represented in yellow, and the phosphorylation sites (T₁₉₅, S₃₃₄, T₃₃₆) are highlighted in purple.

Fig. 3. Dephosphorylation of Nfs1/Isd11 with protein phosphatase decreases cysteine desulfurase activity

Nfs1/Isd11 sulfide production was measured by the methylene blue method in the presence of DTT [6, 58]. Recombinant Nfs1-His₆ (0.2 nmol) copurified with Isd11 was incubated in the absence (bar 1) or presence (bar 2) of 100 U calf intestinal alkaline phosphatase (AP). Recombinant Nfs1_C421A/Isd11 (0.2 nmol) was similarly treated without (bar 3) or with (bar 4) AP. Buffer controls without (bar 5) or with AP (bar 6) were tested. Nfs1/Isd11 in isolation buffer 50 mM Hepes/KOH pH 7.5, 150 mM NaCl (no AP) was tested (bar 7). Sulfide production determined from a standard curve was expressed in nmol of sulfide produced per minute. Error bars represent standard deviations.

Fig. 4. Phenotypes of the Nfs1 phosphomutants

(A) Alanine substitutions were introduced into a yeast expression plasmid carrying Nfs1 (T₁₉₅A, S₃₃₄A, T₃₃₆A, S₃₃₄A/T₃₃₆A) and the mutated plasmids were transformed into a promoter-swap yeast strain in which Nfs1 was placed under control of the carbon sensitive GAL1/10 promoter. Serial 1:5 dilutions of the transformants were plated on agar plates with various non-inducing carbon sources: from left to right, glucose, glucose plus 0.5 mM H₂O₂, glycerol, or ethanol. Plates were photographed after 3 days of growth at 30 °C. (B) Ferric reductase activity from the phosphomutant strains (nmol Fe²⁺/10⁶ cells/h). (C) Ferrous iron uptake activity from the phosphomutant strains (pmol/10⁶ cells/hour). Δaft1 and nfs1-14 strains were included as controls. (D) Cysteine desulfurase activity. The set of Nfs1 phosphomutants (T₁₉₅A, S₃₃₄A, T₃₃₆A, S₃₃₄A/T₃₃₆A) was expressed as the Nfs1_His6/Isd11 complex in E. coli, purified, and assayed for cysteine desulfurase activity using the methylene blue assay. (For graphs B, C, & D error bars indicate standard deviation, and asterisks indicate statistical difference of each mutant versus WT, calculated using Student’s t-test.) (E) The same set of phosphomutants was tested for Nfs1 persulfide forming activity by incubation with ³⁵S-cysteine, TCA precipitation, and non-reducing SDS-PAGE. The upper gel shows the ³⁵S autoradiogram; the lower gel shows the protein stain. The radiolabeled Nfs1 persulfide is indicated as Nfs1-S-³⁵SH, and the Nfs1 protein is indicated as Nfs1. (F) Aconitase loading. Intact mitochondria were labeled with ³⁵S-cysteine for the indicated times at 30 °C in the presence of iron (10 μM ferrous ascorbate) and nucleotides (4 mM ATP, 1 mM GTP, and 4 mM NADH). Following sonication and native gel separation, the gel was exposed to film. Lanes 1 and 2, WT mitochondria; Lanes 3 and 4, T₁₉₅A mutant; Lanes 5 and 6, S₃₃₄A mutant; Lanes 7 and 8, T₃₃₆A mutant; Lanes 9 and 10, S₃₃₄A/T₃₃₆A mutant.

Fig. 5. Purification of mitochondrial kinase activity

(A) Nfs1_His6 complexes. A yeast strain was engineered to overexpress Nfs1_His6. Mitochondria were purified, and proteins were passed over a Ni-NTA agarose column. Left panel shows the protein staining by Coomassie brilliant blue of the elution fraction (1 μg); the middle panel shows the ³²P autoradiogram of a kinase assay containing 0.1 μg of the eluate; the right panel indicates the immunodecorated sample with the indicated antibodies. (B) The Nfs1_His6 eluate fraction was subjected to mass spectrometry analysis. Total MS/MS spectra count is shown for selected proteins including the casein kinase homolog Yck2. (C) Yfh1_His6 complexes A yeast strain was engineered to overexpress Yfh1_His6. Mitochondria were isolated and proteins were passed over a Ni-NTA agarose column, eluting four major components involved in the assembly of Fe-S clusters: Nfs1-Isd11-Isu1-Yfh1. Left panel shows the protein staining by Coomassie brilliant blue of the eluate fraction (10 μg) and the right panel shows the immunoblot with α-Nfs1, α-Yfh1, α-Isu1, and α-Isd11 antibodies. (D) The μg) was radiolabeled with [γ-³²P] ATP in an in vitro kinase Yfh1_His6 eluate fraction (5 reaction and separated by SDS-PAGE. Proteins were transferred to a membrane and immunoblotted with α-Nfs1 antibody (lane 2). (E) The Yfh1_His6 eluate fraction was subjected to mass spectrometry analysis. Total MS/MS spectra count is shown for selected proteins including the casein kinase homolog Yck2.

Fig. 6. Yck2 is a candidate mitochondrial kinase able to phosphorylate Nfs1

(A) Specific phosphorylation of Nfs1/Isd11 by recombinant Yck2. Yck2 (ΔN45, lacking the first 45 amino acids of the open reading frame), was expressed and purified from E. coli and 1.5 pmol was used in a phosphorylation assay by adding [γ-³²P] ATP, recovering the protein(s), separating by SDS-PAGE and exposing to film. The autoradiogram is shown. Lane 1, Yck2 alone; Lane 2, Nfs1/Isd11 alone; Lane 3, Yfh1 alone; Lane 4, Yck2 and Nfs1/Isd11; Lane 5, Yck2 and Yfh1. (B) Phosphorylation of Nfs1/Isd11 with recombinant Yck2 increases desulfurase activity. Recombinant Yck2 was used to phosphorylate Nfs1/Isd11 (10 pmol) with [γ-³²P] ATP. Proteins were recovered by TCA precipitation and separated by SDS-PAGE (upper panel, ³²P Auto). In a parallel experiment, unlabeled ATP was used for phosphorylation and ³⁵S-cysteine was used to label the persulfide on Nfs1 (middle panel, ³⁵S Auto). The polyacrylamide gel was fixed and stained for protein (lower panel, Stain). Lane 1, Yck2 alone; Lane 2, Nfs1/Isd11 alone; Lanes 3–6, increasing amounts of Yck2 from 0.012 to 1.5 pmol added to 10 pmol Nfs1/Isd11; Lane 7, 1.5 pmol Yck2 added to Nfs1/Isd11 omitting ATP; Lane 8, Yck2 added to active site mutant Nfs1_C421A/Isd11.

Fig. 7. Persulfide formation on endogenous Nfs1 in WT or Δyck2 mitochondria

(A) Intact mitochondria were depleted of nucleotides and incubated with ³⁵S-cysteine for 15 minutes at 30 °C prior to separation on a non-reducing SDS-PAGE. Mitochondria from WT (1x =100 μg, 2x=200μg) or Δyck2 were examined. (B) Time course at lowered temperatures. WT or Δyck2 mitochondria were labeled with ³⁵S-cysteine for 5, 10 or 15 min at 20 °C (upper panel) or on ice (lower panel). Nfs1-S-³⁵SH indicates the radiolabeled Nfs1 persulfide. The rate of formation of Nfs1 persulfide was decreased Δyck2 mitochondria at lowered temperature.

Fig. 8. Yck2 is found in the mitochondrial matrix fraction and is able to phosphorylate Nfs1

(A) Soluble fraction isolation. Mitochondria from the Yck2-TAP strain were subjected to freeze, thaw, sonication in 20 mM Hepes/KOH pH 7.5, 150 mM NaCl three times. The lysed mitochondria were centrifuged at 100,000 x g and the supernatant was recovered and analyzed by immunoblotting. Yck2-TAP was found in both the pellet and supernatant fractions. (B) Carbonate extraction. Mitochondria from the Yck2-TAP strain were exposed to carbonate buffer pH 11 for 10 min followed by centrifugation at 100,000 x g for 1 h, and the supernatant and pellet were separated by SDS-PAGE and developed with antibodies to Yck2-TAP, Nfs1, and Mir1. The Yck2-TAP signal was present in both soluble and pellet fractions. (C) Protease protection. Mitochondria were exposed to increasing concentrations of trypsin (0, 20, 40, 100 μg/ml) in isotonic (0.6 M sorbitol), hypotonic (0.12 M sorbitol), or detergent containing (1% Triton X-100) buffers. Proteins were recovered and subjected to SDS-PAGE and immunoblotting with the indicated antibodies. Yck2-TAP was partially protected in isotonic or hypotonic conditions but digested in the lysed mitochondria.

Fig. 9. Nfs1 phosphorylation by soluble matrix fraction

WT or Δyck2 mitochondria were subjected to freeze, thaw, sonication as above, and the soluble fraction (30 μg) was used in a phosphorylation assay in the absence (lanes 1, 2) or presence (lanes 3, 4) of added Nfs1-His₆/Isd11. These samples were passed over Ni-NTA to retrieve the phosphorylated Nfs1-His₆ from WT or Δyck2 (lanes 10, 11). Control samples were Nfs1-His₆ and Yck2-His₆ used in direct phosphorylation (lane 5) or Ni pull-down following phosphorylation (lane 12).