MMS19 Links Cytoplasmic Iron-Sulfur Cluster Assembly to DNA Metabolism

See allHide authors and affiliations

Science  13 Jul 2012:
Vol. 337, Issue 6091, pp. 243-245
DOI: 10.1126/science.1219664

MMS19 Joins the CIA

Iron-sulfur (Fe-S) proteins play a critical role in cell metabolism and particularly in DNA repair and replication. Mutants in eukaryotic gene MMS19 are particularly sensitive to DNA damaging agents, suggesting that it is involved in DNA repair, but the mutations can also have other wide-ranging effects on the cell (see the Perspective by Gottschling). Now, Stehling et al. (p. 195, published online 7 June) and Gari et al. (p. 243, published online 7 June) show that in both yeast and humans, MMS19 functions as part of the cytosolic Fe-S protein assembly (CIA) machinery. The MMS19 is part of a specialized CIA targeting complex that plays a role late in cytosolic Fe-S protein assembly to direct Fe-S cluster transfer from the CIA scaffold complex to a subset of Fe-S proteins, including a number associated with DNA metabolism.


The function of many DNA metabolism proteins depends on their ability to coordinate an iron-sulfur (Fe-S) cluster. Biogenesis of Fe-S proteins is a multistep process that takes place in mitochondria and the cytoplasm, but how it is linked to nuclear Fe-S proteins is not known. Here, we demonstrate that MMS19 forms a complex with the cytoplasmic Fe-S assembly (CIA) proteins CIAO1, IOP1, and MIP18. Cytoplasmic MMS19 also binds to multiple nuclear Fe-S proteins involved in DNA metabolism. In the absence of MMS19, a failure to transfer Fe-S clusters to target proteins is associated with Fe-S protein instability and preimplantation death of mice in which Mms19 has been knocked out. We propose that MMS19 functions as a platform to facilitate Fe-S cluster transfer to proteins critical for DNA replication and repair.

MMS19 is a highly conserved HEAT repeat protein that was first identified in Saccharomyces cerevisiae as being required for the removal of ultraviolet radiation (UV)–induced pyrimidine dimers (1). mms19Δ cells are impaired for both nucleotide excision repair (NER) and RNA polymerase II transcription, which is complemented in vitro by addition of the NER/transcription complex TFIIH but not by purified Mms19 (2). Consistently, mms19Δ cells display reduced protein levels of the TFIIH component Rad3, and overexpression of Rad3 can restore NER proficiency, suggesting a function for Mms19 in stabilizing Rad3 rather than a direct role in DNA metabolism (3). Human MMS19 also interacts with two components of TFIIH: XPB and the Rad3 homolog XPD (4). Moreover, MMS19 and XPD interact independently of TFIIH and were proposed to play a role in mitotic spindle formation and chromosome segregation (5). Schizosaccharomyces pombe Mms19 is part of a different complex composed of Rik1, Dos1, and the catalytic subunit of DNA polymerase ε, which links DNA replication to heterochromatin silencing (6). A molecular explanation of how MMS19 affects so many different processes has remained elusive.

We found MMS19 as an interaction partner of the Regulator of Telomere Length protein RTEL1. RTEL1 is a helicase implicated in telomere-length regulation and the maintenance of genomic stability (7, 8); it acts as an anti-recombinase to counteract toxic recombination and limit crossing over during meiosis (7, 9). RTEL1 belongs to the same family of Rad3-like SF2 helicases as XPD, the Fanconi Anemia protein J (FANCJ), and the Warsaw Breakage Syndrome protein ChlR1, which all bind an iron-sulfur (Fe-S) cluster (10).

Mass spectrometry (MS) analysis identified MMS19 as the most abundant interaction partner of RTEL1 (fig. S1A). This interaction was confirmed by means of reciprocal co-immunoprecipitation experiments (Fig. 1A) and was confined to the cytoplasm, principally because the vast majority of MMS19 is cytoplasmic (Fig. 1B and fig. S1B).

Fig. 1

Interaction network of MMS19. (A) Reciprocal co-immunoprecipitations of RTEL1 and MMS19 overexpressed in HEK 293T cells. (B) Fractionation of HeLa cells stably overexpressing RTEL1Flag into cytoplasm (CP), nucleoplasm (NP), and chromatin (Chr) and immunoprecipitation with M2-Flag beads. (C) Schematic of experiment. SYPRO ruby (Molecular Probes, Eugene, OR)–stained gel showing FlagMMS19-containing (Flag19) and control (mock) complexes. Asterisk indicates FlagMMS19. (D) Immunoprecipitation of FlagMMS19 (Flag19) stably overexpressed in HEK 293 cells and verification of endogenous interaction partners by means of Western blot.

To gain insight into the function of MMS19, we purified cytoplasmic MMS19-containing complexes and identified additional interaction partners by using MS and co-immunoprecipitation (Fig. 1, C and D). The cytoplasmic MMS19 complex contained many other nuclear genome stability factors, including XPD, FANCJ, DNA polymerase δ, DNA primase (Pri2), and DNA2, all of which are known to coordinate an Fe-S cluster (1013). In total, we identified 12 known Fe-S proteins in the cytoplasmic MMS19 complex (fig. S1C). This complex also contained CIAO1, IOP1 (also known as NARFL), and MIP18 (also known as FAM96B) (Fig. 1D and fig. S1C), three members of the cytoplasmic Fe-S assembly (CIA) machinery (14, 15). Co-immunoprecipitation and gel filtration studies confirmed that MMS19 forms a stable complex with MIP18, CIAO1, and IOP1 (Fig. 2, A and B, and fig. S2, A and B). Furthermore, small interfering RNA (siRNA)–mediated depletion of MMS19 resulted in a dramatic down-regulation of the CIA component MIP18, suggesting that MIP18 is unstable in the absence of MMS19 (Fig. 2C). Taken together, these data raised the possibility that MMS19 may physically link the CIA machinery to target Fe-S cluster proteins.

Fig. 2

MMS19 is part of the CIA machinery. (A) Co-immunoprecipitations of FlagMMS19 and members of the CIA machinery overexpressed in HEK 293T cells. (B) Superdex 200 10/300 GL (GE Healthcare, Uppsala, Sweden) gel filtration analysis of complexes containing FlagMIP18, MMS19, and CIAO1 overexpressed in Sf9 insect cells. In [Superdex 200 input], eluate of Flag-immunoprecipitation. (C) Effect of MMS19 depletion on MIP18 and CIAO1. 19, siRNA MMS19; con, siRNA control. (D) Incorporation of 55Fe into target proteins measured by means of liquid scintillation counting. Galactose-driven promoters replace the endogenous promoters of Mms19 (left) and Cia1 (right); expression was switched on (+ galactose) or switched off (+ glucose) for 40 hours. Western blots show immunoprecipitated Fe-S proteins and depletion efficiency of Mms19 and Cia1 in WCEs. Counts were normalized to levels of immunoprecipitated proteins. Error bars indicate SDs from three independent experiments.

To address whether MMS19 could play a direct role in Fe-S cluster biogenesis, we used a well-established assay to study Fe-S metabolism in Saccharomyces cerevisiae, which can be grown in Fe-free conditions and subsequently pulse-labeled with radioactive 55Fe (16). Known Fe-S proteins were immunoprecipitated in the presence or absence of Mms19, and 55Fe incorporation was measured by means of liquid scintillation counting. In the absence of Mms19, incorporation of 55Fe into target proteins was reduced between three- and 10-fold (Fig. 2D and fig. S2C), which was comparable with loss of Cia1, the yeast homolog of CIAO1 [Fig. 2D; compare left with right and (17)]. These data indicate that Mms19 is indeed required for Fe-S cluster biogenesis.

We next tested whether human XPD could incorporate 55Fe when overexpressed in yeast. This was indeed the case, and 55Fe incorporation was reduced when Mms19 was depleted (Fig. 3A). However, the interpretation of this result was complicated by the fact that considerably less GFPXPD was immunoprecipitated in the absence of Mms19, possibly because Fe-S cluster transfer by Mms19 is required for protein folding and stability of XPD. This instability is reminiscent of the reduced levels of Rad3 observed in mms19Δ yeast cells (3). We then addressed whether the absence of MMS19 had a similar impact on XPD stability in human cells. In an inducible XPD overexpression system, siRNA-mediated depletion of MMS19 led to a decrease in the levels of XPD, corresponding to a reduction to 10% relative to control siRNA (Fig. 3B). The protein level of the point mutant C190S XPD, which is mutated in one of the four cysteines that coordinate the Fe-S cluster and is therefore predicted to display diminished Fe-S cluster incorporation, was also reduced when compared with wild-type XPD, and it was further decreased upon depletion of MMS19 (Fig. 3B).

Fig. 3

XPD requires MMS19 function for Fe-S cluster incorporation and protein stability. (A) Incorporation of 55Fe into ectopically expressed human GFPXPD in yeast. A galactose-driven promoter replaces the endogenous promoter of Mms19; expression was switched on (+ galactose) or switched off (+ glucose) for 40 hours. Western blots show depletion efficiency of Mms19 and the levels of XPD in WCEs. Graph depicts 55Fe incorporation and amount of immunoprecipitated protein. Error bars indicate SDs from three independent experiments. (B) Tetracycline-induced expression of stably integrated FlagXPD or FlagC190S XPD after treatment of HeLa cells for 7 days with MMS19 (19) or control (con) siRNAs. (C) Fractionation of HeLa WCEs by means of Superdex 200 gel filtration and Western blot analysis of CIA proteins and TFIIH components. (D) FlagMMS19 complexes were pulled down from WCEs of HEK 293 cells treated with or without 125 μM deferoxamine for 18 hours and analyzed by means of Western blot. Numbers below blots indicate fold increase in deferoxamine-treated samples as compared with control samples (from three independent experiments).

Cell fractionation of HeLa whole-cell extracts (WCEs) showed that XPD was mainly associated with the TFIIH factors XPB and Cdk7, whereas MMS19 comigrated predominantly with CIAO1 and MIP18 (Fig. 3C). In the absence of MMS19, however, the association of XPD with TFIIH was reduced, and elevated levels of XPD comigrated with CIAO1 (fig. S3A). These results suggest that XPD only transiently associates with the CIA machinery—possibly as an apoform before Fe-S cluster incorporation—but that this interaction is stabilized in the absence of MMS19 when Fe-S cluster transfer to XPD is impaired. If this were the case, then cellular depletion of Fe and hence impaired Fe-S biogenesis should also result in a more stable association of the CIA machinery with target Fe-S proteins. Indeed, in human embryonic kidney (HEK) 293 cells treated with the Fe chelator deferoxamine, up to fourfold higher levels of target Fe-S proteins were found to co-immunoprecipitate with MMS19 (Fig. 3D), with the effect on DNA polymerase δ being most pronounced. Previous studies in yeast had established a synthetic lethal connection between mms19Δ and pol3-13, a mutant strain bearing a single missense mutation in the catalytic subunit of DNA polymerase δ (18), which diminishes Fe-S cluster binding (11).

These data also imply that the function of MMS19 is not limited to XPD but instead extends to multiple Fe-S proteins involved in DNA metabolism. Indeed, siRNA-mediated depletion of MMS19 led to a marked decrease in the endogenous levels of the Fe-S proteins DNA polymerase δ, XPD, and FANCJ (Fig. 4A and fig. S4A). Consistent with an impact on DNA replication, mms19Δ yeast cells were sensitive to hydroxyurea (HU), which inhibits ribonucleotide reductase and reduces nucleotide pools (fig. S4B). Cell-cycle analysis further showed that wild-type yeast accumulate in S phase after HU treatment, whereas mms19Δ yeast are impaired for S phase entry (fig. S4C). Under conditions of HU treatment, Rad53 phosphorylation was induced in wild-type strains but was largely impaired in the absence of Mms19 (Fig. 4B). By extension, whereas HeLa cells treated with control siRNA induced phosphorylation of Chk1S345 and RPA in response to HU treatment, MMS19-depleted cells failed to do so (Fig. 4C). These data suggest that MMS19 deficiency in yeast or human cells confers an inability to efficiently enter S phase under limiting nucleotide pools. This may be due to the reduced levels of DNA polymerase δ (Fig. 4A) and potentially additional Fe-S proteins, such as Pri2, DNA2, and DNA polymerase ε. In agreement with an essential role in DNA metabolism, Mms19-β-GEO gene trap mice are embryonically dead before the implantation stage (fig. S5), and attempts to isolate Mms19-deficient embryos or cells were unsuccessful, suggesting that Mms19 is essential for cell viability in mice.

Fig. 4

MMS19 impacts on DNA metabolism. (A) Levels of Fe-S proteins in WCEs of HeLa cells treated with four different siRNAs against MMS19. Western blots show a representative experiment, and numbers below blots indicate protein stability in percent relative to control siRNA. (B) Rad53 phosphorylation in response to increasing doses of HU. Wild-type and mms19Δ yeast cells were treated for 3 hours with 0, 2, 20, and 200 mM HU, TCA-precipitated, and analyzed by means of Western blot. Arrowhead indicates Rad53 phosphorylation. (C) Chk1 and RPA phosphorylation in Hela cells treated with control and MMS19 siRNA in response to increasing doses of HU (0, 0.05, 0.1, 0.25, 0.5, and 1 mM HU for 20 hours). Arrowhead indicates RPA phosphorylation.

Collectively, these data identify MMS19 as a key factor involved in Fe-S cluster assembly, which affects the stability of multiple Fe-S proteins. We propose that MMS19 functions as part of the CIA machinery to facilitate Fe-S cluster transfer to target Fe-S proteins [as depicted in our model (fig. S6)]. The fact that MMS19 affects multiple proteins required for genome stability provides a plausible explanation for DNA replication problems and the early embryonic death of mice in which Mms19 has been knocked out. This study also provides molecular insight to explain the previously reported phenotypes associated with MMS19 deficiency (5, 6) and why defects in mitochondrial Fe-S cluster biogenesis confer genome instability (19).

Supplementary Materials

Materials and Methods

Figs. S1 to S6

Table S1

References and Notes

  1. Acknowledgments: Research in the laboratories of S.J.B. and M.J.S. is supported by Cancer Research UK. The laboratory of S.J.B. is also funded by a European Research Council Advanced Investigator Grant (RecMitMei). S.J.B. is a recipient of a Royal Society Wolfson Research Merit Award. K.G. is funded by long-term fellowships from the Federation of European Biochemical Societies and the Human Frontier Science Program. The authors declare no conflict of interest.
View Abstract

Navigate This Article