Research Article

MMS19 Assembles Iron-Sulfur Proteins Required for DNA Metabolism and Genomic Integrity

See allHide authors and affiliations

Science  13 Jul 2012:
Vol. 337, Issue 6091, pp. 195-199
DOI: 10.1126/science.1219723

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.


Instability of the nuclear genome is a hallmark of cancer and aging. MMS19 protein has been linked to maintenance of genomic integrity, but the molecular basis of this connection is unknown. Here, we identify MMS19 as a member of the cytosolic iron-sulfur protein assembly (CIA) machinery. MMS19 functions as part of the CIA targeting complex that specifically interacts with and facilitates iron-sulfur cluster insertion into apoproteins involved in methionine biosynthesis, DNA replication, DNA repair, and telomere maintenance. MMS19 thus serves as an adapter between early-acting CIA components and a subset of cellular iron-sulfur proteins. The function of MMS19 in the maturation of crucial components of DNA metabolism may explain the sensitivity of MMS19 mutants to DNA damage and the presence of extended telomeres.

Maintaining genomic stability is a key cellular function, and its impairment has been implicated in a variety of diseases, including cancer (13). The process has been connected to mitochondrial function and to the biogenesis of iron-sulfur (Fe-S) proteins, and it may be relevant for the neurodegenerative disorder Friedreich’s ataxia (46). The observation that multiple DNA replication and repair enzymes require Fe-S clusters for function has suggested a link between genomic stability and Fe-S protein biogenesis (712); however, in vivo evidence is lacking and the molecular basis of these connections is unclear.

The synthesis of Fe-S clusters and their assembly into proteins as inorganic cofactors cannot occur without a dedicated and conserved biosynthetic pathway (13, 14). Mitochondrial Fe-S proteins are matured by the iron-sulfur cluster (ISC) assembly machinery. Extramitochondrial Fe-S protein biogenesis depends on both the ISC and CIA (cytosolic iron-sulfur protein assembly) machineries. Cytosolic Fe-S clusters are first assembled on the CIA scaffold complex CFD1-NBP35 (1518) and are then transferred to apoproteins with the help of Cia1 [human CIAO1 (19)] and Nar1 [human IOP1 (20, 21)].

Mutants in eukaryotic MMS19 (also known as MET18 in yeast) show a variety of phenotypes, including defects in methionine synthesis, sensitivity to genotoxic stress [e.g., by methyl methanesulfonate (MMS)], and the presence of extended telomeres (2225). A molecular function explaining these diverse cellular roles is unknown. Previous proteomic studies identified an interaction between yeast Mms19 and (putative) CIA components potentially linking MMS19 to Fe-S protein biogenesis (26, 27). Here, we show that MMS19 is a component of the CIA machinery and acts as part of the “CIA targeting complex” that transfers Fe-S clusters to various DNA metabolism–associated Fe-S proteins. These findings explain the previously described MMS19 mutant phenotypes.

Yeast Mms19 is a late-acting CIA component. The interaction of yeast Mms19 (gene YIL128w) with two late-acting members of the CIA machinery, Cia1 and Cia2 [gene YHR122w (19, 2628)], was validated by coimmunoprecipitation (29) (fig. S1A). The potential role of Mms19 in cellular Fe-S protein biogenesis was examined by depleting the protein in a GAL promoter–regulatable MMS19 yeast strain and measuring 55Fe incorporation into known Fe-S target proteins (Leu1, Rli1, and Ntg2) by immunoprecipitation and scintillation counting (30). Mms19 depletion resulted in decreased 55Fe-S cluster binding to these proteins (Fig. 1A). The decrease in 55Fe binding by Leu1 correlated with loss of its enzymatic activity, whereas the activity of the non–Fe-S protein alcohol dehydrogenase was unchanged (Fig. 1B). These results were similar to those for depletion of the CIA proteins Nbp35 and Cia1 (19). The defect in Leu1 activity did not result from impaired methionine biosynthesis in mms19 mutants, because other methionine synthesis mutants showed normal Leu1 activity (fig. S2A). Mms19 was also required for 55Fe-S cluster assembly and for the sulfite reductase activity of the Met5-Met10 complex, which supplies sulfur for methionine biosynthesis in a CIA-dependent manner (Fig. 1, A to C, and fig. S2B). Together, these findings explain the methionine auxotrophy of mms19 mutants and identify Mms19 as a CIA component. As previously found for CIA defects, both mitochondrial Fe-S protein maturation (fig. S3, A and B) and cellular iron metabolism (fig. S3C) were unaffected by Mms19 deficiency. To clarify when Mms19 acted in the CIA pathway (13), we investigated its requirement for Fe-S cluster assembly on the early-acting CIA proteins Cfd1 and Nbp35. Mms19 depletion had no detrimental effect on the 55Fe-S maturation of these factors (Fig. 1D), which suggests that Mms19 functions late in cytosolic Fe-S protein assembly as a partner of Cia1 and Cia2.

Fig. 1

Yeast Mms19 is a CIA component acting late in cytosolic Fe-S protein maturation. (A) 55Fe incorporation into cytosolic Fe-S proteins. Gal-GFP-MMS19 yeast cells producing Rli1-HA, Ntg2-HA, or Met10-HA were grown in minimal medium (SC) containing galactose (Gal) or glucose (Glc). After radiolabeling with 55Fe, cell extracts were analyzed for the indicated Fe-S proteins by immunoblotting, and associated 55Fe was quantified by immunoprecipitation and scintillation counting. (B) Enzyme activities of the cytosolic Fe-S proteins isopropylmalate isomerase (Leu1) and sulfite reductase (SiR), and of alcohol dehydrogenase (ADH), were measured in extracts of indicated cells grown for 36 hours in glucose-containing SC and plotted relative to wild-type cell activities. (C) SiR activity was measured in wild-type (WT) and Δmms19 cells without and with a Mms19-encoding plasmid on SC (SC-Bi) or YP glucose plates containing bismuth ammonium citrate and sodium sulfite. Sulfide produced by SiR after growth for 3 days yields Bi2S3 (brown precipitate). (D) Mms19 acts late in biogenesis. 55Fe incorporation into the indicated CIA proteins was measured as in (A) for Gal-GFP-MMS19 cells with plasmids encoding Cfd1-TAP and Nbp35-TAP. Values are means ± SD. **P < 0.01, ***P < 0.001.

Yeast Mms19 assembles and interacts with Fe-S proteins involved in DNA metabolism. To explain the sensitivity of mms19 mutants to DNA-damaging agents (22), we analyzed the requirement of yeast Mms19 for the maturation of several Fe-S cluster–containing DNA repair enzymes. 55Fe radiolabeling showed that the Rad3 DNA helicase bound a Fe-S cluster in vivo (Fig. 2A) (7) and that its maturation was dependent on Mms19 and the ISC CIA machinery. Similar findings were made for the human Rad3 ortholog XPD (xeroderma pigmentosum protein D) when expressed in yeast (Fig. 2B). Mutations in Fe-S cluster–coordinating residues of XPD or the disease-relevant residue Arg112 destroyed Fe-S cluster binding (fig. S4A). Mms19 specifically coimmunoprecipitated with Rad3, Met10 (Fig. 2, C and D), and several other Fe-S proteins, including the DNA helicase/nuclease Dna2 (Fig. 2D and fig. S4B), Rli1, and the DNA glycosylase Ntg2 (fig. S1B).

Fig. 2

Yeast Mms19 interacts with and assembles Fe-S clusters into DNA helicases. (A and B) 55Fe incorporation into plasmid-encoded Rad3-TAP (A) and human XPD-TAP (B) was measured in yeast strains deficient in the indicated ISC and CIA proteins. The TAP-tagged proteins were visualized by immunostaining. Values are means ± SD. For all depletions, P < 0.001. (C) MMS19 binds to Rad3 and Cia2. Extracts from overnight cultures of GalL-HA-MMS19 (top) and GalL-HA-RAD3 cells (bottom) grown in SC galactose medium were used for immunoprecipitation with anti-HA beads. Cell extracts (CE) and immunoprecipitates (IP:HA) were immunostained for the indicated proteins. (D) Interaction of Mms19 with target Fe-S proteins. Extracts of Gal-GFP-MMS19 cells producing C-terminally HA-tagged Dna2, Cia2, or Met10 were used for immunoprecipitation with anti-HA beads. Bound proteins were analyzed by immunostaining (anti-HA, top; anti-GFP, bottom).

Human MMS19 matures only a subset of Fe-S proteins. Mammalian MMS19 has been identified as part of a protein complex including XPD and two putative CIA components, CIAO1 and FAM96B (homolog of yeast Cia2). It has been functionally implicated in DNA repair, chromosome segregation, and transcription (24, 31, 32), yet no connection to Fe-S protein assembly has been proposed. To examine whether human MMS19 plays an evolutionarily conserved role in Fe-S protein biogenesis, we depleted MMS19 by RNA interference (RNAi) in HeLa cells, using three different small interfering RNA (siRNA) oligos either alone or as a pool (33). Efficient MMS19 silencing was achieved by three consecutive siRNA transfections performed at 3-day intervals (fig. S5, A and B). We first tested the possible role of MMS19 in the maturation of two well-studied Fe-S proteins, cytosolic aconitase IRP1 (iron regulatory protein 1) and GPAT [glutamine phosphoribosylpyrophosphate amidotransferase (16)]. For IRP1, the enzyme activity, protein level, and binding capacity to iron-responsive RNA elements (IREs) were unchanged upon MMS19 depletion (Fig. 3A and fig. S6, A to C). Similarly, no effects on GPAT protein levels were observed (Fig. 3A and fig. S7). Because Fe-S cluster binding is essential for GPAT stability, GPAT abundance is a reliable measure of its maturation (16).

Fig. 3

Human MMS19 interacts with and assembles Fe-S proteins involved in DNA metabolism. HeLa cells were RNAi-treated three times at 3-day intervals or mock-treated. Cell extracts were prepared by digitonin lysis (figs. S5 to S10). (A) After 9 days of MMS19 depletion, IRP1 (cytosolic aconitase) activities (normalized to lactate dehydrogenase; LDH) were measured (lower panel), and IRP1, GPAT, and actin protein levels were determined by immunostaining (upper panel) and quantitative densitometry (middle panel). Values are means ± SD. ***P < 0.001. (B) The enzyme activity of DPYD (upper panel) was measured by thin-layer chromatography and autoradiography (middle panel) and quantitated by phosphorimaging (lower panel). As controls, HeLa cells were RNAi-depleted for frataxin (FXN), NFS1, and NBP35. Values are means ± SD. For all depletions, P < 0.001. (C and D) Levels of DPYD (C) and POLD1 (D) protein were measured by immunostaining in cell extracts depleted for the indicated proteins. (E) Interaction of human MMS19 with Fe-S proteins and CIA components. Flp-In–TREx-293 cells lacking or stably expressing inducible 3xHA-3xFLAG-MMS19 were induced with doxycycline (500 ng/ml) overnight. Whole-cell extracts (WCE) were subjected to anti-HA immunoprecipitation (IP:HA) followed by immunostaining for the indicated proteins.

These negative results prompted us to hypothesize that MMS19 plays a specialized role in Fe-S cluster maturation of a subset of target proteins. We therefore developed two additional assays for cytosolic Fe-S protein biogenesis. First, we examined the enzymatic activity of the Fe-S protein dihydropyrimidine dehydrogenase [DPYD (34)] by following the conversion of [4-14C]thymine into [4-14C]dihydrothymine by thin-layer chromatography and autoradiography (Fig. 3B and fig. S8, A to C). In contrast to IRP1 and GPAT, DPYD activity was severely impaired upon MMS19 depletion. This decrease was also observed upon depletion of the ISC proteins Nfs1 and frataxin and the CIA component Nbp35, consistent with this effect being a Fe-S cluster assembly defect. Although DPYD protein levels were also depleted (Fig. 3C and fig. S8, D and E), the respective decreases were less pronounced relative to enzyme activities. The lower DPYD levels were likely indirect effects of apoprotein degradation. Next, we measured the amounts of the POLD1 subunit of DNA polymerase δ, the homolog of yeast Fe-S protein Pol3 (10). Because impaired Fe-S protein assembly frequently results in apoprotein degradation (Fig. 3C) (16), Fe-S protein levels can be used to estimate their biogenesis. POLD1 levels were strongly decreased in MMS19-depleted cells (Fig. 3D and fig. S9). Similar effects were observed during RNAi-mediated depletion of Nfs1, frataxin, and Nbp35. Because CIA depletion should not affect mitochondrial Fe-S proteins (fig. S3) (16, 21), we measured the levels and activities of mitochondrial aconitase (mtAco) and succinate dehydrogenase (SDH). They remained unchanged upon MMS19 depletion (fig. S10). Collectively, these results suggest that MMS19 is required for Fe-S cluster assembly of DPYD and POLD1 and may act as a specialized CIA factor with specificity for a subset of cytosolic-nuclear Fe-S proteins.

Human MMS19 is part of the CIA targeting complex. To define those Fe-S proteins that require MMS19 function, we used a proteomic approach to identify MMS19 interaction partners. MMS19-associated protein complexes were affinity-purified from a human embryonic kidney (HEK) 293 cell line stably expressing hemagglutinin- and FLAG-tagged MMS19 (HA-FLAG-MMS19) and were then analyzed using multidimensional protein identification technology [MudPIT (35)]. We identified a wide range of putative MMS19-interacting proteins including known and putative Fe-S proteins and the characterized and putative CIA components CIAO1, IOP1, and FAM96B (table S4). For some of these proteins, the MMS19 interaction was confirmed by coimmunoprecipitation of HA-FLAG-MMS19 from stable HEK293 cells followed by immunoblotting with specific antibodies. Endogenous CIAO1, FAM96B, IOP1, FANCJ, XPD, RTEL1, and POLD1 copurified only in extracts containing HA-FLAG-MMS19 (Fig. 3E). The interaction between XPD and MMS19 was confirmed for endogenous levels (fig. S11). These data validate the proteomic data and show that MMS19 associates with both CIA factors and Fe-S target proteins, including many involved in DNA metabolism.

The strong interactions of MMS19, CIAO1, and FAM96B, as assessed by normalized spectral abundance factors (table S4) and coimmunoprecipitation studies (Fig. 3E), indicate that these three proteins likely form a complex that we term the CIA targeting complex. The results for human MMS19 agree with our yeast data and suggest that this complex acts late in Fe-S protein biogenesis to facilitate Fe-S cluster transfer from the CIA scaffold complex Cfd1-Nbp35 to Fe-S target proteins.

Fe-S protein assembly defects augment sensitivity of cells to DNA damage. The association of MMS19 with multiple DNA replication and repair proteins led us to examine whether the integrity of Fe-S protein biogenesis might be a general requirement for efficient cellular DNA damage repair. We first tested this idea in yeast and assessed whether depleting ISC or CIA components elicited effects on the DNA damage response pathway. Rad53 phosphorylation was observed in strains defective in Fe-S protein assembly after exposure to low levels of MMS (36). These MMS levels were not sufficient to trigger Rad53 phosphorylation in wild-type yeast, which suggests that the former strains accumulated higher levels of DNA damage (Fig. 4A). We then analyzed the effects of Fe-S protein deficiencies on the expression of RNR3 and HUG1, two genes transcriptionally induced in response to DNA damage (37). Reporter plasmids in which the RNR3 and HUG1 promoters controlled luciferase expression were introduced into different regulatable ISC and CIA strains. A factor of 2 to 6 increase in luciferase activity was detected in ISC- and CIA-depleted strains relative to wild-type cells (Fig. 4, B and C). These results indicate that inactivation of Mms19 or other Fe-S protein assembly components leads to up-regulation of the DNA damage response. We finally tested whether the requirement of intact Fe-S protein assembly for DNA damage repair was conserved in humans. HEK293 cell lines expressing short hairpin RNA (shRNA) constructs for silencing MMS19, FAM96B, the CIA protein IOP1, or the mitochondrial scaffold protein ISCU were generated (fig. S12). Depletion of ISC or CIA factors severely diminished the survival of HEK293 cells after ultraviolet (UV) irradiation or MMS treatment (Fig. 4D). Together, these data suggest that the previously observed DNA repair and metabolism defects of MMS19-defective cells may be the consequence of impaired Fe-S protein biogenesis and may not be related to a dedicated function of MMS19 in DNA maintenance (24, 25).

Fig. 4

Defects in Fe-S protein assembly show increased DNA damage sensitivity. (A) Activation of Mec1-dependent DNA damage pathway in yeast. Wild-type yeast (WT) and strains deficient in the indicated proteins were treated with the indicated concentrations of MMS. Extracts were immunostained for phospho-Rad53 (Rad53-P). (B and C) Induction of DNA damage–inducible RNR3 or HUG1. Yeast strains depleted for the indicated proteins were transformed with plasmids encoding RNR3 or HUG1 promoter-regulated luciferase and were exposed to 0.25 mM MMS. Values are means ± SD. For all depletions, P < 0.01. (D) Depletion of the indicated proteins by shRNAs in HEK293 cells results in increased sensitivity to UV- and MMS-induced DNA damage. Stable knockdown cell lines were treated with UV (20 J/m2) or 20 μM MMS. Cell viability relative to the corresponding untreated cell line (No target) was measured using the MTS cell proliferation assay (Promega) 7 days after treatment. Values are means ± SEM (n = 3); for all depletions, P < 0.05. (E) Fe-S protein biogenesis is required for incorporation of XPD into TFIIH. HEK293 cell extracts (WCE) from (D) were used for immunoprecipitation with XPB antibodies (IP:XPB) followed by immunoblotting for the indicated proteins.

The increased DNA damage sensitivity in cells with impaired Fe-S protein biogenesis may include the loss of nucleotide excision repair because maturation of XPD is defective. The requirement of the Fe-S cluster of XPD for its DNA helicase activity in vitro (7) led us to investigate whether inactive apo-XPD is integrated into the transcription factor complex TFIIH in ISC- or CIA-depleted cells. XPD could be detected in TFIIH immunoprecipitated with endogenous XPB from control cells but not from cells depleted of MMS19, FAM96B, IOP1, or ISCU; these findings suggest that the inability to assemble Fe-S clusters on XPD prevented its incorporation into the TFIIH complex (Fig. 4E). The inability of cells impaired in Fe-S protein biogenesis to form functional TFIIH complexes may provide, at least in part, the mechanistic basis for the increased sensitivity of MMS19 mutants to UV- and MMS-induced DNA damage. Other Fe-S cluster–containing substrates of the CIA targeting complex, such as DNA2, FANCJ, and RTEL1, play key roles in maintaining genome stability and the response to other types of DNA damage [i.e., DNA double-strand breaks and DNA interstrand cross-links (3)]. Hence, impairment of Fe-S protein biogenesis (e.g., by mutations in ISC and CIA proteins) may lead to the simultaneous inactivation of multiple DNA repair pathways and thereby promote genomic instability.

Our study clarifies the functional role of MMS19 in DNA maintenance and provides insights into the mechanism of cytosolic Fe-S protein biogenesis (11, 13). MMS19 exerts its function as part of a CIA targeting complex involved in the maturation of a subset of Fe-S proteins, including those with functions in DNA replication, DNA repair, and telomere stability (fig. S13). By undergoing direct interaction with target Fe-S proteins, MMS19 (and its putative functional partners CIAO1 and FAM96B) may serve an adapter function between Fe-S cluster synthesis and insertion into apoproteins. This conserved function of MMS19 can explain possibly all of the previously described phenotypes associated with MMS19 defects. Mitochondria perform an essential role in cellular Fe-S protein biogenesis and, as shown here, in nuclear DNA metabolism. These functions, and not adenosine triphosphate production, may explain the maintenance of these endosymbiotic organelles even in anaerobic eukaryotes (38). Moreover, the crucial role of mitochondria in DNA metabolism and genome maintenance may be relevant to neurodegenerative phenotypes associated with mitochondrial diseases, including Friedreich’s ataxia.

Supplementary Materials

Materials and Methods

Figs. S1 to S13

Tables S1 to S4

References (3964)

References and Notes

  1. See supplementary materials on Science Online.
  2. Acknowledgments: We thank D. R. Dean and D. E. Gottschling for fruitful discussions, and C. Doré, S. A. Freibert, M. Funke, G. Köpf, B. Niggemeyer, R. Rösser, M. Stümpfig, and H. Webert for experimental help. Supported by Deutsche Forschungsgemeinschaft grants (SFB 593), Gottfried-Wilhelm Leibniz program and GRK 1216, von Behring–Röntgen Stiftung, Max-Planck Gesellschaft, Feldberg Foundation, Fonds der chemischen Industrie (all to R.L.), Rhön Klinikum AG (A.J.P. and R.L.), NIH grant GM089778 (J.A.W.), the University of California Cancer Research Coordinating Committee (J.A.W.), and the Jonsson Cancer Center at UCLA (J.A.W.).
View Abstract

Stay Connected to Science

Navigate This Article