Qri2/Nse4, a component of the essential Smc5/6 DNA repair complex

Authors


E-mail barry.panaretou@kcl.ac.uk; Tel. (+44) 207 848 4003; Fax (+44) 207 848 4003.

Summary

We demonstrate a role for Qri2 in the essential DNA repair function of the Smc5/6 complex in Saccharomyces cerevisiae. We generated temperature-sensitive (ts) mutants in QRI2 and characterized their properties. The mutants arrest after S phase and prior to mitosis. Furthermore, the arrest is dependant on the Rad24 checkpoint, and is also accompanied by phosphorylation of the Rad53 checkpoint effector kinase. The mutants also display genome instability and are sensitive to agents that damage DNA. Two-hybrid screens reveal a physical interaction between Qri2 and proteins that are non-Smc elements of the Smc5/6 DNA repair complex, which is why we propose the name NSE4 for the open reading frame previously known as QRI2. A key role for Nse4 in Smc5/6 function is likely, as overexpressing known subunits of the Smc5/6 complex suppresses nse4ts cell cycle arrest. The nse4ts growth arrest is non-lethal and unlike the catastrophic nuclear fragmentation phenotype of smc6ts mutants, the nucleus remains intact; replicative intermediates and sheared DNA are not detected. This could imply a role for Nse4 in maintenance of higher order chromosome structure.

Introduction

In the budding yeast Saccharomyces cerevisiae around 120 genes are expressed towards the end of G1. This group of genes is referred to as the CLN2 cluster, after one of its most conspicuous members (Spellman et al., 1998). This cluster includes sets of genes involved in DNA synthesis (CDC6, POL1, RNR1), DNA repair (SMC6, RAD27), sister chromatid cohesion (SMC1, SMC3), spindle pole body duplication (TUB4, SPC42) and bud emergence (SVS1, BN14) (Spellman et al., 1998). QRI2 is an essential S. cerevisiae open reading frame (ORF) of unknown function that is a member of the CLN2 cluster. This ORF's name reflects its location next to QRI1 on chromosome 4, and is not a descriptor of the function of the corresponding protein. We show that Qri2 is essential for cell cycle progression, and that it is part of the essential Smc5/6 DNA repair complex. The SMC (structural maintainance of chromosomes) family of proteins play essential roles in genome integrity. There are six classes of Smc protein in eukaryotes, forming three pairs of heterodimers. Each pair associates with additional non-Smc proteins to form high molecular mass complexes. Qri2 is a non-Smc element that associates with the Smc5/6 complex, and in keeping with the established nomenclature we have renamed this ORF as NSE4, and refer to the ORF using this name from now on. The name change has been agreed with the group that originally deleted QRI2 (Simon et al., 1994). We have registered NSE4 with the Saccharomyces Genome Database (SGD) – as an alias for QRI2.

NSE4 is an essential S. cerevisiae ORF that is conserved throughout the eukaryotic lineage. Like many genes expressed at the G1/S transition, NSE4 contains a perfect MCB (MluI cell cycle box) element in its promoter. Transcription from MCB elements is controlled by MBF (MCB binding factor), a heterodimeric transcription factor composed of Swi1, which is tethered to DNA by Swi4. MBF mediated transcription activation is cell cycle regulated as MBF itself is activated post-translationally by Cln3-Cdc28 (Valdivieso et al., 1993). Also, microarray profiling shows that NSE4 transcription is extensively repressed in a time-dependent manner when cells are arrested in G1 by α-factor mating pheromone (Roberts et al., 2000). Periodic variation of transcript levels through the cell cycle, and rapid transcriptional inhibition after cell cycle arrest, imply that Nse4 is required for passage through the cell cycle. Furthermore, the recent proteome wide localization of yeast proteins reveals that Nse4 is a nuclear protein (Huh et al., 2003). This is in keeping with the location of most of the proteins encoded by the CLN2 cluster genes, as they play roles predominantly in DNA replication and repair.

The systematic yeast gene deletion project showed that Nse4 is essential (Simon et al., 1994). In this report, we present functional analysis of Nse4. We show that cells require this protein for the metaphase to anaphase transition. More specifically, we demonstrate that cells lacking functional Nse4 arrest in a manner that is dependent on activation of a DNA damage checkpoint. We show that Nse4 plays a role in repairing chromosomal aberrations that appear during normal progression of the cell cycle, at a point after DNA replication. Furthermore, we identify the DNA repair complex that it is abrogated by loss of Nse4 function.

NSE4 has been conserved throughout the eukaryotic lineage, with murine, worm, fly and plant orthologues identified by blast searches. This is not surprising given the overall conservation of all three SMC heterodimers and their non-SMC components. In humans there are two Nse4 orthologues – the product of the AAH27612 locus (NCBI Accession AAH27612) and FLJ20003 (NCBI Accession NP_060085). EST databases reveal the latter is expressed in most (if not all) human tissues; an alignment of Nse4 with FLJ20003 (revealing 22% similarity) is presented in Fig. 1 (the other human orthologue and a murine orthologue are also included in the alignment). Analysis of the primary sequence of Nse4 does not reveal information about its biochemical properties. Our functional characterization of this protein was performed by assessing the arrest caused by conditional lethal alleles of NSE4.

Figure 1.

Structural and functional conservation of Nse4.
A. Alignment of Nse4 from S. cerevisiae, against murine (NCBI Accession NP_079775) and two human orthologues: (i) FLJ20003; NCBI Accession NP_060085 and (ii) product of locus AAH27612 (NCBI Accession AAH27612). The amino acid sequence of the yeast orthologue is numbered. Sites mutated in ts mutants are underlined [nse4-1ts D103N R127K and S195F; nse4-2ts S329D; nse4-3 ts G224D; nse4-4ts G224S and S231D; (*) conserved residues; (:) strongly similar residues; (.) residues of weak similarity].

Results

Isolation of NSE4 temperature-sensitive ( ts) mutants

We used hydroxylamine mutagenesis and the plasmid shuffle technique to isolate point mutations in NSE4 that confer a ts growth phenotype. We initially constructed a haploid yeast strain (BP12 in Table 1) in which NSE4 was deleted from the genome. NSE4 is an essential gene, so viability was maintained by a copy of this gene controlled by its native promoter, on a centromeric URA3 vector. This strain was used to select for ts nse4 mutants, generated by hydroxylamine treatment of a centromeric LEU2 vector bearing NSE4 expressed from its native promoter. Eight ts mutants were isolated – and the Nse4 coding regions were sequenced. Some of the mutants were isolated more than once, giving four unique mutants (nse4-[1–4]ts). As expected, the mutants resulted from G to A or C to T transitions, giving the following amino acid substitutions: nse4-1ts D103N, R127K and S195F; nse4-2ts S*329D; nse4-3ts G*224D; nse4-4ts G*224S and S231D (the residues marked with an asterisk are conserved in the S. cerevisiae, murine and human Nse4 orthologues). All four mutants were isolated by virtue of their inability to grow at 37°C (Fig. 2A). At 25°C, however, nse4-3ts and nse4-4ts cells grew at the same rate as cells bearing the wild-type gene, whereas nse4-1ts and nse4-2ts exhibited a slow growth phenotype, with doubling times of 3 and 4 h, respectively, in contrast to the 2 h doubling time of wild-type. For each mutant, the ts phenotype does not result in cell death at the non-permissive temperature, as returning the cells to 25°C results in restoration of growth – the phenotype at the non-permissive temperature is therefore a growth arrest (Fig. 2B).

Table 1.  S. cerevisiae strains used in this study.
StrainGenotypeSource
  1. EUROSCARF, European SaccharomycescerevisiaeArchive for Functional analysis.

BP12 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBP01 (NSE3/URA3)This study
BH01 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBH1 (nse3-1ts/URA3)This study
BH02 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBH2 (nse3-2ts/URA3)This study
BH03 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBH3 (nse3-3ts/URA3)This study
BH04 MAT a; ura3-52; leu2D1; nse3::kanMX4; pBH4 (nse3-4ts/URA3)This study
BH21 MAT a; ura3Δ0; leu2Δ0; his3Δ1; nse3::HygB; pBH1 (nse3-1ts/URA3)This study
BH30 MAT a; ura3Δ0; leu2Δ0; his3Δ1; lys2Δ0; rad24::kanMX4; nse3::HygB; pBH1 (nse3-1ts/URA3)This study
BH31 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBP01 (NSE3/URA3); YEplac181 (LEU2)This study
BH32 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBH1 (nse3-1ts/URA3); YEplac181 (LEU2)This study
BH33 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBH1 (nse3-1ts/URA3); pBH101 (YDR288W/LEU2)This study
BH34 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBH1 (nse3-1ts/URA3); pBH102 (MMS21/LEU2)This study
BH35 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBH1 (nse3-1ts/URA3); pBH103 (NSE1/LEU2)This study
BH36 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBH1 (nse3-1ts/URA3); pBH104 (SMC5/LEU2)This study
BH37 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBH1 (nse3-1ts/URA3); pBH104 (RHC18(SMC6)/LEU2)This study
BH38 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBH4 (nse3-4ts/URA3); YEplac181 (LEU2)This study
BH39 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBH4 (nse3-4ts/URA3); pBH10 1 (YDR288W/LEU2)This study
BH40 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBH4 (nse3-4ts/URA3); pBH102 (MMS21/LEU2)This study
BH41 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBH4 (nse3-4ts/URA3); pBH103 (NSE1/LEU2)This study
BH42 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBH4 (nse3-4ts/URA3); pBH104 (SMC5/LEU2)This study
BH43 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBH4 (nse3-4ts/URA3); pBH104 (RHC18(SMC6)/LEU2)This study
BH45 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBH2 (nse3-2 ts/URA3); YEplac181 (LEU2) This study
BH46 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBP01 (NSE3/URA3) pDK243 (1ARS/LEU2/ADE3)This study
BH47 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBP01 (NSE3/URA3) pDK368-7 (7ARS/LEU2/ADE3)This study
BH48 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBH2 (nse3-2 ts/URA3); pDK243 (1ARS/LEU2/ADE3)This study
BH49 MAT a; ura3-52; leu2Δ1; nse3::kanMX4; pBH2 (nse3-2 ts/URA3); pDK368-7 (7ARS/LEU2/ADE3)This study
PJ69-4α MAT α; trp1-901; leu2-3112; ura3-52; his3-200; gal4Δ; gal80Δ; LYS2::GAL1-HIS3; GAL2-ADE2; met2::GAL7-lacZ Uetz et al. (2000)
10590D Mata/α; ura3-52/ura3-52; his3Δ200/HIS3; leu2Δ1/LEU2; LYS2/LYS2; trp1D63/TRP1; nse4(YDL105W)::kanMX4/YDL105wEUROSCARF
Y23802 Mata/α; his3Δ1/his3Δ1; leu2Δ0/leu2Δ0; lys2Δ0/LYS2; MET15/met15Δ0; ura3Δ0/ura3Δ0; nse4(YDL105W)::kanMX4/NSE4(YDL105W)EUROSCARF
Y16169 Matα; his3Δ1; leu2Δ0; lys2Δ0; ura3Δ0; YER173W(rad24)::kanMX4EUROSCARF
Figure 2.

Growth arrest phenotype of nse4 conditional mutants.
A. Growth curves at permissive and non-permissive temperature, of nse4 deletes, bearing a vector containing the wild-type gene or one of four temperature-sensitive (ts) alleles. Cultures were grown to mid-log stage at 25°C, split into two aliquots, with one maintained at 25°C and the other shifted to 37°C; OD600 was measured every 2 h.
B. Growth arrest displayed by nse4ts conditional mutants is reversible. NSE4 wild-type and mutant strains were streaked on YPD plates and incubated at 25°C for 3 days or 37°C for 2 days. The plate incubated at 37°C was subsequently incubated at 25°C for 3 days.

nse4 ts mutants arrest at G2/M

We characterized the growth defect by flow cytometry [fluorescence activated cell sorting (FACS)], examining the DNA content of nse4ts mutant cells at non-permissive conditions. Asynchronous log phase cultures grown at 25°C were shifted to 37°C and samples were taken every 3 h. Cells were stained with propidum iodide and cell cycle distribution was assessed by FACS. After 3 h at 37°C, all four mutants arrested with a 2C DNA content, indicating arrest between late S phase and prior to the exit from mitosis (Fig. 3A). We subsequently assessed the cellular distribution of DNA, the state of the mitotic spindle and cell morphology. Exponential phase cultures of wild-type and all four nse4ts mutants grown at 25°C were shifted to 37°C for 6 h; > 80% of nse4ts cells were arrested with large buds and a short bipolar mitotic spindle, the 2C nucleus was located near the bud neck (Fig. 3B).

Figure 3.

At non-permissive conditions, nse4ts mutants are blocked at G2/M transition.
A. DNA flow cytometry (FACS) analysis of nse4ts mutant cells. Wild-type and nse4ts mutants were grown to exponential phase at 25°C. After shifting to 37°C samples were taken at the time intervals indicated, fixed and stained with propidium iodide and sorted for DNA content by FACS.
B. Cell cycle arrest examined by nuclear and microtubule staining. Wild-type and nse4ts mutants were grown to exponential phase at 25°C and shifted to 37°C for 6 h. Cells were fixed and stained with DAPI and antitubulin antibody.
C. At non-permissive conditions DNA replication is complete in nse4ts mutants. Wild-type and mutant cells were grown at 25°C and split into two aliquots, one was maintained at 25°C, the other was shifted to 37°C. After 4 h, chromosomes from each culture were separated by pulse field gel electrophoresis, and then blotted and probed with a DNA fragment of SAC6. As controls, exponential phase wild-type cells grown at 25°C were incubated with either 0.2 M HU, a DNA replication inhibitor that arrests cells in S phase, or 5 µg ml−1α-factor that arrests cells in G1. Less DNA is visible in the nse4ts cells at 37°C because of cell cycle arrest, whereas wild-type cells continue to divide.

Flow cytometry cannot be used to distinguish between a block in late S phase from a block at G2/M boundary (Chakraverty et al., 2001). TOP3, CDC17, HYS2, RFC2 are all genes involved in DNA synthesis, but conditional mutations in these genes are ‘leaky’ and FACS analysis shows they arrest with a G2/M content of DNA under non-permissive conditions, instead of the expected arrest in (or just prior to) S phase (Chakraverty et al., 2001; Sugimoto et al., 1995; Hennessy et al., 1991; Noskov et al., 1998). Although our FACS data suggests that the bulk of DNA appears to be replicated in nse4ts mutants under non-permissive conditions, there could be significant amounts of unreplicated DNA, sufficient to cause cell cycle arrest via activation of the replication checkpoint or via a defect in DNA replication per se. To determine whether DNA was completely replicated, chromosomes of arrested nse4ts cells were analysed by pulse f ield gel electrophoresis (PFGE). In this assay, only fully replicated DNA enters a gel – whereas incompletely replicated DNA, bearing nicks, gaps and/or stalled replication forks, is retarded in wells.

As controls, wild-type cells were incubated with agents that block in S phase (hydroxyurea) or G1 (α-factor). As expected, chromosomes from wild-type cells treated with α-factor entered the gel. Chromosomes from wild-type cells blocked in S phase with hydroxyurea (a DNA replication inhibitor), failed to enter the gel and remained in the well (Fig. 3C). Growing cultures of asynchronous cells should have both fully replicated chromosomes and replication intermediates from cells undergoing S phase. As expected, both types of DNA were detected in exponential phase wild-type cells and mutant cells grown at 25°C. The replication intermediates, however, are entirely absent from arrested mutant cells (Fig. 3C), indicating that at non-permissive conditions nse4ts mutants have completed DNA replication and are arrested at a point in the cell cycle after S phase but prior to the onset of anaphase. We performed the PFGE analysis using the two mutants with, respectively, the least severe (nse4-4ts) and most severe (nse4-2 ts) phenotype as judged by growth rate at permissive temperature, and, arrest at 37°C (Fig. 2A).

The checkpoint effector kinase Rad53 is activated in nse4ts mutants

The arrest displayed by nse4ts mutants could be attributed to activation of a checkpoint signal. There are two major checkpoints operating after the bulk of S phase has been completed. Those that ensure faithful replication and repair of DNA, and, the mitotic spindle checkpoint that guarantees proper segregation of chromosomes during anaphase. However, the increase in cell size when nse4ts mutants acquire 2C content of DNA would be consistent with activation of the G2/M damage checkpoint (Chakraverty et al., 2001).

The MAD2 spindle checkpoint pathway blocks sister chromatid separation (Alexandru et al., 1999). Quantization of cell cycle stages, as judged by flow cytometry, examination of cell morphology, DAPI staining and staining for the mitotic spindle, revealed that mad2Δnse4-1ts cells arrest in the same way as cells bearing the nse4-1ts mutation only (data not shown). Accordingly, NSE4 is epistatic to MAD2, and the spindle checkpoint does not play a role in the G2/M arrest of nse4ts mutants.

Rad53 is the effector protein kinase required for cell cycle arrest in response to both disturbed DNA replication and DNA damage. The resulting aberrant DNA structures result in hyperphosphorylation and activation of Rad53 (Carr, 2002). Phosphorylated Rad53 is detected as a smear of slowly migrating forms of the protein on Western blots immunoprobed with anti-Rad53 antisera. As controls, we incubated cells with methylmethanesulphonate, a DNA damaging agent. As expected, this led to the appearance of hyperphosphorylated forms of Rad53, in both wild-type and nse4-2ts cells, at both permissive and non-permissive temperatures (Fig. 4A). In wild-type cells a single band was detected at both 25°C and 37°C. In contrast, incubation of nse4-2ts cells at the non-permissive temperature (in the absence of extrinsic DNA damaging agents) led to the appearance of hyperphosphorylated forms of Rad53, suggesting that DNA structures that activate DNA damage checkpoints are present (Fig. 4A). These hyperphosphorylated forms were also present, albeit to a lesser extent, in nse4-2 ts cells at permissive temperatures also, this was not surprising as there is a slow growth defect at 25°C (Figs 4A and 2A).

Figure 4.

Cell cycle arrest of nse4ts mutants is checkpoint-dependent.
A. The Rad53 checkpoint effector kinase is phiosphorylated. Wild-type or nse4-2 ts cells were grown to early log phase at 25°C and then split into equal aliquots. One was grown at 25°C and the other shifted to 37°C. After incubation for 3 h, each culture was split again. MMS to a final concentration of 0.03% was added to one aliquot, with no additions made to the other aliquot. Cultures were incubated for a further 2 h. Rad53 was visualized by probing Western blots with anti-Rad53 antibody. The positions of unphosphorylated Rad53 and phosphorylated Rad53 (Rad53-P) are shown on the right.
B. DNA content was determined by FACS analysis for nse4-1ts and nse4-1tsΔrad24 strains, at 25°C and at the times indicated after shift to 37°C.
C. Quantification of cell cycle distribution. Cells from strains listed in B were shifted from 25°C to 37°C for 6 h, fixed and then stained with DAPI. Cells were divided into different cell cycle stages as depicted at the top of the table, and expressed as a percentage of the total number of cells examined, with 300 cells assessed per strain. This was repeated three times with percentage values not varying by ±5%.

nse4ts arrest is dependent on the Rad24 DNA damage checkpoint

Hyperphosphorylation of Rad53 suggests activation of the DNA damage checkpoint. In budding yeast Rad24 is specifically required for response to DNA damage, and does not play a role in the response to stalled replication forks (Pellicioli et al., 1999; Boddy and Russell, 2001; Melo and Toczyski, 2002). Rad24 is instrumental in loading a PCNA-like complex, composed of Mec3, Rad17 and Ddc1, to lesions in DNA (Majka and Burgers, 2003). We performed epistasis analysis using a rad24Δnse4-1ts strain. Cultures growing exponentially at 25°C were shifted to 37°C for 6 h and cell cycle properties were examined. In contrast to the nse4-1ts mutant, flow cytometric analysis showed the appearance of a G1 peak in the rad24Δnse4-1ts suggesting nse4ts arrest was mediated at least in part by the Rad24 epistasis group (Fig. 4B). Assessing cell cycle distribution of these mutants also supports a role played by Rad24 in nse4ts G2/M arrest. After a shift in temperature to 37°C for 4 h, 82% of nse4-1ts cells were arrested at G2/M. In contrast, the rad24Δnse4-1ts double mutants showed a significant reduction in the percentage of cells arrested at this point (Fig. 4C). A defect in Nse4 function may cause DNA damage, which activates the DNA damage checkpoint.

nse4ts mutants are hypersensitive to DNA replication block and DNA damage

One explanation for results presented to date is that nse4ts cells accumulate aberrant DNA structures. Nse4 function may contribute to mechanisms that sense DNA damage or mechanisms that repair damage per se. We tested for sensitivity of nse4ts mutants to the DNA alkylating agent methyl methanesulfonate, MMS. Growth of nse4-1ts and nse4-2 ts was severely impaired on solid media containing MMS at 25°C, 30°C and 32°C; nse4-3 ts and nse4-4ts exhibited severe growth defects at 30°C and 32°C (Fig. 5A). That nse4-3ts and nse4-4ts are less sensitive than the other two mutants at 25°C is not surprising, as the former pair of mutants exhibit growth kinetics identical to the wild-type at 25°C (Figs 5A and 2A). The nse4ts mutants are also sensitive to the replication inhibitor hydroxyurea (HU), presumably because stalled replication intermediates are sensed by the cell as abnormal DNA structures (Forbes and Enoch, 2000). None of the mutants were sensitive to the microtubule destabilizing drug benomyl (not shown), implying the role of Nse4 is unconnected to the spindle checkpoint, in agreement with the results from epistasis analysis of the mad2Δnse4-1ts double mutant.

Figure 5.

The sensitivity of nse4ts mutants to DNA replication block and DNA damage.
A. Wild-type and nse4ts mutants were grown at 25°C to exponential phase and diluted to equal cell density. Sixfold serial dilutions were spotted across YPD, or YPD containing 50 mM HU or 0.025% MMS, followed by incubation at the indicated temperatures for 3 days.
B. Exponential phase wild-type and nse4ts mutants diluted to equal cell density were incubated with 0.02% MMS or 200 mM HU; aliquots were removed at the indicated intervals, diluted, and spread over YPD plates. Percentage viability was determined from the number of colonies that appeared after incubation at 25°C for 3 days – as a ratio against the number of colonies that appeared for each strain at the zero time point.

In quantitative assays, viability of nse4ts mutant cells in the presence of MMS and HU is lost rapidly; nse4-2 ts is highly sensitive to killing by MMS, the other nse4ts mutants exhibiting a milder sensitivity to MMS compared with wild-type (Fig. 5B). An identical spectrum of sensitivity across our series of mutants is exhibited against HU (Fig. 5B). In wild-type cells, early origins of replication fire in the presence of HU, but late origins are repressed and stalled forks are stabilized (Lopes et al., 2001; Tercero and Diffley, 2001). Cell death of nse4ts mutants incubated with HU infers defects in late origin firing and/or stabilization of replication forks.

Nse4 is required for maintenance of minichromosomes

Cells respond to DNA damage with a combination of mechanisms designed to repair the damage while delaying cell cycle progression. If any of these mechanisms fail, then genome instability can result. A role for Nse4 in repair of DNA damage is suggested by the sensitivity to MMS exhibited by nse4ts mutants. Moreover, nse4ts arrest at non-permissive temperatures is dependent on the DNA damage checkpoint. These mutants should also exhibit defects in genome stability. To test this we examined mitotic loss rate of pDK243, a minichromosome containing a single ARS. The nse4-2 ts mutant was significantly impaired in its ability to support minichromsome maintenance, with a plasmid loss rate six times higher than wild-type at 32°C (Fig. 6A and B). Minichromosome loss has previously been associated with defects in replication initiation, and in such cases addition of multiple ARS elements suppresses the mitotic loss rate (Hogan and Koshland, 1992). In nse4ts mutants, however, the elevated rate of plasmid loss was not suppressed significantly by the addition of multiple ARS elements to the pDK243 plasmid (Fig. 6A and B), suggesting a role for Nse4 in maintaining integrity of the minichromosome, as opposed to a role associated with initiation of replication. Furthermore, we monitored S phase progression in wild-type and nse4-4ts cells by synchronizing cells via G1 phase arrest using α-factor at 25°C, and then release through removal of α-factor at 37°C. As expected, monitoring subsequent DNA replication by FACS revealed that S phase progression in nse4-4ts cells was not defective (data not shown). This, in combination with our PFGE data showing completion of S phase by nse4ts mutants at non-permissive temperature, indicates that loss of Nse4 function leads to a decrease in genomic stability – which is not caused by either defective initiation of DNA replication or defective progression of DNA replication per se.

Figure 6.

Elevated mitotic loss of minichromosomes displayed by nse4-2 ts. The pDK243 and pDK368-7 minichromosomes, containing one and seven ARS inserts, respectively, were transformed into the wild-type (A) and nse4-2 ts strains (B). Cells were grown to exponential phase at 25°C. The initial fraction of cells that contained the plasmid was determined by plating dilutions of the culture onto non-selective and selective media. Then, cells were diluted into non-selective media and grown at 25°C, 30°C or 32°C. After 10 generations, the final fraction of cells containing the plasmid was determined by plating aliquots on media that were selective and non-selective for the minichromosomes. Mitotic loss rate is plotted against the temperature at which cells were incubated in non-selective media.

Two-hybrid screening with Nse4 reveals interaction with the product of the ORF YDR288w

To identify novel Nse4 interactors, a bait plasmid encoding full-length Nse4 fused to the GAL4 DNA binding domain, was used to screen a library of S. cerevisiae genomic DNA fragments fused to the GAL4 transcription activation domain (GAL4-AD). The screen was carried out twice, with only ‘double hits’, i.e. those identified in both screens, being counted as positive interacting partners. Only one interactor was identified, the product of YDR288w, an ORF of previously unknown function. The interaction between Nse4 and Ydr288W was particularly strong, promoting rapid growth in the presence of 4 mM 3-AT, a drug used to titrate strength of interaction, as it is an inhibitor of the HIS3 reporter product. The two-hybrid interaction between Nse4 and Ydr288W was also revealed by the S. cerevisiae proteome wide two-hybrid screen (Uetz et al., 2000).

A second (reverse) screen, this time using GAL4AD-YDR288w as bait, identified 21 possible interactors. Generation of false-positives is a drawback of the two-hybrid screen. For this reason we screened all 21 interactions against increasing concentrations of 3-AT, and we present the five interactions that were maintained at the highest 3-AT concentration (Table 2). Not surprisingly, Nse4, the bait from the initial forward screen is one of these interactors. One of the interactors, YCL073c, is a putative membrane protein the function and cellular location of which are unknown. The remaining three interactors are of known function. Ubp9 is a ubiquitin carboxyl-terminal hydrolase and Rei1 is a recently identified component of a mitotic signalling network required for isotropic bud growth (Iwase and Toh-e, 2004). The remaining interaction is with Nse1, previously identified as a component of the Smc5/6 complex. Smc6 was originally isolated as the product of rad18+ from fission yeast, as part of a collection of radiation-sensitive mutants (Lehmann et al., 1995). Point mutations in rad18/Smc6 lead to defective DNA repair pathways; in addition, rad18/Smc6 is required to maintain cell cycle arrest in the presence of DNA damage (Verkade et al., 1999). Our data suggest a role for Nse4 in DNA damage repair, so our subsequent work focused on determining if there was a functional interaction between Nse4 and the known members of the Smc5/6 DNA repair complex.

Table 2.  Strong two-hybrid interactions with Ydr288w/Nse3.
ORFGeneFunction
  1. A ‘bait’ fusion to the Gal4 DNA binding domain (BD) was screened against an array of yeast Gal4 activation domain (AD) protein fusions that represent the entire yeast proteome. Interacting partners were identified as reproducible two-hybrid positives that were observed twice. Only interactors that allowed growth in the presence of high 3-AT levels are shown (3-AT is an inhibitor of the reporter product for this screen).

YCL073C Protein of unknown function
YDL105W QRI2 (NSE4)Member of SMC5–6 DNA repair complex
YER098W UBP9 Ubiquitin carboxyl-terminal hydrolase
YLR007W NSE1 Forms a complex with Smc5 and Rhc18 (Smc6); required for DNA repair
YBR267W REI1 Required for isotropic bud growth

Functional interactions between members of the Smc5/6 complex

Overexpressing Smc5, one of the components of the core Smc5/6 heterodimer, rescued the ts phenotype of nse4-1ts, though overexpression of other members of the Smc5/6 complex (Mms21, Nse1 and Smc6) did not. We repeated this analysis, using the nse4-4ts mutant. In this case, the growth defect at 37°C was rescued by Nse1 (Fig. 7A). The product of the YDR288w ORF is also a known component of the S. cerevisiae Smc5/6 complex (Hazbun et al., 2003). Overexpresssion of YDR288w rescues the ts phenotype of both nse4-1ts and nse4-4ts.

Figure 7.

Nse4 is a component of the Smc5–6 complex.
A. YDR288w/NSE3, NSE1 and SMC5 are muticopy suppressors of nse4ts mutants. The nse4-1ts and nse4-4ts mutants were transformed with multicopy vectors bearing the ORFs indicated under the control of the MET25 promoter. Gene expression via MET25 was induced by inoculating on selective media that did not contain methionine; the plates were incubated at 37°C for 3 days.
B. Composite schematic combining our two-hybrid data with two-hybrid and TAP-tag interaction data described elsewhere. Two-hybrid interactions identified from work presented here are indicated by double-headed solid lines, with the two-hybrid bait represented by a solid circle. Proteins identified by copurification of TAP tagged Ydr288w/Nse3 are circled (Hazbun et al., 2003). The two-hybrid interaction between Nse4 (used as bait) and Ydr288W/Nse3 has also been described previously (Uetz et al., 2000). Functional interactions from Fig. 7A are represented by solid bars.

Overexpression of NSE1, SMC5 or YDR288w did not rescue the ts phenotype of nse4-2ts, most likely because of this allele bearing the mutation with the most potent debilitating effect on function, as evidenced by nse4-2 ts exhibiting the poorest growth rate at permissive temperature (Fig. 2A).

Discussion

Nse4 is essential for progression of the cell cycle

Under non-permissive conditions, conditional nse4 mutants arrest as large budded cells with a 2C nucleus at the bud neck and a short mitotic spindle. The 2C nucleus represents a genuine post-S-phase arrest, as replicative intermediates are not detected when chromosomes are resolved by PFGE. In addition, there is no chromosome breakage as we do not detect sheared DNA by PFGE. The phenotype of nse4ts mutants at non-permissive temperatures has the hallmarks of an arrest caused by activation of the DNA damage checkpoint. In budding yeast damage induced arrest occurs prior to anaphase but is often termed G2/M arrest.

Phosphorylation of the effector checkpoint kinase Rad53 is clearly detectable in arrested nse4ts cells. We demonstrated a requirement for an intact RAD24 checkpoint for the nse4ts cell cycle arrest. This in combination with our FACS and PFGE data, showing complete replication of DNA in arrested cells, suggests Rad53 activation in nse4ts cells is more likely a consequence of activation of the damage checkpoint and not the intra-S checkpoint. Rad53 phosphorylation infers that loss of Nse4 function leads to disturbance in chromosomal structure. Not surprisingly all four of our nse4ts mutants are hypersensitive to agents that generate aberrant DNA structures, namely stalled replicative intermediates generated by the replication inhibitor HU, and DNA damage caused by the methylation agent, MMS.

Physical interaction of Nse4 with components of the Smc5/Smc6 DNA repair complex

There are numerous complexes involved in DNA repair. Nse4 may play a role in the function of one of these, and we anticipated that two-hybrid screening would identify the complex. Use of Nse4 as bait reveals a strong interaction with Ydr288w, a protein of unknown function. As expected, the reverse screen using Ydr288w as bait, picks up Nse4 and a strong interaction with Nse1 (Non-Smc element 1) a component of a known DNA repair complex, Smc5/6 (Lehmann et al., 1995; Verkade et al., 1999). As is the case for nse4ts alleles, all mutants in the core Smc5/6 complex plus mutants in the known non-Smc elements, confer sensitivity to DNA damaging agents. This includes mutations in: S. cerevisiae SMC6/RHC18 and its S. pombe orthologue, rad18 (Lehmann et al., 1995; Onoda et al., 2004), the S. cerevisiae and S. pombe NSE1 orthologues (Fujioka et al., 2002; McDonald et al., 2003) and NSE2 orthologues (Prakash and Prakash, 1977; McDonald et al., 2003). Additionally, like nse4-2 ts cells, smc6ts mutants also exhibit a mitotic instability phenotype (Onoda et al., 2004). Consistent with a nuclear role, GFP tagged Smc5, Smc6, Nse4, Ydr288w and Mms21 (the S. cerevisiae orthologue of S. pombe Nse2) all localize to the nucleus, as does Nse1 (Fujioka et al., 2002; Huh et al., 2003; McDonald et al., 2003).

The mutant phenotype data plus data from our two-hybrid screens suggest a role for Nse4 in Smc5/6 function, which was corroborated by functional interaction between Nse4 and members of the Smc5/6 complex. Growth arrest of nse4-1ts was suppressed by overexpression of SMC5; nse4-4ts growth arrest was suppressed by overexpression of NSE1. That neither SMC5 nor NSE1 overexpression rescued both mutants is intriguing, undoubtedly reflecting the different positions of the amino acid substitutions in the two nse4ts mutants and consequently the role these sites play in the function of the Smc5/6 holocomplex. Both nse4ts mutants were rescued by overexpression of YDR288w. A very recent publication has identified the S. pombe orthologue of Ydr288w as a component of the Smc5/6 complex; this S. pombe orthologue has been named Nse3 (Pebernard et al., 2004). This protein is essential in both budding and fission yeast. In S. pombe, Nse3 is necessary for mitotic chromosome segregation and cellular resistance to genotoxic agents. We have shown that all of these properties are shared by cells lacking Nse4, the S. cerevisiae binding partner for Ydr288w/Nse3. Furthermore, the interaction between Nse4 and Ydr288w/Nse3 is conserved in eukaryotes as the recently completed two-hybrid interaction map for Drosophila melanogaster reveals an interaction between the fly orthologues of these proteins (Giot et al., 2003). We propose renaming YDR288w as NSE3 (non-SMC element 3), in keeping with the name proposed for the S. pombe orthologue of this protein by Pebernard and coworkers. We have registered the name NSE3 as an alias for YDR288w, at the SGD.

Pebernard and coworkers point out that Nse3/Ydr288w is orthologous to the melanoma antigen (MAGE) family in humans. The function of MAGE has been obscure, until the work presented by Pebernard et al. proposed a role in the Smc5/6 DNA repair complex. Humans contain over 20 members of the MAGE family, though Drosophila and yeast contain one (Pebernard et al., 2004). In humans, MAGE has been implicated in cell cycle regulation and inhibition of apoptosis (Barker and Salehi, 2002).

Recent data from proteomic studies substantiates the role we suggest for Nse4 in Smc5/6 function. TAP tagged Nse3/Ydr288w copurifies with a complex containing Nse4, Smc5, Smc6, Nse1 and Mms21 (the S. cerevisiae orthologue of S. pombe Nse2) (Hazbun et al., 2003). Some of these proteins are undoubtedly the orthologues of the unidentified non-Smc proteins in S. pombe that bound to epitope tagged Smc6 (Fousteri and Lehman, 2000). Our two-hybrid data, plus two-hybrid and TAP-tag interaction data generated by Hazbun and coworkers, demonstrate that Nse4 and Nse3/Ydr288w are physically part of the Smc5/6 complex. We have also shown that Nse4 plays a functional role in this complex, as overexpression of Smc5/6 components suppresses the phenotype of nse4 conditional mutants. A composite schematic combining our two-hybrid data with two-hybrid and TAP-tag interaction data described elsewhere, plus our functional interaction data, is shown in Fig. 7B. Recent data on Rad62, the S. pombe orthologue of NSE4 are in agreement with our results, indicating a role in DNA repair. The S. pombe rad62-1 mutant is hypersensitive to genotoxic agents, and is defective in repair of double-strand breaks; in addition, Rad62 coimmunoprecipitates with epitope tagged Smc5 (Morikawa et al., 2004).

Our two-hybrid screen also revealed interactions between Nse3/Ydr288w and a putative membrane protein (YCL073c) a ubiquitin carboxyl-terminal hydrolase (Ubp9) and a component of a mitotic signalling network required for isotropic bud growth (Rei1). None of these proteins are involved in DNA repair or maintenance of higher-order chromosome structures, so functional interaction with Nse3/Ydr288w is not immediately apparent. Our two-hybrid screen using Nse4 as bait identified Nse3/Ydr288w as the only binding partner. This interaction was also revealed by a genome-wide two-hybrid screen, which also identified seven other binding partners (Uetz et al., 2000). Three of these, Hsc82, Sti1 and Cpr6, are components of the Hsp90 chaperosome (Pearl and Prodromou, 2000). This chaperosome is responsible for creating and maintaining the active conformation of key regulatory proteins, and may play a similar role in maintenance of Smc5/6 holocomplex activity. The remaining Nse4 interactors were a repressor of transcription (Tup1) a component of a MAP kinase cascade that responds to osmotic shock (Ssk2) and an enzyme that catalyses the first step of GMP biosynthesis (Imd2). A functional interaction between these proteins and Nse4 is not immediately apparent.

Overall, the function of SMC complexes is to establish the higher-order structure of chromosomes. Smc1/3 (cohesin) is the glue that ensures sister chromatids are maintained as a pair; Smc2/4 (condensin) mediates mitotic chromosome condensation (Hirano, 2000). Higher-order function of the Smc5/6 heterodimer remains poorly understood. Overall, it seems maintenance of higher-order structures is of importance to DNA repair and replication. Mutations in cohesin subunits lead to increased sensitivity to genotoxic agents (Tatebayashi et al., 1998; Walowsky et al., 1999). The same is true for condensin subunits (Aono et al., 2002; Chen et al., 2004). Studies in S. pombe indicate a role for Smc5/6 in homologous recombination based DNA repair (Lehman et al., 1995). A conditional mutant of S. cerevisiae SMC6 is unable to induce interchromosomal recombination in response to DNA damage (Onoda et al., 2004). However, both SMC5 and SMC6 (like NSE4) are essential for progression of the undisturbed cell cycle, in the absence of extrinsic agents that damage DNA. It has been suggested that the Smc5/6 complex holds together broken DNA molecules in the vicinity of double-strand breaks, so that repair by recombination is allowed to take place (Fousteri and Lehman, 2000). This is plausible given that occasional double-strand breaks occur during DNA replication (Muris et al., 1996). It is not surprising that a complex involved in maintaining higher-order chromosome structure plays a role in recombination. In human cells, Smc1/3 promotes repair of DNA lesions by homologous recombination (Jessberger et al., 1996). Indeed, Smc5/6 seems to facilitate various DNA repair processes, which would explain why S. pombe smc6 mutants are sensitive to a wide range of DNA damaging agents (Harvey et al., 2004). Furthermore mutants in smc6 are synthetically lethal with a mutation in DNA topoisomerase 2 (Verkade et al., 1999). All of this suggests that Smc5/6, like the Smc1/3 and Smc2/4 complexes, plays a role in chromatin organization. In cells treated with the genotoxic agents HU and  MMS, nse4ts  cells  lose  viability.  It  is  not  activation of a checkpoint that is lost here, because the Rad53 effector kinase is activated in the mutants, even at non-permissive temperatures. HU causes stalling of replication forks, which also occurs when S phase cells encounter DNA alkylated by MMS (Tercero et al., 2003). Cell death under these circumstances implies a role for stabilization of stalled forks by Nse4, and by the Smc5/6 complex as a whole. The association of human Smc5/6 with DNA during interphase, and its exclusion from DNA during mitosis, is in agreement with this (Taylor et al., 2001).

Unlike mutants in smc5 and smc6, nse4 mutants arrest at a discrete stage in the cell cycle

In S. pombe, cells deleted for either Smc5 or Smc6 display a range of lethal terminal phenotypes such as elongated cells, many with several septa with a single nucleus, or a dispersed distribution of DNA or a cut phenotype. Chromosome missegregation has also been reported for a conditional smc6 mutant in S. cerevisiae (Onoda et al., 2004). Similar phenotypes are exhibited by the ts mutants in the non-Smc subunits, Nse1 and Nse2 (Fujioka et al., 2002; McDonald et al., 2003). Our nse4ts mutants, however, display a terminal phenotype that is far less extreme, namely a growth arrest prior to onset of anaphase. Moreover, this arrest is not lethal, even when the cells are incubated at non-permissive conditions for 3 days. Also, the nucleus is not fragmented at all, and is arrested at a stage prior to anaphase. In FACS analysis of smc6 ts mutants, the G2 peak becomes broader and flatter after 10 h at non-permissive conditions, reflecting the chromosome missegregation that occurs (Onoda et al., 2004). This does not occur in nse4ts mutants, the G2 peak persisting for at least 24 h. If the Smc5/6 complex repairs a distinct type of genomic lesion, characterization of this lesion will be easier in nse4ts mutants. Moreover, the catastrophic terminal phenotype of Smc5/6 mutants implies a collapse in the structure of the complex itself, leading to amorphous distribution of nuclear DNA and ‘cut’ morphology. This is not surprising because Smc5/6 forms the core of the complex. In nse4ts mutants, loss of function does not result in large-scale chromosomal aberrations, implying the central role of the Smc5/6 core remains intact. The nse4ts mutant phenotype may represent a ‘snapshot’ of the Smc5/6 mechanism of action.

In S. pombe, Smc5/6 functions in tandem with Rad60 (Morishita et al., 2002; Boddy et al., 2003). The budding yeast orthologue of Rad60 is Esc2, named for its role in chromatin silencing (Dhillon and Kamakaka, 2000). We have noticed that we are unable to synchronize the most severe nse4ts mutant (nse4-2 ts) with α-factor. This could be because of a silencing defect, giving rise to simultaneous expression of MATα and MATa genes. It is possible that Nse4 is involved in repair of damage, such as double-strand breaks, perhaps through assembly of newly repaired DNA into chromatin. We note that mutants in components of the Smc5/6 complexes exhibit deficiencies in recombination. PFGE of these cells would reveal sheared DNA. However, we do not see sheared DNA in arrested nse4ts cells. This is why we suggest a possible role for Nse4 in chromatin assembly post DNA repair. Moreover, S. cerevisiae which lack Asf1, a protein that mediates chromatin assembly, are – like nse4ts cells – sensitive to MMS and HU, and cannot be synchronized with α-factor (Tyler et al., 1999; Hu et al., 2001). The next key issue to be addressed is the exact nature of the chromosomal aberration that exists in arrested nse4ts cells. We propose to do this by investigating genetic and biochemical interaction between Nse4 and proteins involved in recombination repair and chromatin assembly.

Experimental procedures

Yeast strains, media and plasmid construction

All yeast strains and plasmids used in this study are listed in Tables 1 and 3 respectively. Combinations of mutants in the same strain were constructed by standard procedures, yeast strains were cultured on rich (YPD) or synthetic (SD) glucose supplemented with the appropriate amino acids and nucleic acid bases (Rose et al., 1990). NSE4 was expressed via its native promoter by sequential cloning of DNA fragments into the S. cerevisiae/Escherichia coli shuttle vector YCplac33 (Gietz and Sugino, 1988). First, a 414 bp Pst I-HindIII fragment containing the ADH1 transcription terminator from pGBT9 (Clontech) was ligated into Pst I-HindIII-cleaved YCplac33, to give YCplac33T. The NSE4 ORF and its promoter were amplified separately from S. cerevisiae genomic DNA and cloned sequentially into YCplac33T, to give pBP01. The NSE4 promoter and ORF from this vector were then subcloned into the centromeric LEU2 vector YCplac111T, to give pBP02; construction of YCplac111T itself is described elsewhere (Panaretou et al., 1999)

Table 3.  Vectors used in this study.
PlasmidInsert/MarkerPromoterCopy no.Source
pBP01 NSE3/URA3 NSE3 CENThis study
pBH1 nse3-1 ts /URA3 NSE3 CENThis study
pBH2 nse3-2 ts /URA3 NSE3 CENThis study
pBH3 nse3-3  ts /URA3 NSE3 CENThis study
pBH4 nse3-4 ts /URA3 NSE3 CENThis study
pBH10 NSE3/LEU2 NSE3 CENThis study
pBH101 YDR288W/LEU2 MET25 2 µmThis study
pBH102 MMS21/LEU2 MET25 2 µmThis study
pBH103 NSE1/LEU2 MET25 2 µmThis study
pBH104 SMC5/LEU2 MET25 2 µmThis study
pBH105 RHC18(SMC6)/LEU2 MET25 2 µmThis study
pBH106 FLJ20003/LEU2 MET25 2 µmThis study
pBH107 FLJ20003/LEU2 NSE3 2 µmThis study
pDK2431ARS/LEU2/ADE3   Hogan and Koshand (1992)
pDK368-77ARS/LEU2/ADE3   Hogan and Koshand (1992)
pBDC-Nse3 YDL105W/NSE3  2 µmThis study
Pbdc-288w YDR288W/NSE4  2 µmThis study

Isolation of ts alleles of NSE4

The centromeric plasmid bearing a LEU2 marker and the NSE4 ORF ligated to its native promoter (pB02), was subjected to hydroxylamine mutagenesis in vitro (Sikorski and Boeke, 1991). Pools of mutants, were transformed into a haploid S. cerevisiae strain (BP12) deleted for NSE4. Viability of this strain was maintained by a centromeric plasmid (pB01) bearing a URA3 marker and the NSE4 ORF ligated to its native promoter. More than 8000 transformants were picked and incubated overnight at 25°C in 384 well plates containing minimal (SD) media without leucine and uracil. Using a 384-pin high density replica tool (Fisher Scientific) cells were transferred to two SD plates without leucine containing 5-FOA (Rose et al., 1990). One plate was incubated at 25°C, the other at 37°C. The 5-FOA permitted growth of cells that acquired the mutated plasmid but had lost the wild-type plasmid. Transformants that did not grow at 37°C but grew at 25°C were selected. To ensure these ts phenotypes were attributed to mutated DNA, plasmids from ts mutants were recovered, transformed into the BP12 S cerevisiae strain and rescreened for the ts phenotype. Nse4 coding sequences, from plasmids that continued to confer the ts phenotyope after rescreening, were subcloned into a centromeric URA3 plasmid, followed by transformation into a diploid strain (10590D, from the EUROSCARF collection) that is a heterozygous delete at the NSE4 locus. Following sporulation and tetrad dissection, haploids bearing the mutated plasmid in the NSE4 delete background were screened for the ts phenotype. The coding regions of nse4ts mutants were sequenced with the Taq DyeDeoxy Terminator Cycle Sequencing kit (Applied Biosystems) and a model 370A automated sequencer (Applied Biosystems).

Sensitivity against DNA damaging agents

To determine sensitivity to HU or MMS, exponential cultures were adjusted so that they were at equal cell density (1 × 107 cells ml−1) and four successive sixfold serial dilutions were spotted on YPD, or YPD containing either 50 mM HU or 0.025% MMS. Plates were incubated at 25°C, 30°C or 32°C for 3 days.

To determine viability in HU, cells were grown at 25°C to log phase in YPD, and HU was added to a final concentration of 200 mM. Aliquots were removed and spread on YPD plates every 2 h. Plates were incubated at 25°C for 3 days. The percent survival compared with that at time zero was determined by counting the resulting colonies. For viability after treatment with MMS, cells grown in YPD to log phase were incubated in 0.025% MMS, aliquots were removed every 2 h and the MMS was inactivated by addition of sodium thiosulphate to a final concentration of 5%. Aliquots were spread on YPD, followed by incubation at 25°C. Colony numbers were determined after 3 days. Survival was expressed as a percentage of the number of colonies obtained from cells that were not treated with MMS.

Fluorescence activator cell sorter (FACS) analysis

Standard procedures were used for FACS analysis, using propidium iodide to stain DNA followed by analysis via FACScan (Becton Dickinson) (Chakraverty et al., 2001).

Pulse field gel analysis

Isogenic strains bearing the ts mutant or wild alleles of NSE4 were grown in YPD to log phase at 25°C and then incubated at 37°C for 4 h. Control samples were (i) aliquots of the same cultures maintained at 25°C or (ii) wild-type cells treated with either α-factor (5 µg ml−1) for 3 h or HU (200 mM) for 2 h. Yeast chromosomal DNA was prepared by a standard procedure, involving embedding cells in agarose followed by treatment of the agarose plugs with proteinase K and lyticase (Hennessy et al., 1991). PFGE was carried out in 1% megabase-agarose (Bio-Rad) in a CHEF-DRTM II electrophoresis cell (Bio-Rad). Electrophoresis was performed for 24 h with a switching time of 70 s at 170 V in 0.5% Tris-borate-EDTA. The separated chromosomes were then acid-nicked and transferred to nylon membrane (Hybond N+, Amersham-Pharmacia) (Hennessy et al., 1991). The blot was probed with DIG-labelled SAC6 fragment and detected with CDP-StarTM kit (Boehringer-Mannheim). This probe to chromosome IV was chosen because analysis of replication state by PFGE is most sensitive when probes are derived from large chromosomes (Hennessy et al., 1991).

Immunofluorescence microscopic analysis

Cells were processed for fluorescence and indirect fluorescence (Bowers et al., 2000). Fixed cells were stained for DNA using DAPI (25 µg ml−1), and for the mitotic spindle, using rat monoclonal anti-tubulin antibody (YOL1/34, abcam) and FITC-conjugated goat anti-rat antibody (abcam) using established methods (Hisamoto et al., 1994).

Protein extraction and Rad53 detection

Yeast protein extracts were prepared from TCA-treated cells as described previously (Pellicioli et al., 1999). Rad53 was detected with a rabbit polyclonal antibody provided by J. Diffley (Tercero et al., 2003).

Assay for mitotic loss of minichromosomses

Plasmid loss assays were performed using the minichromosome vectors pDK243 or pDK368-7, bearing one or seven ARS elements respectively (Hogan and Koshland, 1992). Selective pressure was maintained by presence of a LEU2 marker. The vectors were transformed into (i) the haploid BP12 S. cerevisiae strain, which is deleted for NSE4 but bears a plasmid with NSE4 and the URA3 marker and (ii) an isogenic strain except the vector bears the nse4-2ts mutant. The initial fraction of cells that contained the plasmid was determined by plating dilutions of the culture onto SD medium with or without leucine. Cells were subsequently diluted into YPD and grown at 25°C, 30°C or 32°C. After 8–10 generations, the final fraction of cells containing the plasmid was determined by plating aliquots on media that were selective (SD media without leucine) and non-selective (YPD) for the minichromosomes. The loss rate was calculated as previously described (Hogan and Koshland, 1992).

Yeast two-hybrid screen

NSE4 and [Nse3/YDR288w] C-terminal ‘bait’ fusions to the Gal4 DNA binding domain (BD) of the pBDC TRP1 vector were generated by homologous recombination within the yeast strain PJ69-4α as described previously (Uetz et al., 2000; Millson et al., 2003). Presence of a BD fusion was confirmed by Western blotting, using anti-Gal4p BD antiserum (Clontech). The bait fusions in PJ69-4α did not self-activate transcription of the integrated GAL1 promoter/HIS3 fusion reporter, as shown by lack of growth on medium containing 2 mM 3-amino-1,2,4-triazole (3-AT), an inhibitor of the HIS3 reporter product.

PJ69-4α bearing (BD) bait fusions were mated by pinning onto the previously described 16-plate, 384 format array of yeast Gal4 activation domain (AD) protein fusions, all of which are carried in a strain of the opposite mating type (PJ69-4a) (Uetz et al., 2000). Protein–protein interactions were identified by pinning the diploids onto medium lacking leucine, tryptophan and histidine, supplemented with 4 mM 3-AT (Uetz et al., 2000). Colonies were scored for growth after 10 days at 30°C. Putative interacting partners were identified as reproducible two-hybrid positives that were observed twice from duplicate screens. It was assumed that interacting proteins identified only once were the result of a non-reproducible false-positive.

Assessing functional interaction between putative members of the Smc5/6 complex by multicopy suppression

Candidate ORFs were overexpressed from the vector YEp81Met, which is a modification of the episomal LEU2 vector YEplac181. The vector bears the MET25 inducible promoter and the transcription termination site from PGK1, separated by a multiple cloning site. Candidate ORFs were amplified from S. cerevisiae genomic DNA and ligated 3′ to the MET25 promoter. The resulting vectors were transformed into the BP12 strain, or an isogenic strain containing a vector borne nse4ts allele instead of the wild-type gene. Transformants were streaked on SD plates without uracil and leucine, and without methionine to induce overexpression of the candidate ORF.

Acknowledgements

We thank Douglas Koshland (Carnegie Institution of Washington) for providing the pDK243 and pDK368-7 vectors; Frank Cooke (University College London) for providing the YEp81Met vector and John Diffley (CRUK, Clare Hall) for providing antisera to Rad53. We thank Katherine Bowers (CIMR, Cambridge) for assistance with immunofluorescence, Lucy Drury (CRUK, Clare Hall) for assistance with detection of phosphorylated forms of Rad53 and David Barford and Lori Passmore (Institute of Cancer Research, London) for valuable comments. This work was supported by the Central Research Fund of the University of London, the Royal Society and the BBSRC (Grant No. BBS/B/04951). B.H. was partly supported by a doctoral scholarship funded by the KC Wong Foundation and the China Scholarship Council.

Ancillary