Saccharomyces cerevisiae LIF1: a function involved in DNA double‐strand break repair related to mammalian XRCC4

Gernot Herrmann, Tomas Lindahl, Primo Schär

Author Affiliations

  1. Gernot Herrmann1,
  2. Tomas Lindahl*,1 and
  3. Primo Schär2
  1. 1 Imperial Cancer Research Fund, Clare Hall Laboratories, South Mimms, Hertfordshire, EN6 3LD, UK
  2. 2 Institute of Medical Radiobiology, University of Zürich, CH‐8029, Zürich, Switzerland
  1. *Corresponding author. E-mail: lindahl{at}
View Full Text


Saccharomyces cerevisiae DNA ligase IV (LIG4) has been shown previously to be involved in non‐homologous DNA end joining and meiosis. The homologous mammalian DNA ligase IV interacts with XRCC4, a protein implicated in V(D)J recombination and double‐strand break repair. Here, we report the discovery of LIF1, a S.cerevisiae protein that strongly interacts with the C‐terminal BRCT domain of yeast LIG4. LIG4 and LIF1 apparently occur as a heterodimer in vivo. LIF1 shares limited sequence homology with mammalian XRCC4. Disruption of the LIF1 gene abolishes the capacity of cells to recircularize transformed linearized plasmids correctly by non‐homologous DNA end joining. Loss of LIF1 is also associated with conditional hypersensitivity of cells to ionizing irradiation and with reduced sporulation efficiency. Thus, with respect to their phenotype, lif1 strains are similar to the previously described lig4 mutants. One function of LIF1 is the stabilization of the LIG4 enzyme. The finding of a XRCC4 homologue in S.cerevisiae now allows for mutational analyses of structure–function relationships in XRCC4‐like proteins to define their role in DNA double‐strand break repair.


Double‐strand breaks in DNA constitute a significant threat to the stability of cellular genomes, and highly efficient repair pathways have evolved to counteract this form of DNA damage. In eukaryotes, two pathways for double‐strand break repair (DSBR) can be distinguished, homologous recombination and non‐homologous end joining (NHEJ). In mammalian cells, NHEJ seems to be the major pathway, whereas yeast preferentially employs homologous recombination for DSBR. Nevertheless, yeast cells possess the ability to perform NHEJ, but the full biological role of this pathway remains to be elucidated. Some of the components involved in NHEJ, like the Ku proteins and DNA ligase IV, are conserved between yeast and man, whereas others, e.g. DNA‐dependent protein kinase (DNA‐PK) or the XRCC4 protein, have not yet been detected in yeast. Thus, the latter may be either highly diverged or not present at all in yeast.

Human XRCC4 was identified by its ability to complement radiation sensitivity and a deficiency in V(D)J recombination of hamster XR‐1 cells, which have been shown to lack both alleles of the XRCC4 gene (Stamato et al., 1983; Li et al., 1995). The XRCC4 gene encodes a polypeptide with a predicted molecular mass of 38 kDa which occurs as a phosphoprotein and migrates at ∼55 kDa on SDS–PAGE (Critchlow et al., 1997; Grawunder et al., 1997; Mizuta et al., 1997). XRCC4 appears to be conserved in mammals, with human and mouse sequences sharing ∼75% identity, but homologues have not been detected in other eukaryotes. Recently, XRCC4 was shown to interact with the C‐terminal region of human DNA ligase IV and to stimulate double‐strand break joining efficiency of this enzyme in vitro (Critchlow et al., 1997; Grawunder et al., 1997). On the basis of the XR‐1 cellular phenotype, these results indicated that a complex involving DNA ligase IV and XRCC4 might be responsible for joining of double‐strand breaks generated by DNA‐damaging treatments or during V(D)J recombination in mammalian cells.

DNA ligase IV has been isolated as one of at least three distinct ATP‐dependent DNA ligases in mammalian cells (Wei et al., 1995; Robins and Lindahl, 1996). These enzymes share highly conserved catalytic core domains but differ completely in their N‐ and C‐terminal regions. These extra‐catalytic domains appear to be required for interactions with other proteins and/or might target the DNA ligases to different DNA metabolic pathways. DNA ligase I is an essential replication factor; its main function is the joining of Okazaki fragments during lagging‐strand DNA synthesis, but it may also play a role in some forms of DNA repair and recombination (Barnes et al., 1990; Waga et al., 1994; Petrini et al., 1995; Mackenney et al., 1997). DNA ligase III interacts strongly with the DNA repair protein XRCC1 and accounts for the final joining step in the major pathway of base excision repair (Kubota et al., 1996; Cappelli et al., 1997; Nash et al., 1997). An alternatively spliced form of DNA ligase III is present in testis and may be active in meiotic recombination (Mackey et al., 1997), and DNA ligase II is another smaller form of DNA ligase III.

Direct evidence for an involvement of DNA ligase IV in DSBR became available with the recent identification of Saccharomyces cerevisiae LIG4, a gene encoding a homologue of mammalian DNA ligase IV. From investigations of the lig4 mutant phenotype in yeast, we and others could show that unlike the yeast DNA ligase I homologue, CDC9, LIG4 is not essential for growth but is involved in non‐homologous DSBR; in addition, it has a function in meiosis (Schär et al., 1997; Teo and Jackson, 1997; Wilson et al., 1997). Similarly to mammalian DNA ligase IV, the yeast enzyme has a characteristic C‐terminal region consisting of two tandemly arrayed BRCT domains which have been implicated in protein–protein interactions (Callebaut and Mornon, 1997; Critchlow et al., 1997; Nash et al., 1997).

To dissect further the DNA ligase IV‐dependent DSBR pathway and its biological role in S.cerevisiae, we set out to isolate and characterize proteins physically interacting with yeast LIG4. Here, we describe the functional analysis of a so far uncharacterized yeast gene, LIF1 (ligase‐interacting factor 1), the product of which specifically binds to the C‐terminal BRCT domain of LIG4. This interaction involves a part of LIF1 which shows sequence homology with the mammalian XRCC4 protein. Genetic examination of the lif1 mutant phenotype revealed a defect in non‐homologous DSBR.


Identification and isolation of LIF1

For two‐hybrid screening, we constructed five bait vectors, each containing a part of LIG4 fused to the DNA‐binding domain (BD) of the yeast GAL4 transcriptional activator (Figure 1A). The constructs LIG4(N) and LIG4(C) were used for initial screening of a S.cerevisiae cDNA library whereas the others were employed for subsequent confirmation and further characterization of identified interactions.

Figure 1.

Identification of LIF1 by yeast two‐hybrid screening. (A) Bait constructs used in this study. Five DNA fragments representing parts of the LIG4 ORF were fused to the GAL4 DNA‐binding domain in the yeast two‐hybrid vector pAS2‐1. LIG4(N) and LIG4(C) were used for screening a S.cerevisiae cDNA library cloned in the GAL4 activation domain vector pACT‐2. All other constructs were used for confirmation and characterization of identified interactions. LIF1 cDNAs were isolated with the C‐terminal LIG4(C) fragment of LIG4. (B) LIF1 interacts with the downstream BRCT domain of LIG4 as determined by quantitative β‐galactosidase assays. β‐Gal activities shown are averages obtained with at least three independent clones analysed for each vector combination. GAL‐AD, GAL4 activation domain; GAL‐BD, GAL4 DNA‐binding domain; T‐ag, SV40 large T‐antigen; p53, murine p53(72–390); pLAM5′‐1, human lamin C(66–230).

Ten positive clones were isolated from a screen of >106 S.cerevisiae cDNA clones with the LIG4(C) bait, the construct expressing both the BRCT domains of LIG4. Eight of them carried overlapping cDNA sequences from an open reading frame (ORF) on chromosome VII (YGL090w) which has coding potential for a 48.3 kDa protein. We designated this S.cerevisiae gene and its product LIF1. One of the LIF1 clones carried the coding sequence for the first 260 amino acids of the protein [pLIF1(1–260)] fused to the GAL4 activation domain (AD), whereas all others started in the region around codon 190 and extended over the entire C‐terminal part of LIF1. In quantitative β‐galactosidase assays, pLIF1(1–260) showed the strongest interaction with LIG4(C). Further two‐hybrid‐based characterization of this interaction with various LIG4 bait constructs revealed that the combination of pLIF1(1–260) and LIG4(C) produced β‐galactosidase activities as strong as our positive control (Figure 1B). High levels of β‐galactosidase activities were also induced with bait constructs expressing either the entire LIG4 ORF [LIG4(F)] or the downstream BRCT domain of LIG4 [LIG4(B2)]. Other combinations, including a series of controls but also the LIG4(N) and LIG4(B1) constructs encoding the N‐terminus and the upstream BRCT domain of the ligase, respectively, showed low to non‐detectable β‐galactosidase activities (Figure 1B). Further analyses revealed that a two‐hybrid construct expressing residues 137–274 of LIF1 as a GAL4–AD fusion protein was able to induce lacZ expression in combination with LIG4(C), but not so with two constructs containing either residues 1–135 or 278–421 of LIF1 (data not shown). These data localize the LIG4‐interacting domain in LIF1 within residues 137 and 274, a region which includes the common sequence encoded by all isolated LIF1 cDNA clones. Finally, full‐length LIF1 expressed as a GAL4–BD fusion protein induced β‐galactosidase activity also with LIG4(C) subcloned in‐frame with GAL4–AD (data not shown). No positive clones were obtained in cDNA library screening with the N‐terminal domain of LIG4 [LIG4(N)].

LIF1 is a protein related to mammalian XRCC4

The nucleotide sequence of LIF1 was confirmed by DNA sequencing and shown to be identical with ORF YGL090w in the S.cerevisiae genome database (accession No. Z72612). The LIF1 gene encodes an acidic protein of 421 amino acids (pI = 4.86) with a calculated molecular mass of 48.3 kDa. Its predicted subcellular localization is nuclear (Horton and Kenta, 1996) and contains a cluster of putative nuclear localization signals between residues 263 and 285 (Figure 2A). Preliminary database searches with the LIF1 amino acid sequence revealed no significant homologies to other known proteins. However, a direct comparison with human XRCC4 showed that LIF1 shares 22% identity and 49% similarity with this human protein (Figure 2A). Interestingly, the highest degree of sequence conservation between the two proteins was found in the region encoded by the overlapping cDNA sequences of the LIF1 clones isolated in the two‐hybrid screen (Figure 2A). Thus, yeast LIF1 and human XRCC4 are most similar in the region which is involved in the interaction of LIF1 with the C‐terminal BRCT domain of yeast LIG4.

Figure 2.

Structural relationships between S.cerevisiae LIF1 and human XRCC4. (A) Amino acid sequence alignment of the yeast LIF1 protein (scLIF1, SGD YGL090W) and human XRCC4 (hsXRCC4, accession No. U40622). Identical residues are boxed and appear dark grey. The light grey shaded box between amino acids 190 and 260 of scLIF1 marks the overlapping region encoded by all cDNA clones isolated in the two‐hybrid screen. The black line above residues 263–284 of scLIF1 indicates putative nuclear localization signals. (B) LIF1 polypeptide migrates anomalously slowly in SDS–PAGE. Recombinant His‐tagged LIF1 was overexpressed in and purified from E.coli. The panel shows the protein content in relevant nickel column fractions by Coomassie staining after SDS–PAGE. Lane 1, molecular weight standard; lanes 2 and 3, loaded and unbound proteins (2.5 μl), respectively; lanes 4–8, consecutive LIF1‐containing fractions (10 μl) eluted from the Ni‐NTA column. (C) A protein similar in size to LIF1 co‐purifies with His‐tagged LIG4 expressed in a lig4 mutant yeast strain. Shown are the relevant nickel column fractions in a silver‐stained SDS–polyacrylamide gel. Lane 1, molecular weight standard; lane 2, 20 μl of a 500 mM imidazole eluate from a nickel column loaded with extract of lig4 mutant cells carrying a LIG4‐expressing plasmid (pPRS154); lane 3, 20 μl of the corresponding fraction from an extract of the lig4 cells carrying the expression vector only (pYES2). Molecular weights are indicated on the left, His‐LIG4 is indicated on the right side and the co‐eluting band(s) at 60 kDa is marked by an arrow.

LIF1 migrated anomalously slowly during SDS–PAGE (Figure 2B). Similar observations have been made for human XRCC4 (Critchlow et al., 1997; Grawunder et al., 1997; Mizuta et al., 1997). Here, an N‐terminal His‐tagged form of the LIF1 protein (expected size, 50 kDa) overexpressed in Escherichia coli migrated at ∼60 kDa (Figure 2B). A split band of LIF1 protein occurred, implicating two slightly different forms of the protein (Figure 2B). Moreover, a protein of similar size co‐eluted with His‐tagged LIG4 protein when the latter was overexpressed in and purified from a lig4 mutant strain (pPRS154 in PRSY003,1) (Figure 2C); in addition, after initial nickel affinity chromatography, the protein co‐eluted with LIG4 in an FPLC‐Mono S step at 1 M NaCl, resisting washes with 600 mM NaCl (data not shown). A separate protein band at ∼75 kDa (Figure 4C) was identified as a proteolytic fragment of LIG4 by ligase adenylation assays and Western blot analysis (data not shown). These results indicate that yeast LIG4 and LIF1 occur as a heterodimer. Human DNA ligase IV was also isolated initially as a heterodimer from HeLa cell nuclei (Robins and Lindahl, 1996), with its partner subsequently identified as XRCC4 (Critchlow et al., 1997; Grawunder et al., 1997).

Figure 3.

Sensitivity of lif1 mutant cells to ionizing radiation. (A) Exponentially growing cultures; (B) stationary phase cultures. Cells were exposed to X‐rays as indicated at a dose rate of 10 Gy/min. Survival data are shown from at least three independent experiments. Plotted are averaged survival data with standard deviations from at least three independent experiments. Matching test and control strains were always assayed in parallel as detailed in Materials and methods. The strains used are listed with genotypes in Table II and are referred to in the figure as: wt, FF18734; lig4, PRSY003,1; lif1, PRSY031,a; lig4 lif1, PRSY033,a; rad52, FF18743; rad52 lig4, PRSY005.

LIF1 is not essential for growth but mutants show defects in radiation resistance and sporulation efficiency

A heterozygous LIF1 gene disruption was generated in a diploid LIG4/lig4::kanMX strain (PRSY030) by replacing the coding sequence for amino acids 13–383 of the LIF1 ORF with an URA3 expression cassette employing a short homology PCR strategy. URA+ transformants were examined for correct heterozygous gene replacement by PCR and genomic Southern blot analysis, and then sporulated to segregate haploid meiotic progeny. Thirty of 40 dissected tetrads produced four viable spores, with the uracil prototrophy and the geneticin resistance segregating 2:2 in 29 cases, suggesting a non‐essential function for the LIF1 gene and a viable phenotype for lig4 lif1 double mutants. After genotyping haploid segregants, isogenic pairs of lif1::URA3 mutants and lig4::kanMX4 lif1::URA3 double mutants (PRSY031–PRSY034, Table II) were isolated and subjected to further phenotypic analyses.

lif1 single and lig4 lif1 double mutants did not show any significant retardation on exponential growth or during cell cycle progression as established by parallel examination of growth rates and distribution patterns of unbudded, small budded and large budded cells in vegetative cultures (data not shown). They were neither temperature‐sensitive (up to 38°C) nor cold‐sensitive (down to 18°C). However, in complete medium (YPD), lif1 and, to a lesser extent, also lig4 mutants consistently grew to higher densities (3–4×108 cells/ml) than isogenic wild‐type cells (2–2.5×108 cells/ml), and this effect was enhanced in the lig4 lif1 double mutant which grew to the unusual density of 5–5.7×108 cells/ml under the same conditions. Examination of cell morphology in late exponential–early stationary phase populations (1×108 cells/ml) revealed an increased fraction of small budded cells in lif1, lig4 and lig4 lif1 cultures, suggesting that the mutant cells re‐initiate S‐phase more frequently than wild‐type cells under limiting growth conditions. In late stationary phase populations (after 5–7 days incubation at 30°C), the majority (>80%) of cells were unbudded in the mutant as well as in the wild‐type cultures.

The effect of LIF1 disruption on cellular resistance to DNA damage induced by ionizing irradiation was investigated. Previous analyses of the lig4 mutant phenotype by several groups had revealed an apparent discrepancy: whereas no significantly increased sensitivity to different DNA‐damaging agents was reported consistently for lig4 single mutants, increased hypersensitivity of rad52 mutants to ionizing irradiation was observed for rad52 lig4 double mutants in two studies (Teo and Jackson, 1997; Wilson et al., 1997) but not in another (Schär et al., 1997). We re‐examined this issue for the lig4 mutants and assessed the radiation sensitivity of lif1 mutants under conditions that minimize the contribution of the rad52 phenotype in the analysis, so that in the absence of homologous recombination a minor contribution to DSBR by a LIG4‐dependent pathway might be uncovered. For this purpose, we irradiated haploid cells in late stationary phase cultures, which are mainly in the G0 phase of the cell cycle. Due to the lack of sister chromatids as homologous partners, such cells are expected to be limited in their capacity for DSBR by homologous recombination and, consequently, non‐homologous DSBR pathways should be more important in cellular resistance against ionizing irradiation.

Employing different experimental strategies and varying irradiation dose rates from 2.5 to 10 Gy/min, we consistently obtained results as illustrated in Figure 3. With exponential and early stationary phase cells (⩽1×108 cells/ml, ≥50% budding), we confirmed our previous observations with lig4 and rad52 lig4 strains (Schär et al., 1997) and found that the radiation sensitivity of lif1 single and lig4 lif1 double mutants was also only insignificantly different from that of lig4 mutants and wild‐type cells (Figure 3A). In late stationary phase cultures (5–7 days at 30°C, >2×108 cells/ml), however, the situation changed. The radiation sensitivity of lif1 (3.3‐fold relative to wild‐type), lig4 (2.1‐fold) and lig4 lif1 (6.0‐fold) mutant cells was enhanced and, as expected, the contribution of RAD52 to cellular resistance was much less pronounced (Figure 3B). At X‐ray doses >100 Gy, the lig4 lif1 double mutant strain was slightly more sensitive than either of the single mutants alone, implying that in the absence of LIF1 some residual LIG4‐dependent repair can take place and vice versa. Most sensitive in the analysis was the rad52 lig4 double mutant (12.4‐fold relative to wild‐type), and the comparison with the rad52 (3.1‐fold) and lig4 (2.1‐fold) single mutants suggests that, in resting cells, RAD52‐dependent DSBR is in a synergistic relationship with the LIG4–LIF1‐dependent pathway.

Figure 4.

Saccharomyces cerevisiae lif1 mutant cells are deficient in recircularizing linearized plasmid DNA. Transformation assays were carried out using either supercoiled, EcoRI‐digested or SmaI‐digested pBTM116 plasmid (see Materials and methods). (A) Schematic map of the yeast replicative plasmid pBTM116, carrying Amp for selection in E.coli and the yeast TRP1 gene as a selectable yeast marker. A multiple cloning sequence including the relevant EcoRI and SmaI sites is located within a stretch of DNA which shares no homology with S.cerevisiae genomic sequences. This DNA is flanked by transcription control elements of yeast ADH1 (pADH1, tADH1). Numbers below the map indicate the distance in base pairs of relevant plasmid elements from the EcoRI site, and the arrows mark the annealing sites of the PCR primers used for analysis of plasmid repair events. (B) Graphic illustration of relative transformation efficiencies (ratios cut/uncut plasmids) obtained with EcoRI‐ (left panel) and SmaI‐cut plasmid (right panel). (C) Graphic illustration of relative transformation efficiencies obtained with EcoRI‐digested plasmid in yeast strains carrying either LIF1 in a galactose‐inducible expression vector or the expression vector alone. Times of galactose‐induced expression prior to transformation are indicated. Data are averages from at least three independent transformation experiments in which matching test and control strains were treated in parallel. Strains used were LIF1 LIG4, FF18734; lig4, PRSY003,1; lif1, PRSY031,a; lif1 lig4, PRSY033,a; LIF1 LIG4/YCpIF3, FF18734 carrying the expression vector only; lif1/YCpIF3, PRSY031,a carrying the expression vector only; lif1/pLIF1, PRSY031,a carrying a LIF1 expressing plasmid (pGEH014).

We also examined the effect of the LIF1 disruption on the capacity of the cells to undergo meiotic differentiation. For this purpose, we sporulated isogenic diploid strains which were either homozygous LIF1/LIF1 wild‐type (FF18734×FF18984, Table II), homozygous lif1/lif1 mutant (PRSY031×PRSY032) or heterozygous LIF1/lif1 mutant (FF19734×PRSY032), and evaluated sporulation efficiency 24 and 48 h post‐induction by counting the number of 4,6'‐diamidino‐2‐phenylindole (DAPI)‐staining bodies in at least 200 randomly selected meiocytes per time point. Similarly to lig4/lig4 diploids (Schär et al., 1997), the homozygous lif1/lif1 strain sporulated very inefficiently (29% completed meioses after 48 h) as compared with the heterozygous LIF1/lif1 mutant (62% after 48 h) or the homozygous wild‐type (69% after 48 h); >95% of lif1/lif1 mutant cells which failed to sporulate after 48 h showed a single DAPI‐staining body, implying that they did not pass meiosis I. Spore viability in the sporulating fraction of homozygous lig4 mutant cells was only marginally reduced (88%) in comparison with wild‐type cells (96%), and segregation of heterozygous marker alleles was normal as evidenced in 42 dissected spore tetrads.

LIF1 is involved in non‐homologous double‐strand break joining

The physical interaction of LIF1 with LIG4 suggests an involvement of LIF1 in non‐homologous repair of DNA double‐strand breaks. To test this possibility, we performed plasmid rescue assays as described previously (Boulton and Jackson, 1996b; Schär et al., 1997) and compared lif1 mutants with isogenic lig4 mutant and lig4 lif1 double mutant yeast strains. Briefly, the yeast plasmid, pBTM116, was linearized by restriction enzyme digestion in a region without sequence homology to chromosomal DNA (Figure 4A); Completion of digestion was verified by Southern blot hybridization. Competent S.cerevisiae cells were then transformed in parallel with limiting amounts of cut or uncut plasmid DNA, and the number of transformants was determined after selection for a plasmid‐expressed genetic marker (TRP1). As plasmid replication required for establishment of a TRP+ phenotype depends on successful religation of the linear plasmids, the relative transformation efficiency obtained with linear versus circular plasmid DNA is a measure of the DSBR capacity of a yeast strain.

Transformation data with EcoRI‐ or SmaI‐digested plasmid DNA are illustrated in Figure 4B. Wild‐type cells were highly efficient in recircularizing linearized plasmids with 5′ overlapping cohesive ends. In contrast, lif1 and, as previously shown, lig4 mutant cells showed a dramatically, 100‐fold reduced relative transformation efficiency with EcoRI‐digested plasmid, whereas transformation with supercoiled plasmid was as efficient as in wild‐type cells. The same result was obtained in several different experiments and was independent of the overall transformation efficiency when limiting amounts of DNA were used. Transformation efficiency was not reduced further in a lif1 lig4 double mutant strain, indicating that LIF1 and LIG4 act in the same pathway for non‐homologous double‐strand break joining. Transformation experiments with SmaI‐digested blunt‐ended plasmid DNA produced essentially the same result, but the effects were less pronounced (Figure 4B); this was mainly because wild‐type cells, but not the mutant cells, were transformed 40‐fold less efficiently with blunt‐ended plasmids than with cohesive‐ended plasmids.

The deficiency in recircularizing EcoRI‐cut plasmid was fully complemented in lif1 mutant cells by expression of the LIF1 ORF from yeast episomal vector pGEH014 under the control of a GAL1 promoter (Figure 4C). However, complementation was only observed after the cells were grown for several hours under GAL1‐inducing conditions (medium containing galactose and raffinose). No complementation was observed in parallel experiments with cells carrying the expression vector only (YCpIF3, Figure 4C). We have been unable to complement the reduced capacity for plasmid religation in lif1 or lig4 mutant strains with human DNA ligase IV or human XRCC4 protein overexpressed in these strains alone or in combination (data not shown).

To characterize the molecular events that led to the establishment of tryptophan prototrophy after transformation of cells with EcoRI‐ or SmaI‐digested pBTM116, we analysed wild‐type, lif1 and lig4 mutant transformants by PCR amplification of a plasmid segment spanning the putative EcoRI–SmaI junction, EcoRI or SmaI digestion and DNA sequencing of the fragments, and by plasmid stability assays and genomic Southern blot analysis. Table I shows that correct religation of the plasmid only represents a small fraction of events in the rare TRP+ transformants of either lif1 or lig4 strains. Different inefficient, mostly homologous recombination‐mediated backup pathways (gap repair, genomic integration, gene conversion) appeared to be responsible for the generation of the majority of TRP+ transformants in these strains. Thus, the actual capacity of lif1 and lig4 mutant cells to religate the plasmid correctly after transformation is even lower than estimated from the comparison of the relative transformation efficiencies obtained with wild‐type and mutant strains. Due to highly efficient plasmid religation by NHEJ in wild‐type cells, wild‐type TRP+ transformants generated through these backup pathways which are dominant in lif1 and lig4 mutant cells were only detectable in experiments with SmaI‐digested blunt‐ended plasmid DNA, which appears to be a suboptimal substrate for the LIG4‐dependent end‐joining pathway (Table I).

View this table:
Table 1. Molecular events establishing TRP+ prototrophs after transformation of yeast with linearized plasmid DNA
View this table:
Table 2. Saccharomyces cerevisiae strains used

LIF1 affects the stability of LIG4 protein

Since efficient complementation of the plasmid repair deficiency in lif1 mutant cells required several hours of GAL1 promoter‐driven LIF1 expression from a yeast episomal vector (Figure 4C), it appeared that LIF1 might have a regulatory effect on the level of its partner, LIG4. Stabilizing functions of protein–protein interactions have been reported for a variety of heterodimeric factors including the Ku80 and Ku70 subunits of the Ku protein (Chen et al., 1996).

Using purified yeast LIG4 enzyme and crude protein extracts prepared from wild‐type and lig4 mutant strains, we observed that a polyclonal antibody (TL14/15; Robins and Lindahl, 1996) directed against a peptide derived from the highly conserved core domain of human DNA ligase IV also recognizes the yeast homologue. Non‐specific cross‐reactions with other yeast proteins could be virtually eliminated by pre‐incubation of the antiserum with lig4 mutant extract (data not shown).

When Western blot analyses were performed on extracts from lif1 mutant cells (PRSY031,a), no LIG4 protein was detectable (Figure 5A). As expected, the protein was not detected in extracts of lig4 mutant (PRSY003,1) and lig4 lif1 double mutant (PRSY033,a) cells, but it was readily observed in extracts of isogenic wild‐type cells (FF18734), as was the 108 kDa His‐tagged human DNA ligase IV overexpressed (pGEH007) in a lig4 mutant strain which served as positive control (Figure 5A). After prolonged GAL1‐induced expression of LIF1 from an episomal plasmid, LIG4 protein was detected in the lif1 mutant strain, whereas in extracts of the vector control cells it remained at undetectable levels (Figure 5B). Thus, the loss of LIG4 protein is a phenotype of the lif1 mutants, and can be complemented by expression of LIF1 protein.

Figure 5.

LIG4 is destabilized in lif1 mutant cells. (A) Western blot analysis of crude cell extracts using a polyclonal anti‐human DNA ligase IV peptide antiserum for detection of yeast LIG4 protein (see Materials and methods). LIG4 protein is detectable in wild‐type cells (lane 1) but not in lig4 (lane 2), lif1 (lane 3) or lig4 lif1 (lane 4) mutant cells. Detection of human DNA ligase IV expressed in the lig4 mutant is shown as positive control in lane 5. (B) Western blot analysis of lif1 mutant cells either expressing the wild‐type LIF1 gene from an episomal plasmid (lif1/pLIF1) or carrying the expression vector only (lif1/YCpIF3). (C) Northern blot analysis of total RNA extracted from different strains (see Materials and methods). LIG4 mRNA is detectable in wild‐type and lif1 mutant strains but not in lig4 nor lig4 lif1 mutants. The lower panel shows hybridization of the same blot with 18S rRNA used as an internal standard. Strains used were wt, FF18734; lig4, PRSY003,1; lif1, PRSY031,a; lif1 lig4, PRSY033,a; lig4/hsLIG4, PRSY003,1 carrying a hsLIG4‐expressing plasmid (pGEH007).

The absence of LIG4 protein in lif1 mutant cells is not associated with down‐regulation of transcription. Figure 5C shows the result of Northern blot analysis using LIG4 and, as an internal control, 18S rRNA‐specific probes. The relative amount of LIG4 mRNA was not decreased in a lif1 mutant strain in comparison with the wild‐type strain, and it may even be slightly up‐regulated in the mutant cells. These data establish that the regulatory effect of LIF1 on LIG4 is due to protein stabilization and not to differences in gene expression. In conclusion, one function of LIF1 is to stabilize DNA ligase IV in yeast. It seems likely that mammalian XRCC4 protein might have a similar stabilizing effect on DNA ligase IV, but this has not been investigated in XR‐1 cells.

LIF1 is not involved in telomere length maintenance

Several proteins that participate in NHEJ, including the yeast Ku homologues YKU70 and YKU80, the silencing factors SIR2, SIR3 and SIR4, and the RAD50, MRE11 and XRS2 functions, have been implicated in telomere length maintenance (Boulton and Jackson, 1996a, 1998; Porter et al., 1996). In contrast, telomeres are maintained as in wild‐type cells in yeast lig4 mutants (Teo and Jackson, 1997; Boulton and Jackson, 1998). To address any involvement of LIF1, the protein partner of LIG4, we examined telomere lengths in lif1 mutants and lig4 lif1 double mutant cells by the same experimental approach as employed in previous studies (Boulton and Jackson, 1998). Figure 6 shows that after culturing cells for >100 generations, telomere lengths were not affected in lif1 mutants or lig4 lif1 double mutants as compared with wild‐type and lig4 mutant strains, whereas telomere shortening was clearly visible in the yku70 mutant. Although the yku70 mutation is in a genetic background (W303) different from that of the other strains examined (A364A), it has been shown previously that XhoI digestion of genomic DNA and Southern hybridization with a poly(GT)20 probe results in comparable telomeric restriction fragments in both strain backgrounds (Lustig and Petes, 1986; Porter et al., 1996). Thus, the present data establish that LIF1, like LIG4, is not involved in telomere length maintenance and has no obvious role in protection of telomere ends from degradation.

Figure 6.

LIF1 is not involved in telomeric maintenance. (A) Schematic illustration of telomeric regions in yeast. The relative location of the C1–3A repeats between X and Y' regions is indicated. XhoI cleavage generates telomeric fragments of ∼1.3 kb in wild‐type strains which include ∼400 bp of C1–3A repeats. (B) Telomere length is not altered in lig4, lif1 or lig4 lif1 mutant strains as compared with wild‐type strains. XhoI‐digested genomic DNA of the strains indicated was analysed by hybridization with end‐labelled poly(GT)20 (see Materials and methods). Weaker signals in lif1 and lif1 lig4 strains are due to less DNA loaded onto the gel. A yku70 strain was used as a control to confirm telomeric instability. Strains used were wt, FF18734; lig4, PRSY003,1; lif1, PRSY031,a; lif1 lig4, PRSY033,a; yku70, yku70α.


Human DNA ligase IV occurs as a heterodimer together with the XRCC4 protein. The phenotype of a hamster XRCC4‐deficient cell line, XR‐1, indicates that the complex plays a role in DSBR and V(D)J recombination. The identification and mutational analysis of a DNA ligase IV homologue in S.cerevisiae corroborated an involvement of the enzyme in DSBR independent of homologous recombination, and suggested the possible existence of a yeast counterpart to mammalian XRCC4. However, S.cerevisiae genome database searches failed to produce evidence for the presence of a gene encoding a related protein in yeast. Following a yeast two‐hybrid‐based screening strategy, we were able to identify LIF1, a previously uncharacterized yeast gene encoding a protein which physically interacts with the C‐terminal BRCT domain of LIG4.

LIF1 is structurally and functionally related to XRCC4

The C‐terminal BRCT domains of mammalian DNA ligase IV interact stably with XRCC4 (Critchlow et al., 1997; Grawunder et al., 1997). The N‐terminal 204 amino acids of XRCC4 are sufficient to reverse the DSBR defect of XR‐1 cells significantly (Leber et al., 1998). These data suggest that the interacting domain is localized within this N‐terminal region. By comparison, the overlapping cDNA sequences present in all LIF1 two‐hybrid clones isolated located the LIG4‐interacting region within the LIF1 protein between residues 191 and 260. A linear alignment of the amino acid sequences of yeast LIF1 and human XRCC4 positions residue 204 of XRCC4 opposite residue 236 of LIF1 and reveals a region of highly conserved amino acids (residues 201–230 of LIF1 and 172–198 of XRCC4) which is present in the XRCC4 and LIF1 sequences required for association with DNA ligase IV and, therefore, may be a likely candidate for mediating this interaction (Figure 2A). Thus, LIF1 interacts with yeast DNA ligase IV in a similar way to mammalian XRCC4 with DNA ligase IV. The overall similarity between the two proteins is moderate (22% identity, 49% similarity) and explains why LIF1 has not been identified in previous database searches. However, identities of ∼20% at the amino acid level are not unusual between mammalian and yeast homologues, and have also been seen in other factors involved in NHEJ, such as the yeast Ku homologues YKU70 and YKU80 in comparison with the human proteins (Beall et al., 1994; Boulton and Jackson, 1996a, b; Milne et al., 1996).

Mutant phenotypes and implications for the physiological roles of LIF1

The LIF1 gene is not essential; lif1 mutant haploid and diploid cells grow at normal rates, are respiratory‐competent and show normal cell cycle progression during exponential growth. It may be concluded that LIF1 is not required for DNA replication or maintenance of mitochondria. However, lif1, lig4 and lig4 lif1 mutant strains consistently grew to 2‐ to 3‐fold higher densities than isogenic wild‐type strains under identical conditions, and this correlated with a higher proportion of small budded cells in early stationary phase cultures of the mutant strains. This phenotype may implicate a role for the LIG4–LIF1 complex in growth regulation and/or cell cycle arrest. Two other experimental observations indirectly support such a model: (i) BRCT motifs are present in LIG4, and such motifs have been identified in other proteins involved in regulation of cell cycle progression which show a phenotypic link to DNA repair, e.g. budding yeast RAD9 and fission yeast Rad4 (Callebaut and Mornon 1997); and (ii) homozygous lif1 and lig4 diploids sporulate inefficiently due to a defect prior to meiosis I, but spore viability in the sporulating fraction of cells is normal. A simple explanation of this phenotype would be that the lig4 and lif1 mutants fail to respond efficiently to the nutritional signals which force wild‐type cells in G1 to enter meiosis. If so, the sporulation inefficiency of the mutants would reflect a regulatory deficiency in G1 of the cell cycle rather than a meiotic problem. Interestingly, a DNA ligase IV homologue of Candida albicans was cloned by complementation of an ime1 mutation in S.cerevisiae, and IME1 is part of a regulatory pathway leading to induction of meiosis in diploid budding yeast (Andaluz et al., 1996).

Strains lacking functional LIG4–LIF1 complexes display slight and conditional sensitivity to ionizing radiation. Whereas exponentially growing mutant and wild‐type cells are not significantly different from each other (Figure 3A), lif1 and lig4 mutant strains become hypersensitive to irradiation in stationary phase (Figure 3B), and this phenotype is enhanced in lig4 lif1 double mutants. The conditional radiosensitivity of yeast lig4, lif1 and lig4 lif1 mutant cells has an interesting parallel in the XRCC4‐defective hamster XR‐1 cell line which displays a unique cell‐cycle‐dependent radiosensitivity, being hypersensitive in G1 phase and early S phase but not in later stages of the cell cycle (Stamato et al., 1988). In stationary phase, the sensitivity of either a lif1 or lig4 mutant is similar to that of a rad52 mutant deficient in homologous recombination. In these cultures, cells are mainly in G0 of the cell cycle; as they are haploid, their ability to repair double‐strand breaks by homologous recombination is minimized and NHEJ pathways may be required. The synergistically enhanced sensitivity of the rad52 lig4 double mutant indicates that RAD52‐ and LIG4–LIF1‐dependent pathways contribute equally to repair of X‐ray‐induced DNA damage in resting yeast cells. Similarly, the enhanced radiosensitivity of the lig4 lif1 double mutant indicates that, to some degree, LIF1 and LIG4 are also able to act independently of each other in protecting cells against radiation damage.

The most striking phenotype of lif1 and lig4 mutant cells is observed in a transformation‐based plasmid double‐strand break rejoining assay. Whereas wild‐type cells are highly efficient in precise recircularization of linearized plasmids after transformation, this capacity is nearly abolished in lif1 or lig4 mutants (Figure 4B, see also Schär et al., 1997). The finding that wild‐type cells religated blunt ends less efficiently (Figure 4B) and less accurately (Table I) than cohesive ends, whereas lif1 or lig4 mutants show no such difference, suggests that the LIG4–LIF1‐dependent DNA joining pathway prefers DNA termini with protruding single strands to blunt ends as substrate. LIF1 and LIG4 appear to be epistatic with respect to this phenotype and, thus, act in the same pathway for non‐homologous plasmid religation (Figure 4B).

In addition to LIG4 and LIF1, an appreciable number of yeast factors have been identified which appear to be required for efficient non‐homologous religation of linearized plasmids after transformation. These include the Ku proteins, YKU70 and YKU80 (Boulton and Jackson, 1996a, b), the silencing factors SIR2, SIR3 and SIR4, and the recombination proteins RAD50, MRE11 and XRS2 (Boulton and Jackson, 1998). LIG4 and LIF1 are the only proteins among these which show an impaired plasmid repair phenotype but do not seem to be involved in telomere length maintenance (Boulton and Jackson, 1998; Figure 6). Comparing the fidelity of plasmid repair in strains transformed with linearized plasmid DNA having protruding single‐strand termini, accurately ligated products can be found in the rare transformants of lig4, rad50, mre11 and xrs2 strains but not in yku70, sir2, sir3 and sir4 mutants, nor in lif1 mutants (Table I; Boulton and Jackson, 1998). Thus, with respect to telomere length maintenance, lif1 mutants seem identical to lig4 mutants, but with respect to accuracy of plasmid religation they appear to be more similar to yku70, sir2, sir3 and sir4 mutant strains. These data imply that LIF1 may fulfil a structural role in association with LIG4, possibly acting as a mediator between factors protecting DNA ends, and the ligase which needs to be recruited to rejoin DNA double‐strand breaks. Similar speculations have been made for the possible role of XRCC4 in mammalian cells (Critchlow et al., 1997). Detailed genetic analysis in the yeast system should clarify the role of the LIF1 protein further.

Materials and methods

Genetic methods, assessment of sporulation efficiency and radiation sensitivity

Yeast complete medium (YPD), pre‐sporulation medium and synthetic drop out media were prepared as described by Sherman et al. (1982). Media for selection of respiration‐proficient cells and sporulation medium were those described by Bähler et al. (1994). Strains were propagated and sporulated at 30°C, unless otherwise indicated. Transformations were performed by a slight modification of the high‐efficiency lithium acetate method (Gietz and Schiestl, 1991). Analysis of transformants was performed as described (Schär et al., 1997). Mitotic growth, sporulation efficiency and spore viability of different strains were also examined as described previously (Schär et al., 1997).

Radiation sensitivity of cells was assessed as follows: 20 ml of YPD cultures were inoculated with 3×106 cells/ml from stationary phase pre‐cultures and grown to late log–early stationary phase (∼108 cells/ml) or late stationary phase (>2×108 cells/ml) at 30°C with vigorous shaking. After plating serial dilutions from 105 cells to 101 cells onto YPD agar plates, cells were exposed to different doses of X‐rays at varying dose rates (2.5, 5 or 10 Gy/min) using a Seifert, Isovolt X‐ray generator (300 kV, 9.0 mA, 0.5 mm Al‐filter). Incubations were at 30°C for 2–3 days before counting surviving clones.

Yeast strains

The S.cerevisiae strains used in this study are listed in Table II. They are all isogenic derivatives of two closely related, congenic series represented by FF18734 and FF18984 in an A364A background (F.Fabre, personal commununication) and were obtained by transformation and crossing within the set as indicated in Table II. Appropriate genotypes were isolated from dissected spore tetrads using standard replica plating techniques. The LIF1 gene disruption was generated by standard gene replacement techniques using a PCR‐derived disruption construct consisting of the URA3 gene flanked by 80 bp of homologous DNA sequences from the 5′ and 3′ ends of the LIF1 ORF (primer sequence available on request). For transformation of PRSY002, 1 μg of gel‐purified PCR fragment was used. URA+ transformants were first selected on media lacking uracil and then genotyped by replica plating onto diagnostic media, by standard genomic PCR and by genomic Southern blot analysis. Verified transformants carrying the expected heterozygous replacement within the LIF1 ORF with the URA3 gene (PRSY030,n) were then sporulated, and haploid progeny with appropriate genotypes were identified in dissected spore tetrads.

Plasmids, DNA manipulations and sequence analyses

Different fragments of LIG4 were PCR‐amplified from S.cerevisiae genomic DNA isolated according to standard procedures (Ausubel, 1994). PCR amplifications were carried out using Pfu DNA polymerase (Stratagene, La Jolla, CA). All constructs were verified by sequencing using an ABI 377 DNA sequencer (Perkin Elmer, Foster City, CA).

Restriction sites were included in the PCR primers for subsequent in–frame cloning into the GAL4‐BD vector pAS2‐1 (Clontech, Palo Alto, CA); pGEH009 contains the entire LIG4 ORF [LIG4(F)] cloned into the NdeI sites of pAS2‐1. The primers to generate the LIG4 fragment were GH72200, 5′‐GCATGCATCATATGATATCAGCACTAGATTCTATACC‐3′ (restriction sites underlined), and GH72199, 5′‐GCATGCATCATATGTCAGTAGTTGACTACGGGG‐3′. pGEH010 is pAS2‐1 containing the N‐terminal 255 amino acids of LIG4 [LIG4(N)] and the primers were: GH72200 (see above) and GH89052, 5′‐GCATGCATCATATGTCAGGCGAATGCAAAGCCAAC‐3′. pGEH011 is pAS2‐1 containing the coding sequence for the C‐terminal 313 amino acids of LIG4 [LIG4(C)] and the primers were: GH72199 (see above) and GH89051, 5′‐GCATGCATCATATGACAGATTGTTACACACTTAACG‐3′. pGEH012 [LIG4(B1)] and pGEH013 [LIG4(B2)] are also pAS2‐1 derived and contain the coding sequences for each of the C‐terminal BRCT domains of LIG4 in the NdeI–BamHI restriction sites. Primers used were GH89051 and GH91938, 5′‐CGGGATCCTCAATCACCCAAACAATCTACCC‐3′ for LIG4(B1), and GH72199 and GH91937, 5′‐GCATGCATCATATGCTGTCATCATTGTATAAATC‐3′ for LIG4(B2). pGEH015 is pBluescript KS(+/−) containing the entire LIF1 ORF which was PCR‐amplified with primers GH91714, 5′‐CGAATTCGAATGTCCCAGCTGACGGAGTTC‐3′ (start codon in italics) and GH91715, 5′‐GCTCTAGAAGGCTATGTTTCTATATCCG‐3′ (stop codon in italics).

pGEH014 was used for complementation experiments in lif1 mutants and was constructed by subcloning the LIF1 ORF from pGEH015 into the SalI–XbaI sites of the galactose‐inducible yeast expression vector YCpIF3 (Foreman and Davis, 1994). pGEH019 is LIF1 subcloned into EcoRI–PstI sites of pAS2‐1 and was used to subclone LIF1 in‐frame with the His tag in expression vector pET‐16b (Novagen, Madison, WI) (pGEH020). pGEH007 contains the entire ORF of human DNA ligase IV (106 kDa form) with seven N‐terminal histidine residues in the EcoRI site of the galactose‐inducible yeast expression vector pYES2 (Invitrogen, Leek, Netherlands) (primer sequence available on request). pPRS154 contains the entire His‐LIG4 ORF in pYES2 and has been described (Schär et al. 1997). pBTM116 is an E.coli–yeast shuttle vector carrying 2 μ sequences for stable maintenance in yeast, the TRP1 gene as a selectable marker and the promoter and terminator sequences of the ADH1 gene flanking a multiple cloning sequence (P.Bartel and S.Fields, unpublished).

For analysis of telomeric length maintenance, 2 μg of genomic DNA derived from yeast cells that were grown for >100 generations since sporulation were digested with 30 U of XhoI, separated on 0.8% agarose gels, Southern blotted and hybridized with a poly(GT)20 probe as described by Boulton and Jackson (1998).

DNA and protein sequence analyses were performed with the Genetics Computer Group program package, Version 8, 1994 (Devereux et al., 1994). For general database searches and comparisons, we used the BLAST, FASTA and ENTREZ services provided at NCBI's web page; for yeast genome database searches we accessed MIPS and SGD through their web pages.

Two‐hybrid screening

Two‐hybrid screening (Fields and Song, 1989) was performed using the Matchmaker two‐hybrid system from Clontech, largely following the instructions provided by the manufacturer. A S.cerevisiae cDNA library was converted to a two‐hybrid library by subcloning into the XhoI sites of the activation domain vector pACT2 (Clontech) and was kindly provided by Dr N.Lowndes, ICRF Clare Hall Laboratories. Initial library screening was performed using pGEH010 and pGEH011. Filter assays and quantitative β‐galactosidase assays were carried out according to standard protocols (Transy and Legrain, 1995; Bai and Elledge, 1997).

Bacterial expression and purification of LIF1

Recombinant His‐tagged LIF1 protein was expressed in E.coli BL21(DE3) from pGEH020 under control of the T7 RNA polymerase promoter. Pre‐cultures of 10 ml of LB broth + 50 μg/ml carbenicillin were inoculated with four freshly transformed BL21(DE3)/pGEH020 colonies, incubated overnight at 37°C with vigorous shaking, and used for inoculation of 1 l cultures (NZY‐Broth + 0.2% casein acid hydrolysate and 50 μg/ml carbenicillin). Cultures were grown at 37°C to an OD600 of 0.5 and expression was induced by adding isopropyl‐β‐d‐thiogalactopyranoside (IPTG) to a final concentration of 1 mM. Cells were then incubated for 6 h at 37°C, harvested by centrifugation and lysed overnight in 6 M guanidine–HCl, 0.1 M Na2HPO4, 0.01 M Tris–HCl; pH 8.0 (bacterial lysis buffer) with constant rotation. After centrifugation of the lysate (20 min, 15 000 g), soluble proteins in the supernatant were batch adsorbed to 1 ml of pre‐equilibrated (with bacterial lysis buffer) SuperFlow Ni‐NTA agarose (Qiagen, Hilden, Germany) at 4°C for 1 h. The Ni‐NTA resin was then packed into a disposable column (Bio‐Rad, Herts, UK) and washed with 30 ml of wash buffer (8 M urea, 0.1 M Na2HPO4, 0.01 M Tris–HCl, 15 mM imidazole, pH 8.0). Finally, bound proteins were eluted with the same buffer containing 250 mM imidazole, and LIF1‐containing fractions were identified by SDS–PAGE.

Expression in yeast and purification of LIG4 protein

Recombinant, His‐tagged LIG4 protein was expressed in PRSY003,1 (lig4) from pPRS154 under control of the GAL1 promoter. Pre‐cultures of 100 ml of SD medium lacking uracil were inoculated with four freshly transformed PRSY003,1/pPRS154 colonies, incubated overnight at 30°C with vigorous shaking, and used for inoculation of 1 l of culture YP‐medium containing 2% raffinose as carbon source. After 2 h at 30°C with vigorous shaking, galactose was added to a final concentration of 2% and cells were grown for a further 6 h and harvested by centrifugation. Cells were then suspended in 1 vol. of 2× yeast lysis buffer (40% glycerol, 200 mM Tris–HCl pH 8.0, 700 mM NaCl, 8 mM β‐mercaptoethanol, 0.25% Tween‐20) and lysed by vortexing in the presence of acid‐washed glass beads. The lysis buffer was supplemented with 2× proteinase inhibitors to minimize degradation of the proteins [leupeptin (500 ng/ml), pepstatin A (2.7 μg/ml), phenylmethylsulfonyl fluoride (PMSF; 340 μg/ml) and benzamidine (660 μg/ml)]. Cell debris was removed by centrifugation (20 min, 4°C, 15 000 g), and soluble proteins in the supernatant were batch adsorbed to 1 ml of pre‐equilibrated (with 1× yeast lysis buffer) SuperFlow Ni‐NTA agarose (Qiagen, Hilden, Germany) at 4°C for 1 h. The Ni‐NTA resin was packed into a disposable column (Bio‐Rad) and washed with 2× 20 ml of wash buffer (1× yeast lysis buffer, without proteinase inhibitors) containing 50 and 70 mM imidazole. Proteins were then eluted in buffer containing 20% glycerol, 100 mM Tris–HCl pH 8.0, 350 mM NaCl, 5 mM dithiothreitol (DTT), 0.25% Tween‐20 and 500 mM imidazole. LIG4 peak fractions were identified by SDS–PAGE after silver staining (SilverStain Plus, Bio‐Rad) and dialysed against 20% glycerol, 50 mM Tris–HCl pH 8.0, 100 mM NaCl, 1 mM EDTA, 0.5 mM DTT, 0.25% Tween‐20.

Western blot analysis

Yeast cells were incubated in 50 ml of YPD medium and grown to late log‐phase (16 h). Crude lysates were prepared as described above, and equal amounts of protein (20 μg/lane) separated by SDS–polyacrylamide gel (8%) electrophoresis. After transfer of the proteins to nitrocellulose membranes (Schleicher and Schuell, Keene, NH) using the Mini‐PROTEAN blotting system (Bio‐Rad), they were blocked in phosphate‐buffered saline (PBS) containing 5% (w/v) non‐fat dried milk, 0.1% Tween‐20 for 16 h at 4°C, then incubated with an anti‐human DNA ligase IV antibody which was raised against highly conserved parts of the core domain of human DNA ligase IV (TL14/15) (Robins and Lindahl, 1996) at a dilution of 1:200 in PBS containing 1% (w/v) non‐fat dried milk, 0.1% Tween‐20 for 1 h at room temperature. After several washes in this buffer, the membrane was incubated with goat anti‐rabbit antibody conjugated to horseradish peroxidase (Pierce, Rockford, IL). After additional wash steps, the membrane was incubated with ECL solution (Amersham, Bucks, UK) and exposed to X‐ray film. To reduce unspecific background, the primary antibody was pre‐incubated with 100 mg of protein extract from a lig4 strain immobilized on a nitrocellulose membrane in PBS containing 1% (w/v) non‐fat dried milk and 0.1% Tween‐20 at 4°C overnight.

Northern blot analysis

Yeast cells were cultured in 50 ml of YPD overnight at 30°C, and 1.5 ml of late‐log phase cells were harvested by centrifugation and lysed in TRIzol™ reagent (Gibco‐BRL) by vortexing in the presence of acid‐washed glass beads. Total RNA was then extracted, and equal amounts of RNA (10 μg/lane) were separated on a 1% denaturing agarose gel and transferred to a Hybond‐N membrane according to the protocol provided by the manufacturer (Amersham, Bucks, UK). Hybridizations were carried out at 42°C (Sambrook et al., 1989) and followed by two washes in 2× SSC, 0.05% SDS at room temperature and two washes in 2× SSC, 0.1% SDS at 50°C. The membrane was then exposed to X‐ray film. The LIG4‐specific probe was generated by NdeI digestion of pGEH009 and purification of the 2.8 kb LIG4 fragment. 32P‐labelling was performed with the Multiprime DNA Labelling Kit (Amersham) and the probe was purified from unincorporated nucleotides on a Sephadex G‐50 column (Pharmacia Biotech).


We thank Walter Burkard for technical support in X‐ray dosimetry and handling of the irradiation equipment, Margret Fäsi and Claudine Schneider for technical assistance and tetrad dissections, respectively, Peter Robins for advice on protein purification, John Sgouros for advice on bioinformatics, Francis Fabre and Wolf‐Dietrich Heyer for yeast strains, Noel Lowndes for providing the S.cerevisiae cDNA library, Deborah Barnes and John Diffley for critical reading of the manuscript and Josef Jiricny and members of his group for stimulating discussions and general support. G.H. was supported by grant He 2675/1‐1 from the Deutsche Forschungsgemeinschaft (DFG).


View Abstract