Spt16 and Pob3 form stable heterodimers in Saccharomyces cerevisiae, and homologous proteins have also been purified as complexes from diverse eukaryotes. This conserved factor has been implicated in both transcription and replication and may affect both by altering the characteristics of chromatin. Here we describe the isolation and properties of a set of pob3 mutants and confirm that they have defects in both replication and transcription. Mutation of POB3 caused the Spt− phenotype, spt16 and pob3 alleles displayed severe synthetic defects, and elevated levels of Pob3 suppressed some spt16 phenotypes. These results are consistent with previous reports that Spt16 and Pob3 act in a complex that modulates transcription. Additional genetic interactions were observed between pob3 mutations and the genes encoding several DNA replication factors, including POL1, CTF4, DNA2, and CHL12. pob3 alleles caused sensitivity to the ribonucleotide reductase inhibitor hydroxyurea, indicating a defect in a process requiring rapid dNTP synthesis. Mutation of the S phase checkpoint gene MEC1 caused pob3 mutants to lose viability rapidly under restrictive conditions, revealing defects in a process monitored by Mec1. Direct examination of DNA contents by flow cytometry showed that S phase onset and progression were delayed when POB3 was mutated. We conclude that Pob3 is required for normal replication as well as for transcription.
POB3 and Spt16/Cdc68 form an abundant, nuclear heterodimer that binds specifically to DNA polymerase α in the yeast Saccharomyces cerevisiae (Wittmeyer and Formosa 1995, 1997; Brewsteret al. 1998; Wittmeyeret al. 1999). SPT16/CDC68 has been implicated genetically in the global regulation of transcription (Prendergastet al. 1990; Maloneet al. 1991; Rowleyet al. 1991; Xu et al. 1993, 1995; Lycanet al. 1994; Brewsteret al. 1998; Evanset al. 1998). Both elevated Spt16 levels and an spt16 mutation increase the production of some transcripts, notably the aberrant messages from transposon-disrupted alleles of HIS4 and LYS2 that lead to the Spt− phenotype (Prendergastet al. 1990; Maloneet al. 1991; Rowleyet al. 1991; Lycanet al. 1994). However, the levels of other, normal transcripts such as those for cyclins decrease in spt16 mutants (Prendergastet al. 1990; Maloneet al. 1991; Rowleyet al. 1991; Lycanet al. 1994) indicating that diminished Spt16 function can either increase or reduce transcription. Due to the global nature of these effects, the similarity of the phenotypes with those caused by mutations in histone genes (Maloneet al. 1991; Winston and Carlson 1992), and genetic interactions with the putative chromatin factor San1 (Schnellet al. 1989; Xuet al. 1993), Spt16 has been proposed to affect transcription by altering the properties of chromatin. Consistent with this, a portion of the total Spt16-Pob3 complex was found to be stably associated with chromatin (Wittmeyeret al. 1999).
Spt16-Pob3 bound to affinity matrices containing the catalytic subunit of DNA polymerase α (Pol1) as the ligand (Wittmeyer and Formosa 1995, 1997) and also partially copurified with the four subunit Pol α/primase complex (Wittmeyeret al. 1999). A variety of additional physical and genetic tests indicates that the interaction between Spt16-Pob3 and Pol α is important in vivo (Wittmeyer and Formosa 1995, 1997; Formosa and Nittis 1999; Wittmeyeret al. 1999), suggesting that Spt16-Pob3 acts in DNA replication. These results do not contradict studies that infer a role for Spt16-Pob3 in transcription, but instead suggest a similar or additional role for this factor in DNA replication. Since chromatin is the substrate for both replication and transcription, a factor that alters the properties of chromatin could easily affect both processes.
Both SPT16 (Maloneet al. 1991) and POB3 (Wittmeyer and Formosa 1997) are essential genes that are highly conserved among eukaryotes (Wittmeyer and Formosa 1997; Evanset al. 1998). Heterodimers of Spt16 and Pob3 homologs have been purified from human and frog cells (Okuharaet al. 1999; Orphanideset al. 1999). One of these, the human FACT complex, allows RNA polymerase II to elongate transcripts on templates that contain nucleosomes, which otherwise block elongation of transcription (Orphanides et al. 1998, 1999). The mechanism has not been determined, but Orphanides et al. (1999) have proposed that FACT might make intranucleosomal histone contacts more flexible as RNA pol II approaches. The frog DUF complex is also composed of Spt16 and Pob3 homologs (Okuharaet al. 1999), and depletion of DUF from egg extracts caused loss of replication competence (Okuharaet al. 1999). Since replication in this system is independent of transcription (but dependent on chromatin formation; Newport 1987), this indicates a direct role for DUF in DNA replication. The activities of the human and frog Spt16-Pob3 homologs therefore support roles in both transcription and replication, possibly by mediating interactions with nucleosomes for both DNA and RNA polymerases.
While spt16 mutations were identified in several screens for transcription factors, mutations in pob3 have been described only briefly (Wittmeyeret al. 1999). To further dissect the role of Spt16-Pob3, we have isolated a set of mutations in POB3 and tested the mutant strains for defects in both transcription and replication. Our results support the model that both replication and transcription depend on the function of POB3.
MATERIALS AND METHODS
Yeast methods: Selective and rich media were prepared as described (Hartwell 1967; Roseet al. 1990). Strains used are listed in Table 1. For maximal permissive temperature (MPT) determinations, strains were grown to stationary phase in rich medium and 2-μl aliquots were placed on several rich agar plates and spread to a portion of each plate with a wire loop. The plates were incubated for 3 days at various temperatures, including either 22° or 26° and a set at 1° intervals from 28° to 37° with the range depending on the strains being tested. The MPT was judged to be the highest temperature that permitted ~10% of the growth observed at the lowest temperature. The breadth of the transition varied with the mutation, but typically viability dropped by at least 10,000-fold within 2° of the MPT. For arrest and release experiments, cells were treated with 4 μg/ml of α-factor (Sigma, St. Louis) in rich media and then were collected by centrifugation and suspended in rich media containing 0.1 mg/ml protease type XIV (Sigma) at 22° or 37°, as described previously (Paulovich and Hartwell 1995).
Mutagenesis: pJW4 (YCp, POB3, URA3) and pJW11 (YCp, POB3, LEU2) contain the 3924-bp KpnI-SphI fragment including POB3 in YCplac33 and YCplac111 (Gietz and Sugino 1988). pJW11 was treated with 0.1 m hydroxylamine for various amounts of time at 75°, essentially as described (Sikorski and Boeke 1991; Formosa and Nittis 1999), and then was introduced into strain 7697 (pob3-Δ) carrying pJW4. This led to the identification of the alleles pob3-10, -11, and -12. POB3 was also amplified from yeast genomic DNA using the primers 5 ′ -CUACUACUACUAGGATCCTTGTAAGTACTTGGCTCA and 5′-CAUCAUCAUCAUGAATTCTGTCTTACACTCACCATGTC under several mutagenic conditions (Zhouet al. 1991; Cadwell and Joyce 1992; Zhanget al. 1998). The resulting 2043-bp PCR product includes 278 bp upstream through 96 bp downstream of the POB3 open reading frame (ORF) flanked by added BamHI and EcoRI sites on a fragment that can be efficiently recovered using the CloneAmp system (Life Technologies). YCplac111 (Gietz and Sugino 1988) and the PCR products were digested with EcoRI and BamHI, ligated to form pTF139 derivatives, and the ligation mixtures were used directly to transform strain 7697 pJW4. In each screen, Leu+ transformants were transferred to media containing 5-fluoroorotic acid (5-FOA; Boekeet al. 1987) to select for loss of pJW4 and then were tested for growth at 13° and 37°. Plasmids were recovered from candidate mutants and retransformed to establish linkage of the mutant phenotype with the plasmid. The POB3 ORF from each mutant was then sequenced to identify alterations. Reconstruction of some mutations (see Table 2) was performed by mutagenizing pTF139 with pairs of primers using the QuikChange method (Stratagene, La Jolla, CA). To delete the 3′ end of POB3, the same upstream primer above was used with 5′-CAUCAUCAUCAUGAATTCCTACTCTTCCTTACTGATGTTGG, which converts residue Q458 (CAG) to a nonsense codon (TAG) and also removes the coding sequence for the remaining 94 residues of Pob3. This product was inserted into YCplac111 (Gietz and Sugino 1988) to form pTF139-CTΔ95.
Analysis of Pob3 and Spt16 proteins: Cultures in log phase were harvested, suspended in SDS sample buffer, and boiled. Extract representing 1–5 × 106 cells (the same number of cells was used for each sample in a given experiment) was separated by SDS-PAGE, and proteins were transferred to nitrocellulose (Harlow and Lane 1988). Pob3 and Spt16 were detected using polyclonal antisera (Wittmeyeret al. 1999) and either insoluble product (Harlow and Lane 1988) or chemiluminescent (Amersham-Pharmacia Biotech, Piscataway, NJ) staining.
Flow cytometry: Cells were fixed in 70% ethanol, washed, stained with propidium iodide, and their DNA contents were measured as previously described (Wittmeyer and Formosa 1997).
Isolation of pob3 mutations: POB3 was mutagenized with hydroxylamine (Sikorski and Boeke 1991) or by PCR amplification under mutagenic conditions (Zhouet al. 1991; Cadwell and Joyce 1992; Zhanget al. 1998). Clones were screened for Ts−, Cs−, and slow growth phenotypes using a standard plasmid shuffle (Sikorski and Boeke 1991). A total of 13 Ts−, 0 Cs−, and 2 slow-growing mutants were obtained. The Ts− mutants carry conditionally functional versions of Pob3 that act adequately at low temperatures but fail at elevated temperatures. We used these strains to analyze the behavior of transcription and replication in cells with minimal Pob3 function, or upon the removal of Pob3 function by shifting growing cultures to a nonpermissive temperature. The DNA sequences of the mutated genes were determined and revealed the changes in amino acid sequence listed in Table 2. In some cases where multiple mutations were identified, subsets were reconstructed by site-directed mutagenesis and retested. A single L78R mutation was found to be responsible for the Ts− phenotype (and for all other phenotypes; see below) caused by pob3-1, while two mutations were required for the Ts− phenotypes of pob3-7 and pob3-11. Single point mutations can therefore produce conditional lethality in POB3, but other alleles are more complex.
Two alleles, pob3-20 and pob3-21, caused a serious defect in the rate of growth at all temperatures in the A364a background (and strong temperature sensitivity in an S288C/A364a hybrid shown in Figure 1). Both mutations were found to alter the extreme C terminus of Pob3, either deleting the final five residues or substituting the sixth residue from the C terminus. The growth defects of pob3-20 and pob3-21 strains are not due to the formation of unstable or dominant interfering proteins, since the Pob3 levels are at least as high as wild type in these strains (Figure 2B) and the growth rate of strains containing both the mutant and wild-type alleles on low copy plasmids is normal. However, the levels of Spt16 protein are diminished in these mutants (Figure 2D; data not shown), suggesting that the C terminus of Pob3 plays a role in stabilizing Spt16.
Three independent alleles (pob3-9, -12, and -13, from two different PCR reactions and a hydroxylamine-mutagenized plasmid) were found to be identical, changing residue Q458 to a stop codon, which appears to delete the final 95 residues of the 552-amino acid Pob3 protein (a fourth allele, pob3-5, also had this mutation along with two additional changes). The Ts− phenotype caused by the Q458-stop mutation varied with different strain backgrounds; very slow growth at 37° was observed in the original A364a background, but the hybrid A364a/S288C strain used in Figure 1 grew normally at this temperature. These results suggested that the C-terminal 95 amino acids of Pob3 are not essential for viability. However, examination of the Pob3 protein in these strains consistently revealed the presence of some full-length Pob3 protein along with the expected truncated form (Figure 2C). The truncated form was observed even under nonpermissive conditions, whereas the full-length form was not (Figure 2C). We constructed a deletion of the C-terminal domain of Pob3 in which Q458 was mutated to a stop codon but the remaining Pob3 sequence was removed (creating pob3-CTΔ95). This allele was unable to complement the lethality of a pob3 deletion (Figure 3). We conclude that the C-terminal domain of POB3 is essential and that the Q458-stop mutation is viable due to translational read-through that is temperature sensitive in a strain-dependent manner, not because of the production of a temperature-sensitive protein. Consistent with this interpretation, the Q458-stop nonsense mutation creates a poor termination context in yeast (Bonettiet al. 1995). It is not clear why this mutation is recovered at such a high frequency.
The remaining pob3 alleles have mutations distributed throughout the gene. Comparing the positions of these mutations to the degree of conservation among 12 Pob3 homologs from GenBank (Altschulet al. 1997) indicates that many residues that are absolutely conserved can be altered without loss of viability. For example, pob3-10 has eight amino acid changes, three of which affect residues that are invariant in all 12 Pob3 homologs. While viable, these strains display slower growth than wild type. The complexity of these alleles prevents us from inferring structure-function relationships, but we note that POB3 is able to tolerate substitution of some highly conserved residues.
Pob3 protein is rapidly lost in Ts− pob3 mutants: Different pob3 alleles caused arrest of growth at different temperatures, but even the tightest alleles allowed two to three divisions to occur after a shift to 37° (although the number of viable cells does not increase; see below). This is likely to be the null phenotype for pob3 mutants since we observed a similar accumulation of cells after germination of haploids carrying a deletion of POB3 (Wittmeyer and Formosa 1997). In addition, all of the alleles that cause a Ts− phenotype also cause the disappearance of intact Pob3 protein as assayed by immunodetection after SDS-PAGE (Figure 2A; data not shown). The levels of Pob3 protein were reproducibly diminished in pob3 Ts− mutants relative to wild-type cells even under conditions permissive for growth, and no Pob3 protein was detected within 30 min after a shift to 37°. Since Pob3 is associated with Spt16, we also determined the stability of Spt16 in pob3 mutants. As shown in Figure 2D, we found that Spt16 levels dropped reproducibly by about twofold upon shifting cells with a wild-type POB3 gene to 37° for 3 hr and disappeared completely in pob3 Ts− mutants under the same conditions. We conclude that pob3 mutations inhibit growth by diminishing the level of essential Spt16-Pob3 heterodimers.
Pob3 defects cause the Spt− phenotype: High copy expression and mutation of SPT16 were found previously to produce the Spt− phenotype (Clark-Adamset al. 1988; Maloneet al. 1991), which results from changes in the selection of transcription initiation sites for a promoter found in the Ty1 δ-element. This relieves the auxotrophy for histidine and lysine normally caused by the his4-912δ and lys2-128δ δ-insertion alleles (Clark-Adamset al. 1988; Maloneet al. 1991). We screened the pob3 mutations to see if they also cause this phenotype. As shown in Figure 1, all of the mutations in POB3 identified for either conditional growth or slow growth also allowed some expression of both his4-912δ and lys2-128δ. The different alleles displayed different levels of Spt− phenotype, as indicated by the different amounts of growth on −His and −Lys media at different temperatures. The strength of the Spt− phenotype did not correlate well with the positions of the mutations within the gene or with the MPT (Table 3). For example, pob3-10 has a higher MPT than pob3-1, but both have a strong Spt− phenotype (both grew well on media lacking lysine). Since the Spt− phenotype has been associated previously with alterations of transcription initiation (Winstonet al. 1984; Hirschmanet al. 1988; Maloneet al. 1991; Winston and Carlson 1992), we infer that pob3 mutations cause a defect in transcription. Since all of the POB3 alleles caused the Spt− phenotype even at temperatures completely permissive for growth, we conclude that normal transcription requires full Pob3 activity levels.
pob3 mutations interact genetically with spt16 mutations: Spt16 and Pob3 are members of the same complex (Wittmeyer and Formosa 1997; Brewsteret al. 1998). We therefore expected the SPT16 and POB3 genes to interact in genetic tests and reported previously that pob3-10, -11, and -12 display a decreased MPT relative to single mutants when combined with the spt16-G132D mutation (Wittmeyeret al. 1999). The spt16-G132D allele was isolated in three independent screens (Prendergastet al. 1990; Maloneet al. 1991; Lycanet al. 1994; Evanset al. 1998) but this residue has been shown to be in a region of SPT16 that can be deleted and still produce viable cells (Evanset al. 1998). We have identified additional conditional alleles of SPT16, including mutations in the essential region of the gene (T. Formosa, unpublished results). We used one of these, spt16-4 (P565S P570L), to assess the effect of a different allele of SPT16 on pob3 mutations. To facilitate rapid screening of large numbers of combinations of mutations, strains were constructed with genomic mutations of SPT16, deletions of POB3, and POB3 plasmids marked with URA3. The set of pob3 alleles on LEU2-marked plasmids was then transformed into these strains, and transformants were screened for the ability to lose the POB3 URA3 plasmid by selecting on media containing 5-FOA (Boekeet al. 1987). Viable double mutants were then tested for growth at various temperatures (as described in materials and methods) to determine whether the MPT was different from the single mutants. While spt16-G132D has a lower MPT than spt16-4, the severity of the synthetic defects with pob3 mutations was much greater with spt16-4. As shown in Figure 3 and Table 3, all pob3 mutations tested were viable in combination with spt16-G132D, but displayed very strong synthetic defects, lowering the MPT as much as 5°. In contrast, 10 of the 13 combinations tested were lethal at all temperatures when spt16-4 was used, and the remainder also had strong synthetic defects. The combinations of pob3 and spt16 mutations available previously (Wittmeyeret al. 1999) were therefore among the weakest effects we have noted. These strong allele-specific defects further support the conclusion that Spt16 and Pob3 act in a complex in vivo. Loss of function in one protein in this case causes a requirement for the optimal function of the other partner.
We also found that elevated levels of POB3 could suppress some spt16 defects. A strain with a genomic spt16-4 mutation is unable to grow at 36° (Figure 4), but providing extra copies of POB3 either on high or low copy vectors partially suppressed this temperature sensitivity. This suppression of the Ts− phenotype was allele specific, since it was not observed with three other spt16 alleles, including the spt16-G132D allele (not shown). Therefore, in a case where two proteins are known to form a strong physical interaction, we observe dramatic evidence for extensive, allele-specific genetic interactions.
pob3 mutants display allele-specific synthetic defects with several DNA replication factors: We previously found that spt16-G132D mutations display synthetic defects with several DNA replication factors, but not with an unrelated set of Ts− mutations (Wittmeyer and Formosa 1997; Formosa and Nittis 1999). We screened the pob3 alleles to test for similar effects and again found several allele-specific synthetic defects consistent with a role for POB3 in DNA replication. As for the spt16 pob3 double mutants, strains were constructed with a deletion of POB3 covered by a URA3-marked plasmid containing the normal POB3 gene and genomic mutations in candidate genes. The MPT of double mutants was then tested; the difference between this value and the MPT of the most stringent single mutant is listed in Table 3. Representative examples of the data are shown in Figure 5 to show the severity of these synthetic defects.
Spt16-Pob3 binds to the Pol1 subunit of DNA polymerase α (Wittmeyer and Formosa 1995, 1997), so double mutants with the POL1 allele cdc17-1 were constructed. Relatively small but reproducible effects were noted with four of the pob3 alleles (Figure 5 and Table 3; pob3-L78R is a derivative of pob3-1, so this pair is counted once), similar to the effects noted with spt16-G132D (Wittmeyer and Formosa 1997). POB3 therefore interacts genetically with DNA polymerase α. Ctf4 protein appears to compete with Spt16-Pob3 for binding to Pol1 (Miles and Formosa 1992b; Wittmeyer and Formosa 1995, 1997), so a ctf4 deletion might be expected to rescue some pob3 mutations since it might reduce the competition for Pol1 binding. In fact, the temperature sensitivity caused by several pob3 alleles was observed to be suppressed by the ctf4 mutation (Table 3). However, as previously noted (Wittmeyeret al. 1999) some pob3 alleles display a synthetic defect with the ctf4 mutation (Table 3 and Figure 5). The alleles pob3-20 and -21 that alter the extreme C terminus of Pob3 (Table 3) are particularly affected. The observation of both suppression and enhancement of the Ts− phenotypes of pob3 mutants upon deletion of CTF4 suggests that the interactions among Pol α, Spt16-Pob3, and Ctf4 are more complicated than predicted by a simple binding-competition model.
DNA2 encodes an essential nuclease/helicase that has been implicated in Okazaki fragment maturation (Buddet al. 1995; Budd and Campbell 1997; Fiorentino and Crabtree 1997; Baeet al. 1998; Formosa and Nittis 1999). The dna2-2 allele alters a residue expected to destroy helicase function (Formosa and Nittis 1999) and causes synthetic lethality with a ctf4 deletion as well as sensitivity to the DNA-damaging agent methyl methanesulfonate (Formosa and Nittis 1999), but does not affect growth at elevated temperatures. An spt16-G132D mutation displayed a strong synthetic defect when combined with dna2-2 (Wittmeyer and Formosa 1997; Formosa and Nittis 1999), so we screened the effect of pob3 mutants in combination with dna2-2. Strong interactions were found, particularly with the pob3 alleles that altered or removed portions of the C terminus of Pob3 (Figure 5 and Table 3).
Deletion of CHL12 is lethal when combined with mutations in either DNA2 or CTF4 (Formosa and Nittis 1999). This gene has strong sequence similarity to the five RFC genes that encode the DNA polymerase processivity clamp loading factor RFC (Kouprinaet al. 1994; Cullmannet al. 1995) and has been implicated in DNA metabolism (Kouprinaet al. 1994). Several pob3 alleles also displayed strong synthetic effects with a deletion of CHL12 (Table 3 and Figure 5). As with ctf4, both suppression and enhancement were noted, with a similar pattern of allele specificity. Taken together, the allele-specific synthetic effects observed with these replication factors provide further evidence of a role for Pob3 in DNA replication.
POB3 deficiencies cause sensitivity to hydroxyurea: Hydroxyurea inhibits ribonucleotide reductase, leading to decreased rates of synthesis of the dNTPs required for DNA replication. If yeast Spt16-Pob3 functions in DNA replication in a manner analogous to the activity displayed by the human FACT complex in transcription (Orphanides et al. 1998, 1999), then DNA replication elongation should be inhibited in pob3 mutants due to difficulty progressing through nucleosomes. If two factors independently impair replication elongation, they should produce a stronger defect when they are combined than when they are applied separately. We therefore examined the set of pob3 mutants for the ability to grow on media containing hydroxyurea (HU). The pob3-7 allele caused sensitivity to HU even at low temperatures (Figure 6). The Ts− phenotype caused by pob3-7 was separable from the HU sensitivity, since reconstructed alleles with only W28R and N69K mutations displayed the full Ts− phenotype but were not HU sensitive at 26° (not shown). Since pob3-7 is one of the most stringent Ts− alleles, we tested to see whether other alleles cause HU sensitivity at temperatures nearer to their MPT. In all cases, we found that pob3 mutants were sensitive to HU at some temperature compared to wild-type POB3 strains (Figure 6). In some cases the effect was small (addition of HU to a pob3-L78R strain caused a >1000-fold decrease in viability at 32°, but at this temperature the viability of this strain in the absence of HU is diminished by ~100-fold relative to 26°, so it is already severely stressed), but in other cases it was robust (addition of HU to a pob3-2 strain caused at least a 10,000-fold effect at 32°, which is well below the MPT of 34° for this strain). Since all pob3 alleles caused HU sensitivity under some conditions, HU enhances the defect caused by pob3 mutations, and this effect is severe with some alleles. Since partial loss of Pob3 function and inhibition of dNTP synthesis have additive effects, we conclude that HU and Pob3 both function in the same essential process. Since the rapid dNTP synthesis inhibited by HU is required only for DNA replication, this common process is likely to be DNA replication.
pob3 mutants depend on a DNA replication checkpoint to maintain viability: As a further test for a role in DNA replication, we determined the effect of disabling DNA damage checkpoints in pob3 mutants. If pob3 mutants fail to perform DNA replication normally, replication checkpoints such as the one promoted by MEC1 should be important for maintaining the viability of pob3 mutants (Elledge 1996; Paulovichet al. 1997; Longheseet al. 1998; Weinert 1998). Strains with mec1-1 and pob3 mutations were therefore examined for changes in viability after inactivating Pob3 function by shifting to a nonpermissive temperature.
Figure 7 shows that cells with a wild-type POB3 gene continue to grow at 37° whether MEC1 is intact or mutated (although the mec1-1 strain initially loses some viability and then grows slowly). While the pob3-1 MEC1 strain failed to grow at 37° as expected, it retained full viability for ~12 hr and then lost ~90% of its viability during the subsequent 12 hr. In contrast, the pob3-1 mec1-1 double mutant began losing viability immediately upon shifting to 37°, and this loss continued throughout the course of the incubation, resulting in an ~1000-fold drop in viability after 24 hr. Therefore, the pob3-1 allele caused a strong dependence on the MEC1 checkpoint for surviving at 37°, such that after 24 hr at this restrictive temperature pob3-1 mec1-1 double mutants had 100-fold less viability than pob3-1 MEC1 single mutants. The timing and total loss of viability for MEC1 and mec1 strains varied in different experiments but was 20- to 100-fold lower for the mec1 strain after 24 hr in several independent trials with both pob3-1 and the reconstructed pob3-L78R. Similar results were obtained with pob3-7 (not shown). Ts− pob3 mutants therefore depend on the MEC1 checkpoint to retain viability under nonpermissive conditions. This is a strong indication that pob3 deficiency causes a defect in DNA replication. This effect is at least somewhat checkpoint specific since similar experiments with a deletion of the G2 checkpoint gene RAD9 (Weinert and Hartwell 1988) did not show a similar response (data not shown).
As noted earlier, pob3 Ts− mutants placed on agar plates at 37° double two to three times to produce four to eight cell bodies. However, pob3-1 mec1-1 cells placed at 37° appeared to arrest immediately. We determined the concentration of cells in the liquid cultures from the experiment shown in Figure 7 by direct examination in a hemacytometer (after sonication). The cultures of POB3 MEC1 and POB3 mec1-1 strains each increased their total cell number by 52-fold in 24 hr, the pob3-1 MEC1 strain increased by 5.2-fold, and the pob3-1 mec1-1 strain increased by 1.4-fold. Since the pob3-1 cells continued to increase in cell number during a period when the number of viable cells was constant or dropping, the increase must be due to the production of inviable cells, and this was observed as an increase in the number of cells that failed to form colonies after being placed at 26° (not shown). We expected the loss of a checkpoint to result in decreased viability as observed, but we also expected the loss of a monitor to be accompanied by increased cell cycle progression (Weinert and Hartwell 1988). Instead, we find that pob3 mec1 double mutants arrest more rapidly than pob3 MEC1 strains. Examination of the nuclear morphology of the cells provides a potential explanation. As shown in Figure 8, pob3-1 cells accumulated an abnormally high level of single-nucleated large-budded cells after a shift to 37° (59% of the large-budded cells had a single nucleus after a 3 hr shift vs. 26% for WT), but pob3-1 mec1-1 double mutants did not (14% of the large-budded cells had a single nucleus). Since the double mutants do not progress significantly into S phase under these conditions (see below), but produce binucleated large-budded cells, it appears that these strains have a “cut” phenotype in which a 1C DNA content is segregated into two nuclei. This could cause both the inviability and the rapid cessation of growth that were observed in the absence of the MEC1 checkpoint.
Loss of the MEC1 checkpoint was found to alter the MPT of pob3 mutants (Table 3 and Figure 5). In some cases (notably pob3-11) combination with the mec1-1 mutation caused a drop in the MPT, indicating that at elevated temperatures these strains were capable of growth only because of the action of the checkpoint, presumably due to the presence of levels of DNA damage that would be lethal if mitosis were permitted. Surprisingly, however, most pob3 alleles displayed an increase in the MPT when combined with the mec1-1 mutation. In several cases the suppression of temperature sensitivity was dramatic (pob3-1 and -7, for example), even though these same alleles caused dependence on the MEC1 checkpoint for retaining viability at 37° (Figure 7; data not shown). These alleles therefore allow growth at higher temperatures when MEC1 is mutated than when it is intact, but cause more rapid death at 37° when MEC1 is mutated. Examination of the cells rescued by the loss of the checkpoint (pob3-1 mec1-1 cells growing at 32° as in Figure 5, for example) reveals that they are abnormal in both cell and colony morphology and that they have reduced plating efficiency. Apparently these cells sustain enough damage at intermediate temperatures like 32° to cause cell cycle arrest through the MEC1 checkpoint pathway, but not enough damage to lead to cell death in every case if this arrest does not occur (see discussion).
pob3 mutants progress slowly through S phase: As a more direct test for proficiency of S phase progression, we examined the DNA content of cells with pob3 mutations by flow cytometry. Rapidly growing cultures shifted to 37° for up to 24 hr did not display dramatic changes in the distribution of DNA contents, although some alleles caused a small shift towards 1C, indicating a slight tendency to arrest in G1 (not shown). This differs from the spt16-G132D allele, which causes ~80% arrest in G1 (Prendergastet al. 1990; Rowleyet al. 1991; Wittmeyer and Formosa 1997), although this is not a property of all spt16 alleles (T. Formosa, unpublished observations). pob3 mutations therefore do not cause a uniform arrest at a unique point in the cell cycle, consistent with the data in Figure 8.
To examine populations synchronously traversing S phase, cells were arrested with the mating pheromone α-factor and then were released from this block into media at permissive or restrictive temperatures. POB3 cells entered S phase at either 22° or 37° after 50–60 min and completed replication by ~80 min (Figure 9, top). The pob3-1 mutants also traversed S phase at 22° with about the same timing as wild type, although fewer cells ultimately released from the block, suggesting a delay at the G1/S boundary. When pob3-1 cells were released to 37°, this G1/S delay was even more pronounced, with only about half of the cells released by 90 min. This shows that pob3-1 cells are less likely to traverse the G1/S boundary under restrictive conditions, reminiscent of the arrest of spt16 mutants that are unable to synthesize sufficient cyclin proteins to surmount this step (Rowleyet al. 1991). The pob3-1 cells that did enter S phase at 37° did so later than wild type and appeared to delay in S phase. Comparing the wild type at 60 and 70 min to pob3-1 at 70 and 80 min at 37°, it appears that the wild-type cells rapidly traverse S phase and accumulate as 2C cells soon after they release from G1, but the pob3-1 cells accumulate as cells with an intermediate DNA content indicating slow progression through S phase. The shape of the profile shown with pob3-1 cells released at 37° for 80 min shows that many cells have entered S phase but have not progressed efficiently. This shape was reproducible in several experiments, and was also observed using the pob3-7 allele (not shown). We conclude that Ts− pob3 mutants have difficulty entering S phase at a restrictive temperature and also progress slowly through S phase.
Since S phase progression is monitored by MEC1 and loss of this checkpoint altered the viability of pob3 mutants, we also tested the progression of replication in pob3 mec1 double mutants (Figure 9, bottom). The double mutants failed to progress significantly at the restrictive temperature, even after several hours of release from α-factor. Since 86% of these cells are binucleate (Figure 8), most of these cells must contain nuclei with less than a 1C DNA content.
Spt16 and Pob3 form a stable heterodimer in S. cerevisiae, and complexes of homologous proteins have also been found in human and frog cells (Okuharaet al. 1999; Orphanideset al. 1999). The human version of this factor (FACT) has been shown to promote elongation of RNA polymerase II on nucleosomal templates (Orphanides et al. 1998, 1999), and the frog version (DUF) has been implicated in DNA replication (Okuharaet al. 1999). Genetic evidence from yeast revealed a role for Spt16 in transcription (Prendergastet al. 1990; Maloneet al. 1991; Rowleyet al. 1991; Xu et al. 1993, 1995; Lycanet al. 1994; Brewsteret al. 1998; Evanset al. 1998), and physical and genetic evidence showed that Spt16-Pob3 interacts with DNA polymerase α (Wittmeyer and Formosa 1995, 1997; Formosa and Nittis 1999; Wittmeyeret al. 1999). We have proposed that Spt16-Pob3 acts both in transcription and replication, possibly by affecting the properties of the chromatin that is the common substrate for both processes. Here we have shown that conditional mutations in POB3 display transcription defects similar to those caused by mutations in SPT16, interact extensively with SPT16 as well as with several replication factors, and cause defects in DNA replication. These results strengthen the proposal that Spt16 and Pob3 act together in both transcription and replication.
Spt16 and Pob3 form a complex required for normal transcription: Both elevated SPT16 copy number and the spt16-G132D mutation cause altered transcription initiation site selection (Clark-Adamset al. 1988; Maloneet al. 1991; Rowleyet al. 1991), which can lead to either increased or decreased expression at different loci (Rowleyet al. 1991; Lycanet al. 1994). While elevated levels of POB3 were also found to cause some transcription defects, this did not include the Spt− phenotype (Brewsteret al. 1998). We isolated a set of pob3 mutants and found that all display the Spt− phenotype, indicating that full Pob3 function is required to produce normal patterns of transcription.
Spt16 and Pob3 were detected in whole cell lysates only in a complex with one another (Brewsteret al. 1998; Wittmeyeret al. 1999), so we expected them to function together. Consistent with this, we previously found that three alleles of POB3 all caused synthetic defects with the spt16-G132D mutation (Wittmeyeret al. 1999). While this allele causes the most stringent Ts− phenotype of any of the spt16 mutations we have isolated (T. Formosa, unpublished observations) and was isolated in four independent genetic screens (see Evanset al. 1998; T. Formosa, unpublished observations), we show here that its genetic interactions with POB3 are mild compared with another allele, spt16-4. While spt16-G132D significantly reduced the MPT when combined with any pob3 mutation, spt16-4 was lethal in 10 of 13 combinations tested. Further, while spt16-G132D was not noticeably affected by increased levels of POB3 copy number, the Ts− phenotype of spt16-4 was strongly suppressed by elevated POB3. SPT16 and POB3 therefore displayed the extensive allele-specific genetic interactions expected for members of a common complex.
Interactions with replication factors: Ctf4 and Spt16-Pob3 each bind to Pol1 (Miles and Formosa 1992a,b; Wittmeyer and Formosa 1997), and CTF4 mutations display synthetic lethality with mutations in the nuclease/helicase encoded by DNA2 and with mutations in a potential effector of polymerase processivity clamp loading encoded by CHL12 (Formosa and Nittis 1999). We found previously that the Ts− phenotypes of some pob3 and spt16 mutations were enhanced by deletion of CTF4 (Wittmeyer and Formosa 1997; Wittmeyeret al. 1999), although the apparent competition between Ctf4 and Spt16-Pob3 for binding to Pol1 might predict that the loss of Ctf4 should suppress the defects in SPT16 or POB3. We now find that the Ts− phenotype caused by some alleles of pob3 is suppressed by loss of Ctf4, indicating a complex relationship among these gene products. A similar pattern of allele-specific suppression or enhancement of temperature sensitivity was found with a chl12 deletion, suggesting that this potential PCNA clamp loading or unloading protein interacts with Spt16-Pob3 in a way similar to Ctf4. Synthetic defects with POL1 itself and with the nuclease/helicase encoded by DNA2 further strengthen ties between POB3 and members of the eukaryotic replication complex.
Ts− mutants result from partially impaired gene products that either become less functional at elevated temperatures or are unable to meet an increased requirement for their function at elevated temperatures. The Ts− pob3 mutations lead to decreased Pob3 protein levels under permissive conditions and undetectable levels at 37°. The effect of these mutations is therefore likely to be the same as complete deletion of the POB3 gene at 37°, but the interpretation of the phenotypes at semipermissive tempteratures is less obvious. The genetic interactions observed could be due to changes in Pob3 protein stability or changes in either Pob3 function or the level of Pob3 function required. These genetic interactions could therefore reflect physical interactions, as suspected in the POB3-SPT16 case, or they could be due to indirect effects. Therefore, the observation that pob3 mutations interact genetically with many replication factors suggests that Pob3 function is needed for normal DNA replication, but we cannot conclude that this is related to the direct Spt16-Pob3:Pol1 interaction detected in vitro. It is noteworthy that mutations in CTF4, DNA2, and CHL12, a set of genes that show mutual synthetic lethality, affected the same pob3 alleles most severely and that these mutations produce stable Pob3 proteins with defective C termini. The physical basis for these allele-specific interactions remains to be determined through further investigation.
HU inhibits dNTP synthesis and can be particularly toxic to cells lacking S phase checkpoints (since they fail to respond to slow progression of replication) and to cells with impaired replication (since two mechanisms that partially inhibit the same process independently cause a more severe defect when combined). pob3-7 caused sensitivity to HU at all temperatures, and other alleles caused HU sensitivity at elevated temperatures. Since all pob3 alleles tested caused the Spt− phenotype, reflecting a defect in transcription at temperatures permissive for growth, but only one allele caused HU sensitivity at low temperatures, HU sensitivity is unlikely to be a secondary effect of a transcription deficiency. Diminished Pob3 function therefore appears to affect the same process as HU, suggesting that Pob3 is required for normal DNA replication.
Detecting errors in replication: Alterations in transcription can be readily detected since a broad range of transcription rates can be tolerated for many genes. DNA replication is not so flexible: all sequences must be accurately copied once per cell cycle. Serious replication errors are therefore lethal, and detecting more subtle defects is difficult. Since many mutations in known replication factors lead to arrest as large-budded cells with a single nucleus and a 2C DNA content, this morphology has been inferred to suggest that replication errors might be present and could reflect the intervention of checkpoints that prevent mitosis until repair can occur (Elledge 1996; Paulovichet al. 1997; Longheseet al. 1998; Weinert 1998). Because of the various forms of repair and the limited capacity of each repair pathway, damaged DNA can lead to increased levels of recombination, increased chromosome loss, and sensitivity to further DNA damage. These phenotypes can therefore reveal some replication errors, but if replication fork progression is simply impeded, the DNA itself would not be considered damaged and these repair-mediated phenotypes would not be expressed even though replication was abnormal. pob3 mutants accumulated increased numbers of large-budded single-nucleated cells, which could reflect a replication defect, but neither spt16-G132D nor any of the pob3 alleles described here caused dramatic plasmid or chromosome fragment loss phenotypes, nor did they lead to sensitivity to DNA damage using MMS or UV (Wittmeyer and Formosa 1997; data not shown). These results suggest that spt16 and pob3 mutants do not require elevated levels of DNA repair, but do not rule out the possibility that progression of replication forks is impaired.
Checkpoints monitor the status of cell cycle events, acting to slow or prevent cell cycle progression when the genome or critical cellular structures like the mitotic spindle are flawed (Elledge 1996; Paulovichet al. 1997; Longheseet al. 1998; Weinert 1998; Amon 1999). The signals detected by the checkpoints are not known, but different genes appear to be required to monitor different aspects of DNA integrity. The MEC1 checkpoint is particularly important for monitoring S phase progression (Paulovich and Hartwell 1995; Desanyet al. 1998). Cells lacking MEC1 are sensitive to hydroxyurea (Desanyet al. 1998) and die rapidly in the presence of mutations in replication factors (Weinertet al. 1994). MEC1 has an essential role aside from its checkpoint function, which is related to deoxynucleotide production (Zhaoet al. 1998). While its precise mechanism is not clear, MEC1 seems to both monitor and participate in DNA replication, but has not been implicated in global regulation of transcription (although it does participate in the signal transduction pathway that induces some genes in response to DNA damage; Elledge 1996; Desanyet al. 1998; Weinert 1998). We therefore used a mec1-1 mutation to determine whether POB3 functions in DNA replication as well as in transcription, reasoning that even subtle errors that prevent normal progression of replication would lead to increased dependence on this checkpoint.
Double pob3 mec1 mutants were found to die more rapidly at 37° than either single mutant, leading to a 20- to 100-fold decrease in the number of viable cells after a 24-hr incubation at 37° when the checkpoint was inactive. The DNA content of these cells and their nuclear morphology were also examined and indicated that loss of the MEC1 checkpoint caused premature mitosis, since most pob3 mec1 cells had two nuclei but only a 1C DNA content. Loss of Pob3 function therefore causes a defect that signals a delay in cell cycle progression through the MEC1 checkpoint, and failure of the checkpoint causes inappropriate progression into mitosis and rapid death. Since MEC1 monitors S phase, the dependence on MEC1 strongly indicates that pob3 mutants cause a defect in DNA replication.
Since pob3 mutants depend on MEC1 function to remain viable at a restrictive temperature, we expected double pob3 mec1 mutants to display a synthetic defect and die at lower temperatures. This was the case for at least one allele, but, in general, inactivation of the checkpoint caused an increase in the maximal permissive temperature. We infer two conclusions from this observation. First, this means that the Ts− pob3 mutants arrest growth at elevated temperatures because of a problem with DNA replication rather than with transcription. Since, for example, pob3-1 strains grow at 32° only if the MEC1 checkpoint is inactive, this checkpoint is part of the mechanism that prevents growth of pob3-1 mutants at 32°. If pob3-1 mutants arrested growth at 32° due to a lethal deficiency in transcription, this should not be affected by loss of an S phase checkpoint. Therefore, since MEC1 monitors DNA replication and detects the pob3 mutant defect, the lethal defect in pob3 mutants is inferred to be in the replication pathway. Second, the MEC1-dependent checkpoint must be able to detect sublethal amounts of damage. As the temperature rises, a pob3-1 mutant encounters increasing amounts of damage until at 31° the signal through the MEC1 pathway is sufficient to prevent further growth. However, if the checkpoint is inactive at least some of the cells continue to progress. Between 31° and 34°, this is often lethal, but enough cells survive to produce significant growth. Above 34°, the amount of damage is always lethal so the checkpoint is needed to survive incubations under these conditions. The checkpoint is therefore able to detect a dangerous situation before it is actually lethal, indicating a prudent but not an essential course of action.
The role of Spt16-Pob3: Chromatin must be altered both to allow assembly of polymerase complexes at initiation sites and to allow passage of polymerases during elongation, so factors that mediate interactions between polymerases and chromatin could affect either initation or elongation. The Spt− phenotype in yeast is most readily explained by altered initiation site selection, while the activity of human FACT indicates a role in elongation. These conclusions appear contradictory, but altering chromatin could affect both phases of transcription.
The results reported with the frog DUF complex and those presented here are also ambiguous concerning the role of Spt16-Pob3 in DNA replication initiation or elongation. Depletion of DUF from extracts blocked replication as assayed both by nucleotide incorporation and by examination of replication intermediates (Okuharaet al. 1999). This could indicate either failure to initiate or a very early block to elongation. Using flow cytometry we found that at 37°, pob3-1 cells entered S phase inefficiently (not all cells participated) and traversed S phase slowly (cells accumulated with less than a 2C DNA content for longer than normal). The failure to release from G1 into S phase efficiently could be due to a transcription defect as observed for spt16-G132D (Rowleyet al. 1991), or it could be due to inefficient initiation at replication origins. Slow progression of S phase could also be caused by inefficient initiation, which would require each active replication fork to copy a larger portion of the genome and lead to a longer S phase, or it could be due to slow progression by individual replication complexes as expected if chromatin structures impeded replication to an abnormal extent. The sensitivity of pob3 mutants to hydroxyurea suggests a role in elongation, but might instead reflect an inability to initiate efficiently in the absence of normal dNTP concentrations. Further analysis of replication intermediates will therefore be needed to determine the role of Spt16-Pob3 in replication. The availability of the set of pob3 mutants described here will be useful for pursuing this approach.
We thank Jacqui Wittmeyer for providing strains, plasmids, and other materials used in initiating this project, and Jennifer Ginn for excellent technical assistance. We thank Jacqui Wittmeyer, David Stillman, and Brad Cairns for valuable discussions and improvements to this manuscript. This work was supported by a grant from the National Science Foundation to T.F.
Communicating editor: F. Winston
- Received December 21, 1999.
- Accepted April 17, 2000.
- Copyright © 2000 by the Genetics Society of America