High-Copy-Number Expression of Sub2p, a Member of the RNA Helicase Superfamily, Suppresses hpr1-Mediated Genomic Instability

Abstract

We report on a novel role for a pre-mRNA splicing component in genome stability. The Hpr1 protein, a component of an RNA polymerase II complex and required for transcription elongation, is also required for genome stability. Deletion of HPR1 results in a 1,000-fold increase in genome instability, detected as direct-repeat instability. This instability can be suppressed by the high-copy-number SUB2 gene, which is the Saccharomyces cerevisiae homologue of the human splicing factor hUAP56. Although SUB2 is essential, conditional alleles grown at the permissive temperature complement the essential function of SUB2 yet reveal nonessential phenotypes. These studies have uncovered a role for SUB2 in preventing genome instability. The genomic instability observed in sub2 mutants can be suppressed by high-copy-number HPR1. A deletion mutant of CDC73, a component of a PolII complex, is also unstable for direct repeats. This too is suppressed by high-copy-number SUB2. Thus, defects in both the transcriptional machinery and the pre-mRNA splicing machinery can be sources of genome instability. The ability of a pre-mRNA splicing factor to suppress the hyperrecombination phenotype of a defective PolII complex raises the possibility of integrating transcription, RNA processing, and genome stability or a second role for SUB2.

All cells have evolved mechanisms to maintain the integrity of their genomes. Homologous and nonhomologous rearrangements or recombinations occur primarily in response to DNA damage during mitotic growth and are mechanisms used to repair DNA lesions. To identify factors that participate in mitotic recombination and to understand the mechanisms of the recombinogenic processes, mutants that display altered rates of mitotic recombination have been isolated (23). In general, hyporecombination mutants are defective in some aspect of the recombination process and are typically associated with mutations in genes encoding components of the recombinational repair machinery. In contrast, mutations that lead to an accumulation of nicked or gapped or broken DNA strands can be identified by a hyperrecombination phenotype. Such mutations may lie within genes encoding proteins that participate in DNA replication, in nonrecombinogenic repair pathways, and in the repression of mitotic recombination. Importantly, hyperrecombination mutants can reveal sources of genome instability.

The HPR1 gene was identified through a mutation that increased the rate of deletion events between directly repeated DNA sequences in Saccharomyces cerevisiae (2). The hyperrecombination phenotype of hpr1Δ is reminiscent of chromosome instability syndromes, such as ataxia-telangiectasia, Werner syndrome, and Bloom syndrome (16, 17, 27). Additionally, hpr1 mutants display increased rates of chromosome and plasmid loss, indicating that Hpr1p is not exclusively involved in stabilizing directly repeated DNA sequences but has a general influence on genome stability (36).

Hpr1p shows sequence homology to yeast topoisomerase I, and an hpr1Δ mutation is lethal in combination with mutations that compromise or eliminate the function of any of the three yeast topoisomerases (3, 14). Interestingly, mutations in each of the topoisomerases can also lead to increased rates of intrachromosomal deletion events (10, 48). Other studies have revealed that the removal of histone H3-H4 (copy I) in hpr1Δ mutants is lethal (14, 53). These observations have led to the suggestion that Hpr1p can influence DNA structure (13).

Unlike other recombination mutants of Escherichia coli, S. cerevisiae, and Homo sapiens, the hpr1 mutant shows no DNA replication or repair defect. In fact, evidence accumulated thus far indicates that hpr1Δ-induced recombination is related to the process of transcription. Mutations in genes encoding transcription factors suppress hpr1Δ-associated phenotypes. These transcription factors include proteins involved in basal transcription such as Rpb2p and Sua7p (yTFIIB) (13); transcription mediators such as Sin4p (H.-Y. Fan and H. L. Klein, unpublished data), Srb2p and Hrs1p (32, 36), and gene-specific activators such as Gcr3p (44). A recent search for high-copy-number suppressors of hpr1Δ uncovered a novel gene, THO2/RLR1, which is involved in RNA polymerase II transcription (33, 50). The hyperrecombination phenotype associated with hpr1Δ mutants was also found to be more severe when direct repeats were highly transcribed (13), and recent evidence indicates that Hpr1p influences transcription elongation (7, 34).

To further explore the relationship between Hpr1p and transcription, we have studied a novel class of transcription factors in which mutations are also hyperrecombinogenic and are lethal in combination with a hpr1Δ mutation. Cdc73p is one example of this class. Cdc73p functions as a transcription factor and interacts with the C-terminal domain of the largest subunit of RNA polymerase II (39). Cdc73p is found in a distinct RNA polymerase II complex which also contains Paf1p, Ccr4p, and Hpr1p (6). cdc73Δ and paf1Δ mutants have multiple phenotypes. They show temperature-dependent growth, affect the abundance of some transcripts, and are hypersensitive to growth on 8 mM caffeine (6). The mutations also result in elevated rates of recombination between direct repeats, similar to hpr1Δ (6). The similarity in phenotype of Cdc73p and Hpr1p and their association in an RNA polymerase II complex are underscored by the finding that the hpr1Δ cdc73Δ double mutant is a temperature-sensitive lethal at 3°C. To understand how genome stability is maintained during transcription, we have isolated a high-copy-number suppressor of the hpr1Δ cdc73Δ double mutant. Further characterization of this suppressor, the SUB2 gene, has revealed that it is a putative RNA helicase involved in pre-mRNA splicing (22, 24, 52). SUB2 was originally isolated as a high-copy-number suppressor of the yeast snRNP biogenesis mutant brr1-1 (42). Precedence for a presumptive RNA helicase being a high-copy-number suppressor of a transcription component comes from the recovery of DHH1 as a high-copy-number suppressor of the transcription regulator complex CCR4 mutants pop2Δ and ccr4Δ (21). Our studies on the SUB2 gene have uncovered a novel role for this factor in the maintenance of genome integrity.

MATERIALS AND METHODS

Yeast strains, growth conditions and recombination rate determinations.

All strains are in the W303 background. Strains in Table 2 were obtained by transforming plasmids into HFY2170-6A (sub2Δ1::TRP1 + pHF68-1 [SUB2 CEN6, URA3]). The vector backbone of pHF68-1 is pRS316. pHF68-1 will be referred to as YCp2-SUB2. Transformants were streaked on fluoro-orotic acid (FOA)-containing leucine dropout medium to select for cells that lost pHF68-1. FOA-resistant cells were crossed to HFY998-1C to acquire the recombination assay system. Recombination rates were calculated as described previously (2).

TABLE 2.

Effect of SUB2 on recombination

Genotype	Plasmida	Recombination ratea (10−6)b	Fold over wild typea
SUB2	YCplac111	2.45 ± 0.09	1
SUB2	YCp1-SUB2	2.48 ± 0.11	1
sub2Δ1	YCplac111	Inviable	NAc
sub2Δ1	YCp1-SUB2	490 ± 160	200
sub2Δ1	YEp-SUB2	11 ± 8	4
sub2Δ1 mud2Δ	None	1,960 ± 710	800

^aYCp1-SUB2 (pHF127-3) contains a sequence encoding full-length Sub2p on the CEN4-based plasmid YCplac111. 

^bRates were calculated as described in footnote c to Table 1. 

^cNA, not applicable. 

Standard media were prepared as described previously (38). 5-FOA was used at 1 mg/ml. To assay gene expression at telomeres, strains were grown in leucine dropout medium to mid-log phase at 30°C, diluted, and spotted on leucine dropout plates with or without FOA. Yeast cells were grown at 30°C for 2 days, with the exception of the sub2 conditional allele strains, which were grown at 25°C until fully grown.

Generation of sub2 and mud2 deletion strains and SUB2 and MUD2 plasmids.

sub2 disruption plasmids pHF15-1 and pHF124-2 were constructed as follows. The HindIII-SacI fragment of YEp-SUB2 (hy41) was cloned into pBS-SK (Stratagene) to form pHF13-2. A SnaBI-PstI fragment containing TRP1 was used to replace the SnaBI-PstI SUB2 fragment from pHF13-2 to form pHF15-1. Plasmid pHF124-2 was generated by inserting HIS3 into the ClaI site of pHF120, which contains the SalI-SacI fragment of SUB2 in pBS-SK. The SacI-XhoI fragment of pHF15-1 and the EcoRI-SacI fragment of pHF124-2 were used to transform the diploid yeast strain HFY2115 to generate sub2Δ1::TRP1 and sub2Δ2::HIS3 mutants, respectively. The resulting diploids were sporulated and dissected. A 2+:2− segregation for growth was observed for both diploids and no Trp⁺ or His⁺ segregants were recovered, indicating that yeast cells carrying these mutations, sub2Δ1::TRP1 and sub2Δ2::HIS3, were inviable. pHF22-3 was generated by subcloning a SalI-SacI SUB2 fragment of YEp-SUB2 into pRS316. A fragment containing the 3′ untranslated region of the SUB2 gene (which was absent in YEp-SUB2) was obtained through PCR using primer oHF028 (5′AACGTTCATGGTCATATG3′) and oHF032 (5′CCGGAATTCTTGAAGAAGGCCTTCACC3′) and ligated into the EcoRI site of pHF22-3 to form pHF68-1 (YCp2-SUB2). Yeast strains heterozygous for sub2Δ alleles were subsequently transformed with YCp2-SUB2, sporulated, and dissected to generate sub2Δ1::TRP1 and sub2Δ2::HIS strains carrying YCp2-SUB2. The resulting strains were called HFY2170-6A and HFY2210-124B. pHF11-1 was constructed by restriction endonuclease digestion of hy41 with SalI and religation to remove an internal SalI fragment containing the YDL085w sequence. pHF81-4, pHF127-3, and pHF80-4 were generated by subcloning the SalI-SacI fragment of pHF68-1 into pRS315, YCplac111, and YEp351, respectively. To confirm that no changes were introduced into the SUB2 gene through PCR amplification, the SUB2 insert in pHF80-4 was completely sequenced and compared to the reported SUB2 sequence in the database. No changes were found. sub2 conditional alleles and a wild-type control SUB2 allele on the pRS315 vector (CEN6 LEU2) were kindly provided by Christine Guthrie and Amy Kistler.

YEp-MUD2 was constructed by insertion of the MUD2 sequence into the high-copy-number plasmid YEplac112-TRP1. Primers 5′CGCGGATCCATAGAACCGCTCCCCATGTC3′ and 5′GCGGGATCCGTCCTTCCATGAAGTTTGCCC3′ were used to amplify the MUD2 coding sequence. The PCR product was digested with BamHI and inserted into the BamHI site of YEplac112. The mud2::URA3 deletion was created by the one-step gene disruption method (35). PCR amplification was used to create the deletion cassette. Each primer consisted of 40 bp at the 5′ end that were homologous to MUD2 followed by 20 bp that were homologous to URA3. The primers used were 5′TATAGGAAAATCAGAAAAGGATGTTGTGCCGATTGAGAACAAAAGATTCATTGTACTGAGAGTGCACCAC3′ and 5′TCGTCCTCATCTATATAAGTACACAGAAC AGTGCGATCGTTGAATTGCGTTGTGCGGTATTTCACACCGC3′. The mud2 deletion retained the first 100 bp and the last 84 bp of the MUD2 coding sequence.

C terminus FLAG-tagged Sub2 protein was generated in vivo using a PCR-generated copy of SUB2 inserted into the high-copy-number vector YEp351. SUB2 was amplified from the yeast genome (Expand Long Template PCR System; Roche Diagnostics) using primers 5′GAAGGGATTCCTCCGTGTAG3′ and 5′CGCGGATCCTTACTTGTCATCGTCGTCCTTGTAGTCATTATTCAAATAAGTGGACGG3′. The latter primer contains the information for the FLAG epitope as well as a BamHI restriction site. This construct was then cloned into YEp351 at the SalI and BamHI restriction sites. Four hundred seventy base pairs directly 3′ of genomic SUB2 were then PCR amplified (primers used were 5′CGCGGATCCAAAAAAGATACGTTTTTATATAG3′ and 5′CGCGAGCTCCGAATTGAAGAAGGCCTTCACC3′) and cloned into the SUB2-FLAG-containing vector using the BamHI and SacI restriction sites.

Site-directed mutagenesis.

sub2-112 and sub2-267 were generated using the QuikChange site-directed mutagenesis kit (Stratagene). pHF81-4 and pHF80-4 were used as templates, and oHF043, oHF044, and oHF048 as well as oHF049 were used as primers for PCRs. oHF043 (5′ GCAAAGTCTGGTTTAGGTAGGACAGCTCTCTTTGTC 3′) introduces an A-to-G mutation at position 335. oHF048 (5′ CTTACAGAATCCATTGAAAATTTTCGTCGATGATG 3′) introduces a G-to-A change at nucleotide 799. oHF044 and oHF049 are complementary to oHF043 and oHF048, respectively.

Determination of plasmid loss rates.

Plasmid loss rates of strains containing pRM102 (CEN6 TRP1) and YEp351 (2μm LEU2) with no insert (YEp-vector), with a SUB2 insert (YEp-SUB2), or with a sub2-112 insert (YEp-sub2-112) were calculated as described previously (9). Briefly, cells were grown in synthetic liquid medium lacking both tryptophan and leucine to mid-log phase. Equal aliquots were taken and plated onto leucine dropout medium and tryptophan-leucine double-dropout medium to determine the percentages of plasmid (pRM102)-containing cells (P1). The culture was then diluted 1:1,000 into leucine dropout liquid medium to release the selection for pRM102 and grown to stationary phase. Again, equal aliquots were plated onto leucine dropout and tryptophan-leucine double-dropout media to determine the percentages of cells that still contained plasmid pRM102 (P2). Plasmid loss rates (m) were calculated using the equation m = 1 − e^{ln (P2/P1)/g}, where g is the number of cell doublings during nonselective growth and is described by the equation g = ln (N₂/N₁)/ln 2, where N₁ equals the number of viable cells per milliliter before nonselective growth and N₂ equals the number of viable cells per milliliter after nonselective growth (viable cells are based on the number of colonies growing on leucine dropout medium).

For experiments using pRM102CYC1ter, the CYC1 transcriptional terminator (30) was amplified by PCR using W303 DNA as a template. Primers were designed to add ApaI restriction sites at the ends of the amplified sequence.

Localization of the Sub2-HA protein.

Indirect immunofluorescence was performed according to the method of Harlow and Lane (20). The Sub2-HA-tagged protein was detected with monoclonal anti-HA antibody (Babco) and visualized with Cy2-conjugated anti-mouse antibody (Jackson Laboratory). Nuclei were stained with 4′,6′-diamidino-2-phenylindole (Boehringer Mannheim).

In vitro synthesis and analysis of Sub2p and Mud2p.

The Sub2-HA and Mud2-Flag proteins were synthesized using the TNT reticulocyte lysate system (Promega). Immunoprecipitation was performed as described by Goto and Meyerowitz (18). Monoclonal anti-Flag antibody (1:1,000; Kodak) was used for immunoprecipitation. Polyclonal anti-HA antibody (1:1,000; Santa Cruz) was used for the Sub2-HA protein detection.

Immunoprecipitation and Western blotting.

Approximately 10⁸ log-phase cells were collected for protein extracts. The cells were resuspended in 0.4 ml of ice-cold lysis buffer (50 mM HEPES-KOH [pH 7.5], 140 mM NaCl, 1.0 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate) with protease inhibitors (1.0 mM phenylmethylsulfonyl fluoride, 1.0 mM benzamidine, 0.05 mg of N-tosyl-l-phenylalanine chloromethyl ketone [TPCK] per ml, 0.25 mM Nα-p-tosyl-l-lysine chloromethyl ketone [TLCK], 0.01 mg of aprotinin per ml, 0.001 mg of leupeptin per ml, 0.001 mg of pepstatin A per ml, 0.002 mg of antipain per ml) and with 1/2 volume of acid-washed glass beads and kept on ice. Tubes were vortexed for 1 min and then transferred back to ice for an additional minute. This was repeated (usually 10 times) until approximately 90% of the cells were broken. The extract was separated from the beads and cell debris and stored at 4°C.

For the immunoprecipitation experiments, extracts from strains expressing Sub2-FLAGp were incubated with 1.0 μg of mouse anti-Rpb3p antibody (Neoclone) on a shaking platform at 4°C for 2 h. A 1.0-μg amount of anti-mouse immunoglobulin G biotin conjugate antibody (Sigma) was then incubated with the extracts for another 2 h at 4°C. Two hundred microliters of streptavidin-coated magnetic beads (Polysciences, Inc.) was subsequently added to the extracts and incubated an additional 3 h at 4°C. After separation with a magnet, the beads were washed three times in lysis buffer for 15 min. The beads were then resuspended in 1% sodium dodecyl sulfate (SDS) Tris-EDTA (TE) and incubated at 65°C for 10 min to release protein from the beads. Equal volumes were mixed with 2× SDS sample buffer, boiled, and electrophoresed on 9% SDS-polyacrylamide gels. Western blot analysis was performed using anti-FLAG M2 monoclonal antibody (Sigma) as the primary antibody and anti-mouse immunoglobulin G-HRP (Santa Cruz Biotechnology) as the secondary antibody.

RESULTS

Genetic interaction between hpr1Δ and cdc73Δ mutations.

Previous studies have suggested a role for HPR1 in transcription (7, 33, 34). Therefore, we examined genetic interactions between an hpr1Δ mutation and mutations in genes encoding transcription factors to further understand the relationship between mitotic recombination and transcription. Although both hpr1Δ and cdc73Δ single mutants are viable at 30°C, an hpr1Δ cdc73Δ double mutation is lethal at 30°C (Fig. 1) (6). Further studies uncovered similarities in the phenotypes of these two mutants. hpr1Δ and cdc73Δ strains both showed temperature-dependent growth and increased instability between directly repeated DNA sequences (Fig. 1 and Table 1) (5), with hpr1Δ and cdc73Δ mutants having increases of 700- and 70-fold, respectively, in recombination rates (Table 1). These observations suggest that Hpr1p and Cdc73p may perform similar functions in parallel pathways or work together in a complex. This prediction was verified by the finding that Hpr1p and Cdc73p are complexed together in a novel RNA polymerase II complex (6).

FIG. 1.

Suppression of growth defect associated with hpr1Δ, cdc73Δ, and hpr1Δ cdc73Δ strains by elevated-copy-numbers of the SUB2 gene. Wild type (HKY579-10A), hpr1Δ (U768-1C), cdc73Δ (HFY580-104), and hpr1Δ cdc73Δ (HFY2051-1D) carrying either YEp-vector (YEp351) or YEp-SUB2 (hy41) were streaked on leucine dropout plates to select for plasmids as indicated and allowed to grow at 30 or 37°C for 2 days. YEp351 is a 2μm-based vector, and hy41 contains the SUB2 gene in the YEp351 plasmid.

TABLE 1.

Effects of overexpressing SUB2 on deletion events between direct repeats^a

Genotypeb	Recombination rate (10−6)c	% Suppression
Wild type + YEp-vector	2.50 ± 0.06
Wild type + YEp-SUB2	3.01 ± 0.46	0
hpr1Δ + YEp-vector	1,800 ± 310
hpr1Δ + YEp-SUB2	190 ± 97	90
cdc73Δ + YEp-vector	170 ± 40
cdc73Δ + YEp-SUB2	20 ± 3	90

^aStrains used in this table are HKY870-12A (wild type), HFY998-2C (hpr1Δ), and HFY2059-1A (cdc73Δ) transformed with either YEp-vector or YEp-SUB2. 

^bYEp-vector, a 2μm-based plasmid YEp351, is the control vector; YEp-SUB2 is the original hy41 isolate from a YEp351 library. pHF80-4, which contains a sequence encoding full-length Sub2p, behaves similarly to hy41. 

^cRates were calculated as described in Materials and Methods from 5-FOA resistance frequencies of strains carrying the duplication leu2-k::ADE2-URA3::teu2-k. Each rate was calculated from three independent fluctuation tests on three strains of the same genotype and is expressed as the mean rate ± the standard deviation of three determinations. 

Since an hpr1Δ cdc73Δ double mutation is lethal, an essential function must be carried out by Hpr1p and Cdc73p at 30°C. To begin to understand what this essential function is, we searched for high-copy-number suppressors that restored the viability of an hpr1Δ cdc73Δ double mutant.

Increased copies of SUB2 rescue the inviability of an hpr1Δ cdc73Δ mutant.

Although an hpr1Δ cdc73Δ double mutant is inviable at 30°C, it grows extremely slowly at 25°C. We used this phenotype to isolate high-copy-number suppressors by transforming an hpr1Δ cdc73Δ strain grown at 25°C with a YEp351-based high-copy-number genomic library and selecting for growth at 30°C. The YEp-SUB2 (hy41) plasmid was isolated based on its ability to restore viability to an hpr1Δ cdc73Δ double mutant at 30°C (Fig. 1). Sequence analysis of this plasmid revealed that it contains nearly the full-length coding sequence of YDL084w, lacking only a sequence encoding three amino acids at the C terminus of YDL084w, and a partial coding sequence of YDL085w (amino acid residues 27 to 535). To confirm that YDL084w and not YDL085w was responsible for the suppression, the YDL085w sequence was completely deleted from hy41. The resulting plasmid, pHF11-1, still rescued the lethality of an hpr1Δ cdc73Δ strain (data not shown). In addition, pHF80-4, a high-copy-number plasmid containing a YDL084w sequence encoding the full-length protein, behaved similarly to hy41 (data not shown). These results demonstrated that the YDL084w sequence on a high-copy-number plasmid can restore viability to an hpr1Δ cdc73Δ double mutant. YDL084w was thereafter termed SOH9 (for suppressor of hpr1Δ), the latest in our series of hpr1Δ suppressors (14). We subsequently learned that the open reading frame YDL084w had been isolated as a high-copy-number suppressor of the yeast splicing mutant brr1-1 mutant and was termed SUB2 (42).

Since HPR1 has been shown to be involved in transcription (7, 34) and the Hpr1 protein is associated with RNA polymerase II complexes (6), it was important to determine whether SUB2 expression was altered in an hpr1Δ mutant. Northern analysis of SUB2 mRNA from HPR1 and hpr1Δ strains showed no difference in amount. This eliminates the trivial explanation for recovery of SUB2 as a high-copy-number suppressor and strengthens the argument for a significant biological interaction between Hpr1p and Sub2p.

Suppression of hyperrecombination by an increase in SUB2 copy number.

YEp-SUB2 suppresses the lethality of an hpr1Δ cdc73Δ double mutant. To determine whether SUB2 suppresses the phenotypes associated with hpr1Δ or cdc73Δ single mutants, we evaluated the effects of SUB2 on growth of hpr1Δ and cdc73Δ strains at 37°C and on recombination rates of deletion events between directly repeated DNA sequences. As shown in Fig. 1, YEp-SUB2 restored the growth of an hpr1Δ strain at 37°C but not the growth of the cdc73Δ single or hpr1Δ cdc73Δ double mutant at 37°C. Additionally, YEp-SUB2 suppressed 90% of the increased recombination observed in the hpr1Δ and cdc73Δ single mutants at 30°C (Table 1). We next determined whether Sub2p has a general influence on spontaneous or transcription-induced recombination (43) between direct repeats in wild-type strains. In contrast to what occurs with the hpr1Δ and cdc73Δ mutant strains, an increase in SUB2 copy number does not suppress the rate of spontaneous recombination in wild-type strains (Table 1) or transcription-induced recombination (Fan and Klein, unpublished). This indicates that Sub2p acts on recombination events occurring in the absence of Hpr1p or Cdc73p but not when both gene products are functional.

Additional hpr1Δ phenotypes suppressed by high-copy-number SUB2

We have found that hpr1Δ strains are compromised in the ability to retain certain YCp plasmids that carry strong yeast promoters. Figure 2 shows pRM102, which carries a 1.7-kb BamHI fragment with the HIS3 gene and fragments of the DED1 and SUP56 genes that include the promoters to these genes. The HIS3 fragment is inserted in the polylinker of the YCp plasmid pRS314. The strains also carry the empty vector YEp351 (YEp-vector) or YEp351 with a SUB2 insert (YEp-SUB2). The pRM102 plasmid is stable in wild-type strains, but stability is decreased 2.6-fold in hpr1Δ strains (Fig. 2, compare wt + YEp-vector to hpr1Δ + YEp-vector [P < 0.001]). The pRM102 plasmid has the strong DED1 promoter transcribing leftwards. Evidence that the plasmid instability is due to transcription comes from the finding that removal of the DED1 sequence or placement of a CYC1 transcription terminator sequence immediately downstream of the DED1 sequence restores plasmid stability to the level observed in wild-type strains. Insertion of the CYC1 terminator sequence decreased plasmid loss rate 7.2-fold in the hpr1Δ strain to an extent such that the plasmid was more stable than the pRM102 plasmid in an HPR1 strain (Fig. 2).

FIG. 2.

Plasmid instability in hpr1Δ strains is suppressed by high-copy-number SUB2. Wild-type (wt) and hpr1Δ strains were transformed with plasmid pRM102 and YEp351 with no insert or with a SUB2 insert. pRM102CYC1ter contains a CYC1 transcriptional terminator inserted into the ApaI restriction site located directly downstream of the ded1 sequence on pRM102. Plasmid loss rates were determined after transfer from selective to nonselective growth conditions for the pRM102 plasmid. Loss rates were determined as described previously (9). Plasmid loss rates were averaged from three independent experiments.

Another way to restore the stability of the pRM102 plasmid in an hpr1Δ strain is to overexpress SUB2. The introduction of YEp-SUB2 into hpr1Δ strains results in increased levels of stability close to those of the wt + YEp-SUB2 strains (Fig. 2, compare wt + YEp-SUB2 to hpr1Δ + YEp-SUB2). The ability of high-copy-number SUB2 to maintain the stability of pRM102 in hpr1Δ strains is dependent upon an intact nucleotide binding motif (GKT) in the putative helicase sequence (see below). The mutation of lysine 112 to arginine results in loss of suppression by SUB2. The additional increase in instability in both wild-type and hpr1Δ strains when the variant sub2-112 is overexpressed is suggestive of a dominant negative phenotype (P < 0.001 for both wt + YEp-sub2-112A and hpr1Δ + YEp-sub2112A compared to wt + YEp-SUB2 or hpr1Δ + YEp-SUB2). Since plasmid stability can be increased by reducing transcription via a transcription terminator or high-copy-number expression of SUB2, this suggests that high-copy-number SUB2 acts on the hpr1Δ strain at some level of transcription.

SUB2 is an essential gene and encodes a putative RNA helicase.

sub2 deletion mutants were generated to determine the requirement of SUB2 for cell growth. Two sub2 disruption constructs were made and used to transform a wild-type diploid strain. Tetrad analyses of the two sub2 disruption constructs, sub2Δ1::TRP1 and sub2Δ2::HIS3, were not viable in haploid spore segregants, indicating that the SUB2 gene is indispensable for cell viability (data not shown), as previously reported by Shiratori et al. (40). To maintain the inviable sub2Δ mutant strains, the wild-type SUB2 gene was segregated into sub2Δ spore segregants as described in Materials and Methods (Fig. 3). This is consistent with the phenotype reported by Lopez et al. (25), Kistler and Guthrie (22), and Zhang and Green (52).

FIG. 3.

SUB2 is an essential gene, as illustrated by complementation of the sub2Δ mutants by different plasmids. Shown is the growth of SUB2 (HKY579-10A), sub2Δ1::TRP1 + YCp2-SUB2 (pHF68-1) (HFY2170-6A) and sub2Δ2:: HIS3 + YCp2-SUB2 (pHF68-1) (HFY2210-114B) on yeast extract-peptone-dextrose and FOA-containing media. To examine the effects of SUB2 or sub2 mutant alleles (on CEN vectors) in complementing the lethality of the sub2Δ1 strain, HFY2170-6A and HFY2210-114B were transformed with various plasmids. The transformants were checked for viability on FOA-containing medium. The genotype of each strain is as indicated. Strains streaked on the FOA-containing medium all carried YCp2-SUB2. Since only uracil-auxotrophic cells can grow on media containing FOA, all cells grown on the FOA-containing medium have lost YCp2-SUB2. SUB2 is an essential gene, as sub2 deletion strains failed to grow on FOA-containing medium which selects against YCp2-SUB2. YCp1-SUB2 (pHF127-3) and YCp2-sub9-267 can substitute for YCp2-SUB2 and rescue the sub2Δ strains, indicating that these plasmids contain a functional SUB2 fragment. YCp2-sub2-112, which contains a mutation in the ATP binding domain of the SUB2 gene, was unable to replace YCp2-SUB2, suggesting that the putative ATPase activity is necessary for the Sub2p function.

SUB2 contains an open reading frame of 1,338 nucleotides and encodes a protein 446 amino acids in length. The predicted amino acid sequence revealed that the Sub2 protein is a putative RNA helicase and belongs to the DECD subfamily of the DEAD-box-containing ATP-dependent RNA helicase family. Members of this subfamily have been found in a variety of species including humans, pigs (31), Drosophila melanogaster (49), Schizosaccharomyces pombe, and S. cerevisiae. Sequence alignments of Sub2p with family members from humans, Drosophila, and S. pombe are shown in Fig. 4. Sub2p is 70% identical and 78% similar to the S. pombe protein (GenBank accession number Z99162), 64% identical and 74% similar to the Drosophila protein WM6/HEL, 66% identical and 74% similar to the human protein UAP56/BAT1, and 66% identical and 76% similar to a second human family member protein (U90426).

FIG. 4.

Sequence alignment of Sub2p and related proteins. Shown are sequence alignments of the S. cerevisiae Sub2p (GenBank accession number Z74132) with related proteins from S. pombe (Z99162), Drosophila (WM6/HEL, X79802), human 1 (UAP56/BAT1, Z37166), and human 2 (U90426). Sequences were aligned according to the Hotun Hein algorithm method with a PAM250 weight table using Lasergene sequence analysis software. Shaded regions contain amino acid identity. Boxed sequences represent the conserved helicase motifs.

This RNA helicase family contains several conserved motifs, including the consensus sequence AXXGXGKT, which functions in ATP hydrolysis. To determine whether the potential ATPase activity of Sub2p was required for its function, we mutated the position 112 lysine residue to arginine in the conserved ATPase motif and studied the phenotype of the resulting sub2 mutant plasmid. A sub2Δ strain containing the pSub2-112 plasmid depends upon YCp2-SUB2 for growth, indicating that the potential ATPase activity of Sub2p is necessary for cell viability (Fig. 3).

Characterization of the Sub2 protein and mRNA.

To determine the subcellular localization of Sub2p, an HA epitope was introduced at the carboxyl terminus of Sub2p, and DNA sequence encoding the Sub2-HA protein was placed downstream of the inducible GAL10 promoter. This fusion protein complemented the sub2Δ mutation, indicating that the HA-tagged Sub2 protein behaves like the endogenous Sub2 protein. Indirect-immunofluorescence analysis demonstrated that the Sub2-HA protein was localized to the nucleus (data not shown). The Drosophila homologue of Sub2p, WM6/HEL, was identified as a suppressor of a wee1⁻ mik1⁻ double mutant of S. pombe, which results in mitotic catastrophe (49), suggesting that Sub2p may play a critical role in cell cycle control. Therefore, we determined whether SUB2 expression fluctuated during the cell cycle. Northern blot analysis indicated that SUB2 produced a single transcript that was expressed constitutively throughout the cell cycle (data not shown.). This is consistent with recent reports on genome-wide analyses of yeast mRNA levels during the cell cycle (8, 41).

Sub2p interacts with Mud2p in vitro.

Sub2p shows high sequence homology to the human protein UAP56/BAT1 (Fig. 4) (see also references 22, 24, and 52). UAP56 has been found to interact with U2AF⁶⁵, which is a branch point recognition protein and is involved in pre-mRNA splicing (15). The yeast Mud2 protein, which resembles the human U2AF⁶⁵ protein, also binds to pre-mRNA and is a component of the pre-mRNA-U1-snRNP complex (1). We examined Mud2p and Sub2p interaction through Flag-tagged Mud2p and HA-tagged Sub2p. In vitro-translated proteins were mixed and examined for interaction through immunoprecipitation. Figure 5A shows that Sub2-HA coimmunoprecipitated with the Mud2-Flag protein, indicating that Sub2p binds to Mud2p in vitro. Interaction of Sub2p with Mud2p has also been noted by another group (22). The interaction strengthens the thesis that Sub2 is the yeast counterpart of UAP56/BAT1. Moreover, Sub2p has recently been demonstrated to be required for pre-mRNA splicing in vitro and in vivo (22, 24, 52).

FIG. 5.

Interaction of Sub2p with Mud2p in vitro and Rpb3p in vivo. (A) Sub2-HA and Mud2-Flag proteins were synthesized in vitro using a coupled transcription and translation system. Sub2-HA and Mud2-Flag proteins were mixed and immunoprecipitated with monoclonal antibodies as indicated. Immune complexes were resolved on an SDS-polyacrylamide gel electrophoresis gel which was transferred and probed with polyclonal anti-HA antibody. Lane 4 indicated that the Sub2-HA and Mud2-Flag proteins coimmunoprecipitated. Lane 1, positive control; lanes 2 and 3, negative controls. The positions of protein size markers are indicated. (B) Coimmunoprecipitation of Sub2p and the RNA polymerase II component Rpb3. Lanes 1 and 2, Western blots of extracts from HPR1 (wt) + YEp351-SUB2-FLAG and hpr1 + YEp351-SUB2-FLAG cells. Extracts were electrophoresed on a 9% SDS-polyacrylamide gel and subjected to Western blot analysis with anti-FLAG antibodies. Lanes 4 and 6 show extracts from the strains used in lanes 1 and 2 after immunoprecipitation using mouse anti-Rpb3 antibody, goat anti-mouse with a biotin conjugate, and then streptavidin-coated beads. Lanes 3 and 5 are controls in which only the goat anti-mouse biotin conjugate and streptavidin beads were added to the same extracts. The immunoprecipitated proteins were electrophoresed on a 9% SDS-polyacrylamide gel and subjected to Western blot analysis with anti-FLAG antibodies.

Suppression of hpr1Δ temperature-dependent growth by loss of MUD2

MUD2 is not an essential gene, whereas SUB2 is essential. This suggests that there may be multiple roles for SUB2. In pre-mRNA splicing, Sub2p has been proposed to aid in the removal of Mud2 from the pre-mRNA-U1-snRNP complex (22). It was of interest to examine the effect of a mud2Δ mutation on hpr1Δ growth, as loss of Mud2p might free up Sub2p to participate in other activities. Figure 6 shows growth of wild-type, mud2Δ, hpr1Δ, and hpr1Δ mud2Δ strains at 30 and 37°C carrying the empty vector YEp-vector or the high-copy-number YEp-SUB2 plasmid. The hpr1Δ strain carrying only the empty vector does not grow at 37°C, but the hpr1Δ mud2Δ double mutant does show some growth at 37°C. High-copy-number SUB2 allows slightly better growth of the hpr1Δ mud2Δ strain at 37°C, showing that this suppression is independent of MUD2. This is consistent with the finding that a MUD2 deletion can bypass the cellular requirement for SUB2 function (22).

FIG. 6.

Suppression of hpr1Δ growth defect is MUD2 independent. Wild type (HKY579-10A), hpr1Δ (HFY824-1A), mud2Δ (RMY161-2C), and hpr1Δ mud2Δ (RMY159-11A) carrying either YEp-vector (YEp351) or YEp-SUB2 (pHF80-4) were streaked on leucine dropout plates to select for plasmids as indicated and allowed to grow at 30 or 37°C for 2 days. YEp351 is a 2μm-based vector, and hy41 contains the SUB2 gene in the YEp351 plasmid.

Effect of SUB2 on recombination.

SUB2 suppresses the hyperrecombination phenotype of hpr1Δ and cdc73Δ mutants, suggesting that Sub2p may be involved in regulating recombination between direct repeats. To investigate a potential role of Sub2p in recombination, we determined the effects of varying the copy number of SUB2 on the maintenance of stability between directly repeated DNA sequences. The recombination rate of a sub2Δ strain containing the pHF127-3/YCp1-SUB2 plasmid (SUB2 [full-length] CEN4) was found to be increased relative to a wild-type strain containing a YCp empty vector (Table 2). This indicates that, although YCp1-SUB2 rescues the lethality of a sub2Δ strain (Fig. 3), it is unable to fully complement all aspects of Sub2p function, including rescue of hpr1Δ growth at 37°C. Indeed, a similar result was obtained when SUB2 was carried on another CEN-based plasmid (CEN6), plasmid YCp2-SUB2 (Fan and Klein, unpublished). We do not know the basis of the lower-copy-number SUB2 recombination phenotype. It may be related to titration of Sub2p by Mud2p. We do not believe that our plasmid versions of SUB2 are missing important transcriptional or translational regulatory elements, although this is always a formal possibility. We have reproduced the hyperrecombination phenotype with a wild-type SUB2 CEN-based plasmid (p297) obtained from C. Guthrie (Table 3). The SUB2 insert in p297 contains more upstream and downstream sequences than does the pHF127-3 SUB2 insert. Increasing the copy number of SUB2 in a wild-type strain had no effect on recombination rates.

TABLE 3.

Effects of sub2 conditional alleles on recombination

Genotype	Plasmida	Recombination ratea (10−6)b	Fold over wild typea
SUB2 HPR1	p297 SUB2	0.93	1
hpr1Δ	p297 SUB2	140	150
sub2Δ1	p297 SUB2	140	150
sub2Δ1 hpr1Δ	p297 SUB2	550	590
SUB2 HPR1	p322 sub2-1	2.8	1
hpr1Δ	p322 sub2-1	250	90
sub2Δ1	p322 sub2-1	4,200	1,500
sub2Δ1 hpr1Δ	p322 sub2-1	610	220
SUB2 HPR1	p326 sub2-6	5.6	1
hpr1Δ	p326 sub2-6	470	85
sub2Δ1	p326 sub2-6	1,200	214
sub2Δ1 hpr1Δ	p326 sub2-6	1,100	200
SUB2 HPR1	p320 sub2-5	2.1	1
hpr1Δ	p320 sub2-5	200	95
sub2Δ1	p320 sub2-5	780	370
sub2Δ1 hpr1Δ	p320 sub2-5	460	220

^aAll plasmids are based on a SUB2 insert into plasmid pRS315 (CEN6 LEU2). 

^bRates were calculated as described in footnote c to Table 1. 

To eliminate a concern that the recombination phenotype of the sub2Δ strain containing pHF127-3/YCp1-SUB2 reflected a difference between the SUB2 allele from the plasmid library versus the SUB2 allele in our wild-type W303 based strains, both SUB2 alleles were sequenced. Two changes from the published sequence were found in the W303 SUB2 chromosome allele. The change at position 1363 of the nucleotide sequence from G to T is a silent change that leaves the Leu residue unchanged. The change at position 1139 from A to G results in a change from Thr to Gly. We tested the effect of the W303 SUB2 allele on recombination by comparing the recombination rate of a sub2Δ strain with the library version YCp1-SUB2 to the rate obtained from a sub2Δ strain with the W303 version YCp1-SUB2. No significant difference was found (1.3 × 10⁻⁴ versus 2.4 × 10⁻⁴), indicating that the recombination phenotype is not the result of strain-specific SUB2 alleles.

We then examined the effects of mutant sub2 alleles on recombination. The nonnull alleles were isolated by A. Kistler and C. Guthrie on the basis of temperature-conditional growth by a plasmid shuffle protocol (A. Kistler and C. Guthrie, personal communication). The results, shown in Table 3, reveal that the conditional sub2 alleles grown at permissive temperature lead to an increase in recombination between direct repeats. The increase in recombination over that of the wild type is of the magnitude of rates we have observed in hpr1Δ strains (Table 1, hpr1Δ + YEp351) (see references 3 and 14). Recombination rates in the sub2Δ1 strains are increased when a mutant allele of SUB2 is on the plasmid, compared to the rates observed with the wild-type allele (p297 versus p322, p326, and p320), indicating a role for functional Sub2p in suppressing recombination. Finally, we have been able to recover viable sub2Δ1 strains by including a mud2Δ mutation in the strain. Viability of the sub2Δ mud2Δ genotype has been reported by Kistler and Guthrie (22). We used this strain to determine the effects of the complete loss of Sub2p in repeat stability. As reported in Table 2, the recombination rate of this strain is 800-fold increased over that of the wild type. The mud2Δ mutation has no effect on recombination.

We note that the low-copy-number plasmids bearing SUB2 partially suppress hpr1Δ recombination (compare Table 1 hpr1Δ + YEp-SUB2 to Table 3 hpr1Δ + p297 YCp-SUB2). The fact that the double-mutant sub2Δ1 hpr1Δ shows no synergistic increase in recombination over the single-mutant strains suggests that these mutants do not act in independent processes to give increased repeat instability.

Interaction of Sub2p with the RNA polymerase II transcription complex.

Since sub2 mutants have recombination phenotypes similar to that of the hpr1Δ mutant, high-copy-number SUB2 suppresses hpr1Δ mutant phenotypes and Hpr1p is found in an RNA polymerase II complex, it was of interest to determine whether we could find Sub2p associated with an RNA polymerase II complex. We used FLAG-tagged Sub2p in coimmunoprecipitation reactions with anti-RPB3. Rpb3p is a subunit of RNA polymerase II. We were able to detect Sub2-FLAG in a coimmunoprecipitate in an hpr1Δ strain but not in the isogenic HPR1 strain (Fig. 5B). This suggests that Sub2p may substitute for Hpr1p in the RNA polymerase II complex.

Involvement of the Sub2 protein in transcriptional repression at telomeres.

The HEL/WM6 protein, the Drosophila homologue of Sub2p, was also identified through a mutation that enhanced position effect variegation (PEV) (12). PEV is the change in gene expression levels when a euchromatic gene is translocated adjacent to heterochromatic regions, from variations in the extent of propagation of heterochromatin-associated proteins into euchromatic structures (for a recent review, see reference 47). The chromatin structure at telomeres and the silent MAT loci can be considered heterochromatin in S. cerevisiae. The silencing of sequences adjacent to telomeres is called telomere position effect in yeast cells. We examined gene expression at the telomeric regions and at the HMR locus in sub2Δ1 strains carrying the YCp1-SUB2 plasmid, using the reporters adh4::URA3-TEL and hmr::TRP1. In wild-type cells, chromatin structure at these two regions results in transcriptional silencing (4, 19); therefore, cells are Ura⁻ (growth on FOA-containing media) and Trp⁻. Strains were allowed to grow on leucine dropout medium to maintain the SUB2-containing plasmids. The amount of growth on an FOA-containing leucine dropout plate indicates the expression level of the reporter gene URA3, while growth on the leucine dropout plate indicates the relative growth rates and the total number of cells plated for each dilution. All strains grew equally well on a leucine dropout plate (Fig. 7A). Wild-type cells carrying YCp1-SUB2 grew as well on FOA plates as did wild-type strains carrying the YCplac111 vector (Fig. 7B), indicating that an increase in SUB2 copy number did not influence transcriptional silencing at telomeres to a noticeable level. Interestingly, the number of colonies of the YCp1-SUB2-containing sub2Δ strain that grew on a FOA plate was reduced compared to that of the wild-type strain carrying YCp1-SUB2, indicating an involvement of SUB2 in the maintenance of a transcriptionally repressed chromatin state at telomere regions. Control experiments using a sub2Δ URA3 + YCp-SUB2 strain showed that this genotype had no effect on growth on 5-FOA-containing medium, the URA3 allele being located at the normal URA3 locus on chromosome V. The influence of SUB2 on telomere position effect could be indirect through altered splicing of genes that more directly affect this process.

FIG. 7.

Involvement of Sub2p in transcriptional silencing at telomeres. Serial dilutions of yeast strains were plated on a leucine dropout plate (A) or an FOA-containing leucine-dropout medium (B). Leucine dropout medium selects for cells containing plasmids, and FOA selects against Ura⁺ cells. Lanes 1 and 2, wild-type cells containing YCp1-vector (YCplac111) and YCp1-SUB2 (pHF127-3), respectively. Lane 3, sub2Δ cells carrying YCp1-SUB2. All the above strains contain the telomere-silencing assay system adh4::URA3-TEL. The ratio of the number of cells grown on panel B to the number of cells grown on panel A is slightly decreased in lane 3 compared to lanes 1 and 2.

A spontaneous mutation of glutamic acid to lysine at position 267 in the Drosophila HEL protein slightly enhanced PEV (12). We introduced the same mutation into SUB2. This mutation, sub2-267, was maintained on a CEN-based plasmid, pSub2-267. The sub2-267 mutation did not disrupt the essential function of the Sub2 protein, as sub2Δ carrying pSub2-267 is viable (Fig. 3). However, the sub2-267 mutation had no effect on either transcriptional repression at telomeres or stability between directly repeated DNA sequences (Fan and Klein, unpublished).

We also examined transcriptional silencing at the silent mating locus hmr::TRP1 by quantifying the growth of wild-type strains and sub2Δ1 strains with YCp1-SUB2 on tryptophan dropout medium. However, no differences significant enough to suggest a role of Sub2p in regulating chromatin structure at the mating loci were detected between wild-type strains and sub2Δ1 strains with YCp1-SUB2 (Fan and Klein, unpublished).

Mutual suppression of HPR1 and SUB2

Overexpressing HPR1 did not rescue sub2Δ lethality, indicating that Hpr1p, even in excess, cannot substitute for the essential function of Sub2p (Fan and Klein, unpublished). We then evaluated the influence of overexpressing HPR1 on the increased recombination rate observed in a sub2Δ1 strain with YCp1-SUB2. Interestingly, YEp-HPR1 (pHF136-3/[HPR1, 2μM]) was able to suppress the hyperrecombination phenotype of the sub2Δ1 + YCp1-SUB2 (SUB2 CEN) strain (Table 4). Therefore, an increase in the amount of Hpr1p can compensate for the function of Sub2p in stabilizing directly repeated DNA sequences, although it cannot fulfill the essential function of Sub2p.

TABLE 4.

Effect of overexpressing HPR1 on sub2

Genotypea	Recombination rate (10−6)b	% Suppression
sub2Δ + YCp1-SUB2 + YEp-vector	490 ± 160	0
sub2Δ + YCp1-SUB2 + YEp-HPR1	47 ± 30	90

^aYCp1-SUB2 (pHF127-3) is SUB2 on YCplac111, and YEp-HPR1 (pHF136-3) is HPR1 on the 2μm-based plasmid YEp-vector (PHF135-12). 

^bRates were calculated as described in footnote c to Table 1. 

Lastly, we asked whether high-copy-number expression of MUD2 interfered with the SUB2-mediated suppression of hpr1Δ recombination by comparing recombination rates obtained with hpr1Δ + YEp-SUB2 + YEp-vector to rates obtained with hpr1Δ + YEp-SUB2 + YEp-MUD2 and found no difference (140 × 10⁻⁶ versus 160 × 10⁻⁶, respectively).

DISCUSSION

In this study we have shown that hyperrecombination phenotypes associated with mutations in the transcription components HPR1 and CDC73 can be suppressed by increased expression of a gene linked to pre-mRNA splicing. This surprising action of a putative RNA helicase factor in genome stability gains further support from our observation that conditional alleles and the null allele of SUB2 result in greatly increased repeat instability or hyperrecombination. The convergence of transcription components and the putative splicing factor to ensure genomic stability suggests that different protein machines may act together or redundantly to avoid damage to the DNA template and that the splicing machinery is closely associated with the RNA polymerase II transcription machinery.

The transcription machinery acts to repress genome instability.

Our results offer direct evidence that defects in the transcriptional machinery can be a source of genome instability. First, we have found that the cdc73Δ mutant, similar to the hpr1Δ mutant, displays a hyperrecombination phenotype. Second, we have observed a novel interaction between the hpr1Δ mutant and the transcription factor cdc73Δ mutant in that the hpr1Δ cdc73Δ double-mutant is a temperature-sensitive lethal. This genetic interaction is distinct from previous results in which transcription mutants were shown to suppress hpr1Δ-associated phenotypes (13, 33, 36, 44). The lethality indicates that these two proteins together carry out an essential biological process or that they are redundant with an essential function that fails to function at high temperature. Cdc73p and Hpr1p have been found together in a complex (5, 6), and Cdc73p has been demonstrated to interact with the C-terminal domain of the largest subunit of RNA polymerase II (39). We do not yet know if Sub2p is found associated with the Hpr1p/Cdc73p complex, particularly when SUB2 is overexpressed. Although absence of either Hpr1p or Cdc73p from a transcription-related complex leads to an increase in genome instability, when both proteins are missing this complex may be crippled and cell death is the result. Alternatively, the hpr1Δ and cdc73Δ mutations may act synergistically to increase genome instability to such an extent that cell growth cannot be supported. Third, we show that plasmid instability in an hpr1Δ strain is directly linked to transcription of a plasmid segment since inhibition of transcription with a transcription terminator restores plasmid stability.

Biological function of the Sub2 protein.

Sub2p is the closest yeast homologue to the human UAP56/BAT1 protein (11) and is suggested to function as an ATP-dependent RNA helicase in pre-mRNA splicing. In vivo and in vitro splicing are dependent on Sub2p (22, 24, 52). The amino acid sequence of Sub2p suggests that it is an ATP-dependent RNA helicase, and we have demonstrated the importance of the putative ATPase activity for its essential function by site-directed mutagenesis. The involvement of Sub2p in splicing is also supported by the interaction that we and others (22) have observed between Sub2p and Mud2p, a protein required for RNA splicing in yeast cells. Furthermore, overexpression of SUB2 has been found to suppress the brr1-1 mutant, which is defective in splicing (42).

The action of SUB2 in the instability of direct repeats is independent of MUD2. This suggests that Sub2p has functions in addition to a role in pre-mRNA splicing. Kistler and Guthrie (22) have proposed a Mud2p-independent role for Sub2p but have not linked Sub2p to repeat stability. hpr1Δ-mediated recombination is directly related to transcription elongation (34). However, unlike the hpr1Δ strains, neither the sub2Δ1 + YCp-SUB2 strains nor the sub2Δ1 + YEp-SUB2 strains showed any increase in sensitivity to 6-azauracil, an indicator of an elongation defect (R. Merker and H. L. Klein, unpublished data). Thus, it is not clear whether the Sub2p function in repeat stability is linked to transcription elongation, although the 6-azauracil sensitivity test does not rule this out. However, our preliminary results show that Sub2p can be coimmunoprecipitated with an RNA polymerase II subunit in hpr1Δ strains, suggesting that under certain conditions, Sub2p can have additional functions in vivo.

Our results show that in genome stability SUB2 has a role that is distinct from its essential function, which is pre-mRNA splicing. We base this conclusion on the following observations. First, a sub2Δ strain carrying the YCp1-SUB2 plasmid is viable, demonstrating that YCp1-SUB2 contains a DNA fragment, the entire coding sequence of SUB2, that complements the sub2Δ lethality. However, this strain displays a hyperrecombination phenotype, indicating that YCp1-SUB2 does not complement all of the sub2Δ deletion defects. Second, the capability of Sub2p to influence genome stability is further supported by its ability to suppress the induced recombination events occurring in the absence of Hpr1p or Cdc73p (Table 1). Third, conditional alleles of SUB2 have greatly increased rates of recombination, increased over the sub2Δ1 + p297-SUB2 strain (Table 3), showing that partial inactivation of Sub2p function increases direct repeat instability. Fourth, complete loss of Sub2p, viable as the sub2Δ1 mud2Δ genotype, results in even higher recombination rates.

Genetic interactions between SUB2, HPR1, and CDC73

The hyperrecombination phenotypes and the genetic interactions between hpr1Δ, cdc73Δ, and SUB2 show that the Hpr1p-Cdc73p complex and Sub2p have some degree of functional overlap, representing two activities that can influence genome stability. The recombination rates of the hpr1Δ sub2Δ + YCp-sub2 strains with the conditional sub2 mutant alleles are not additive or synergistic but rather show rates close to the sub2 mutant rate (except for p322 sub2-1) (Table 3). This suggests that both hpr1Δ and the sub2 mutants contribute to repeat instability in a single process. The finding that the SUB2 hyperrecombination can be suppressed by high-copy-number expression of HPR1, similar to SUB2 high-copy-number suppression of hpr1Δ hyperrecombination, strengthens this view. We have also shown that lack of Cdc73p leads to hyperrecombination, and this phenotype too can be suppressed by increasing the copy number of SUB2.

Mechanism of Hpr1p-Cdc73p and Sub2p function.

Absence of Hpr1p or Cdc73p or partial function (in the conditional alleles) or complete loss of function (the sub2Δ mud2Δ genotype) of Sub2p leads to hyperrecombination between direct repeats. Evidence accumulated thus far indicates that Hpr1p and Cdc73p function during transcription. The mutual suppression of hyperrecombination suggests that Hpr1p and Sub2p functions are intimately related in their ability to maintain genome stability. The hyperrecombination phenotype associated with hpr1Δ mutants has been suggested to originate in blocked transcription elongation (13, 34). Since the effect of SUB2 on repeat instability is MUD2 independent, we suggest that free Sub2p can have functions outside of pre-mRNA splicing. One of these functions may be to aid in transcription elongation as an RNA helicase. We suggest that during elongation Hpr1p aids in avoiding transcription pauses, possibly by removing accumulated secondary structures in the nascent RNA. In the absence of Hpr1p, available free Sub2p may be able to remove transcription blocks through the RNA helicase activity, removing the secondary structure. This could explain the mutual high-copy-number suppression by HPR1 and SUB2 and the absence of synergism or additivity in recombination rates. The transcription blocks, if not removed by Hpr1p or Sub2p, are eventually processed into recombination substrates. Indeed, it has been proposed that deletions between direct repeats may result from the convergence of a stalled transcription complex with a DNA replication complex (32, 45). We suggest that the action of Hpr1p is during elongation and not splicing, as no defect in splicing is observed in the hpr1Δ mutant (37).

Sub2p could be directly involved in genome stability. An intriguing possibility is that Sub2p does not function exclusively in splicing but may have a more general function in resolving RNA structures, a function that may be Mud2 independent. Evidence strongly suggests that transcription by RNA polymerase II and pre-mRNA processing are temporally coupled (26, 29). Furthermore, the largest subunit of RNA polymerase II has been found to physically associate with spliceosome components, including SR proteins and snRNPs (28, 46, 51). Therefore, it is reasonable to speculate that Sub2p may not be specifically targeted to intron-containing genes but is delivered to all genes transcribed by RNA polymerase II. A unifying hypothesis would be that Sub2p is tethered to the splicing apparatus through Mud2p and to the transcription apparatus through other proteins and functions to unwind a diversity of problematic RNA structures. The lethality observed in a sub2Δ strain may then be directly related to a loss of Sub2p function in splicing, while the hyperrecombination phenotype displayed in the sub2Δ1 + YCp-SUB2 strains may result from an effect on a more general aspect of RNA production. The amount of free Sub2 protein may be controlled by the amount of Mud2 protein in the pre-mRNA-U1 complex.