Role of Transcription in Plasmid Maintenance in the hpr1Δ Mutant of Saccharomyces cerevisiae

Abstract

The Saccharomyces cerevisiae hyperrecombination mutation hpr1Δ results in instability of sequences between direct repeats that is dependent on transcription of the repeat. Here it is shown that the HPR1 gene also functions in plasmid stability in the presence of destabilizing transcription elongation. In the hpr1Δ mutant, plasmid instability results from unchecked transcription elongation, which can be suppressed by a strong transcription terminator. The plasmid system has been used to examine in vivo aspects of transcription in the absence of Hpr1p. Nuclear run-on studies suggest that there is an increased RNA polymerase II density in the hpr1Δ mutant strain, but this is not accompanied by an increase in accumulation of cytoplasmic mRNA. Suppression of plasmid instability in hpr1Δ can also be achieved by high-copy expression of the RNA splicing factor SUB2, which has recently been proposed to function in mRNA export, in addition to its role in pre-mRNA splicing. High-copy-number SUB2 expression is accompanied by an increase in message accumulation from the plasmid, suggesting that the mechanism of suppression by Sub2p involves the formation of mature mRNA. Models for the role of Hpr1p in mature mRNA formation and the cause of plasmid instability in the absence of the Hpr1 protein are discussed.

The hpr1 mutant was originally identified in a genetic screen to identify mutants with increased intrachromosomal recombination events between direct repeat sequences in yeast (2). The majority of these events in the hpr1 mutant resulted in a deletion of the intervening sequence between the direct repeats, leaving only one copy of the original repeat (1). Genetic suppressors of this phenotype cluster in mutations in components of the RNA polymerase II transcriptional machinery. rpb2, sua7 (TFIIB), sin4 (13, 14), srb2 (29), gcr3 (45), and gal11 (31) all suppress hpr1-stimulated hyperrecombination to some degree. This finding has led to the idea that HPR1 has a functional role in RNA polymerase II transcription. Indeed, hpr1-stimulated hyperrecombination has been demonstrated to be dependent on transcription (13) and specifically dependent on transcriptional elongation (7, 32).

Additionally, Hpr1p has been demonstrated to associate with two different protein complexes, each further linking HPR1 to transcription. Hpr1p associates with Paf1p, Cdc73p, and Ccr4p as part of an RNAPII complex that is a distinct form of RNA polymerase II, characterized by the absence of Srb proteins (6). Both cdc73Δ and paf1Δ mutants display hyperrecombination phenotypes, although the phenotypes are not as strong as that of the hpr1Δ mutant. Consistent with the assembly of Hpr1p into this RNAPII complex, the hpr1Δ mutation is lethal when combined with either a paf1Δ or a ccr4Δ mutation, and the double-mutant hpr1Δ cdc73Δ strain displays temperature-sensitive growth at 30°C. Hpr1p has also been demonstrated to associate with a separate protein complex, termed the THO complex, that contains Tho2p, Mft1p, and Thp2p (8). The exact function of this complex is not clear; however, it appears to be involved in linking transcription to mRNA export and is associated with the Yra1p, Tex1p, and Sub2p factors (41), which have been linked to mRNA export (34, 40). Importantly, tho2Δ, mft1Δ, and thp2Δ mutants display the hyperrecombination phenotype observed in the hpr1Δ mutant and have impaired transcriptional elongation through the bacterial lacZ sequence (8, 30). It has been proposed that the THO protein complex acts as a functional unit connecting transcription elongation with mitotic recombination.

Other genetic data suggest that HPR1 in some way influences the structural integrity of DNA. First, a full complement of topoisomerase activities is necessary for viability in hpr1Δ strains (3). Second, the hpr1Δ mutation is lethal when combined with a deletion of histones H3 and H4 (hht1-hhf1Δ) (14). Consistent with these findings, overexpression of either HHT1-HHF1 or HTA-HTB (H2A-H2B) results in a severe growth defect when combined with an hpr1Δ mutation (51). Overexpression of all four histones has no effect on the growth of an hpr1Δ mutant strain.

Although there is evidence that Hpr1p promotes transcriptional elongation, the exact function it performs that allows it to promote elongation is not known and its recent association with mRNA export factors complicates the picture (40, 41). Previous experiments have demonstrated that Hpr1p-deficient cells have trouble transcribing through both the bacterial lacZ sequence and the bacterial ori-amp sequence of pBR322, resulting in the production of truncated transcripts (7, 32). Transcriptional run-on experiments have also provided some evidence that transcriptional blocks may be localized to a specific region of lacZ (7). Importantly, when these sequences are located between direct repeats, transcriptional elongation through them is linked to the recombination events observed between the repeats in hpr1Δ mutants. If elongation can be stopped prior to entering these sequences with the use of a strong transcription terminator, the recombination phenotype is not observed (7, 32). Therefore, there appears to be a direct link in hpr1Δ cells between the difficulty in transcribing these sequences and the hyperrecombination phenotypes observed. However, more recent results suggest that the defective transcription of lacZ segments in Hpr1p-deficient cells cannot be correlated with a particular DNA sequence or structure (9). Hpr1p seems to be required for transcription of particularly long DNA sequences or G+C-rich sequences, and the transcription impairment through such sequences is correlated with their ability to stimulate direct repeat recombination. This transcription defect can be suppressed by high-copy-number Sub2p (18). Transcription impairment is associated with a decrease in the steady-state accumulation of full-length mRNA. These observations have been extended to decreased transcription of endogenous DNA sequences that are long and G+C rich and direct repeat stimulation by these same sequences in an hpr1Δ mutant (9). However, exactly how Hpr1p functions in transcription elongation is not understood.

We have reported that hpr1Δ strains have a defect in nuclear mRNA export (34). We have also observed that high-copy expression of SUB2, a DEAD box family RNA helicase motif-containing gene involved in prespliceosome formation (19, 22, 24, 50), suppresses hpr1Δ-stimulated recombination and other hpr1Δ-associated phenotypes (15) and that sub2 mutants have a similar hyperrecombination phenotype that can be suppressed by high-copy-number HPR1 (15). Interestingly, SUB2 has recently been implicated in nuclear mRNA export of intron-containing and intronless transcripts.

Here we describe a situation in which transcriptional elongation through the bacterial ori-amp sequence from a yeast promoter present on a CEN-based plasmid contributes to its own instability in an hpr1Δ mutant strain. We have used the plasmid loss phenotype to gain further insight into the transcriptional defects present in hpr1Δ strains and have examined the effects of high-copy-number SUB2 on the stability and transcription of the unstable plasmid in hpr1Δ mutant strains to learn more about the mechanism of suppression by Sub2p and the role of Hpr1p in transcription and mRNA maturation.

MATERIALS AND METHODS

Strains and growth media.

The strains used in this study are listed in Table 1. All of the strains are isogenic to the W303 RAD5⁺ background. Nonselective synthetic complete medium (SC) and SC lacking the indicated amino acids were prepared by standard methods (37). All strains were grown at 30°C. For experiments requiring 6-azauracil (6-AU; Sigma), 100 μg of 6-AU per ml was added to liquid SC lacking tryptophan and uracil (SC-Trp-Ura) after autoclaving.

TABLE 1.

Yeast strains used in this study

Strain	Genotypea	Source
RMY057	MATα W303(pRS314)	This study
RMY063	MATα hpr1Δ::HIS3 W303(pRS314)	This study
RMY065	MATa W303(pRM101)	This study
RMY067	MATα hpr1Δ::HIS3 W303(pRM101)	This study
RMY064	MATa W303(pRM102)	This study
RMY066	MATα hpr1Δ::HIS3 W303(pRM102)	This study
RMY083	MATα W303(pRM103)	This study
RMY085	MATα hpr1Δ::HIS3 W303(pRM103)	This study
RMY079	MATa W303(pRM104)	This study
RMY081	MATα hpr1Δ::HIS3 W303(pRM104)	This study
RMY078	MATα hpr1Δ::HIS3 W303(pRM110)	This study
RMY076	MATα hpr1Δ::HIS3 W303(pRM111)	This study
RMY206	MATα W303(pRM115)(YEp-SUB2)	This study
RMY207	MATα hpr1Δ::HIS3 W303(pRM115) (YEp531-SUB2)	This study

^aAll strains are in the W303 genetic background (leu2-3,112 his3-11,15 trp1-1 ura3-1 ade2-1 can1-100 RAD5).

Plasmid constructions.

The plasmids used in this study are described in Table 2. pRM102 was constructed by amplifying yeast genomic DNA with primers 5′-ATGGATCCGCTGCACGGTCCTGTTCCCTA-3′ and 5′-TTGTCGACACCACCAGAACGGCCGTTAGA-3′ and inserting the amplified sequence into pRS314 (38) at the BamHI and SalI restriction sites. This sequence includes the entire HIS3 gene, as well as the first 236 bp of the PET56 gene and the first 259 bp of the DED1 gene. The HIS3 promoter in this amplified sequence was modified such that it contained both an EcoRI site and a ClaI site inserted between the TATA box and the translation start site. This modification had no detectable effect on HIS3 expression and is mentioned solely to indicate its presence. The pet56 and ded1 sequences are oriented such that both promoters directly flank and are directed away from the HIS3 gene. pRM101 was made by inserting the same amplified HIS3 sequence into the pGEM-11Zf (Promega) vector with the BamHI and SalI restriction sites. The resulting plasmid was then digested with ApaI and SacI. The resulting pet56-HIS3-ded1 fragment was purified from the digest and then reinserted into pRS314 at the ApaI and SacI restriction sites, resulting in an inversion of the inserted yeast sequence.

TABLE 2.

Plasmids used in this study

Name	Descriptiona	Source
pRS314	CEN6 ARSH TRP1 E. coli ori-amp	R. Sikorski
pRM101	1.7-kb ded1-HIS3-pet56 fragment in pRS314 with PET56 promoter oriented toward ori-amp	This study
pRM102	1.7-kb ded1-HIS3-pet56 fragment in pRS314 with DED1 promoter oriented toward ori1-amp	This study
pRM103	pRM102 with ded1 and its promoter removed	This study
pRM104	pRM102 with HIS3-pet56 removed	This study
pRM110	pRM102 with CYC1 transcription terminator inserted directly after ded1	This study
pRM111	pRM102 with CYC1 transcription terminator inserted directly in front of GEN6	This study
pRM112	pRM102 with CYC1 transcription terminator inserted between ori and amp	This study
pRM115	pRM102 with 630 bp of E. coli dnaB sequence, inserted 3′ to ded1	This study
YEp-SUB2	SUB2 cloned into YEp351	H. Fan

^aSee Materials and Methods for details of each plasmid.

pRM103 is identical to pRM102, except that the ded1 fragment, along with the promoter, has been excised from the plasmid with the BssHII and SalI restriction sites. It is important to note that this procedure does not remove the major HIS3 mRNA termination site (42). The restriction sites were processed with Klenow fragment to generate blunt ends that were subsequently ligated together. pRM104 is identical to the pRM102 plasmid, except that the HIS3-pet56 region has been removed from the plasmid with the BssHII and BamHI restriction sites. Again, the overhanging ends were processed with Klenow fragment and the plasmid was religated at these blunt ends. Therefore, pRM104 contains the ded1 fragment together with the DED1 promoter but lacks the HIS3 and pet56 sequences.

pRM110 is pRM102 with a CYC1 transcriptional terminator (26) inserted directly after the ded1 sequence at the SalI and ApaI restriction sites. The CYC1 terminator was amplified by PCR from the yeast genome with primers 5′-CGCGGGTCCTGTCGACCGATATCATGTAATTAGTTATGTCACGCTTAC-3′ and 5′-CGCAGGACCCGGGCCCGCGACGATGAGAGTGTAAACTGCGAAGCTTGC-3′, which contain SalI and ApaI restriction sites, respectively, at the 5′ end. The orientation of the terminator is identical to that of wild-type CYC1 relative to the direction of transcription. Both of the primers also contain a Psp511 restriction site, which was used to construct pRM111 with the PCR-amplified CYC1 transcriptional terminator. pRM111 contains the CYC1 terminator inserted into the Psp511 restriction site of pRM102, almost directly between the Escherichia coli amp and yeast CEN6 sequences. pRM112 has the CYC1 terminator inserted between the ori and amp sequences of pRM102 at the NcoI restriction site.

pRM115 is pRM102 with 630 bp of the E. coli dnaB sequence inserted into the SalI restriction site. The dnaB sequence was PCR amplified with the primers (each containing a SalI site at the 5′ end) 5′-CATGGTCGACCCCTTCAACAAACAGCAGGCTGAAC-3′ and 5′-CATGGTCGACGTTGAGATCGTCATAACCGGTGTTTAC-3′ and inserted into the SalI site of pRM115 such that the sequence was in the reverse orientation relative to the direction of transcription. This plasmid was constructed so that the dnaB sequence could be used to probe for plasmid DED1 promoter transcription in the presence of high-copy hpr1 suppressor SUB2.

Plasmid loss rates.

Plasmid loss rates of strains containing pRM102 (CEN TRP) and its derivatives were calculated as described previously in detail (10). Briefly, cells were grown in liquid SC lacking tryptophan (SC-Trp) to mid-log phase. Equal aliquots were taken and plated onto both SC and SC-Trp plates to determine the percentage of plasmid-containing cells. The culture was then diluted 1:1,000 into liquid SC to release the plasmid selection and allowed to grow to stationary phase. Again, equal aliquots were plated onto SC and SC-Trp to determine the percentage of cells that still contained the plasmid. Plasmid loss rates (m) were calculated with the formula m = 1 − e^ln(P2/P1)/g, where P₁ and P₂ are the percentages of plasmid carrying cells before and after the release of selection, respectively. g is the number of doublings during nonselective growth and is described as g = ln(N₂/N₁)/ln2. 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 cells growing on SC). The significance of differences between the plasmid loss rates was analyzed with the Student t test.

Transcriptional run-on analysis.

Transcriptional run-on analysis was performed as previously described (46), with some modifications (7). Cultures were grown under plasmid selection in liquid medium to mid-log phase and separated into two aliquots, one for DNA analysis and one for transcriptional run-on analysis. For the run-on analysis, 3 × 10⁷ cells were chilled by being poured over sterile ice and collected by centrifugation at 2,860 × g and 4°C. Cells were then washed in 5 ml of cold (0°C) TMN buffer (10 mM Tris [pH 7.4], 100 mM NaCl, 5 mM MgCl₂) and collected by centrifugation. The cells were resuspended in 0.95 ml of 0°C sterile H₂O and transferred to a microcentrifuge tube. A 0.05-ml volume of 10% N-lauroyl sarcosine was added, and the tube was left on ice for 15 min. The tube was gently centrifuged at 4°C (1,830 × g for 1 min), and the supernatant was removed. The cells were then resuspended in 0.15 ml of a reaction mixture containing 20 mM Tris (pH 7.7); 200 mM KCl; 32 mM MgCl₂; 2 mM dithiothreitol; 0.5 mM (each) ATP, CTP, and GTP; and 120 μCi of [³²P]UTP (>3,000 Ci/mmol) and incubated for 7.5 min at room temperature. The reaction was stopped by addition of 1 ml of cold (4°C) TMN buffer containing 1 mM unlabeled UTP. Cells were collected, and RNA was isolated with hot acidic phenol (20).

Probes (5 μg of each) for the run-on analysis were immobilized to positively charged Hybond-XL nylon membranes (Amersham Pharmacia Biotech) as previously described (11). Each probe was generated via PCR amplification and gel purified. The primers used to generate probes for pRM102 were 5′-CTGTCGTGCCAGCTGCATTAATG-3′ and 5′-CCGAAGGGAGAAAGGCGGACAGG-3′, 5′-AGCGTGGCGCTTTCTCAATGCTC-3′ and 5′-CTCATGACCAAAATCCCTTAACGTGAG-3′, 5′-TTATCAAAAAGGATCTTCACCTAG-3′ and 5′-ATACACTATTCTCAGAATGACTTGG-3′, 5′-GGCGACCGAGTTGCTCTTGC-3′ and 5′-GTAAAATCACAGGATTTTCGTG-3′, and 5′-TTTACTTATCGTTAATCGAATG-3′ and 5′-ATGGTGCACTCTCAGTACAATC-3′.

The rRNA gene control probe (RDN25) used in these experiments was also generated via PCR amplification with primers 5′-CAACCGGGATTGCCTTAGTAAC and TTACGCGTATGGGTTTTACACC-3′. The primers used to generate the probes for pRM119 were 5′-CAGCTTTTGTTCCCTTTAGTG-3′ and 5′-CCGAAGGGAGAAAGGCGGACAGG-3′, 5′-AGCGTGGCGCTTTCTCAATGCTC-3′ and 5′-ATACACTATTCTCAGAATGACTTGG-3′, and 5′-GGCGACCGAGTTGCTCTTGC-3′ and 5′-ATGGTGCACTCTCAGTACAATC-3′. The ACT1 control probe was generated with the primers 5′-ACTAACATCGATTGCTTCATTC-3′ and 5′-TTCTCAAAATGGCGTGAGGTAG-3′.

Isolated RNAs containing [³²P]UTP from the run-on experiments were hybridized to the filters with the immobilized probes in 5× SSC (0.75 M NaCl, 0.075 M Na citrate)-5× Denhardt's solution (0.1% Ficoll 400, 0.1% polyvinylpyrrolidone, 0.1% bovine serum albumin fraction V)-50% formamide-1% sodium dodecyl sulfate (SDS)-100 ng of single-stranded salmon sperm DNA at 42°C overnight.

Northern and Southern analyses.

RNA was isolated from exponential-phase cultures growing in the appropriate selective medium with hot acidic phenol (20). Northern analysis was performed with formaldehyde agarose gels by standard procedures (33). RNAs were transferred to and immobilized on Hybond-XL nylon membranes (Amersham Pharmacia Biotech). The appropriate probe was hybridized to the transferred RNAs in 5× SSC-5× Denhardt's solution-50% formamide-1% SDS-100 μg of single-stranded salmon sperm DNA per ml at 42°C overnight.

Genomic and plasmid DNAs were isolated from yeast in the logarithmic growth phase as previously described (37). Southern analysis was performed in accordance with standard procedures (33). Gels were transferred to and immobilized on Hybond-XL nylon membranes. The appropriate probes were hybridized to the transferred material in 5× SSC-5× Denhardt's solution-1% SDS-100 μg of single-stranded salmon sperm DNA per ml at 68°C overnight.

The DED1 promoter plasmid transcript was detected with a probe synthesized from a fragment (generated by a PCR with primers 5′-CTGTCGTGCCAGCTGCATTAATG-3′ and 5′-CCGAAGGGAGAAAGGCGGACAGG-3′) of the plasmid bacterial ori sequence. This same probe was also used to detect plasmid DNA by Southern analysis. To detect the DED1 promoter plasmid transcript with pRM115, a probe was synthesized from the SalI dnaB fragment of pRM115. Detection of actin mRNA on Northern blots, as well as the ACT1 DNA sequences on Southern blots, was done with the same ACT1 sequence as that used for the run-on analysis. All probes were generated by random priming and labeled with [α-³²P]dCTP. All blots, including those from the run-on analysis, were visualized with a Bio-Rad Molecular Imager FX and quantitated with NIH Image 1.61.

Quantitative RT-PCR methods.

Reverse transcription (RT)-PCRs were performed such that the transcript produced from the DED1 promoter present on pRM102 could be detected and quantitated from the five different RT primers shown in Fig. 5. As a control, each of the RT reaction mixtures contained the same amount of RNA isolated from E. coli overexpressing dnaB. This was done so that the variation between each RT and the subsequent PCR could be measured and proper quantitation of the DED1 promoter transcript from pRM102 could be achieved. Overexpression of dnaB in E. coli strain BL21(DE3) was achieved with expression plasmid pET3C containing a copy of dnaB (43).

FIG. 5.

Quantitative RT-PCR analysis of transcription from the DED1 promoter of pRM102. (A) Schematic of the locations of the five RT primers used to analyze the DED1 promoter transcript. (B). Southern analysis of the RT-PCR products of the five reactions from wild-type and hpr1Δ cells carrying pRM102. The position of the dnaB transcript used as a PCR efficiency control is indicated. Relative levels of plasmid and total mRNAs present in each reaction were determined by Southern and Northern analyses, respectively. (C). Quantitated RT-PCR levels of hpr1Δ(pRM102) cells relative to those of wild-type cells carrying pRM102 are shown. The results are from three independent experiments for each reaction. The error bars indicate standard deviations. Absolute RT-PCR levels were corrected for the amount of plasmid present in each strain and for the total amount of mRNA. chrom., chromosome.

All yeast cultures were grown to mid-log phase in liquid SC-Trp. Each culture was split in half so that RNA could be isolated from one half and DNA could be isolated from the other. RNA was isolated with hot acidic phenol (20). DNA was isolated by a standard protocol (37) and used to determine the relative amount of plasmid present in each strain. RT-PCR was performed with the Thermoscript RT-PCR System (Life Technologies). The RT reaction mixture contained 1 μg of total yeast RNA; 0.1 μg of total E. coli RNA; 0.2 μM 3′ primer; 1× Thermoscript cDNA synthesis buffer; 1× Thermoscript RNaseOUT; 250 μM each dATP, dCTP, dGTP, and dTTP; and 15 U of Thermoscript reverse transcriptase. The RNA and 3′ primer were first incubated at 65°C in 10 μl of diethyl pyrocarbonate-treated water for 5 min to denature the RNA and the primer. The rest of the components were added to create a 20-μl reaction mixture, which was then incubated at 50°C for 60 min. The reaction was terminated by incubation at 85°C for 5 min. Two units of RNase H was added, and the mixture was incubated at 37°C for 20 min. Two microliters of this RT reaction mixture was subsequently used for the PCR along with 0.2 μM each primer; 100 μM each dATP, dCTP, dGTP, and dTTP; 1× manufacturer's PCR buffer; and 2.5 U of Expand Long Template polymerase mixture (Boehringer Mannheim). Normal cycling times were 93°C for 30 s, 55°C for 30 s, and 72°C for 1 min 30 s. Most reactions were cycled 35 times. Five different 3′ RT primers were used and designed such that they would hybridize to the transcript produced from the DED1 promoter present on pRM102. These primers were as follows: RT primer 1, 5′-GTTTCCCGACTGGAAAGCGGGCAGTG-3′; primer 2, 5′-CTCATGACCAAAATCCCTTAACGTGAG-3′; primer 3, 5′-ATAGACTGGATGGAGGCGGATAAAG-3′; primer 4, 5′-ATACACTATTCTCAGAATGACTTGG-3′; primer 5, 5′-TGTATCCGCTCATGAGACAATAACCC-3′. The dnaB RT control primer, 5′-GCAATTTCCAGCGTACGGTTATCGG-3′, was also added to each tube.

The subsequent five PCRs were performed with one each of the above 3′ primers and a single 5′ primer (5′-CATATTATGGCTGAACTGAGCGAAC-3′) that hybridizes to the DED1 fragment present on pRM102, 110 bp downstream from the translation start site. In addition, to each of the five PCRs, specific control dnaB primers 5′-CCCTTCAACAAACAGCAGGCTGAAC-3′ and 5′-GCAATTTCCAGCGTACGGTTATCGG-3′ were also added.

Exactly the same amount of each PCR mixture was run on 1% agarose gels and subjected to Southern analysis. Each RT-PCR was visualized with a radiolabeled probe made from gel-purified PCR amplifications of the plasmid backbone sequence on pRM102 with primers 5′-GGCCCGGTACCCAGCTTTTG-3′ and 5′-GTTTCCCGACTGGAAAGCGGGCAGTG-3′ (Fig. 5). dnaB was visualized with a radiolabeled probe made from gel-purified PCR amplifications of dnaB with primers 5′-CCCTTCAACAAACAGCAGGCTGAAC-3′ and 5′-GCAATTTCCAGCGTACGGTTATCGG-3′.

Northern analysis probing for actin mRNA was performed on the material initially isolated to determine the relative levels of total RNA. Radiolabeled probe was synthesized from purified ACT1 PCR amplifications with primers 5′-ACTAACATCGATTGCTTCATTC-3′ and 5′-TTCTCAAAATGGCGTGAGGTAG-3′.

The relative amount of plasmid present in each strain was determined by Southern analysis. Approximately 1.0 μg of DNA from each isolation was digested with EcoRI for 4 h at 37°C and resolved on a 1% agarose gel. EcoRI cuts the plasmid once and cuts chromosomal DNA such that a 4.0-kb fragment containing ACT1 is produced. DNA blots were simultaneously hybridized with radiolabeled probes designed to recognize plasmid pRM102, as well as a single-copy chromosomal sequence, ACT1. pRM102 was visualized with the same probe used to detect the above-described RT-PCR products. ACT1 DNA was visualized with the same probe used as described above to detect actin mRNA.

Relative levels of the RT-PCR amplification products were quantitated and normalized against relative actin mRNA levels, the relative amount of plasmid present, and the relative efficiency of the dnaB RT-PCR. All hybridizations were visualized with a Bio-Rad Molecular Imager FX and quantitated with NIH Image 1.61 software.

DNA enzymatic manipulations.

All restriction digests were performed in accordance with the manufacturer's recommendations. All PCR amplifications were performed with the Expand Long Template PCR System and standard protocols (Roche Diagnostics). All PCR products were gel purified with 1% agarose gels and the QIAquick Gel Extraction Kit from Qiagen.

RESULTS

Transcription affects plasmid stability in hpr1.

Two plasmids were constructed by inserting a 1.7-kb yeast chromosomal sequence into the multiple cloning site of single-copy yeast plasmid pRS314. The two plasmids differ only in the orientation of the insert, which contains the entire HIS3 gene, along with approximately 250 bp of the 5′ ends of both DED1 and PET56. DED1 and PET56 directly flank HIS3 and are oriented such that both promoters are directed away from HIS3 (Fig. 1). These two plasmids differed markedly in stability when carried by an hpr1Δ strain, while both were stable in the wild-type strain (Fig. 1). To quantitate these observations, the rate of loss of each plasmid in each strain was determined (Table 3). We have previously reported plasmid loss rates for pRM102 in both the wild-type and hpr1Δ strains (15). We repeated these experiments and expanded upon these observations.

FIG. 1.

A yeast chromosomal insert confers differential plasmid stability in an hpr1Δ background, dependent on its orientation in the plasmid. pRM101 and pRM102 differ in the orientation of a 1.7-kb yeast chromosomal insert. The PET56 and DED1 genes are both oriented away from HIS3 and are designated pet56 and ded1 solely to indicate that they are truncations. The photograph above each plasmid diagram gives a qualitative representation of its respective stability in hpr1Δ cells. hpr1Δ cells containing the plasmid were taken from selective (SC-Trp) plates, streaked onto nonselective rich medium (YPD) plates, and then replica plated back onto selective (SC-Trp) plates. An example of a colony that has lost the plasmid is indicated by the white arrow.

TABLE 3.

Plasmid loss rates per cell per generationa

Genotype	Avg plasmid loss rate, 10−2 (SD)
HPR1(pRS314)	1.7 (1.4)
hpr1Δ(pRS314)	3.3 (0.5)
HPR1(pRM101)	2.3 (1.1)
hpr1Δ(pRM101)	2.2 (1.2)
HPR1(pRM102)	1.9 (0.4)
hpr1Δ(pRM102)	16.2 (0.8)
HPR1(pRM103) (pRM102 minus ded1)	1.6 (1.9)
hpr1Δ(pRM103) (pRM102 minus ded1)	2.7 (1.3)
HPR1(pRM104) (pRM102 minus HIS3-pet56)	3.8 (1.3)
hpr1Δ(pRM104) (pRM102 minus HIS3-pet56)	29.1 (2.1)
hpr1Δ(pRM110) (pRM102 with CYC1 terminator after ded1)	1.4 (0.8)
hpr1Δ(pRM111) (pRM102 with CYC1 terminator before CEN6)	8.1 (0.7)
hpr1Δ(pRM112) (pRM102 with CYC1 terminator between ori and amp)	1.8 (0.5)

^aPlasmid loss rates of strains tested in this study are shown. The values averages of three independent experiments., pRM102 is a pRS314 vector with a 1.7-kb chromosomal insert that flanks the yeast HIS3 gene. Plasmid pRM103 is missing the ded1 sequence and promoter. Plasmid pRM104 is missing the HIS3-pet56 sequence, leaving only the ded1 sequence and promoter. pRM110 terminates DED1 promoter transcription prior to entering ori-amp sequences. pRM111 terminates transcription after elongation through the ori-amp sequence. pRM112 terminates transcription after ded1 and before CEN6, between the ori and amp sequences.

pRM101 has the same stability when it is present in both wild-type and hpr1Δ cells, and the loss rate in each of these strains is not significantly different from that of the empty vector (pRS314) alone (Table 3). However, pRM102 is more than eight times decreased in stability when present in hpr1Δ cells than when present in wild-type cells (P < 0.001). To determine the cause of the instability of pRM102 in an hpr1Δ mutant strain, we deleted various regions of the 1.7-kb insert. Deletion of the PET56 fragment along with its promoter had no effect on the instability of pRM102 in an hpr1Δ strain (data not shown). Furthermore, removal of HIS3 with the PET56 fragment from pRM102 (to create pRM104) also did not rescue the instability of this plasmid in an hpr1Δ mutant strain. In fact, pRM104 has an average loss rate of 29.1 × 10⁻² and is much more unstable in hpr1Δ cells than is pRM102 (P < 0.001) (Table 3). Since pRM104 has only the DED1 fragment and its promoter inserted in the plasmid, this suggests that the main cause of the instability of pRM102 in hpr1Δ mutant cells is the DED1 promoter. To confirm this, we deleted the DED1 fragment, along with its promoter, from the pRM102 plasmid to create pRM103. This cured the instability of this plasmid in hpr1Δ cells (Table 3).

To determine why the DED1 promoter oriented toward the bacterial ori-amp sequence in pRM102 causes this plasmid to be unstable in hpr1Δ cells, we examined the transcripts from the DED1 promoter in pRM102 and transcripts from the PET56 promoter in pRM101 in both wild-type and hpr1Δ cells (Fig. 2). These Northern assays indicate that the DED1 promoter is capable of producing an approximately 2.5-kb transcript when it is oriented toward the ori-amp sequence (pRM102) in both wild-type and hpr1Δ cells. However, when the PET56 promoter is oriented toward the ori-amp sequence, this large transcript is not produced. A 2.5-kb transcript corresponds to a transcript starting from the DED1 promoter and terminating in front of the CEN6 ARSH4 region of the plasmid. This was confirmed by Northern analysis with probes flanking the CEN6 ARSH4 sequence. Only a probe in front of CEN6 would hybridize to the transcript (data not shown). Therefore, the DED1 promoter present on pRM102 is capable of producing a transcript that can progress at least through the bacterial ori-amp sequence on the plasmid. These experiments suggest that the DED1 promoter is relatively stronger than the PET56 promoter, consistent with previously published results (44).

FIG. 2.

Analysis of transcription through the bacterial ori-amp sequence on both pRM102 and pRM101. Northern analysis was performed to identify transcripts originating from the DED1 promoter (pRM102, lanes 1 through 6) and the PET56 promoter (pRM101, lanes 7 through 12) with a DNA probe from the bacterial ori sequence on the plasmid. RNA was isolated from three independent cultures grown to mid-log phase in selective (SC-Trp) medium for both wild-type (wt) and hpr1Δ strains each containing either pRM102 or pRM101. The arrow points to the major transcript (approximately 2.5 kb) that is generated from the DED1 promoter on pRM102 (lanes 1 to 6). This major transcript is not observed when the PET56 promoter is oriented toward the bacterial ori-amp sequence (pRM101, lanes 7 to 12). Relative actin mRNA levels from each culture are shown at the bottom

The data in Fig. 2 suggest that transcription elongation through the ori-amp sequence may be the cause of the plasmid instability in hpr1Δ cells. To test this hypothesis, we inserted a strong CYC1 transcriptional terminator at three different sites of pRM102. First, we inserted the terminator directly 3′ to the DED1 fragment on pRM102 to create pRM110 (Fig. 3). Qualitative observations indicated that this completely stabilized the plasmid in hpr1Δ cells. Quantitation of the rate of plasmid pRM110 loss confirmed this and showed that this transcriptional terminator completely suppresses the instability of pRM102 in hpr1Δ cells (Table 3). Northern analysis confirmed that the CYC1 terminator efficiently blocked all detectable transcription initiating from the DED1 promoter on pRM110 (data not shown). We next inserted the CYC1 transcriptional terminator directly in front of the CEN6 sequence on pRM102 to create pRM111 (Fig. 3). pRM111 remained relatively unstable in hpr1Δ cells compared to pRM110 (Table 3) (P < 0.001). Northern analysis of the transcript generated from pRM111 confirmed that transcription progressed through the ori-amp sequence and terminated at the CYC1 terminator sequence. Finally, the CYC1 transcription terminator was inserted between the ori and amp sequences to create pRM112 (Fig. 3). Plasmid instability was suppressed in hpr1Δ cells (Table 3) (P < 0.001). Therefore, termination of transcription from the DED1 promoter prior to progression through the bacterial ori-amp sequence on pRM102 suppresses the instability in an hpr1Δ strain. However, allowing transcription elongation to pass through the ori-amp sequence and terminate downstream maintains the instability of pRM102 in hpr1Δ cells. The finding that the pRM112 plasmid with a transcription terminator between the ori and amp sequences is stable in hpr1Δ cells rules out aberrant chromatin changes during the initiation of transcription as the cause of plasmid instability.

FIG. 3.

Early termination of transcription from the DED1 promoter suppresses plasmid instability in hpr1Δ cells. CYC1 transcription terminators were inserted either directly after the ded1 sequence on pRM102 (pRM110), directly in front of the CEN6 sequence (pRM111), or between the ori and amp sequences (pRM112) (see Materials and Methods). Given are qualitative assessments of the relative stability of each plasmid in an hpr1Δ background.

Plasmid transcript levels are not reduced in hpr1Δ cells.

Previous studies with plasmid-based systems have determined that Hpr1p-deficient cells have difficulty with transcription elongation through a variety of DNA sequences and often result in a transcription block that is correlated with genetic instability between direct repeats (7, 32). If transcription from the DED1 promoter on pRM102 in hpr1Δ cells were somehow blocked within the ori-amp sequence, then we would expect to see a reduction in steady-state levels of the full-length transcript. The Northern analysis in Fig. 2 suggests that although the 2.5-kb transcript can be detected in hpr1Δ cells carrying pRM102, the levels are reduced compared to those in the wild type. However, proper quantitation of this transcript by correction against the amount of plasmid present, as well as the level of actin mRNA and actin DNA, reveals that the levels of the DED1 promoter transcript are not reduced in hpr1Δ cells compared to those in wild-type cells (Fig. 4). Figure 4 shows the results of three independent experiments. The absolute levels of the DED1 promoter transcript are not reduced in hpr1Δ cells relative to those in wild-type cells because the actual levels of pRM102 are very low in hpr1Δ cells, even under selection, because of its inherent instability. On the contrary, these results demonstrate that there might be a slight increase in the absolute levels of this transcript in hpr1Δ cells relative to those in wild-type cells. Although these increases are small, we suspect that they are not trivial. We have observed this slight increase in numerous other Northern experiments, as well as in quantitative RT-PCR experiments designed to detect levels of the DED1 promoter transcript from pRM102.

FIG. 4.

Quantitation of the major transcript produced from the DED1 promoter on pRM102 in both wild-type (wt) and hpr1Δ strains. (A) Northern analysis of transcription from the DED1 promoter present on pRM102 in both wild-type and hpr1Δ stains. RNA was isolated from three independent cultures of both wild-type and hpr1Δ strains, each containing pRM102, grown to mid-log phase in liquid SC-Trp. The transfers were probed with the same bacterial ori probe as in Fig. 2. The arrow points to the major transcript (approximately 2.5 kb) that is generated from the DED1 promoter on pRM102. (B) Southern analysis to determine the relative amount of plasmid present in each strain. DNA was isolated from each of the same cultures used for the above Northern assay. The DNAs were digested with BamHI, and equal amounts were loaded onto a 0.9% agarose gel. The transferred materials were simultaneously probed for plasmid (with the bacterial ori probe), as well as a single-copy yeast gene (ACT1). (C) Quantitation of the DED1 promoter transcript levels present in hpr1Δ cells relative to those in the wild type. Relative levels of transcript produced from the DED1 promoter were normalized against relative actin mRNA levels, as well as the relative amount of plasmid present in each strain. The relative amount of plasmid was determined by comparing the plasmid levels against that of a single-copy chromosomal sequence (ACT1).

Quantitative RT-PCR was performed to show that the transcripts observed in Fig. 4 were derived from the DED1 promoter on pRM102. These experiments also confirmed the quantitation obtained by Northern analysis. The locations of five RT primers used in this experiment are shown in Fig. 5. Each primer was used with the 5′ PCR primer to generate five different RT-PCR products, which are shown in Fig. 5B. Quantitation of the results of three separate experiments showed that the steady-state level of the pRM102 DED1 promoter transcript in the hpr1Δ strain does not differ from that observed in the wild-type strain (Fig. 5C).

Therefore, although we determined that transcription elongation from the DED1 promoter on pRM102 is required to observe the plasmid instability in an hpr1Δ strain, we could not detect any evidence of a transcriptional defect by Northern analysis or RT-PCR.

Plasmid instability in hpr1Δ cells is increased in the presence of 6-AU.

To further demonstrate that the instability of pRM102 in hpr1Δ cells is directly related to transcription, we tested the stability of the plasmid in the presence of 6-AU. 6-AU decreases intracellular pools of GTP and UTP by inhibiting components of their biosynthetic pathways (12, 17). Consequently, 6-AU sensitivity is often correlated with mutations in components of the RNA polymerase II transcriptional machinery involved in elongation (17). In the presence of 6-AU, pRM102 was extremely unstable in hpr1Δ cells. The slow growth of this strain made the calculation of a plasmid loss rate impractical. To demonstrate instability, we calculated the fraction of cells possessing plasmid both in the presence and in the absence of 6-AU while the cells were grown in liquid medium selecting for the plasmid. pRM101 is relatively stable in hpr1Δ cells and is not affected by the addition of 6-AU (Table 4). Consequently, pRM102 is much more unstable in hpr1Δ cells when 6-AU is present in the growth medium.

TABLE 4.

Plasmid stability in cultures containing 6AU

Genotype	Avg proportion of cells with plasmid (SD)a
	With 6-AU	Without 6-AU
HPR1(pRM101)	0.91 (0.07)	0.91 (0.03)
hpr1Δ(pRM101)	0.78 (0.02)	0.88 (0.05)
HPR1(pRM102)	0.81 (0.08)	0.62 (0.06)
hpr1Δ(pRM102)	0.05 (0.01)	0.38 (0.02)

^aProportions of wild-type and hpr1Δ strains carrying plasmids pRM101 and pRM102 in the presence and absence of 6-AU are shown. All strains are URA3⁺ and were grown in liquid SC-Trp-Ura with or without 100 μg of 6-AU per ml to mid-log phase. Cells were then plated on SC plates and allowed to grow for 3 days at 30°C. The plates were then replica plated to SC-Trp plates to determine the proportion of colonies still containing pRM102. The values shown are averages of at least three independent experiments.

High-copy-number SUB2 suppresses plasmid instability in hpr1Δ cells and increases steady-state plasmid DED1 promoter transcript levels relative to those in the wild type.

DEAD box-containing putative RNA helicase-encoding gene SUB2, when present on a high-copy-number plasmid, suppresses hpr1Δ-stimulated recombination events between direct repeats and the increased plasmid loss rate of pRM102 (15). This suppression is extremely strong and results in plasmid loss rates that are close to the wild-type levels. SUB2 suppression is easily visualized with Southern analysis by comparing the relative plasmid levels in hpr1Δ mutant and wild-type cells (Fig. 6). Since high-copy-number SUB2 has such a strong stabilizing effect on pRM102 in hpr1Δ cells, we wanted to determine its effect on the steady-state levels of the plasmid DED1 promoter transcript. For this, we inserted a short unique DNA sequence directly 3′ to the ded1 sequence on pRM102 (to create pRM115, which is identical to pRM102 with regard to plasmid instability in hpr1Δ cells) to use as a probe for the DED1 promoter transcript. We used the unique sequence as a probe since the plasmid containing SUB2 (YEp-SUB2) contains the same bacterial ori sequence used to probe for the pRM102 DED1 promoter transcript. Surprisingly, in the presence of high-copy-number SUB2, hpr1Δ cells displayed significantly increased steady-state levels of the DED1 promoter transcript from pRM115 compared to those in wild-type cells (P < 0.01) (Fig. 6). This suggests that the DED1 promoter on pRM115 may actually be more accessible to the RNA polymerase II transcriptional machinery in hpr1Δ cells. We do not know how high-copy-number SUB2 stimulates transcription in an hpr1Δ strain. Presumably, the putative RNA helicase activity is involved in this process. We have previously shown that Sub2p associates with RNA polymerase II when it is overexpressed in hpr1Δ cells (15).

FIG. 6.

High-copy-number SUB2 suppresses plasmid instability and increases the relative amount of the major transcript produced from the DED1 promoter on pRM1115 in hpr1Δ cells. pRM115 is identical to pRM102 except that 630 bp of the E. coli dnaB sequence is inserted directly 3′ to the ded1 sequence so that the plasmid DED1 promoter transcript could be uniquely probed in the presence of the high-copy-number YEp-SUB2 plasmid (YEp351 backbone). (A) Northern analysis of transcription from the DED1 promoter present on pRM115 in both wild-type (wt) and hpr1Δ stains in the presence of high-copy-number YEp-SUB2. RNA was isolated from three independent cultures grown to mid-log phase in liquid SC-Trp-Leu. The same amount of RNA was loaded onto each lane, and transfers were probed with the bacterial dnaB probe. The blots were subsequently stripped and reprobed for actin mRNA. (B) Southern analysis to determine the relative amount of plasmid pRM115 present in each strain. DNA was isolated from each of the same cultures used for the above Northern assay. DNAs were digested with EcoRI. The transfer was simultaneously probed for pRM115 (dnaB probe) and a single-copy yeast gene (ACT1). (C) Quantitation of the DED1 promoter transcript levels present in hpr1Δ cells relative to those in wild-type cells. Relative levels of transcript produced from the DED1 promoter were normalized against relative actin mRNA levels, as well as the relative amount of plasmid present in each strain.

Transcriptional run-on analysis of plasmid transcripts in hpr1Δ cells.

To further elucidate the relationship between transcriptional elongation through the ori-amp sequence in pRM102 and its instability in hpr1Δ cells, we performed a transcriptional run-on analysis with probes covering the entire ori-amp sequence, as well as the downstream CEN6-ARSH4 sequence (Fig. 7C). The results of these experiments are shown in Fig. 7A. A representative blot is shown on the left, while the averages and standard deviations of four independent experiments are shown on the right. These results suggest that an increased RNA polymerase II density is present throughout the length of these sequences in hpr1Δ cells compared to that in wild-type cells. The relative levels on each blot were corrected for both the amount of plasmid present in the cells and the total amount of RNA (25S rRNA). Transcription across ACT1 was also used to normalize for the total amount of RNA in other run-on experiments with similar results (data not shown).

FIG. 7.

Transcriptional run-on analysis of the plasmid DED1 promoter transcript in hpr1Δ and wild-type cells. Cells were grown under selection in liquid medium (SC-Trp) to mid-log phase prior to run-on analysis. (A) Radiolabeled RNA from both wild-type (wt) and hpr1Δ cells, each containing pRM102, bound to each probe. Transcription of rRNA was used to normalize transcription levels. The arrow indicates the direction of transcription. The location of each probe from the plasmid is shown in panel C. The bar graph shows the average level of transcription across each probe from four independent experiments for hpr1Δ cells relative to wild-type cells, normalized against both the total amount of RNA (25S rRNA) and the relative amount of plasmid present in each strain (B). Error bars represent the standard deviations. (B) Southern analysis to determine the relative amount of plasmid pRM102 present in each strain. DNA was isolated from each of the same cultures used for the above run-on analysis. DNAs were digested with EcoRI. The transfer was simultaneously probed for pRM102 (bacterial ori probe), as well as a single-copy yeast gene (ACT1). chrom., chromosome.

These results are surprising because previous transcriptional run-on experiments have shown a loss of RNA polymerase II density of transcription of the bacterial lacZ sequences in Hpr1p-deficient cells (7). These same data also provided evidence of a nonspecific transcription block within this sequence, suggesting that RNA polymerase II transcription may be more susceptible to transient pauses in hpr1Δ. In our experiments, however, we saw an increase in RNA polymerase II density, as well as no evidence of a transcription block when a DED1 promoter was used to drive transcription across the ori-amp sequence.

Plasmid copy number in hpr1Δ cells.

One possible explanation for the plasmid instability in hpr1Δ cells is that unchecked transcription might continue through the CEN6 region and disrupt its centromere activity. Our Northern analyses did not show evidence of transcription in this region. However, disruption of the centromere function on pRM102 should make it segregate as a yeast replicating plasmid without a centromere, a YRp plasmid, which is known to be unstable because of missegregation (28, 39). In this model, pRM102 plasmids would accumulate in mother cells, thereby increasing the plasmid copy number in cells containing plasmid pRM102. On the basis of our Southern analyses, we calculated that, under selective conditions, the copy number of plasmid pRM102 in hpr1Δ cultures is 0.26 ± 0.04 of the copy number found in wild-type cultures. From the plasmid loss rate determinations, we found that the proportion of cells in an hpr1Δ culture that carry plasmid pRM102 is 0.28 ± 0.08. This indicates that there is approximately one copy of plasmid pRM102 in hpr1Δ cells that have plasmid DNA. This is inconsistent with a missegregation model in which cells that contain the plasmid have a greatly increased copy number, up to 50 plasmids per cell.

We also tested a sequestration model of hpr1Δ-mediated plasmid loss. Under this model, a frozen or blocked RNA polymerase II-plasmid DNA-nascent mRNA ternary complex might be complexed with the nuclear membrane such that the sequestered plasmid would not segregate correctly to the daughter cell. The trapped complex might make it difficult to recover plasmid DNA by standard isolation methods, such that the results of Southern analyses done to estimate plasmid copy number are underestimates. To test this, we isolated and quantitated the plasmid copy number by a method that can extract DNA from membrane fractions (16). This method did not increase the pRM102 plasmid yield in hpr1Δ cells.

Role of homologous recombination in plasmid loss.

Another explanation for the plasmid loss in hpr1Δ cells is that aberrant transcription generates a damaged DNA substrate, in this case, plasmid pRM102, which is a target for homologous recombination. Recombination between sister chromatids of the plasmid would form a dicentric plasmid, which is known to be unstable because of mitotic nondisjunction events (21, 25). If this were the case, then elimination of recombination through a rad52Δ mutation should increase plasmid stability. We tested plasmid stability in hpr1Δ rad52Δ and hpr1Δ rad50Δ strains and found that plasmid stability was unaltered from that observed in an hpr1Δ strain, indicating that increased sister chromatid recombination is not the explanation for hpr1Δ-mediated plasmid instability.

DISCUSSION

These studies have shown that hpr1Δ cells have defects in the transcription of certain DNA sequences that can be related to problems in transcription elongation. These conclusions are similar to those of previous studies that found that hpr1Δ cells have difficulties in transcribing through portions of the bacterial lacZ sequence and parts of the plasmid ori-amp sequence (7, 32). The authors of those studies found evidence of transcriptional stalling, particularly when transcription was driven from the strong GAL1 promoter under inducing conditions (7). We have observed that transcription through the plasmid ori-amp sequence results in instability of the plasmid when transcription is driven by the relatively strong constitutive DED1 promoter. Deletion of the DED1 promoter or insertion of a transcription terminator immediately 3′ to the promoter restored plasmid stability.

However, in spite of the plasmid instability resulting from transcription elongation problems, we did not observe a decrease in message accumulation in hpr1Δ cells by Northern analysis. On the contrary, we observe equal, if not slightly higher, levels of the destabilizing transcript from the pRM102 DED1 promoter in hpr1Δ cells, while the nuclear run-on experiment with the DED1 promoter showed a relative increase in RNA polymerase II density in hpr1Δ cells. A possible explanation for this observation may involve our previous finding that polyadenylated RNA export from the nucleus to the cytoplasm is defective in hpr1Δ cells (34, 41). Northern analysis that measures mRNA accumulation would not reflect the RNA polymerase II transcripts that remain trapped in the nucleus and are a target for degradation (5, 48, 49). The increase in RNA polymerase II density in the hpr1Δ mutant could arise from increased accessibility of the DED1 promoter to RNA polymerase II assembly in that promoter region. Increased accessibility in the hpr1Δ mutant could result from negative supercoiling at the promoter region behind the multiple RNA polymerase II complexes assembled on a circular DNA template at a strong promoter. We propose that some of the nascent mRNA transcripts in the hpr1Δ mutant are incomplete, being either aborted or degraded because of aberrant structures. These messages would not be exported to the cytoplasm. Thus, the increased transcription relative to that in the wild type does not result in increased message accumulation.

We believe that this structural explanation is plausible, given the synthetic lethality hpr1Δ displays when combined with either a top1Δ, a top2(Ts), or a top3Δ mutation (3) or with deletions in histones H3 and H4 (hht1-hhf1Δ) (14). Plasmids in yeast can become negatively supercoiled in a transcription-dependent manner when they are present in both top1Δ and top1Δ top2(Ts) mutant strains (4, 35). Indeed, plasmids can become negatively supercoiled in wild-type strains in a transcription-dependent manner given the proper conditions of a strong promoter on the plasmid producing a relatively long (7-kb) transcript (27). An elongating RNA polymerase complex generates an accumulation of positive supercoils in front of the complex and an accumulation of negative supercoils behind it (4, 23, 47). It is possible that in an hpr1Δ mutant, localized negative supercoiling may accumulate behind an elongating RNA polymerase II complex, which in turn would make the corresponding promoter more accessible.

We have previously reported that high-copy-number expression of the pre-mRNA splicing component SUB2 suppresses hpr1Δ defects, including instability of plasmid pRM102 (15), and this has been confirmed by more recent studies (18). Our studies may give some insight into the mechanism of suppression. By Northern analysis, there is a three- to fourfold increase in the pRM102 DED1 promoter transcript level in the hpr1Δ mutant carrying a high-copy-number SUB2 plasmid, relative to that in the wild type. This suggests that high-copy-number SUB2 does not suppress by returning transcription to a wild-type level. Rather, high-copy-number SUB2 appears to make the defective transcripts into productive transcripts that can be exported to the cytoplasm. The increase in cytoplasmic message accumulation most likely reflects the increased transcription that occurs in hpr1Δ cells. In a mechanism that we do not understand, Sub2p overcomes the transcription defect of hpr1Δ cells, allowing production of full-length messages that can be processed and transported to the cytoplasm. We would expect nuclear run-on analysis of hpr1Δ plus high-copy-number SUB2 cells to show the same increase in RNA polymerase II density as hpr1Δ cells lacking the high-copy-number SUB2 plasmid. However, we have not been able to do this experiment, as the run-on analysis requires hybridization to the ori-amp sequence and adjacent regions of the plasmid backbone that are also present on the SUB2 plasmid. Nonetheless, we surmise that suppression occurs by a resolution of the mRNA processing-export defect of hpr1Δ transcription and not through a bypass of the transcription process. High-copy-number Sub2p could act as an RNA helicase to unwind inhibitory secondary structure that may be present in the nascent RNA molecule. Alternatively, Sub2p could act as an RNA-DNA helicase to unwind the RNA-DNA hybrid just behind RNA polymerase II, thereby allowing the RNA polymerase complex to regress and any blocking structure to be removed. However, in pre-mRNA splicing, Sub2p is not hypothesized to directly contact the RNA molecule. Rather, Sub2p contacts proteins that are bound to the 3′ splice site region. Sub2p has been proposed to act as an RNPase to aid in the removal of these proteins from the 3′ splice site in a protein complex swap (19, 22, 36, 50). Sub2p may have a similar role in hpr1Δ suppression by aiding in the removal of a complex from the nascent RNA or the DNA template or the RNA-DNA hybrid region. It will be of interest to see if other hpr1Δ suppressors, in particular the recessive suppressors that are mutations in the basic transcription apparatus, act in a fashion similar to that of high-copy-number Sub2p by nuclear run-on studies.

We have proposed and tested several models of hpr1Δ-mediated plasmid instability. The aberrant transcription in an hpr1Δ strain does not pass through the plasmid centromere region and cause instability though missegregation. Sister chromatid recombination to form a dicentric plasmid also seems to be ruled out by our studies with recombination-deficient strains. We also found no evidence of plasmid sequestration in a membrane fraction. The most likely explanation is that the aberrant transcription and possible transiently altered plasmid supercoiling in some manner affects efficient replication at the yeast DNA replication ARS sequence on the plasmid. Once the transcription problem can be overcome, for example, by high-copy-number SUB2, even though the level of transcription remains high, plasmid stability is restored. This suggests that it is not the high transcription rate per se that affects origin firing but the defective transcription. Perhaps the presence of a trapped mRNA processing-export complex, for example, the TREX complex (41), or trapping of the plasmid in a mRNA export nuclear compartment occludes or prevents the DNA replication complex from assembling on the plasmid. In this way, a failure of DNA replication would account for the reduced plasmid stability and the low plasmid copy number in cells that have retained the plasmid.