RNA polymerase II (RNAPII) in eukaryotic cells drives transcription of most messenger RNAs. RNAPII core enzyme is composed of 12 polypeptides where Rpb1 is the largest subunit. To further understand the mechanisms of RNAPII transcription, we isolated and characterized novel point mutants of RPB1 that are sensitive to the nucleotide-depleting drug 6-azauracil (6AU). In this work we reisolated the rpo21-24/rpb1-E1230K allele, which reduces the interaction of RNAPII–TFIIS, and identified five new point mutations in RPB1 that cause hypersensitivity to 6AU. The novel mutants affect highly conserved residues of Rpb1 and have differential genetic and biochemical effects. Three of the mutations affect the “lid” and “rudder,” two small loops suggested by structural studies to play a central role in the separation of the RNA–DNA hybrids. Most interestingly, two mutations affecting the catalytic center (rpb1-N488D) and the homology box G (rpb1-E1103G) have strong opposite effects on the intrinsic in vitro polymerization rate of RNAPII. Moreover, the synthetic interactions of these mutants with soh1, spt4, and dst1 suggest differential in vivo effects.

RNA polymerase II (RNAPII) is composed of 12 polypeptides with a high degree of structural conservation from yeast to humans (Hahn 2004). In Saccharomyces cerevisiae, Rpb1 and Rpb2 form the central part of RNAPII and share an ample contact surface. The interaction between both subunits shapes several domains of the enzyme, such as the active center, which is formed by the “active site” and the “hybrid-binding” regions of Rpb1 and Rpb2, respectively (Cramer et al. 2001). The functional complexity of RNAPII is reflected in its structure, revealing the existence of a large number of regions with very specific functions. The core polymerase requires the association of a number of initiation factors for promoter recognition and initiation of RNA synthesis. Once in elongation mode, the crosstalk between RNAPII and a number of associated factors assures a proper mRNA synthesis (Sims et al. 2004). To gain further insight into the process of transcription, we have isolated mutants of RPO21/RPB1, encoding the largest subunit of the RNAPII in yeast, obtained by a random mutagenesis. The mutants were selected by their sensitivity to 6-azauracil (6AU), a drug that decreases GTP and UTP pools (Exinger and Lacroute 1992). We reasoned that leaky mutants of RNAPII might be susceptible to imbalances in the intracellular nucleotide pools at steps in initiation, elongation, or termination.

Sensitivity to 6AU is a well-documented phenotype associated with transcription-elongation mutants. For example, the 6AUS mutant rpb2-10 (P1018S) is intrinsically arrest prone and has a slower polymerization rate (Powell and Reines 1996; Mason and Struhl 2005). Further, yeast knockouts of the genes DST1/PPR2 and SPT4, encoding the transcription elongation factors TFIIS (Fish and Kane 2002) and Spt4 (Hartzog et al. 1998), are highly sensitive to 6AU. As expected, the rpb1-E1230K (rpo21-24) point mutant, which decreases the binding of TFIIS to RNAPII (Archambault et al. 1992; Wu et al. 1996), is also 6AUS. Biochemical analyses have shown that TFIIS stimulates transcription elongation, increases the fidelity of incorporation of ribonucleotides, and is essential for the reactivation of arrested RNAPII in vitro (Fish and Kane 2002). Apart from the well-characterized role of TFIIS in elongation in vitro, reports from different laboratories show that TFIIS is also involved in transcription initiation (Davie and Kane 2000; Malagon et al. 2004; Adelman et al. 2005; Prather et al. 2005). Thus, it is possible that some of the sensitivity to 6AU in a dst1 knockout is due to defects in initiation caused by lack of TFIIS, and it might be argued that some mutants in RNAPII that are sensitive to 6AU would be arrest prone or compromised in initiation of transcription. Although no single rpb1 or rpb2 6AUS mutant is affected in initiation, the rpb2-101 (G369S) mutation is 6AUS and has an altered transcription initiation in the presence of a mutant TFIIB initiation factor (Chen and Hampsey 2004).

We report here the isolation and characterization of several novel 6AUS alleles of RPB1. These include alterations of conserved domains of RNAPII near the active site, the point where the DNA–RNA hybrid separates (the lid and rudder domains), and the region where the template and nontemplate strands of the DNA downstream of the active site separate. In addition, we reisolated the rpb1-E1230K allele that blocks TFIIS binding (Wu et al. 1996). Biochemical characterization of these 6AUS alleles demonstrates that they have different consequences on elongation. Similarly, genetic characterization shows that they have different dependency on other transcription factors. Finally, these mutations have different consequences for the expression of some genes normally involved in response to 6AU. Combined, these results suggest that 6AU sensitivity can be caused by defects in several different aspects of transcription and that 6AUS rpb1 mutants can be obtained that reveal these different functions.


Media, yeast manipulations, strains, plasmids, and oligonucleotides:

Media, growth conditions, and yeast manipulations were as previously described (Malagon et al. 2004). Sensitivity to 6AU was scored on AA-Ura + 6AU plates by replica plating (100 μg/ml 6AU) or serial dilutions (10 μg/ml 6AU). All strains used are direct derivatives or closely related to the BY series of the yeast knockout collection into which we introduced different alleles of RPB1 (see Table 1). Strains GRY3019–GRY3029 are all his3Δ leu2Δ lys2Δ met15Δ trp1Δ∷hisG URA∷CMV-tTA. GRY3100 and GRY3101 are his3Δ leu2Δ met15Δ trp1Δ∷hisG URA∷CMV-tTA dst1Δ∷natMX4. Strains GRY3030–GRY3040 are his3 leu2 lys2Δ met15Δ trp1 can1 pep4∷HIS3 prb1Δ1.6R RPB3∷6xHis URA∷CMV-tTA and are related to the BY yeast knockout collection and to BJ5464 MATα can1 his3Δ200 leu2Δ1 trp1 ura3-52 pep4∷HIS3 prb1Δ1.6R GAL+ (American Type Culture Collection Yeast Genetic Stock Center). The PtetRPB1 allele and the tTA transactivator were introduced by crosses with the strain YSC1180-7428981 (OpenBiosystems; Mnaimneh et al. 2004). All yeast strain relevant genotypes are described in Table 1. All oligonucleotides used are shown in Table 2. Plasmid pL-RPB1 is a LEU2-based centromeric plasmid containing the RPB1 gene from position −595 to +5754 relative to the start of the open reading frame. The plasmids containing the rpb1 mutations were named pL-rpb1-x (x represents the specific allele).

View this table:

Yeast strains

View this table:


RPB1 mutagenesis and sequencing of the mutations:

Plasmid pL-RPB1 was mutagenized using the mutator XL1-Red competent cells kit, following the supplier's recommendations (Stratagene, La Jolla, CA). The genotype of the Escherichia coli strain XL1-red is endA1 gyrA96 thi-1 hsdR17 supE44 relA1 lac mutD5 mutS mutT Tn10 (Tetr). Sequencing a region covering the RPB1 gene from positions −509 to +5641 identified the mutations.

Integration of RPB1 mutants:

The wild-type and mutant alleles of RPB1 were introduced by homologous recombination. For that purpose, strains GRY3019 and GRY3030 were transformed with linear fragments from the pL-RPB1 series obtained by digestions with AvaI (located in the RPB1 promoter proximal polylinker) and EagI, XbaI, KpnI, or HindIII, depending on the location of the mutation of interest. Recombinants replace the G418rPtet with the natural RPB1 promoter and were therefore selected by growth in doxycycline-containing media and subsequently screened for sensitivity to G418. The proper integrations were then confirmed by sequencing.

Northern analysis:

Northern analysis was done basically as described (Shaw and Reines 2000). Briefly, total RNA was extracted with acid phenol from cells growing exponentially in AA-Ura media. The final concentration of 6AU was 75 μg/ml, and samples were taken at times 0, 30, and 120 min. RNA transfer and hybridization were done using the NorthernMaxR kit (Ambion, Austin, TX). Radioactive labeled probes for IMD2, SED1, ACT1, and RDN25-1 were obtained from PCR fragments, using the corresponding primers (Table 2). The SSM1 probe was obtained from a 0.5-kb BsmI–PvuI internal fragment of the gene from plasmid YEplac181-SSM1 (F. Malagon, unpublished data).

Transcription in vitro:

RNAPII purification and transcription complex reconstitution were done as previously described (Kireeva et al. 2003). Briefly, RNAPII was purified from yeast cell extracts by attachment of hexahistidine-tagged Rpb3 to Ni2+-NTA agarose beads. The 5′ radioactively labeled RNA (rna9) and the template DNA strand (TDS45G) oligonucleotides were incubated with the immobilized RNAPII. The nontemplate DNA strand (NDS45G) was subsequently added. Elongation of the RNA was allowed to proceed by adding NTPs at a final concentration of 10 μm. The products were resolved in 20% denaturing polyacrylamide gels.

Sequence alignment analysis:

Sequence alignment was done using the Clustal W program (Thompson et al. 1994).

Crystal structure visualization:

RNAPII transcription complex PDB:1SFO (Westover et al. 2004) visualization and localization of specific residues were done using Protein Explorer (www.proteinexplorer.org/) (Martz 2002).


Because the largest subunit of RNAPII (encoded by RPB1) is essential and we wanted to be able to identify recessive mutations, we developed a screen for rpb1 mutants on the basis of a strain (GRY3019) that contains an RPB1 allele (PtetRPB1) that can be turned off by addition of doxycycline to the media (Mnaimneh et al. 2004). GRY3019 does not survive in the presence of doxycycline, but can be rescued by a low-copy plasmid, pL-RPB1, carrying the RPB1 gene with its own promoter (Figure 1A). We introduced a mutagenized pool of pL-RPB1 into GRY3019 and then screened for novel phenotypes that were revealed when the chromosomal RPB1 allele was turned off.

Figure 1.—

Isolation of plasmid-borne rpb1 mutants. (A) Schematic representation and in vivo validation of the strategy used to isolate plasmid-borne rpb1 recessive mutants. Shown are serial 50-fold dilutions, containing different combinations of RPB1 alleles as indicated, plated in rich media without or with doxycycline. The striped and solid boxes represent a doxycycline repressible tet promoter and the RPB1 natural promoter, respectively. The shaded arrow represents the RPB1 open reading frame. YEPD, rich media; DOX, doxycycline. (B) The growth phenotype of plasmid-borne thermosensitive rpb1 mutants isolated using the strategy depicted in A. YEPD, rich media; DOX, doxycycline.

Isolation of temperature-sensitive rpb1 mutants:

To validate the mutant isolation strategy, we first looked for mutants with a robust growth phenotype in rich media at 30° but that were unable to grow at 37°. We chose this phenotype because of the ease of scoring it and the fact that a number of temperature-sensitive (TS) rpb1 mutants have already been described. Of 32,000 independent clones analyzed, 11 showed the TS phenotype. Due to the relatively long size of the RPB1 ORF, 5.2 kb, a high level of comutations could potentially complicate the identification of the relevant mutations. The sequencing of the candidate clones revealed five single, five double, and one triple mutant. Therefore, this mutagenesis level is sufficient to obtain the desired clones without the inconvenience of having too many secondary mutations. In agreement with the lethal phenotype of deletions of the carboxyl-terminal domain of rpb1 (Nonet et al. 1987), none of the mutations were out-of-frame deletions or insertions. The sequencing data uncovered unambiguously the temperature-sensitive alleles rpb1-C67Y, rpb1-C70Y, and rpb1-H80Y, isolated from two, five, and three independent clones, respectively (Figure 1B). An additional mutation, causing the residue change C103Y, was also isolated but was not further analyzed due to the presence of secondary mutations. To avoid possible artifacts due to the plasmid-borne expression of the rpb1 mutants or low-level expression of the PtetRPB1 allele, the mutations were integrated in the chromosomal RPB1 locus and tested for growth at 30° and 37° (Figure 2). It is interesting that all 11 TS alleles isolated in this screen were in the zinc-binding domain. Temperature-sensitive alleles of RPB1 that involve changes in amino acids C67, C70, or H80 were previously isolated by directed mutagenesis (Donaldson and Friesen 2000).

Figure 2.—

Isolation and identification of rpb1 mutants sensitive to 6AU. Shown are phenotypes of mutations in RPB1 causing sensitivity to 6AU, isolated using the strategy shown in Figure 1 or a related strategy (see text), after integration in the chromosome. Integrated versions of the rpb1 thermosensitive mutants are also shown. The allele names indicate the amino acid change in Rpb1. Mutation positions are relative to the first nucleotide in the open reading frame of RPB1 and the changes shown correspond to the nontranscribed strand. All the mutants contain a single-base substitution in RPB1. 6AU, 6-azauracil.

Isolation of 6AU-sensitive rpb1 mutants:

We used the same library of mutagenized pL-RPB1 transformed into GRY3019 to isolate mutants sensitive to 6AU. Transformants were screened for the ability to grow in the presence of doxycycline, but for lack of growth in the presence of doxycycline plus 6AU. Among ∼30,000 independent transformants analyzed in this screen, we found four 6AUS mutations, rpb1-D261N, rpb1-R320C, rpb1-E1103G, and rpb1-E1230K. Two additional 6AUS alleles, rpb1-D260N and rpb1-N488D, were obtained from a related screen of ∼30,000 independent transformants done in a strain lacking TFIIS, GRY3100. The 6AUS phenotype was reproduced after integration of the mutant alleles in the chromosome (Figure 2). As noted above, the rpb1-E1230K allele has been previously isolated as rpo21-24, a mutant with a reduced interaction of RNAPII with TFIIS (Wu et al. 1996). The other alleles are novel and alter highly conserved regions of Rpb1 (Figure 3) and are located in the vicinity of the RNA–DNA hybrid (Figure 4). Amino acids D260, D261, and R320 map in the lid and rudder domains of Rpb1. The lid and rudder are located in the upstream limit of the RNA–DNA hybrid and have been proposed to have a role in separating the RNA from the template strand (Cramer et al. 2001; Gnatt et al. 2001; Kettenberger et al. 2004; Westover et al. 2004). The mutants described here are the first eukaryotic mutants in those regions and their isolation as 6AUS indicates that they are important for transcription in vivo. Amino acid N488 is located in the proximity of the invariant motif NADFDGD that coordinates one of two Mg2+ ions (metal A) in the active center of the enzyme downstream of the RNA–DNA hybrid (Cramer et al. 2001; Gnatt et al. 2001). N488 is also located remarkably close to the basic residues N445 and R446 in the RNAPII structure (Figure 4C). Mutations affecting N445 have a strong effect in transcription initiation, specifically in start site selection (Berroteran et al. 1994; Archambault et al. 1998). Amino acid E1103 is located near the position where the nontemplate strand is separated from the template strand downstream of the active site in a region defined by sequence homology (Jokerst et al. 1989) that has been shown to control the lateral mobility of the elongation complexes in bacteria (Bar-Nahum et al. 2005).

Figure 3.—

Evolutionary conservation of Rpb1-mutated residues. Sequence alignments covering the entire lid and rudder (A), regions in the active site (B), homology box G (C), and TFIIS binding (D) are shown. The numbers refer to the codon of the S. cerevisiae open reading frame. The S. cerevisiae original (blue) and mutant (red) Rpb1 residues are indicated. The black striped lines show residues involved in the coordination of Mg2+ (metal A) and in the physical interaction with TFIIS. Independent alignments in A–D show the homology among eukaryotes (first row) and between eukaryotes, archaebacteria, and eubacteria (second row). No relevant homology was found for bacteria in D. Blue line, eukaryote; green line, archaebacteria; orange line, eubacteria; *, identical residue; :, conserved substitution; ., semiconserved substitution; Sc, S. cerevisiae; Sp, Schizosaccharomyces pombe; Ce, Caenorhabditis elegans; Dm, Drosophila melanogaster; Hs, Homo sapiens; Pf, Pyrococcus furiosus; Ta, Thermus aquaticus; Ec, Escherichia coli.

Figure 4.—

The novel rpb1 mutants sensitive to 6AU affect residues located in the vicinity of the RNA–DNA hybrid. Depiction of Rpb1 (blue) and the RNA–DNA hybrid (light–dark green) in a three-dimensional structure of the RNAPII transcription complex (1SFO.PDB) is shown. For simplicity other subunits of RNAPII are not shown. The position of the residues affected in the rpb1 mutants sensitive to 6AU (red), the aspartic acids coordinating metal A (white), and two additional residues of interest (purple) are highlighted. (A) Right side view. (B) Top view. (C) Zoom showing the residues of interest closest to the RNA–DNA hybrid. Right and top are arbitrary but consistent with previous reports in the literature. The representation uses van der Waals spheres with the exception of the second pictures (zooms centered in the RNA–DNA hybrid) of A and B, where the Rpb1 backbone is represented by ribbons for better visualization of specific residues. The two pictures in C are rotated ∼180° in the y-axis. All the novel 6AU mutations alter residues in the internal surface of the RNAPII and are not in direct contact with any other subunit.

Altered regulation of IMD2 and SSM1 in the 6AU-sensitive rpb1 mutants:

Yeast cells respond to changes in the pools of nucleosides caused by 6AU by inducing genes involved in nucleoside metabolism like IMD2/PUR5 and SSM1/SDT1. Imd2 participates in the first step of the GMP anabolic pathway (Escobar-Henriques and Daignan-Fornier 2001) and Ssm1 is a pyrimidine nucleotidase required for detoxification of 6AU (Nakanishi and Sekimizu 2002). One reason cells can be sensitive to 6AU is because of a failure to induce these genes. Mutants that are sensitive to 6AU often fail to induce the IMD2/PUR5 and SSM1/SDT1 genes (Shaw and Reines 2000; Shimoaraiso et al. 2000). Therefore we tested the expression levels of IMD2 and SSM1 in our mutants and their response to 6AU. As loading controls, we also monitored the constitutively expressed ACT1 and SED1 genes and the level of RNA polymerase I-dependent rRNA25S ribosomal RNA. As shown in Figure 5, IMD2 is clearly upregulated in rpb1-D260N and rpb1-N488D (and to a lesser extent in rpb1-E1230K) before the addition of 6AU. In other words, these mutants behave as if they are starved for nucleotides even in the absence of 6AU. In contrast, rpb1-D261N, rpb1-R320C, and rpb1-E1103G fail to induce IMD2 or SSM1 upon addition of 6AU. These results indicate an abnormal regulation of genes involved in nucleotide metabolism in the rpb1 mutants tested and suggest that rpb1 mutants can be sensitive to 6AU for very different reasons, perhaps reflective of defects in different steps of transcription.

Figure 5.—

rpb1 alleles sensitive to 6-azauracil have altered transcription regulation of IMD2 and SSM1 genes. (A) Gene expression of IMD2 and SSM1 in response to 6AU determined by Northern analysis. ACT1, SED1, and rRNA are loading controls. For simplicity, rpb1 mutant alleles are identified just by the position of the amino acid change. The triangles represent from left to right 0, 30, and 120 min in the presence of 6AU. (B) Averaged quantitation of two independent sets of RNA preparations and Northerns compared to the expression ratio of IMD2 normalized to the expression of SED1 in the absence of 6AU.

rpb1-N488D decreases and rpb1-E1103G increases RNAPII polymerization rate in vitro:

To test the effect of the different amino acid substitutions on the biochemical properties of RNAPII, the mutations were integrated into strain GRY3030, which contains a hexahistidine-tagged Rpb3 that can be used for affinity purification of the RNAPII (Kireeva et al. 2003). The mutant RNA polymerases were purified and the polymerization capabilities were tested in vitro, using factor-independent assembly of transcription elongation complexes (Kireeva et al. 2003). TFIIS is not present or required in this assay and, as expected, rpb1-E1230K does not affect the polymerization properties of the RNAPII tested in this assay (not shown). We also did not detect a difference between RNAPII from wild type vs. that from rpb1-D261N (not shown) or rpb1-R320C in this assay (Figure 6). On the other hand, this assay clearly shows that the rpb1-N488D mutation causes a change in the intrinsic elongation properties of the enzyme, resulting in an RNAPII that is slower than the wild-type polymerase. In contrast, the RNAPII from rpb1-E1103G is faster than the wild-type enzyme. For unknown reasons, we were unable to introduce rpb1-D260N into GRY3030. These results again indicate that 6AU sensitivity can be caused by very different kinds of defects in RNAPII.

Figure 6.—

RNAPII in vitro transcription rate is decreased and increased by the Rpb1 amino acid changes N488D and E1103G, respectively. RNAPII transcription complexes assembled in vitro elongate a short priming RNA upon addition of nucleotides. The Rpb1 amino acid change of the complexes is indicated. The solid line with an asterisk represents the RNA radioactively labeled at the 5′ end; the shaded lanes represent DNA strands; the triangles represent from left to right 0, 10, 20, 40, 90, and 180 sec in the presence of nucleotides.

Synthetic interactions with SOH1, SPT4, and DST1 distinguish 6AU-sensitive rpb1 alleles in vivo:

The 6AUS mutants exhibited different biochemical defects and showed different alterations to the expression of IMD2 and SSM1. To see whether these defects correlated with sensitivity to loss of specific transcription factors, we tested the genetic interaction of the 6AUS rpb1 alleles with the transcription elongation genes DST1 and SPT4 and with MED31/SOH1, a gene encoding a subunit of the Mediator transcription initiation complex (Guglielmi et al. 2004; Linder and Gustafsson 2004). The results, summarized in Figure 7, again show that these 6AUS alleles have different defects. The rpb1-E1230K mutant, presumably due to its inability to interact with TFIIS, has a synthetic lethal phenotype with soh1, as expected on the basis of the synthetic lethal phenotype of soh1 dst1 mutants (Malagon et al. 2004). The novel mutants rpb1-D260N, rpb1-D261N, rpb1-R320C, and rpb1-N488D are also synthetic lethal with a soh1 deletion, suggesting a role in transcription initiation or early elongation phases. In contrast, the rpb1-E1103G allele shows a synthetic lethal phenotype with dst1 and spt4. Two other rpb1 mutations, rpb1-221 (H1367D) and rpb1-244 (E1351K), isolated as suppressors of a cryosensitive spt5 mutant, have been described as 6AUS and synthetic with dst1 (Hartzog et al. 1998). The synthetic interaction of rpb1-E1103G with DST1 and SPT4 suggests an in vivo requirement of residue E1103 for a proper transcription elongation.

Figure 7.—

rpb1 alleles show differential synthetic phenotypes with transcription factor mutants. (A) Examples of tetratypes revealing the four meiotic products from selected crosses. The open bars indicate the double mutants rpb1-N488D dst1 (top left tetrad), rpb1-N488D soh1 (top right tetrad), rpb1-E1103G dst1 (bottom left tetrad), and rpb1-E1103G soh1 (bottom right tetrad). (B) Table showing the complete panel of growth phenotypes of the rpb1 alleles in the absence of TFIIS (dst1), Spt4 (spt4), and Soh1 (soh1). For simplicity, rpb1 mutant alleles are identified just by the position of the amino acid change.


In this work we describe the isolation and initial characterization of novel rpb1 mutations that are highly sensitive to the nucleotide-depleting drug 6AU. These mutants define several classes of defects that are distinguished by their behavior in an initiation factor independent in vitro transcription elongation assay, their regulation of genes involved in nucleotide metabolism induction (IMD2 and SSM1), and their genetic interaction with mutants defective in elongation (dst1 and spt4) or initiation (soh1). Three of the mutations (rpb1-D260N, rpb1-D261N, and rpb1-R320C) alter the lid or rudder loops that are postulated to have roles in separating the RNA–DNA hybrid and have a synthetic phenotype with soh1. Two of the mutants near the catalytic center (rpb1-N488D) and the homology box G downstream of the active site (rpb1-E1103G) alter the intrinsic properties of RNAPII in vitro and show differential synthetic interactions in vivo.

Lid and rudder mutants:

The lid and rudder loops of Rpb1, along with the Rpb2 fork loop 2, are located in the upstream limit of the RNA–DNA hybrid, forming a “strand-loop network” with dynamic complex interactions (Westover et al. 2004). Structural studies in yeast and bacteria indicate that these loops may facilitate the separation of the RNA from the template DNA strand (Korzheva et al. 2000; Westover et al. 2004). It has been suggested that the formation of the strand-loop network occurs during a transcriptional pause provoked by the clash of the 5′ end of the RNA with TFIIB during transcription initiation (Bushnell et al. 2004; Westover et al. 2004). This led to the proposal that during promoter escape the formation of the strand-loop network allows further chain elongation, causing a displacement of TFIIB in eukaryotes and of σ-factor in bacteria (Vassylyev et al. 2002; Westover et al. 2004). We show here in vivo phenotypes caused by alterations in the lid and rudder of RNAPII caused by the rpb1-D260N, rpb1-D261N, or rpb1-R320C mutations. Amino acids D260, D261, and R320 of Rpb1 in yeast and the correspondent amino acids in bacteria are located close to one another in the RNA polymerase crystal structures (Vassylyev et al. 2002; Westover et al. 2004) (see Figure 4C), suggesting possible roles in the formation of the strand-loop network. The rpb1-D261N and rpb1-R320C mutations cause a defect in the induction of IMD2 and SSM1 genes that may be sufficient to explain the sensitivity to 6AU. In contrast, the rpb1-D260N mutant expresses IMD2 and SSM1 even in the absence of 6AU. It will be interesting to determine whether these alterations in the expression levels of these genes involved in nucleotide metabolism reflect a direct effect on initiation or some indirect effect. We detected no defect in elongation efficiency for rpb1-D261N or rpb1-R320C in the transcription factor independent assay used here, which is consistent with the view that their defect is at some other step. Indeed, each of these lid and rudder mutants is unable to survive when combined with a defect in initiation caused by loss of a subunit of the Mediator initiation complex, soh1. Further experiments will be required to determine whether they are specifically defective in initiation and whether that defect is manifested at particular genes.

The active site mutant, rpb1-N488D:

Treatment with 6AU alters the nucleotide pools and causes a decrease in the rate and processivity of RNAPII in vivo (Exinger and Lacroute 1992; Mason and Struhl 2005). Two mutations in RNAPII that decrease the RNA elongation rate in vitro have been described: the 6AU-sensitive rpb2-10 allele in S. cerevisiae (Scafe et al. 1990; Powell and Reines 1996) and the Drosophila melanogaster C4 mutation corresponding to a change in RPB1-R726 in yeast (Coulter and Greenleaf 1985). Similar to the 6AU treatment, Rpb2-P1018S (the rpb2-10 mutation) decreases the polymerization rate and the processivity of RNAPII in vivo (Mason and Struhl 2005). A lower speed of RNAPII theoretically can increase the probability of transcriptional arrest, an irreversible state of RNAPII in vitro that can be rescued only by TFIIS and that has been invoked to explain transcription-associated recombination and mutation (Aguilera 2002). Surprisingly, although rpb2-10 mutants show some synthetic interaction with dst1, as shown by the reduced levels of poly(A) RNA in a rpb2-10 dst1 double mutant compared to the single mutations, rpb2-10 mutants are not synthetic lethal with TFIIS (Lennon et al. 1998). Similarly, we found that Rpb1-N488D had a decreased RNA elongation rate in vitro and rpb1-N488D mutants were hypersensitive to 6AU and were not synthetic lethal with dst1 on YEPD. We did note that the rpb1-N488D dst1 double mutant showed slower growth than either single mutant on minimal media (data not shown). Similarly to the lid and rudder mutants, rpb1-N488D has a strong synthetic phenotype with soh1. Soh1 is a bona fide subunit of the transcription initiation Mediator complex (Guglielmi et al. 2004; Linder and Gustafsson 2004) originally isolated in a screen for suppressor of hyperrecombination mutants (Fan and Klein 1994). We believe that the simplest explanation for the synthetic interaction of rpb1-N488D with soh1 is that, in addition to its possible role in transcription elongation highlighted by its similarities with rpb2-10, the Rpb1 residue N488 also plays a role in transcription initiation. This interpretation is supported by the fact that the rpb1 mutations sua8-1 (rpb1-N445S) and sit1-278 (rpb1-N445T) alter amino acids that are located adjacent to N488 in the RNAPII structure (see Figure 4C) and affect transcription start site selection in vivo (Berroteran et al. 1994; Archambault et al. 1998). It remains to be determined whether the increased level of expression of IMD2 and SSM1 in the absence of 6AU caused by the rpb1-N488D mutation reflects an alteration in initiation.

The downstream mutant rpb1-E1103G:

The 6AUS mutant rpb1-E1103G causes an alteration in the regulation of IMD2 and SSM1 so that they fail to induce in response to 6AU. Rpb1-E1103G exhibited an increased RNA polymerization rate in vitro in our transcription factor independent assay. The position of the residue E1103 in the G loop, a region that has been suggested to modulate the catalytic activity of bacterial RNA polymerase, gives some insights into the effect of the mutation. Recently, Bar-Nahum and collaborators described a mutation in this G loop of bacterial RNA polymerase also showing an associated increase in the polymerization rate (Bar-Nahum et al. 2005). A “fast” RNAPII has also been described in D. melanogaster, the S1 mutant that altered the DNA–RNA hybrid-binding region of the Rpb2 homolog (Chen et al. 1996). Changes in the speed of the RNAPII may interfere with a series of tightly coupled mRNA processes occurring during elongation, as illustrated by the correlation between RNAPII elongation rate and efficiency of mRNA splicing in eukaryotes (de la Mata et al. 2003; Howe et al. 2003). The rpb1-E1103G was unique in our collection in demonstrating a synthetic lethal interaction with the deletion of DST1 or SPT4. Both TFIIS and the Spt4/Spt5 complex affect RNA splicing (Howe et al. 2003; Lindstrom et al. 2003; Xiao et al. 2005) and rpb1 mutants synthetic with dst1 were previously isolated by their genetic interaction with SPT5 (Hartzog et al. 1998). Defects in mRNA processing caused by an increase in chain elongation rate may explain the high dependence on TFIIS and Spt4 for cell viability in the rpb1-E1103G mutant.


The collection of rpb1 mutants described here, although originally isolated as sensitive to 6AU, exhibits several very different biochemical and genetic interaction phenotypes. RNAPII has multiple roles in transcription including initiation, promoter escape, elongation, splicing, transcription-coupled repair, and termination. The results presented here are consistent with the view that several of those roles can be rendered sensitive to nucleotide pool levels by mutations in different domains of Rpb1.


We thank Alison Rattray and Dwight Nissley for critical reading of the manuscript and Deanna Gotte for sequencing assistance. This work was supported by the Intramural Research Program of the National Institutes of Health, National Cancer Institute, Center for Cancer Research. The contents of this publication do not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organization imply endorsement by the U.S. government.


  • Communicating editor: M. Hampsey

  • Received October 14, 2005.
  • Accepted January 30, 2006.


View Abstract