Phenotypic analysis of Paf1/RNA polymerase II complex mutations reveals connections to cell cycle regulation, protein synthesis, and lipid and nucleic acid metabolism

Paf1 is an RNA polymerase II-associated protein in yeast, which defines a complex that is distinct from the Srb/Mediator holoenzyme. The Paf1 complex, which also contains Ctr9, Cdc73, Hpr1, Ccr4, Rtf1 and Leo1, is required for full expression of a subset of yeast genes, particularly those responsive to signals from the Pkc1/MAP kinase cascade. We have extensively characterized the pleiotropic phenotypes of deletion mutants for factors present in the Paf1 complex, identifying more than a dozen new phenotypes, and, in some cases, establishing possible molecular explanations for the growth defects. For example, paf1 Delta causes sensitivity to hydroxyurea; this phenotype correlates with a reduction in RNR1 transcript abundance and is suppressed by over-expression of RNR1. In contrast, the resistance of paf1 Delta cells to the transcription elongation inhibitors 6-azauracil and mycophenolic acid correlates with its ability to derepress the IMD2 transcript. We tested the hypothesis that Paf1 communicates with some promoters through the DNA-binding factors Swi4, Mbp1 or Rlm1. The phenotypes of mutations in Paf1 complex components are exacerbated in the swi4 Delta background, suggesting that the complex acts in a pathway parallel to that controlled by Swi4. Conversely, the fact that mbp1 Delta and rlm1 Delta mutations do not enhance the phenotypes suggests that the Paf1 complex may function in the same regulatory pathway(s) with Mbp1 and Rlm1.

Identifying a core RNA polymerase surface critical for interactions with a sigma-like specificity factor

Cyclic interactions occurring between a core RNA polymerase (RNAP) and its initiation factors are critical for transcription initiation, but little is known about subunit interaction. In this work we have identified regions of the single-subunit yeast mitochondrial RNAP (Rpo41p) important for interaction with its sigma-like specificity factor (Mtf1p). Previously we found that the whole folded structure of both polypeptides as well as specific amino acids in at least three regions of Mtf1p are required for interaction. In this work we started with an interaction-defective point mutant in Mtf1p (V135A) and used a two-hybrid selection to isolate suppressing mutations in the core polymerase. We identified suppressors in three separate regions of the RNAP which, when modeled on the structure of the closely related phage T7 RNAP, appear to lie on one surface of the protein. Additional point mutations and biochemical assays were used to confirm the importance of each region for Rpo41p-Mtf1p interactions. Remarkably, two of the three suppressors are found in regions required by T7 RNAP for DNA sequence recognition and promoter melting. Although these essential regions of the phage RNAP are poorly conserved with the mitochondrial RNAPs, they are conserved among the mitochondrial enzymes. The organellar RNAPs appear to use this surface in an alternative way for interactions with their separate sigma-like specificity factor, which, like its bacterial counterpart, provides promoter recognition and DNA melting functions to the holoenzyme.

A complex containing RNA polymerase II, Paf1p, Cdc73p, Hpr1p, and Ccr4p plays a role in protein kinase C signaling

Yeast contains at least two complex forms of RNA polymerase II (Pol II), one including the Srbps and a second biochemically distinct form defined by the presence of Paf1p and Cdc73p (X. Shi et al., Mol. Cell. Biol. 17:1160-1169, 1997). In this work we demonstrate that Ccr4p and Hpr1p are components of the Paf1p-Cdc73p-Pol II complex. We have found many synthetic genetic interactions between factors within the Paf1p-Cdc73p complex, including the lethality of paf1Delta ccr4Delta, paf1Delta hpr1Delta, ccr4Delta hpr1Delta, and ccr4Delta gal11Delta double mutants. In addition, paf1Delta and ccr4Delta are lethal in combination with srb5Delta, indicating that the factors within and between the two RNA polymerase II complexes have overlapping essential functions. We have used differential display to identify several genes whose expression is affected by mutations in components of the Paf1p-Cdc73p-Pol II complex. Additionally, as previously observed for hpr1Delta, deleting PAF1 or CDC73 leads to elevated recombination between direct repeats. The paf1Delta and ccr4Delta mutations, as well as gal11Delta, demonstrate sensitivity to cell wall-damaging agents, rescue of the temperature-sensitive phenotype by sorbitol, and reduced expression of genes involved in cell wall biosynthesis. This unusual combination of effects on recombination and cell wall integrity has also been observed for mutations in genes in the Pkc1p-Mpk1p kinase cascade. Consistent with a role for this novel form of RNA polymerase II in the Pkc1p-Mpk1p signaling pathway, we find that paf1Delta mpk1Delta and paf1Delta pkc1Delta double mutants do not demonstrate an enhanced phenotype relative to the single mutants. Our observation that the Mpk1p kinase is fully active in a paf1Delta strain indicates that the Paf1p-Cdc73p complex may function downstream of the Pkc1p-Mpk1p cascade to regulate the expression of a subset of yeast genes.

A multiplicity of mediators: alternative forms of transcription complexes communicate with transcriptional regulators

The already complex process of transcription by RNA polymerase II has become even more complicated in the last few years with the identification of auxiliary factors in addition to the essential general initiation factors. In many cases these factors, which have been termed mediators or co-activators, are only required for activated or repressed transcription. In some cases the effects are specific for certain activators and repressors. Recently some of these auxiliary factors have been found in large complexes with either TBP, as TBP-associated factors (TAFs) in the general factor TFIID, or with pol II and a subset of the general factors, referred to as the 'holoenzyme'. Although the exact composition of these huge assemblies is still a matter of some debate, it is becoming clear that the complexes themselves come in more than one form. In particular, at least four forms of TFIID have been described, including one that contains a tissue-specific TAF and another with a cell type-specific form of TBP. In addition, in yeast there are at least two forms of the 'holoenzyme' distinguished by their mediator composition and by the spectrum of transcripts whose expression they affect. Genetic and biochemical analyses have begun to identify the interactions between the components of these complexes and the ever increasing family of DNA binding regulatory factors. These studies are complicated by the fact that individual regulatory factors often appear to have redundant interactions with multiple mediators. The existence of these different forms of transcription complexes defines a new target for regulation of subsets of eukaryotic genes.

Identification of three regions essential for interaction between a sigma-like factor and core RNA polymerase

The cyclic interactions that occur between the subunits of the yeast mitochondrial RNA polymerase can serve as a simple model for the more complex enzymes in prokaryotes and the eukaryotic nucleus. We have used two-hybrid and fusion protein constructs to analyze the requirements for interaction between the single subunit core polymerase (Rpo41p), and the sigma-like promoter specificity factor (Mtf1p). We were unable to define any protein truncations that retained the ability to interact, indicating that multiple regions encompassing the entire length of the proteins are involved in interactions. We found that 9 of 15 nonfunctional (petite) point mutations in Mtf1p isolated in a plasmid shuffle strategy had lost the ability to interact. Some of the noninteracting mutations are temperature-sensitive petite (ts petite); this phenotype correlates with a precipitous drop in mitochondrial transcript abundance when cells are shifted to the nonpermissive temperature. One temperature-sensitive mutant demonstrated a striking pH dependence for core binding in vitro, consistent with the physical properties of the amino acid substitution. The noninteracting mutations fall into three widely spaced clusters of amino acids. Two of the clusters are in regions with amino acid sequence similarity to conserved regions 2 and 3 of sigma factors and related proteins; these regions have been implicated in core binding by both prokaryotic and eukaryotic sigma-like factors. By modeling the location of the mutations using the partial structure of Escherichia coli sigma70, we find that two of the clusters are potentially juxtaposed in the three-dimensional structure. Our results demonstrate that interactions between sigma-like specificity factors and core RNA polymerases require multiple regions from both components of the holoenzymes.

Cdc73p and Paf1p are found in a novel RNA polymerase II-containing complex distinct from the Srbp-containing holoenzyme

The products of the yeast CDC73 and PAF1 genes were originally identified as RNA polymerase II-associated proteins. Paf1p is a nuclear protein important for cell growth and transcriptional regulation of a subset of yeast genes. In this study we demonstrate that the product of CDC73 is a nuclear protein that interacts directly with purified RNA polymerase II in vitro. Deletion of CDC73 confers a temperature-sensitive phenotype. Combination of the cdc73 mutation with the more severe paf1 mutation does not result in an enhanced phenotype, indicating that the two proteins may function in the same cellular processes. To determine the relationship between Cdc73p and Paf1p and the recently described holoenzyme form of RNA polymerase II, we created yeast strains containing glutathione S-transferase (GST)-tagged forms of CDC73, PAF1, and TFG2 functionally replacing the chromosomal copies of the genes. Isolation of GST-tagged Cdc73p and Paf1p complexes has revealed a unique form of RNA polymerase II that contains both Cdc73p and Paf1p but lacks the Srbps found in the holoenzyme. The Cdc73p-Paf1p-RNA polymerase II-containing complex also includes Gal11p, and the general initiation factors TFIIB and TFIIF, but lacks TBP, TFIIH, and transcription elongation factor TFIIS as well as the Srbps. The Srbp-containing holoenzyme does not include either Paf1p or Cdc73p, demonstrating that these two forms of RNA polymerase II are distinct. In confirmation of the hypothesis that the two forms coexist in yeast cells, we found that a TFIIF-containing complex isolated via the GST-tagged Tfg2p construct contains both (i) the Srbps and (ii) Cdc73p and Paf1p. The Srbps and Cdc73p-Paf1p therefore appear to define two complexes with partially redundant, essential functions in the yeast cell. Using the technique of differential display, we have identified several genes whose transcripts require Cdc73p and/or Paf1p for normal levels of expression. Our analysis suggests that there are multiple RNA polymerase II-containing complexes involved in the expression of different classes of protein-coding genes.

A novel collection of accessory factors associated with yeast RNA polymerase II

A relatively simple subset of general transcription factors is sufficient for transcript initiation by RNA polymerase II. However, a recently identified "holoenzyme" contains additional accessory proteins required for mediating signals from some activators (Y-J. Kim et al., 1994, Cell 77, 599-608; A. Koleske and R. Young, 1994, Nature 368, 466-469). By immobilizing RNA polymerase II and associated proteins (RAPs) from a transcriptionally active yeast extract, we have identified a novel collection of proteins distinct from those found in the holoenzyme. The eluted RAP fraction did not contain the holoenzyme components Srb2,4,5 + 6p, Gal11p, or Sug1p, but did include the known transcription factors TFIIB and TFIIS and the three subunits of yeast TFIIF (Ssu71p/Tfg1p, Tfg2p, and Anc1p/Tfg3p). Also isolated as RAPs are two proteins (Cdc73p and Paf1p) with interesting connections to gene expression. Mutations in CDC73 and PAF1 affect cell growth and the abundance of transcripts from a subset of yeast genes (X. Shi et al., Mol. Cell. Biol., 1996 16, 669-676). The RAP fraction may therefore define one or more functional forms of RNA polymerase II distinct from the activator-mediating holoenzyme.

Transcriptional corepression in vitro: a Mot1p-associated form of TATA-binding protein is required for repression by Leu3p

Signals from transcriptional activators to the general mRNA transcription apparatus are communicated by factors associated with RNA polymerase II or the TATA-binding protein (TBP). Currently, little is known about how gene-specific transcription repressors communicate with RNA polymerase II. We have analyzed the requirements for repression by the saccharomyces cerevisiae Leu3 protein (Leu3p) in a reconstituted transcription system. We have identified a complex form of TBP which is required for communication of the repressing signal. This TFIID-like complex contains a known TBP-associated protein, Mot1p, which has been implicated in the repression of a subset of yeast genes by genetic analysis. Leu3p-dependent repression can be reconstituted with purified Mot1p and recombinant TBP. In addition, a mutation in the Mot1 gene leads to partial derepression of the Leu3p-dependent LEU2 promoter. These in vivo and in vitro observations define a role for Mot1p as a transcriptional corepressor.

Paf1p, an RNA polymerase II-associated factor in Saccharomyces cerevisiae, may have both positive and negative roles in transcription

Regulated transcription initiation requires, in addition to RNA polymerase II and the general transcription factors, accessory factors termed mediators or adapters. We have used affinity chromatography to identify a collection of factors that associate with Saccharomyces cerevisiae RNA polymerase II (P. A. Wade, W. Werel, R. C. Fentzke, N. E. Thompson, J. F. Leykam, R. R. Burgess, J. A. Jaehning, and Z. F. Burton, submitted for publication). Here we report identification and characterization of a gene encoding one of these factors, PAF1 (for RNA polymerase-associated factor 1). PAF1 encodes a novel, highly charged protein of 445 amino acids. Disruption of PAF1 in S. cerevisiae leads to pleiotropic phenotypic traits, including slow growth, temperature sensitivity, and abnormal cell morphology. Consistent with a possible role in transcription, Paf1p is localized to the nucleus. By comparing the abundances of many yeast transcripts in isogenic wild-type and paf1 mutant strains, we have identified genes whose expression is affected by PAF1. In particular, disruption of PAF1 decreases the induction of the galactose-regulated genes three- to fivefold. In contrast, the transcript level of MAK16, an essential gene involved in cell cycle regulation, is greatly increased in the paf1 mutant strain. Paf1p may therefore be required for both positive and negative regulation of subsets of yeast genes. Like Paf1p, the GAL11 gene product is found associated with RNA polymerase II and is required for regulated expression of many yeast genes including those controlled by galactose. We have found that a gal11 paf1 double mutant has a much more severe growth defect than either of the single mutants, indicating that these two proteins may function in parallel pathways to communicate signals from regulatory factors to RNA polymerase II.

The yeast EGD2 gene encodes a homologue of the alpha NAC subunit of the human nascent-polypeptide-associated complex

Two Saccharomyces cerevisiae proteins of 21 and 27 kDa co-purify with a novel enhancer of Gal4p DNA binding activity (Egdp) [Parthun et al., Mol. Cell. Biol. 12 (1992) 5683-5689]. Mutations in the EGD1 gene encoding the 21-kDa protein (Egd1p) have been shown to affect the kinetics and extent of the Gal4p-mediated, galactose-induced activation of the GAL genes. Egd1p is homologous to human BTF3b, recently identified as the beta subunit of the heterodimeric nascent-polypeptide-associated complex (NAC) involved in ensuring signal-sequence-specific protein sorting and translocation [Wiedmann et al., Nature 370 (1994) 434-440]. We have cloned and characterized EGD2 encoding the 27-kDa protein and found that Egd2p is strikingly similar to the alpha subunit of human NAC. Yeast, therefore, contains a complex composed of Egd1p and Egd2p very similar to the NAC complex described in human cells. Disruption of EGD2, alone or in combination with an EGD1 disruption, causes no obvious phenotypes. The lack of phenotype, the high levels of EGD1 and EGD2 expression, and the identification of multiple human genes encoding NAC subunits suggest that the yeast EGD genes may be members of multigene families with redundant function.

Isolation of yeast transcription factor IIA using a functional transcription assay

TFIIA was extensively purified from a whole-cell transcription extract from yeast. Activity was followed throughout isolation utilizing a functional transcription assay. Transcription activity was found to copurify with polypeptides of 43 and 12.5 kDa, consistent with a previous purification that utilized a TBP/DNA gel mobility shift assay (J. Ranish and S. Hahn, J. Biol. Chem. 266, 19320-19327, 1991). The Stoke's radius of the purified protein was determined by gel filtration chromatography to be 44 A under native conditions. The solution molecular weight derived from this measurement, 110 kDa, is consistent with a heterotetrameric structure of TFIIA.

Release of the yeast mitochondrial RNA polymerase specificity factor from transcription complexes

The yeast mitochondrial RNA polymerase is composed of two nuclear encoded subunits, a catalytic core (Rpo41p), which resembles the enzymes from bacteriophage T7 and T3, and a specificity factor required for promoter recognition (Mtf1p), which is similar to members of the eubacterial sigma factor family. Using mitochondrial RNA polymerase reconstituted from highly purified subunits, we have determined that Rpo41p and Mtf1p interact to form a holoenzyme in solution prior to DNA binding and promoter recognition. We analyzed the composition of the polymerase during and after the initiation of transcription and found that, like the eubacterial sigma factors, Mtf1p is released after initiation and is available to catalyze transcription on a second template. By analyzing gel mobility shift complexes of the RNA polymerase and DNA at different stages of the transcription reaction, we found that both subunits were associated with DNA prior to initiation and after the formation of two phosphodiester bonds. After the formation of a 13-nucleotide transcript, Mtf1p is no longer associated with Rpo41p on the DNA. These data establish that Mtf1p is functionally as well as structurally similar to eubacterial sigma factors.

MTF1, encoding the yeast mitochondrial RNA polymerase specificity factor, is located on chromosome XIII

MTF1 is a nuclear gene that encodes the promoter recognition factor of the yeast mitochondrial RNA polymerase. The MTF1 gene was physically mapped to chromosome XIII. Genetic mapping data indicate that the gene is closely linked to RNA1.

Glucose repression of yeast mitochondrial transcription: kinetics of derepression and role of nuclear genes

Yeast mitochondrial transcript and gene product abundance has been observed to increase upon release from glucose repression, but the mechanism of regulation of this process has not been determined. We report a kinetic analysis of this phenomenon, which demonstrates that the abundance of all classes of mitochondrial RNA changes slowly relative to changes observed for glucose-repressed nuclear genes. Several cell doublings are required to achieve the 2- to 20-fold-higher steady-state levels observed after a shift to a nonrepressing carbon source. Although we observed that in some yeast strains the mitochondrial DNA copy number also increases upon derepression, this does not seem to play the major role in increased RNA abundance. Instead we found that three- to sevenfold increases in RNA synthesis rates, measured by in vivo pulse-labelling experiments, do correlate with increased transcript abundance. We found that mutations in the SNF1 and REG1 genes, which are known to affect the expression of many nuclear genes subject to glucose repression, affect derepression of mitochondrial transcript abundance. These genes do not appear to regulate mitochondrial transcript levels via regulation of the nuclear genes RPO41 and MTF1, which encode the subunits of the mitochondrial RNA polymerase. We conclude that a nuclear gene-controlled factor(s) in addition to the two RNA polymerase subunits must be involved in glucose repression of mitochondrial transcript abundance.

Accurate initiation of mRNA synthesis in extracts from Schizosaccharomyces pombe, Kluyveromyces lactis and Candida glabrata

We demonstrate the successful adaptation to other yeast species of a protocol previously described for production of transcriptionally active whole cell extracts from Saccharomyces cerevisiae (Woontner and Jaehning, 1990, J. Biol. Chem. 265, 8979-8982). Extracts prepared from Schizosaccharomyces pombe, Kluyveromyces lactis and Candida glabrata were all capable of initiating transcription from a template containing the S. cerevisiae CYC1 TATA box fused to a G-less cassette. Transcription in all of the extracts was sensitive to inhibition by alpha-amanitin, indicating that it was catalysed by RNA polymerase II, and was dramatically stimulated by the chimeric activator GAL4/VP16. The different extracts used different subsets of a group of three initiation sites.

Gal4 protein binding is required but not sufficient for derepression and induction of GAL2 expression

The Saccharomyces cerevisiae GAL2 gene upstream activator sequence (UAS) region was examined for protein bound in vivo by chromatin footprinting at high resolution. Gal4 transcriptional activator protein binds to the two consensus UAS sites whether GAL2 expression is induced, uninduced, or repressed by growth with different carbon sources. Although wild type strains show loss of the Gal4 protein-specific footprint in repressing media containing glucose, constitutive high level expression of Gal4 protein restores the GAL2 UAS footprints without fully derepressing GAL2 transcription. Thus binding of the Gal4 activator to target sites in the DNA is required but not sufficient for GAL2 derepression and induction. Gal4-independent protein-DNA complexes were also detected in the region, including one over the previously noted centromere-binding protein (CP1) site upstream of the Gal4 complexes.

Resolution of transcription factors from a transcriptionally active whole-cell extract from yeast: purification of TFIIB, TBP, and RNA polymerase IIa

We describe techniques for production and chromatographic fractionation of a transcriptionally active whole-cell extract from Saccharomyces cerevisiae. The procedure is suitable for large-scale isolation of the factors involved in mRNA synthesis. Both yeast transcription factor IIB and TATA-binding protein were purified from the extract as single species using an immunoblot assay. In addition, the three previously described isoforms of yeast RNA polymerase II were resolved and form IIa, the intact, unphosphorylated isoform proposed to be involved in initiation, was purified to apparent homogeneity.

Mitochondrial transcription: is a pattern emerging?

Despite the striking similarities of RNA polymerases and transcription signals shared by eubacteria, archaebacteria and eukaryotes, there has been little indication that transcription in mitochondria is related to any previously characterized model. Only in yeast has the subunit structure of the mitochondrial RNA polymerase been determined. The yeast enzyme is composed of a core related to polymerases from bacteriophage T7 and T3, and a promoter recognition factor similar to bacterial sigma factors. Soluble systems for studying mitochondrial transcript initiation in vitro have been described from several organisms, and used to determine consensus sequences at or near transcription start sites. Comparison of these sequences from fungi, plants, and amphibians with the T7/T3 promoter suggests some intriguing similarities. Mammalian mitochondrial promoters do not fit this pattern but instead appear to utilize upstream sites, the target of a transcriptional stimulatory factor, to position the RNA polymerase. The recent identification of a possible homologue of the mammalian upstream factor in yeast mitochondria may indicate that a pattern will eventually be revealed relating the transcriptional machineries of all eukaryotic mitochondria.

The EGD1 product, a yeast homolog of human BTF3, may be involved in GAL4 DNA binding

A variety of techniques, including filter binding, footprinting, and gel retardation, can be used to assay the transcriptional activator GAL4 (Gal4p) through the initial steps of its purification from yeast cells. Following DNA affinity chromatography, Gal4p still bound DNA selectively when assayed by filter binding or footprinting. However, the affinity-purified protein was no longer capable of forming a stable complex with DNA, as assayed by gel retardation. Mixing the purified Gal4p with the flowthrough fraction from the DNA affinity column restored gel retardation complex formation. Gel retardation assays were used to monitor the purification of a heat-stable Gal4p-DNA complex stabilization activity from the affinity column flowthrough. The activity coeluted from the final purification step with polypeptides of 21 and 27 kDa. The yeast gene encoding the 21-kDa protein was cloned on the basis of its N-terminal amino acid sequence. The gene, named EGD1 (enhancer of GAL4 DNA binding), encodes a highly basic protein (21% lysine and arginine) with a predicted molecular mass of 16.5 kDa. The amino acid sequence of the EGD1 product, Egd1p, is highly similar to that of the human protein BTF3 (X. M. Zheng, D. Black, P. Chambon, and J. M. Egly, Nature [London] 344:556-559, 1990). Although an egd1 null mutant was viable and Gal+, induction of the galactose-regulated genes in the egd1 mutant strain was significantly reduced when cells were shifted from glucose to galactose.

In vitro transcriptional activation by a metabolic intermediate: activation by Leu3 depends on alpha-isopropylmalate

In the absence of the leucine biosynthetic precursor alpha-isopropylmalate (alpha-IPM), the yeast LEU3 protein (Leu3p) binds DNA and acts as a transcriptional repressor in an in vitro extract. Addition of alpha-IPM resulted in a dramatic increase in Leu3p-dependent transcription. The presence of alpha-IPM was also required for Leu3p to compete effectively with another transcriptional activator, GAL4/VP16, for limiting transcription factors. Therefore, the addition of alpha-IPM appears to convert a transcriptional repressor into an activator. This represents an example in eukaryotes of direct transcriptional regulation by a small effector molecule.

A transcriptionally active form of GAL4 is phosphorylated and associated with GAL80

The GAL4 activator and GAL80 repressor proteins regulate the expression of yeast genes in response to galactose. A complex of the two proteins isolated from glucose-grown cells is inactive in an in vitro transcription reaction but binds DNA and blocks activation by the GAL4-VP16 chimeric activator. The complex purified from galactose-grown cells contains a mixture of phosphorylated and unphosphorylated forms of GAL4. The galactose-induced form of GAL4 activates in vitro transcription to levels similar to those seen with GAL4-VP16. The induced GAL4 complex is indistinguishable in size and apparent shape from the uninduced complex, consistent with a continued association with GAL80. These results confirm in vivo analyses that correlate GAL4 phosphorylation with galactose induction and support a model of transcriptional activation that does not require GAL80 dissociation.

The yeast mitochondrial RNA polymerase specificity factor, MTF1, is similar to bacterial sigma factors

We have purified the protein that confers selective promoter recognition on the core subunit of the yeast mitochondrial RNA polymerase. The N-terminal sequence of the 43-kDa specificity factor identified it as the product of the MTF1 gene described by Lisowsky and Michaelis (1988). We confirmed that MTF1 encoded the specificity factor by analyzing extracts from a yeast strain bearing a disruption of the gene. The extracts contained normal levels of core RNA polymerase but lacked selective transcription activity; adding the purified 43-kDa protein restored selective transcription. Comparison of the MTF1 protein sequence to the family of bacterial sigma factors has revealed striking similarity to domains identified with--10 promoter recognition, promoter melting, and holoenzyme stability.

The yeast IMP1 gene is allelic to GAL2

We have found that many laboratory strains of yeast are defective in galactose metabolism owing to a recessive mutation in the previously characterized nuclear gene, IMP1. This defect leads to a requirement for mitochondrial function for growth on, and metabolism of, galactose. Genetic background affects the degree to which cells are defective. In particular, alleles of GAL3 affect the ability to score the Imp phenotype. We have found that in imp1 strains, transcriptional induction of the galactose inducible genes (GAL1, 2, 7 + 10, MEL1) is normal, but galactose transport is reduced in both rho+ and rho0 cells. This phenotype is normally associated with mutations in GAL2, the galactose permease. Although the growth phenotypes of gal2 and imp1 mutants are distinct, we found that the transformation of imp1 rho0 strains with a plasmid containing the GAL2 gene allows these strains to grow on galactose. Initial genetic analyses did not demonstrate linkage between the GAL2 and IMP1 genes owing to the effects of an unlinked gene on the Imp phenotype. By disrupting the GAL2 gene in an Imp+ background, we have shown that IMP1 and GAL2 segregate as tightly linked genes. Based on these data, we believe that imp1 is a partially defective allele of the GAL2 gene.

Transcriptional activation in an improved whole-cell extract from Saccharomyces cerevisiae

We report an improved in vitro transcription system for Saccharomyces cerevisiae. Small changes in assay and whole-cell extraction procedures increase selective initiation by RNA polymerase II up to 60-fold over previous conditions (M. Woontner and J. A. Jaehning, J. Biol. Chem. 265:8979-8982, 1990), to levels comparable to those obtained with nuclear extracts. We have found that the simultaneous use of distinguishable templates with and without an upstream activation sequence is critical to the measurement of apparent activation. Transcription from any template was very sensitive to the concentrations of template and nontemplate DNA, extract, and activator (GAL4/VP16). Alterations in reaction conditions led to proportionately greater changes from a template lacking an upstream activation sequence; thus, the apparent ratio of activation is largely dependent on the level of basal transcription. Using optimal conditions for activation, we have also demonstrated activation by a bona fide yeast activator, heat shock transcription factor.

Accurate initiation by RNA polymerase II in a whole cell extract from Saccharomyces cerevisiae

We have developed a simple procedure for isolating a transcriptional extract from whole yeast cells which obviates the requirement for nuclear isolation. Detection of accurate mRNA initiation by RNA polymerase II in the extract requires the use of a sensitive assay, recently described by Kornberg and co-workers (Lue, N. F., Flanagan, P. M., Sugimoto, K., and Kornberg, R. D. (1989) Science 246, 661-664) that involves activation by a GAL4-VP16 fusion protein and a template lacking guanosine residues in the coding strand. The extract is prepared from fresh or frozen yeast cells by disruption with glass beads and fractionation of proteins by ammonium sulfate precipitation. The alpha-amanitin-sensitive transcripts synthesized in the assay were identical to those produced in a parallel assay using a yeast nuclear extract. The activity of the whole cell extract is lower per mg of protein than a nuclear extract but proportional to the volume of the nucleus relative to the whole cell. The optimal ranges for several reaction components including template, mono- and divalent cations, and nucleotide substrate concentration were determined. Under optimal conditions the whole cell extract produced a maximum of approximately 1 X 10(-2) transcripts/template molecule in 30 min.

Purification and characterization of the yeast transcriptional activator GAL4

We have purified extensively the transcriptional activator, GAL4, from a yeast strain overexpressing the gene product from the ADH1 promoter. Our purification followed GAL4 activity by its binding to a specific DNA target sequence, using filter binding assays. No specific binding activity was detected in extracts from a strain containing a disrupted copy of the GAL4 gene. The purification protocol included fractionation of a whole cell extract by ion-exchange and DNA-affinity chromatography on a column containing a 17-base pair oligomer encoding a near consensus GAL4 binding site. Two polypeptides co-eluted with the GAL4 DNA binding activity from the DNA-affinity column. One had an apparent molecular mass of 99 kDa (the predicted size of the GAL4 protein) and cross-reacted with antibodies raised against GAL4 epitopes from fusion proteins expressed in bacterial cells. The second polypeptide did not cross-react with the anti-GAL4 antibody and is presumed to be the GAL80 transcriptional repressor based on its size (48 kDa) and known physical association with the GAL4 protein. GAL4 binding activity elutes from a gel filtration column as a 155-kDa species suggesting that it exists in solution in a heterodimer complex of one GAL4 and one GAL80 molecule. The dissociation constant of the DNA-affinity-purified GAL4-GAL80 complex for a 900-base pair DNA fragment containing the UASGAL element from the GAL1-GAL10 divergent promoter was, Kd(effective) (0.15 M KCl) = 2.4 x 10(-9) M.

Use of yeast nuclear DNA sequences to define the mitochondrial RNA polymerase promoter in vitro

We have extended an earlier observation that the TATA box for the nuclear GAL10 gene serves as a promoter for the mitochondrial RNA polymerase in in vitro transcription reactions (C. S. Winkley, M. J. Keller, and J. A. Jaehning, J. Biol. Chem. 260:14214-14223, 1985). In this work, we demonstrate that other nuclear genes also have upstream sequences that function in vitro as mitochondrial RNA polymerase promoters. These genes include the GAL7 and MEL1 genes, which are regulated in concert with the GAL10 gene, the sigma repetitive element, and the 2 microns plasmid origin of replication. We used in vitro transcription reactions to test a large number of nuclear DNA sequences that contain critical mitochondrial promoter sequences as defined by Biswas et al. (T. K. Biswas, J. C. Edwards, M. Rabinowitz, and G. S. Getz, J. Biol. Chem. 262:13690-13696, 1987). The results of these experiments allowed us to extend the definition of essential promoter elements. This extended sequence, -ACTATAAACGatcATAG-, was frequently found in the upstream regulatory regions of nuclear genes. On the basis of these observations, we hypothesized that either (i) a catalytic RNA polymerase related to the mitochondrial enzyme functions in the nucleus of the yeast cell or (ii) a DNA sequence recognition factor is shared by the two genetic compartments. By using cells deficient in the catalytic core of the mitochondrial RNA polymerase (rpo41-) and sensitive assays for transcripts initiating from the nuclear promoter sequences, we have conclusively ruled out a role for the catalytic RNA polymerase in synthesizing transcripts from all of the nuclear sequences analyzed. The possibility that a DNA sequence recognition factor functions in both the nucleus and the mitochondria remains to be tested.

Two forms of RPO41-dependent RNA polymerase. Regulation of the RNA polymerase by glucose repression may control yeast mitochondrial gene expression

We have identified two chromatographically separable forms of mitochondrial RNA polymerase from Saccharomyces cerevisiae which utilize different DNA templates. One form is only active in a nonselective assay utilizing a poly[d(A-T)] template. The other form selectively initiates from a mitochondrial promoter consensus sequence. Both enzymes can be extracted from yeast mitochondria and all components are encoded by nuclear genes. The possibility that these two activities represent core and holoenzyme forms of the multicomponent mitochondrial RNA polymerase is supported by our observation that both enzymes are absent from a strain bearing a disrupted copy of the RPO41 gene (Greenleaf, A. L., Kelly, J. L., and Lehman, I. R. (1986) Proc. Natl. Acad. Sci. U. S. A. 83, 3391-3399). The two enzyme activities are differentially regulated by carbon source; the nonselective enzyme is repressed during growth on glucose relative to the selective enzyme. The 5-fold increase in RNA polymerase activity on a nonrepressing carbon source correlates with the increased level of transcript production from mitochondrial DNA. These results suggest that the mitochondrial RNA polymerase and, in consequence, mitochondrial transcription are regulated by carbon catabolite control.

Transcription of multiple copies of the yeast GAL7 gene is limited by specific factors in addition to GAL4

High levels of the GAL7 gene in the yeast cell appear to titrate regulatory factors and to impair transcription of related sequences. To investigate the role that the GAL regulatory factors GAL4 and GAL80 have in this process we have compared the accumulation of mRNA transcribed from single-copy (plasmid-borne GAL7 and chromosomal GAL10) and high-copy (plasmid-borne GAL7) genes in several GAL regulatory mutants. Our results show that functional GAL4 gene product is required for induction of transcription from the single- and high-copy genes. In a strain containing the GAL4 gene fused to the high expression ADH1 promoter, glucose can replace galactose to induce high levels of transcription of GAL7 and GAL10 genes, although the kinetics of accumulation induced by the two sugars are distinctly different. In the presence of high levels of GAL4, maximum accumulation of mRNA from single and high copy genes is elevated two-fold; disruption of the gal80 gene in combination with high levels of GAL4 results in a further two-fold increase in transcription. In this genetic background, galactose-induced transcription of the high copy GAL7 gene results in a greater than 50-fold increase in the levels of GAL7 mRNA, representing 30%-50% of the total cellular mRNA. Our results are consistent with a cooperative effect of saturation of multiple GAL4 DNA binding sites and with a limiting factor, in addition to GAL4, that is required for transcription of the GAL genes.

mRNA transcription in nuclei isolated from Saccharomyces cerevisiae

We developed an improved method for the isolation of transcriptionally active nuclei from Saccharomyces cerevisiae, which allows analysis of specific transcripts. When incubated with alpha-32P-labeled ribonucleoside triphosphates in vitro, nuclei isolated from haploid or diploid cells transcribed rRNA, tRNA, and mRNAs in a strand-specific manner, as shown by slot blot hybridization of the in vitro synthesized RNA to cloned genes encoding 5.8S, 18S and 28S rRNAs, tRNATyr, and GAL7, URA3, TY1 and HIS3 mRNAs. A yeast strain containing a high-copy-number plasmid which overproduced GAL7 mRNA was initially used to facilitate detection of a discrete message. We optimized conditions for the transcription of genes expressed by each of the three yeast nuclear RNA polymerases. Under optimal conditions, labeled transcripts could be detected from single-copy genes normally expressed at low levels in the cells (HIS3 and URA3). We determined that the alpha-amanitin sensitivity of transcript synthesis in the isolated nuclei paralleled the sensitivity of the corresponding purified RNA polymerases; in particular, mRNA synthesis was 50% sensitive to 1 microgram of alpha-amanitin per ml, establishing transcription of mRNA by RNA polymerase II.

A transcription map of a yeast centromere plasmid: unexpected transcripts and altered gene expression

YCp19 is a yeast centromere plasmid capable of autonomous replication in both yeast and E. coli (J. Mol. Biol., 158: 157-179, 1982). It is stably maintained as a single copy in the yeast cell and is therefore a model yeast "minichromosome" and cloning vector. We have located the positions and measured the abundance of the in vivo yeast transcripts from YCp19. Transcripts from the selectable marker genes TRP1 and URA3 were present at increased levels relative to chromosomal copies of the genes. Unanticipated transcripts from the yeast CEN4 and E. coli pBR322 sequences were also found. Although much of the plasmid vector is actively transcribed in vivo, the regions around the most useful cloning sites (BamHI, EcoRI, SalI) are free of transcripts. We have analyzed transcription of BamHI inserts containing promoter variants of the HIS3 gene and determined that although initiation events are accurate, plasmid context may alter levels of gene expression.

A multicomponent mitochondrial RNA polymerase from Saccharomyces cerevisiae

Using a whole cell extract from Saccharomyces cerevisiae (bakers' yeast) we have been able to detect a selective RNA polymerase activity originally identified in purified yeast mitochondria (Levens, D., Morimoto, R., and Rabinowitz, M. (1981) J. Biol. Chem. 256, 1466-1473). We have shown that in in vitro transcription reactions this activity recognizes a consensus mitochondrial promoter sequence ATA-TAAGTA (Osinga, K. A., DeHaan, M., Christianson, T., and Tabak, H. F. (1982) Nucleic Acids Res. 10, 7993-8006) in the upstream region of the nuclear GAL10 gene as well as promoters from yeast mitochondrial DNA. Using these promoter-containing templates for in vitro assays, we have chromatographically separated the mitochondrial specific RNA polymerase activity from the three nuclear RNA polymerases (I, II, and III). Further characterization has revealed that this preparation has distinctive properties on two different types of DNA templates, poly[d(AT)] and cloned DNA containing mitochondrial promoters. Salt and divalent cation optima and substrate saturation kinetics are different for the two types of templates. Using promoter-containing DNA as an assay template, we have chromatographically dissociated the RNA polymerase activity into two nonfunctional components. Selective transcription of the GAL10 template is restored when the two components are recombined. It is possible that the RNA polymerase active on poly[d(AT)] is a nonspecific component of the selective transcription apparatus or that two distinct RNA polymerases are present in the preparation.

Expression of the Saccharomyces cerevisiae GAL7 gene on autonomous plasmids

We used a combination of cloned DNA fragments encoding the GAL7 gene, yeast plasmid vectors, and chromosomal gal7 deletions to characterize the in vivo transcription of the GAL7 gene on autonomously replicating plasmids. Our results demonstrated that a plasmid-borne 3.1-kilobase DNA fragment containing the GAL7 gene provides sufficient information to mimic the regulated expression of the chromosomal location. Normal expression of GAL7 could occur in the absence of DNA encoding the functional genes of the GAL cluster region and was not altered when the gene was adjacent to other plasmid elements such as autonomously replicating sequences or centromeres. The chromosomal and single-copy centromeric plasmid locations of GAL7 were indistinguishable in their response to growth conditions (induction by galactose, repression by glucose) and positive and negative regulatory factors (GAL4 and GAL80). Increasing the gene dosage to more than 200 copies per cell resulted in constitutive expression of the GAL7 mRNA; fully induced mRNA levels were increased more than 10-fold at these high gene dosages. When cells were shifted from noninducing to inducing conditions, the initial time of appearance and the rate of accumulation of GAL7 mRNA were altered in cell populations containing multiple GAL7 genes. The induction kinetics and final accumulation of the chromosomal GAL10 mRNA were also affected by the presence of multiple copies of the GAL7 gene; these results are consistent with a model involving limiting amounts of regulatory factors.

Altered promoter selection by a novel form of Bacillus subtilis RNA polymerase

Bacillus subtilis RNA polymerase holoenzyme prepared by several standard methods utilizes bacteriophage T7 DeltaD111 DNA as an efficient template. The major RNA products are specific transcripts from T7 promoters A(1) and C; these promoters are also efficiently utilized by RNA polymerases purified from a wide range of other bacterial species [Wiggs, J., Bush, J. & Chamberlin, M. (1979) Cell 16, 97-109]. In contrast, B. subtilis RNA polymerase preparations purified by a modification of the method of Burgess and Jendrisak (designated fraction 5) utilize T7 DeltaD111 promoters A(1) and C and an additional promoter site, J, which has been located at 90.6% on the standard T7 physical map. This promoter is not used by B. subtilis core RNA polymerase or by RNA polymerase from any other bacterial species we have tested. Sodium dodecyl sulfate/polyacrylamide gel electrophoresis of fraction 5 RNA polymerase shows that it contains B. subtilis components sigma and delta and a polypeptide of M(r) 92,000 in addition to the B. subtilis beta, beta', and alpha subunits. Chromatography of fraction 5 on single-stranded DNA-cellulose gives an enzyme fraction, Bs I, that is indistinguishable from B. subtilis RNA polymerase holoenzyme both in its peptide composition (betabeta'alpha(2)sigma) and in the selective transcription of only T7 RNAs A(1) and C. Chromatography of fraction 5 on phosphocellulose yields an enzyme fraction, Bs II, devoid of sigma subunit but containing the M(r) 92,000 peptide and traces of delta. This fraction synthesizes predominantly T7 J RNA in vitro together with traces of T7 A(1) and C RNAs. Hence, B. subtilis RNA polymerase fraction Bs II appears to contain a form of RNA polymerase that can transcribe selectively without detectable amounts of B. subtilis sigma subunit and that utilizes a promoter site not used by other known bacterial RNA polymerases. The structural basis for this specificity is not yet known.

Purification and subunit structure of deoxyribonucleic acid-dependent ribonucleic acid polymerase III from the posterior silk gland of Bombyx mori

DNA-dependent RNA polymerase III was purified from the posterior silk gland of the moth Bombyx mori by chromatography on DEAE-cellulose, DEAE-Sephadex, CM-Sephadex, and phosphocellulose and by sedimentation in sucrose density gradients. The specific activity of this chromatographically homogeneous enzyme was comparable to that reported for other purified eukaryotic RNA polymerases. Sucrose gradient sedimentation analysis suggested a molecular weight of approximately 590,000 to 660,000 for B. mori RNA polymerase III. Analysis of subunit composition by polyacrylamide gel electrophoresis under denaturing conditions showed that the chromatographically purified RNA polymerase III contained subunits with molecular weights of 155,000 (IIIa), 136,000 (IIIb), 67,000 (IIIc), 62,000 (IIId), 49,000 (IIIe), 39,000 (IIIf), 36,000 (IIIg), 31,000 (IIIh), 28,000 (IIIi), and 18,000 (IIIj). Molar ratios were close to unity for all subunits except for IIIj, which was present in an approximate molar ratio of 2. As has been observed for mammalian class III enzymes, the B. mri RNA polymerase III can be resolved into two components upon electrophoresis under nondenaturing conditions. Comparative studies of the class III enzymes from B. mori and from higher eukaryotic cells show that many of the general chromatographic and catalytic properties, as well as the overall subunit compositions, are similar for the various enzymes. However, unlike the mammalian class III enzymes, B. mori RNA polymerase III is completely resistant to high concentrations of alpha-amanitin, and it does not contain an 89,000-dalton subunit. The data are discussed in terms of the function and regulation of RNA polymerase III in lower and higher eukaryotes.

Viral RNA synthesis and levels of DNA-dependent RNA polymerases during replication of adenovirus 2

The rates of RNA synthesis in cultured human KB cells infected by adenovirus 2 were estimated by measuring the endogenous RNA polymerase activities in isolated nuclei. The fungal toxin alpha-amanitin was used to determine the relative and absolute levels of RNA polymerases I, II, and III in nuclei isolated during the course of infection. Whereas the level of endogenous RNA polymerase I activity in nuclei from infected cells remained constant relative to the level in nuclei from mock-infected cells, the endogenous RNA polymerase II and III activities each increased about 10-fold. These increases in endogenous RNA polymerase activities were accompanied by concomitant increases in the rates of synthesis in isolated nuclei of viral mRNA precursor, which was quantitated by electrophoretic analysis on polyacrylamide gels. The cellular RNA polymerase levels were measured with exogenous templates after solubilization and chromatographic resolution of the enzymes on DEAE-Sephadex, using procedures in which no losses of activity were apparent. In contrast to the endogenous RNA polymerase activities in isolated nuclei, the cellular levels of the solubilized class I, II, and III RNA polymerases remained constant throughout the course of the infection. Furthermore, no differences were detected in the chromatographic properties of the RNA polymerases obtained from infected or control mock-infected cells. These observations suggest that the increases in endogenous RNA polymerase activities in isolated nuclei are not due to variations in the cellular concentrations of the enzymes. Instead, it is likely that the increased endogenous enzyme activities result from either the large amounts of viral DNA template available as a consequence of viral replication of from replication or from functional modifications of the RNA polymerases or from a combination of these effects.

DNA-dependent RNA polymerase levels during the response of human peripheral lymphocytes to phytohemagglutinin

The cellular levels of the various RNA polymerases have been monitored in resting human peripheral lymphocytes and in lymphocytes stimulated by phytohemagglutin. Activity was measured in the presence of exogenous templates following solubilization and chromatographic resolution of the different RNA polymerases. Resting lymphocytes contain Class I, II, and III RNA polymerases, although the respective levels of activity are very low compared to the levels in metabolically active cell types. During the PHA-induced transformation of resting lymphocytes, the Class I, II, and III enzyme levels rise dramatically. During four days exposure to PHA, the levels of RNA polymerases I and III (which synthesize, respectively, rRNA and the transfer and 5S RNAs) increase 17 fold, while the level of RNA polymerase II (which synthezies heterogeneous nuclear RNA) increase 8 fold. The possible relationship between enzyme levels and the regulation of gene expression is discussed.