Mechanism and regulation of transcriptional elongation by RNA polymerase II

Abstract

Over the past few years, biochemical and genetic studies have shed considerable light on the structure and function of the RNA polymerase II (pol II) elongation complex and the transcription factors that control it. Novel elongation factors have been identified and their mechanisms of action characterized in increasing detail; new insights into the biological roles of elongation factors have been gained from genetic studies of the regulation of mRNA synthesis in yeast; and intriguing links between the pol II elongation machinery and the pathways of DNA repair and recombination have emerged.

Introduction

Eukaryotic mRNA synthesis is catalyzed by the multisubunit RNA polymerase II (pol II) and proceeds through multiple stages designated pre-initiation, initiation, and elongation. Although substantial evidence argues that expression of most genes is controlled, at least in part, at the level of transcription elongation [1,2,3^•,4,5], efforts to elucidate the mechanisms by which specific elongation factors regulate elongation by pol II in cells have been hampered by lack of suitable assays. Many transcriptional regulatory proteins, such as sequence-specific DNA binding transactivators and the general initiation factors, function through interactions with DNA sequences in the promoter-regulatory regions of genes. In contrast, most of the known elongation factors appear to expedite transcription either by promoting efficient elongation by pol II through chromatin (for example, FACT [facilitates chromatin transcription] and SWI/SNF family members [6^•,7^•]) or by targeting pol II directly and suppressing transient pausing (for example, TFIIF, Elongin, ELL, ELL2, CSB [Cockayne syndrome B] [8^•], and Tat stimulatory factor 1 [Tat-SF-1] [9^•]) or arrest (for example, SII, P-TEFb [positive transcription elongation factor b]) [4,5,6^•–9^•]. Because specific DNA sequences that mediate the activity of the known elongation factors have not been identified, development of DNA-sequence-based reporter assays for elongation factor activity in cells has not been possible. Thus, little information about how the known elongation factors contribute to regulation of transcription of specific genes is available. Nevertheless, ongoing biochemical studies are leading to the discovery of a growing list of potential elongation factors and genetic studies are beginning to provide substantive insights into the roles of these proteins in control of gene expression in cells. In addition, recent studies suggest that the pol II elongation complex interacts functionally with proteins involved in DNA repair and recombination and serves as a recruitment platform for factors involved in the capping, splicing, and polyadenylation of nascent transcripts (see reviews by Bentley, pp 347–351 and Minivielle-Sebastia and Keller pp 352–357, this volume).

This short article is a commentary on recent developments in biochemical and genetic studies on the mechanism and regulation of transcriptional elongation by pol II. Pol II elongation factors and mechanisms have recently been reviewed in detail [4,5]. Here, we do not attempt a comprehensive review of the literature but, instead, describe recent findings that are shaping our current view of the elongation stage of eukaryotic mRNA synthesis.

Emerging evidence for functions of specific elongation factors in cells

Over the past few years, studies on the roles of elongation factors in cells have taken advantage of the experimentally tractable simple eukaryote Saccharomyces cerevisiae and the HIV-1 viral model of Tat-dependent elongation. Tat is a sequence-specific RNA-binding protein encoded by HIV. Tat binds to the transactivation response (TAR) element at the 5′-end of HIV transcripts and functions together with cellular proteins to promote efficient elongation of HIV transcripts by pol II. These studies have brought to light genetic interactions among genes encoding subunits of pol II and elongation factors SII, elongator, and DRB sensitivity inducing factor (DSIF) (Spt4–Spt5) and have provided new insights into the roles of these and other elongation factors in gene expression (Figure 1).

Figure 1.

Genetic interactions between pol II and elongation factors in yeast. Genetic interaction (arrows) between pol II subunits (rectangles), elongation factors (square, triangle) and chromatin-related transcription factors (circle) that have been recently shown in S. cerevisiae are depicted. A factor involved in elongation and DNA recombination (HPR1, oval) has also been shown to interact with RNA pol II in yeast.

SII

SII was initially identified biochemically by its ability to promote synthesis of long transcripts by purified pol II [10]. SII enables pol II to transcribe through a variety of transcriptional impediments, including intrinsic arrest sites in DNA and nucleoprotein complexes. Mechanistic studies have shown that SII accomplishes this by interacting with arrested pol II and activating cleavage of the nascent transcript by a latent endoribonuclease intrinsic to the polymerase [11]. Following cleavage, pol II can re-extend the nascent transcript, allowing repeated attempts at elongation through arrest sites and, eventually, clearance of the impediment. In light of evidence that pol II arrests when the 3′-OH of the nascent transcript loses contact with the polymerase catalytic site, it has been proposed that SII-induced nascent transcript cleavage reactivates arrested pol II by realigning polymerase catalytic site residues with the 3′-OH of the nascent transcript [11].

The S. cerevisiae gene encoding SII (designated DST1 or PPR2) is not essential for viability; however, growth of yeast strains lacking SII is weakly sensitive to drugs such as 6-azauracil (6-AU) and mycophenolic acid, which are known to reduce intracellular NTP pools [12,13] and thus are expected to decrease overall rates of elongation by pol II in cells and to increase the likelihood that pol II will suffer arrest. Genetic evidence for a functional interaction between pol II and SII has come from several observations. Overexpression of SII rescues the drug sensitivity of a yeast strain harboring a pol II mutant that exhibits a decreased binding affinity for SII in vitro [12,14]. In addition, Archambault and coworkers [15] identified an allele of RPB1, encoding the largest pol II subunit, that confers synthetic lethality in yeast lacking SII. Finally, Lennon and coworkers [16^•] recently found that combination of two mutations (deletion of DST1 and the pol II mutant rpb2-10), each of which confers weak 6-AU and mycophenolic acid sensitivity upon yeast, leads to a dramatic decrease in growth rate in the presence of drug. The rpb2-10 allele contains a point mutation that changes a conserved proline residue near the pol II catalytic pocket to serine. Pol II with the rpb2-10 mutation elongates transcripts more slowly in vitro and is more prone to arrest than wild-type pol II [17], suggesting that the phenotype of the double mutant may be due to an inability of the mutant pol II enzyme to overcome arrest induced by nucleosomes, by other DNA-bound proteins, or by intrinsic arrest sites in the DNA.

Evidence for a broad role for SII in regulation of transcription in vivo is provided by the observation that the levels of total poly A+ RNA — as well as of a number of specific transcripts — have been shown to decrease dramatically following treatment of the rpb2-10, dst1 deletion strain with 6-AU [16^•]. It is not yet clear whether SII directly regulates expression of many genes or whether its primary targets are a few important transcriptional regulators. It is noteworthy, however, that the kinetics of reduction of steady state mRNA levels upon addition of drug to the rpb2-10, dst1 strain are very similar to those observed when yeast bearing temperature-sensitive alleles of genes encoding basal transcription factors or subunits of the pol II holoenzyme are shifted to the non-permissive temperature. The development of SII-deficient strains with severe phenotypes and the identification of potential target genes should provide a foundation for future studies on the role of SII in gene regulation in cells.

Elongator

Elongator is a three-subunit complex composed of polypeptides of ~150, ~90, and ~60 kDa. Elongator was purified from yeast chromatin preparations containing the elongating form of pol II. Evidence suggests that the interaction of elongator with pol II requires phosphorylation of the pol II carboxy-terminal domain (CTD) [18^•,19^•]. Like the gene encoding SII, the ELP1 gene, encoding the largest subunit of elongator, is not essential for yeast viability. Yeast lacking ELP1 are weakly sensitive to 6-AU but yeast lacking both DST1 and ELP1 are hypersensitive to the drug, suggesting a functional interaction between the two proteins. A variety of genes, including GAL1-10, PHO5, and INO2 — but not HIS3 and CHA1 — exhibit delayed activation in cells lacking the ELP1 gene [19^•]. On the basis of these observations, together with previous work from several laboratories demonstrating that DNA binding transactivators can control the efficiency of elongation [1,2,3^•], it has been proposed that Elp1 might function as a co-activator in post-initiation events [19^•].

DSIF, P-TEFb and regulation of elongation by HIV Tat

The nucleotide analog DRB potently inhibits synthesis of cellular heterogeneous nuclear (hn)RNA in vivo and of long transcripts in crude and partially fractionated transcription systems [20,21]. In addition, expression of the Tat-activated HIV-1 polyprotein gene is exquisitely sensitive to DRB [22,23]. DRB inhibits RNA synthesis in these systems by inducing elongating pol II to arrest shortly after initiating transcription. Recent studies on the mechanism of DRB-sensitive elongation have led to the identification of a negatively-acting DRB-sensitivity inducing factor, DSIF [24^•], and a positively-acting factor, P-TEFb, which counteracts DSIF activity.

Mammalian DSIF is composed of ~14 and ~160 kDa polypeptides, which are homologs of the S. cerevisiae Spt4 and Spt5 proteins, respectively [24^•]. Early genetic experiments implicated Spt4 and Spt5, as well as a genetically related gene product, Spt6, in transcription and modification of chromatin in yeast [25]. Consistent with the idea that DSIF (Spt4–Spt5) functions as a negative elongation factor in cells, some conditional spt5 mutations are suppressed when combined with the elongation-defective pol II rpb2-10 mutant or when cells are grown in 6-AU [26^•]. In addition, mutant alleles of SPT4, SPT5 and SPT6 confer 6-AU sensitivity on yeast, and combinations of mutations in SPT5 and RPB1, SPT5 and RPB2, SPT4 and DST1, SPT5 and DST1, and SPT6 and DST1 have been shown to produce synthetic phenotypes in yeast [26^•] (Figure 1). Together with results implicating Spt4, Spt5, and Spt6 in chromatin modification, these results raise the possibility that DSIF and SII may help to enable pol II to transcribe through nucleosomes [26^•].

The inhibitory effect of DSIF on productive elongation by pol II in vitro is antagonized by pTEFb [27^•], a DRB-sensitive pol II CTD kinase composed of cdk9 (also known as PITALRE) and cyclin T [28^•,29–32]. Phosphorylation of pol II CTD by P-TEFb appears to block functional interactions between DSIF and pol II, on the basis of findings indicating that a mutant P-TEFb lacking kinase activity is transcriptionally inactive, that the pol II CTD is essential for P-TEFb activity [27^•,31,32], and that both DSIF transcription inhibitory activity and the binding of DSIF to pol II are blocked upon pol II CTD phosphorylation by P-TEFb [27^•].

A possible explanation for the extreme DRB sensitivity of Tat-dependent elongation of the HIV-1 polyprotein transcript has come from studies implicating the pol II CTD kinase P-TEFb in this process. Tat-dependent transcription depends on the presence of an intact CTD [33–35], and immunodepletion experiments demonstrate that P-TEFb is required for Tat-dependent transcription in vitro [28^•,36]. In addition, a high throughput screen for drugs that block Tat-dependent HIV-1 transcription led to the identification of compounds that specifically inhibit P-TEFb kinase activity [37^•].

Genetic support for the idea that P-TEFb plays a critical role in Tat function comes from the observation that expression of human cyclin T1 rescues Tat function in rodent cells, which are normally refractory to HIV-1 infection and to Tat-dependent transcriptional activation [38^•]. Human but not murine cyclin T1 is capable of forming a ternary complex with Tat and the transactivation response (TAR) element [38^•,39^•], suggesting that Tat may subvert normal cellular elongation control mechanisms by interacting with P-TEFb and recruiting it to the TAR element, thereby specifically activating elongation of the HIV-1 transcript.

Links between elongation and DNA repair and recombination

Substantial evidence links the process of transcription elongation to a variety of DNA transactions, including transcription-coupled DNA repair, somatic hypermutation of immunoglobulin genes, homologous DNA recombination, and maintenance of genome stability [40–44,45^•]. Although the mechanisms linking these processes are poorly understood, proteins with roles in both transcription elongation and either DNA repair or recombination have been identified.

The gene encoding the mammalian CSB protein was initially identified by its requirement in transcription-coupled nucleotide excision repair of damaged DNA [46]. Recently, the CSB protein was found to function in vitro as an elongation factor that interacts directly with transcribing pol II [47] and stimulates the overall elongation rate [6^•].

Mutations in the S. cerevisiae HPR1 and THO2 genes induce a hyper-recombination phenotype [48,49,50^•]. Evidence that the Hpr1 and Tho2 proteins have roles in transcription by pol II has come from findings indicating that the hyper-recombination phenotype of HPR1 and THO2 mutant yeast strains can be suppressed by mutations in genes encoding pol II subunits, the mediator component SRB2, and initiation factor TFIIB [51] and further that the Hpr1 protein can be purified in association with a pol II-containing complex [52]. Evidence that the hyper-recombination phenotype is due to a defect in transcription elongation has come from several findings. These indicate, first, that the hyper-recombination phenotype of HPR1 and THO2 mutant yeast strains can also be suppressed by insertion of transcription termination signals upstream of recombinogenic regions of transcribed DNA [53^•]. Second, that on the basis of the results of nuclear run-on experiments transcription elongation by pol II is impaired in yeast strains lacking HPR1 or THO2 [50^•,53^•], and third, that the hyper-recombination phenotype of HPR1 and THO2 mutant yeast strains is enhanced when cells are grown in 6-AU [45^•]. On the basis of these observations, it has been proposed that the hyper-recombination phenotype of HPR1 or THO2 mutant strains results from recruitment of the recombination machinery to the vicinity of inappropriately paused elongation complexes [50^•]. Whether paused pol II elongation complexes play a similar role in normal recombination events remains unclear; however, it is intriguing that somatic hypermutation of an artificial immunoglobulin gene is enhanced upon insertion of a DNA sequence predicted to induce transcriptional pausing [41].

Conclusions and prospects for the future

Over the past few years, our understanding of the mechanism and regulation of elongation by pol II has improved substantially. During this time, novel elongation factors have been identified, their mechanisms of action in vitro characterized in increasing detail and, in some cases, their cellular roles defined. In the future, we expect that investigations of transcription elongation will focus more and more on understanding the mechanisms by which pol II transcribes its natural template, chromatin. These studies are now well underway with the development of purified, reconstituted chromatin transcription systems and the discoveries of important roles for the SWI/SNF chromatin remodeling complex and for novel elongation factors such as FACT in transcription of nucleosomal templates by pol II in vitro [7^•,8^•,54–56]. It is likely that these new avenues of research will continue to gain momentum and will bring us closer to a mechanistic understanding of the elongation stage of eukaryotic mRNA synthesis.

Abbreviations

6-AU

6-azauracil

CSB

Cockayne syndrome B

CTD

carboxy-terminal domain

DRB

5,6-dichloro-1-β-D-ribofuranosylbenzimidazole

DSIF

DRB sensitivity inducing factor

ELL

11–19 lysine-rich in leukemia

FACT

facilitates chromatin transcription

pol II

RNA polymerase II

P-TEFb

positive transcription elongation factor b

TAR

transactivation response

Tat-SF1

Tat-stimulatory factor 1