Coupling Termination of Transcription to Messenger RNA Maturation in Yeast

See allHide authors and affiliations

Science  10 Apr 1998:
Vol. 280, Issue 5361, pp. 298-301
DOI: 10.1126/science.280.5361.298


The direct association between messenger RNA (mRNA) 3′-end processing and the termination of transcription was established for theCYC1 gene of Saccharomyces cerevisiae. The mutation of factors involved in the initial cleavage of the primary transcript at the poly(A) site (RNA14, RNA15, andPCF11) disrupted transcription termination at the 3′ end of the CYC1 gene. In contrast, the mutation of factors involved in the subsequent polyadenylation step (PAP1,FIP1, and YTH1) had little effect. Thus, cleavage factors link transcription termination of RNA polymerase II with pre-mRNA 3′-end processing.

Polyadenylation signals at the 3′ end of pre-mRNA are required for the termination of transcription in higher eukaryotes (1), budding yeast (2), and fission yeast (3). This ensures that transcription will only terminate after RNA polymerase II (pol II) has read beyond the end of the mRNA sequence. The mechanism of pre-mRNA 3′-end formation (sequential endonucleolytic cleavage and polyadenylation) and many of the factors involved in catalyzing these reactions are very similar between higher eukaryotes and S. cerevisiae (4). Reconstitution of this reaction in vitro in S. cerevisiae has allowed the fractionation of factors [cleavage factor IA (CF IA), CF IB, CF II, polyadenylation factor I (PF I), and poly(A) polymerase (PAP)] required for each step in the 3′-end formation reaction (5). The use of temperature-sensitive (ts) mutants has facilitated the determination of the molecular composition of these factors and revealed interactions between them (6). Here, we used yeast strains carrying these ts mutant alleles to demonstrate that some of these factors are also involved in pol II transcription termination.

We used transcription run-on analysis to measure transcription termination at the 3′ end of the S. cerevisiae CYC1 gene (Fig. 1), for which signals that direct 3′-end formation have been well characterized (7). To achieve high transcription, we transformed yeast cells (8) with the multicopy plasmid pGCYC, in which theCYC1 promoter has been replaced by the GAL1/10promoter region (9). Reverse transcription polymerase chain reaction (RT-PCR) analysis confirmed that transcripts initiating at theGAL promoter are polyadenylated at the same sites as found for the intact CYC1 gene (10, 11). The distribution of run-on transcript over the contiguous single-stranded probes 1 to 6 showed that transcription stops efficiently, soon after the CYC1 poly(A) site (located at the 3′ end of probe 2), with only small amounts of run-on transcript detected beyond probe 3 (Fig. 1, B and D). This finding is in agreement with previous in vivo data showing that signals 100 base pairs (bp) beyond the CYC1 poly(A) site are required to direct the termination of transcription (2). The background signal detected in the upstream GAL probe indicates that transcription begins at the GAL promoter. A similar distribution of polymerases was observed with the genomic copy of theCYC1 gene, although the signal was 5% of that detected with the pGCYC plasmid. No signals were detected using the CYC1deletion strain (B-7467) (12). In all strains tested, polymerase density showed a strong correlation with the amount of stable CYC1 pre-mRNA (Fig. 1C). Run-on analysis was also performed in the rpb1-1 mutant strain that carries a ts mutation in the large subunit of RNA pol II (13). Inactivation of pol II activity, achieved by culturing cells at the restrictive temperature (37°C) for 60 min, abolished signals overCYC1 (1 to 6) and ACT1 probes that detect pol II transcription, whereas the amount of pol III–generated transcript, detected over the tRNA probe, remained unchanged (Fig. 1D).

Figure 1

Pol II transcription terminates close to the CYC1 poly(A) site. (A) Schematic representation of plasmid pGCYC (9) showing the arrangement of M13 probes (20) used in run-on analysis relative to the CYC1 poly(A) site (position 502). Locations of CYC1 probes 1 to 6 and the upstream GALprobe (G) are shown relative to the CYC1 ORF start site (position 1). (B) Run-on analysis across the 3′ end of theCYC1 gene in strains indicated. Hybridization of transcripts to actin (A) and M13 control (M) probes are shown. (C) Northern blot analysis showing amount ofCYC1 mRNA in strains shown. The CYC1 probe was generated by random-primer labeling of a 382-bp fragment spanning M13 probes 1 and 2. The filter was stripped and reprobed forACT1 mRNA as a loading control. The ACT1 probe was generated by random-primer labeling of a 567-bp fragment spanning the M13 actin probe. (D) Run-on analysis across the 3′ end of the CYC1 gene performed in the pol II temperature-sensitive strain (rpb1-1) under permissive (25°C) or restrictive (37°C) conditions. Hybridization of pol III transcripts to the tRNA probe (T) is shown. Transcription run-on and Northern blot analysis were performed as described (3). The conditions used in the run-on assay (0.5% Sarkosyl) inhibit transcription initiation but allow elongation over a short distance. Incorporation of [α-32P]uridine triphosphate into nascent transcripts reflects the relative density of active RNA-polymerase complexes. After partial hydrolysis, hybridization of these end-labeled primary transcripts to single-stranded probes reveals where polymerase density falls to zero, indicating that transcription has stopped.

Next, we mutated the CYC1 polyadenylation signal to test for an effect on transcription termination. The 38-bp deletion mutation of pGcyc1-512, which prevents CYC1 polyadenylation (14), also affected the termination of pol II transcription, confirming previous in vivo data (2). Steady-state RNA analysis (Fig. 2A) showed that the amount of stable CYC1 transcript decreased in the strain carrying pGcyc1-512, but this analysis did not reveal the location or efficiency of transcription termination. In contrast, run-on analysis showed a marked shift in the distribution of polymerase density when the polyadenylation signal was mutated, with an increase in active polymerase complexes located well beyond the normal site of transcription termination (Fig. 2B).

Figure 2

Pre-mRNA 3′-end formation signal directs transcription termination. (A) Northern blot analysis showing CYC1 mRNA in strains carrying either pGCYC1 or pGcyc1-512. The filter was stripped and reprobed for ACT1mRNA as a loading control. (B) Run-on analysis across 3′ end of the CYC1 gene in strains indicated.

Because the efficiency of mammalian pre-mRNA 3′-end formation has been correlated with the degree of transcription termination, 3′-end formation factors have been implicated in the coupling of pol II termination with RNA processing (15). We therefore analyzed yeast strains carrying ts mutations in the 3′-end RNA processing factors (Fig. 3). Consistent with the results of previous studies (6, 16), switching growth from permissive (25°C) to restrictive conditions (37°C) severely affected the extent of pre-mRNA 3′-end formation observed in all of the mutant strains, with the amount of stable CYC1 transcript decreasing substantially within 25 min (Fig. 3A). However, only a subset of these mutations—rna14, rna15, andpcf11 (protein 1 of CF I)—also affected transcription termination (Fig. 3B), with considerable run-on transcript detected over probes 4, 5, and 6. With the pap1 [poly(A) polymerase], fip1 [factor interacting with poly(A) polymerase], and yth1 (yeast 30-kD homolog) mutant strains, we observed little difference in polymerase density after culture in the permissive and restrictive conditions. Slightly slower decay profiles were observed with the fip1-1 and yth1-1alleles (Fig. 3A); however, even after extended incubation at 37°C, no difference was observed in the polymerase profiles (11). The major poly(A) binding protein, Pab1p, is associated with CF IA and functions in controlling the length of the poly(A) (17). However, extended incubations of the ts mutant allelepab1-F364L (18) in nonpermissive conditions does not influence the polymerase profile at the 3′ end of theCYC1 gene (11).

Figure 3

Pre-mRNA 3′-end cleavage factors direct transcription termination. (A) Northern blot analysis showing amount of CYC1 mRNA under permissive conditions (25°C), defined as 0, or after a shift to restrictive conditions (37°C) for 5, 25, or 60 min in strains indicated. (B) Run-on analysis performed under permissive conditions and after a shift to restrictive conditions for 25 min in strains indicated (8). (C) Schematic representation of proteins involved in pre-mRNA 3′-end processing in yeast. Only the proteins corresponding to the mutants tested in this study are designated.

RNA14, RNA15, and PCF11 are components of CF IA (Fig. 3C), a factor required for both the cleavage of the pre-mRNA and subsequent polyadenylation (6). In contrast,YTH1 and FIP1, components of PF I andPAP1, are involved in the addition of the poly(A) (Fig. 3C) (6). This analysis therefore demonstrates that only factors involved in the pre-mRNA cleavage reaction play a role in directing termination of pol II transcription. Factors involved in polyadenylation do not appear to be required for termination.

The mammalian cleavage and polyadenylation specificity factor (CPSF) and cleavage stimulation factor (CstF) are associated with the COOH-terminal domain of the largest pol II subunit (19). This raised the possibility that the termination of transcription might be signaled through the release of these 3′-end RNA processing factors from the COOH-terminal domain to the poly(A) signals on the nascent RNA. Therefore, the ts mutations in the cleavage factors might prevent termination through loss of poly(A) site binding. TheRNA15 gene product, Rna15p, is the only subunit of CF 1A that cross-links to RNA, binding to uridine-rich sequences like those found in yeast poly(A) signals (6). It is also homologous to mammalian cleavage factor CstF 64, which is known to interact with the pre-mRNA poly(A) signal (4). Therefore, the binding properties of this protein were compared with the ts mutant Rna15 protein. In vitro translated wild-type and ts mutant proteins both bound tightly to poly(U)-agarose (Fig.4A, lanes 1 and 2), but not to poly(A), poly(C), or poly(G) (11). After incubation at 45°C for 10 min, the mutant Rna15-1p did not exhibit altered affinity for poly(U) compared with that of the wild-type protein (Fig. 4A, lanes 4 and 5). Furthermore, the COOH-terminal domain of Rna15p carrying therna15-1 mutation that replaces a leucine with a proline (CTA to CCA) is dispensable for its binding capacity (Fig. 4A, lanes 3 and 6). We finally tested whether the rna15-1mutation affected the stability of the protein after a shift to the nonpermissive temperature. Even after incubation for 60 min at 37°C, no loss in RNA15 gene product was detectable (Fig. 4B).

Figure 4

Poly(U) binding of the wild-type and mutant Rna15 proteins. (A) The Rna15-1p produced by in vitro transcription-translation of the ts allele rna15-1was assayed for its ability to bind poly(U) (21) before or after incubation at 45°C for 10 min. Lanes 1 and 4, wild-type Rna15p; lanes 2 and 5, mutant Rna15-1 protein; lanes 3 and 6, COOH-terminal truncated version of the Rna15 protein corresponding to the RNA binding domain only (154 amino acids). (B) Protein immunoblot analysis showing the Rna15 mutant protein extracted from LM91 cells shifted to the nonpermissive temperature (37°C) for the times indicated. The blot was probed with antibody to Rna15p at a 1:10,000 dilution.

The fact that the rna15-1 mutation does not affect RNA binding or stability of the protein argues against models directly connecting termination with the transfer of poly(A) factors from the pol II COOH-terminal domain to nascent RNA. Instead, a subset of factors involved in catalyzing the cleavage of the primary transcript must be functionally active to direct the termination of transcription. Endonucleolytic cleavage of the nascent transcript may therefore be required for efficient pol II termination in yeast.

  • * Present address: Department of Biological Chemistry, 240 D Med Sci I, College of Medicine, University of California, Irvine, CA 92697, USA.

  • To whom correspondence should be addressed. E-mail: nicholas.proudfoot{at}


View Abstract

Stay Connected to Science

Navigate This Article