Portal de Periódicos da CAPES

Nascent transcription from the nmt1 and nmt2 genes of Schizosaccharomyces pombe overlaps neighbouring genes

1998; Springer Nature; Volume: 17; Issue: 11 Linguagem: Inglês

10.1093/emboj/17.11.3066

1460-2075

Karen Tranberg Hansen, Charles E. Birse, Nicholas Proudfoot,

Plant biochemistry and biosynthesis

Article1 June 1998free access Nascent transcription from the nmt1 and nmt2 genes of Schizosaccharomyces pombe overlaps neighbouring genes Karen Hansen Karen Hansen Present address: Centre for Plant Biochemistry and Biotechnology, University of Leeds, Leeds, LS2 9JT GB Search for more papers by this author Charles E. Birse Charles E. Birse Present address: Department of Biological Chemistry, 240 D Med Sci I. College of Medicine, University of California, Irvine, CA, 92697 USA Search for more papers by this author Nick J. Proudfoot Nick J. Proudfoot Sir William Dunn School of Pathology, University of Oxford, Oxford, OX1 3RE GB Search for more papers by this author Karen Hansen Karen Hansen Present address: Centre for Plant Biochemistry and Biotechnology, University of Leeds, Leeds, LS2 9JT GB Search for more papers by this author Charles E. Birse Charles E. Birse Present address: Department of Biological Chemistry, 240 D Med Sci I. College of Medicine, University of California, Irvine, CA, 92697 USA Search for more papers by this author Nick J. Proudfoot Nick J. Proudfoot Sir William Dunn School of Pathology, University of Oxford, Oxford, OX1 3RE GB Search for more papers by this author Author Information Karen Hansen2, Charles E. Birse3 and Nick J. Proudfoot1 1Sir William Dunn School of Pathology, University of Oxford, Oxford, OX1 3RE GB 2Present address: Centre for Plant Biochemistry and Biotechnology, University of Leeds, Leeds, LS2 9JT GB 3Present address: Department of Biological Chemistry, 240 D Med Sci I. College of Medicine, University of California, Irvine, CA, 92697 USA The EMBO Journal (1998)17:3066-3077https://doi.org/10.1093/emboj/17.11.3066 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info We have determined the extent of the primary transcription unit for the two highly expressed genes nmt1 and nmt2 of Schizosaccharomyces pombe. Transcription run-on analysis in permeabilized yeast cells was employed to map polymerase density across the 3′-flanking region of these two genes. Surprisingly, polymerases were detected 4.3 kb beyond the nmt1 polyadenylation [poly(A)] site and 2.4 kb beyond the nmt2 poly(A) site, which in each case have transcribed through an entire convergent downstream transcription unit. However, the steady-state levels of both downstream genes were unaffected by the high level of nmt1 or nmt2 nascent transcription. Analysis of nmt1 and nmt2 RNA 3′ end formation signals indicates that efficient termination of transcription requires not only a poly(A) signal but also additional pause elements. The absence of such pause elements close to the poly(A) sites of these genes may account for their extended nascent transcripts. Introduction The 3′ end of most eukaryotic mRNAs is formed by coupled endonucleolytic cleavage and polyadenylation of the nascent transcript synthesized by RNA polymerase II (pol II). Transcription termination, defined as the dissociation of the ternary complex into its constituent parts, occurs further downstream and is required to release the polymerase for subsequent rounds of transcription, as well as to prevent transcriptional interference. Transcriptional interference arises when an elongating polymerase fails to terminate transcription and reads into a co-transcribed downstream transcription unit resulting in reduced expression of the downstream gene through disruption of interactions at the promoter (Cullen et al., 1984; Bateman and Paule, 1988; Henderson et al., 1989; Irniger et al., 1992). Interference can be relieved by cloning 3′ end formation signals between the two tandem promoters (Irniger et al., 1992; Eggermont and Proudfoot, 1993; Greger et al., 1998). For convergent transcription units, the potential problem of polymerases collision or hybridization between nascent transcripts that could target the transcripts for degradation may also arise. In yeast, transcriptional interference is believed to be of particular significance due to the compact nature of the yeast genome (Oliver et al., 1992). Furthermore, yeast autonomously replicating sequences (ARSs) and centromere (CEN) elements, which are both required for chromosome maintenance, as well as sites of meiotic gene conversion, are also sensitive to transcriptional interference (Hill and Bloom, 1987; Snyder et al., 1988; Rocco et al., 1992). Since transcription units have been identified in the Saccharomyces cerevisiae genome which direct transcription into ARSs or CEN elements without apparently affecting their function (Snyder et al., 1988; Hegemann and Fleig, 1993; Tanaka et al., 1994; Chen et al., 1996), efficient transcription termination signals must exist to prevent such interference. The signals directing cleavage/polyadenylation have been studied extensively in both mammals and the yeast S.cerevisiae, and are now well-characterized (Guo and Sherman, 1996; Wahle and Keller, 1996). In mammals, the almost invariant AAUAAA hexanucleotide is located 15–20 nucleotides upstream of the cleavage site, and a less well-defined GU or U-rich element is located downstream of the cleavage site (for review see Wahle and Keller, 1996). In S.cerevisiae, the polyadenylation [poly(A)] signals are more degenerate and comprise an upstream efficiency element (EE; consensus UAUAUA) which promotes 3′ end formation at downstream sites, and a positioning element (consensus AAAAAA or AAUAAA) which directs cleavage at preferred sites [characterized by Y(A)n] 15–30 nucleotides downstream (Irniger and Braus, 1994; Guo and Sherman, 1995, 1996). The development of in vitro processing extracts has also led to the identification and cloning of many of the trans-acting factors involved in cleavage/polyadenylation (Keller and Minvielle-Sebastia, 1997). In contrast, far less is known about the process of pol II transcription termination, partly due to the instability of the primary transcript and lower transcript levels compared with genes transcribed by RNA polymerases I and III. In higher eukaryotes, transcription run-on (TRO) assays have been employed to map sites of nascent transcription in isolated nuclei or whole cells (Citron et al., 1984; Hagenbüchle et al., 1984; Connelly and Manley, 1988; Maa et al., 1990). Generally, transcription termination occurs in broad, poorly defined regions up to 4 kb beyond the poly(A) signal. Two signals are believed to be necessary for pol II transcription termination: a functional poly(A) signal and a downstream element (DSE) at the site of termination (Logan et al., 1987). Mutation of the poly(A) signal results in continued transcription beyond the normal termination site (Whitelaw and Proudfoot, 1986; Logan et al., 1987; Connelly and Manley, 1988), and the in vitro processing efficiency of the poly(A) signal correlates with termination efficiency (Edwalds-Gilbert et al., 1993). The nature of the DSE appears variable, but a common feature may be the pausing of the polymerase by perturbing the normal mode of elongation. The ability to induce polymerase pausing may be intrinsic to the DNA sequence being transcribed (Kerppola and Kane, 1990; Enriquez-Harris et al., 1991) or be induced by a bound protein (Connelly and Manley, 1989a,b; Ashfield et al., 1991, 1994) which may bend the DNA (Ashfield et al., 1994). In S.cerevisiae, most studies have employed indirect assays to identify putative transcription termination signals. Such assays rely on monitoring transcription-induced changes in plasmid topology (Osborne and Guarente, 1988) or increased plasmid stability conferred by cloning sequences that prevent transcriptional interference between the highly expressed GAL1 promoter and a CEN element (Russo and Sherman, 1989; Russo, 1995). These studies indicate that sequences required for 3′ end formation of the CYC1 gene are also required for transcription termination. A more direct study using a TRO assay (Osborne and Guarente, 1989) showed that an 83 bp fragment from the 3′-flanking region of the CYC1 gene confers efficient transcription termination. Also, in vitro transcription studies have detected transcription up to 150 nucleotides beyond the GAL7 and ADH2 poly(A) sites and are consistent with polymerase pausing in this region (Hyman and Moore, 1993). In the fission yeast, Schizosaccharomyces pombe, the signals required for 3′ end formation have only been defined for the ura4 gene and comprise two site-determining elements (SDEs) which position the major and minor cleavage sites and a downstream EE required for efficient 3′ end formation at the upstream SDEs (Humphrey et al., 1994). The same 3′ end formation signals are also required for transcription termination, in association with a DSE which acts by pausing the polymerases (Birse et al., 1997). In this latter study, termination of ura4 transcription was shown to occur 180–380 bp downstream of the poly(A) site using a TRO assay in permeabilized yeast cells. The aim of this study was to extend these observations on ura4 to other S.pombe genes by mapping the polymerase density across the 3′-flanking region of the co-ordinately regulated S.pombe genes, nmt1 and nmt2, and so identify where transcription termination occurs and which signals are responsible. The nmt1 and nmt2 genes were selected because they are highly transcribed and repressed when thiamine is included in the growth medium (nmt: no message thiamine) (Maundrell, 1990; Manetti et al., 1994). Also, the nmt1 promoter and 3′-flanking sequences are present in a widely used series of vectors for the expression of genes in S.pombe (Maundrell, 1990, 1993). Information regarding the signals required for efficient nmt1 3′ end formation and transcription termination are therefore of general interest. Results Identification of genes positioned downstream from nmt1 and nmt2 Previous analysis of the nmt1 gene (Maundrell, 1990) provided a sequence (DDBJ/EMBL/GenBank databases accession No. J05493) extending 1.25 kb beyond the stop codon, but no information was available as to the nature, location and direction of any downstream transcription units. In contrast, available sequence extended 2.45 kb beyond the nmt2 stop codon (DDBJ/EMBL/GenBank accession No. X82363; additional sequence kindly provided by K.Maundrell) and included a convergently transcribed open reading frame (ORF), named avn2, of unknown function (Manetti et al., 1994). To extend the nmt1 3′-flanking sequence and identify any downstream transcription unit, cosmid clones were obtained from the Reference Library Data Base (RLDB; Maier et al., 1992; Hoheisel et al., 1993). Restriction fragments identified by Southern analysis were subcloned and sequenced. In total, 3.24 kb of novel sequence was obtained (DDBJ/EMBL/GenBank accession No. Y14993). A convergent ORF was identified (called gut2 for glycerol utilization) which is predicted to encode a mitochondrial glycerol-3-phosphate dehydrogenase based upon the observed 56.3% identity to the S.cerevisiae GUT2 gene (Rønnow and Kielland-Brandt, 1993). Mapping of nmt1 and nmt2 poly(A) sites The major nmt1 poly(A) site was originally mapped by S1 nuclease protection analysis to 142 bp downstream of the nmt1 stop codon (Maundrell, 1990). To confirm this result and to map the previously uncharacterized poly(A) site of nmt2, RT–PCR analysis was employed. Total RNA was isolated from the wild-type S.pombe strain 972h− and subjected to RT–PCR using a gene-specific primer and a phased oligo(dT) primer. DNA products were then subcloned and sequenced. Specific products were only obtained from yeast grown in the absence of thiamine, when nmt1 and nmt2 are expressed (not shown). As indicated in Figure 1A, the RT–PCR analysis confirms that nmt1 possesses a single major poly(A) site at a position which matches that obtained using S1 nuclease protection analysis (data not shown; Maundrell, 1990). Four cDNAs also mapped to different, minor sites nearby. Figure 1.nmt1 poly(A) signals. (A) RT–PCR mapping of the nmt1 poly(A) sites. The RT–PCR products were subcloned and 13 clones sequenced. The vertical arrows indicate the cleavage site positions. Where the cDNA ended with an adenosine, it was not possible to assign the cleavage site unambiguously since the adenosine may be template derived or added by poly(A) polymerase, indicated by brackets. The numbers above each arrow and the arrow length reflect the frequency with which each site was obtained. An A-rich sequence upstream of the poly(A) site is underlined. (B) A schematic diagram of the convergently transcribed nmt1 and gut2 genes is shown. The major poly(A) sites mapped by RT–PCR are indicated. The test fragments indicated were subcloned (in both orientations) into the StuI site of plasmid p41, to assess their ability to direct 3′ end formation in competition with the downstream wild-type nmt1 poly(A) site. The constructs were used to transform nmt1 disruption strain Sp204, and total RNA was prepared, which was resolved on a 1.6% formaldehyde–agarose gel overnight, transferred to a nylon membrane and probed with a random-primed probe of the XhoI–NcoI nmt1-specific fragment. The XhoI–NcoI fragment has been replaced by ura4 sequences in the host strain Sp204, so no background from the chromosome is detected. Truncated transcripts (labelled 3′ end) are due to 3′ end formation directed by the test fragment inserted at the StuI site in the nmt1 coding region. Readthrough transcripts (labelled nmt1) are due to 3′ end formation directed by the wild-type nmt1 poly(A) signals located downstream. The '% 3′ end' indicates the percentage truncated transcripts of the total detected using the nmt1-specific probe. The inclusion of thiamine in the growth medium is indicated. In the presence of thiamine, transcription from the nmt1 promoter is repressed. Download figure Download PowerPoint For nmt2, two clusters of cleavage sites were identified (Figure 2A), which were again confirmed by S1 nuclease mapping (data not shown). The cleavage sites revealed by the RT–PCR mapping are consistent with the existence of poly(A) signals specifying a region in which cleavage occurs at preferred sites, usually characterized by the sequence Y(A)n, although less favourable sites nearby may also be utilized. The situation observed in S.pombe thus appears similar to that in S.cerevisiae where multiple positioning elements are found directing cleavage at several clustered sites 15–30 nucleotides downstream (Heidmann et al., 1992; Russo et al., 1993; Guo and Sherman, 1995). An inspection of the sequence surrounding the mapped poly(A) sites for nmt1 and nmt2 (also avn2, data not shown) reveals the presence of an A-rich sequence located ∼15 nucleotides upstream. A comparison with the proposed consensus positioning element (AAAAAA or AAUAAA) in S.cerevisiae (Guo and Sherman, 1995) and the similar positioning with respect to the cleavage site suggests an A-rich element is a candidate SDE in S.pombe which may act to define the cleavage site. Indeed, the 16 nucleotide minimal element defined for the S.pombe ura4 SDE2 also contains an A-rich sequence (Humphrey et al., 1994). Figure 2.nmt2 poly(A) signals. (A) RT–PCR mapping of the nmt2 poly(A) sites. The RT–PCR products were subcloned and nine clones sequenced. The vertical arrows indicate the cleavage site positions as in Figure 1. An A-rich sequence upstream of each major poly(A) site is underlined. (B) The upper part shows a schematic diagram of the convergently transcribed nmt2 and avn2 genes. The poly(A) sites mapped by RT–PCR are indicated. As in Figure 1, nmt2 test fragments were subcloned into the StuI site of plasmid p41 to assess their ability to direct 3′ end formation by Northern blotting. Download figure Download PowerPoint Analysis of nmt1 and nmt2 poly(A) signals The signals required for efficient 3′ end formation of S.pombe mRNA have only been characterized extensively for the ura4 gene (Humphrey et al., 1994). To begin analysis of the poly(A) signals for nmt1 and nmt2, test fragments from the 3′-flanking region were cloned into the unique StuI site within the nmt1 coding region located in plasmid p41. Constructs were used to transform S.pombe nmt1 disruption strain Sp204, total RNA isolated and subjected to Northern analysis. The ability of the test fragment to direct 3′ end formation in competition with the downstream wild-type nmt1 poly(A) site is assessed by measuring the ratio of truncated transcripts, due to 3′ end formation directed by the test fragment, compared with readthrough transcripts, due to 3′ end formation at the nmt1 poly(A) site (labelled nmt1). Typical Northern blots are presented in Figures 1B (nmt1) and 2B (nmt2). For nmt1, most of the signals required for efficient 3′ end formation appear to be contained within a 256 bp fragment. This fragment contains the mapped poly(A) sites and directs 3′ end formation with 92% efficiency in this context, in an orientation-specific manner. The greater efficiency observed for the larger 434 bp fragment (99%) may simply reflect increased spacing. All Northern blot signals were sensitive to the presence of thiamine. For nmt2 (Figure 2B), fragment 2 (259 bp) containing the mapped nmt2 poly(A) sites was sufficient to direct 100% 3′ end formation in competition with the downstream nmt1 poly(A) site. However, when fragment 2 is subdivided, the 5′ fragment 1 (155 bp) which still contains the mapped nmt2 poly(A) sites showed reduced efficiency (21%), while the 3′ fragment 3 (104 bp) had no RNA 3′ end-forming activity. Therefore, sequences downstream of the nmt2 poly(A) site contribute to efficient 3′ end formation, as for the S.pombe ura4 gene (Humphrey et al., 1994). This contrasts to the situation in the majority of S.cerevisiae genes examined where no role for downstream sequences has been observed (Abe et al., 1990; Egli et al., 1997). It should also be noted that both fragments 1 and 2 function in the reverse orientation as efficient poly(A) signals. The ability of poly(A) signals to function in both orientations has also been observed in a number of S.cerevisiae poly(A) signals (Irniger et al., 1991). No cDNAs with 3′ ends mapping within fragment 4 were identified, suggesting that the ability of this fragment to direct 3′ end formation is due to cryptic signals within the fragment, not normally utilized in the natural context. A 220 bp fragment [Figure 2B, compare 5(+) with 5(−)] in the sense orientation with respect to nmt2 did not direct formation of a truncated transcript but in the reverse orientation directed 3′ end formation at two major sites (and one minor) which correspond to the two major 3′ ends mapped by RT–PCR analysis for avn2 (data not shown). Extended nascent transcription at the 3′ end of the nmt1 and nmt2 genes A TRO assay was employed to map pol II density across the 3′-flanking region of the nmt1 and nmt2 genes. Briefly, detergent-permeabilized yeast cells are incubated in a moderately high salt transcription buffer containing [α-32P]UTP for a short period (2–5 min). The conditions permit transcriptionally engaged polymerases to elongate a short distance and incorporate radioactive label (Birse et al., 1997). Total RNA is then isolated, partially hydrolysed (to sub-probe length fragments) and hybridized to immobilized single-stranded M13 probes to localize the active polymerase complexes. The signal for each probe is proportional to the average polymerase density across the DNA fragment, within the yeast population. If the pulse-labelled RNA is hybridized to contiguous single-stranded DNA probes spanning the 3′-flanking region, a profile of polymerase density is obtained. TRO analysis was performed on the wild-type strain 972h− grown in the presence or absence of thiamine. Figure 3 shows the results of a representative experiment when the pulse-labelled RNA is hybridized to probes spanning the 3′-flanking region of nmt1. High signals are observed for the nmt1 probes when the yeast are grown in the absence of thiamine (−T), but the signals fall to background levels when the yeast are grown in the presence of thiamine (+T), confirming that the signals are due to transcription from the nmt1 promoter. A signal was detected for the H2A probe both in the presence and absence of thiamine. The signals (−T) were quantitated, corrected for the M13 background (M13mp18 without insert) and for the U content of the nascent transcript that hybridizes to each probe, and were expressed relative to the 5′ nmt1 probe (1-1). The polymerase profile obtained reveals that polymerases transcribe a distance of 4.3 kb beyond the poly(A) site of the endogenous nmt1 gene, reading entirely through the convergent gut2 transcription unit. This is surprising since all previous studies in yeast have indicated that transcription termination occurs close to the poly(A) site. Indeed, termination of transcription close to the poly(A) site is intuitively expected, given the compact nature of yeast genomes and the need to avoid transcriptional interference. A build-up in polymerase density is observed over the probes spanning the gut2 ORF which may in part reflect a hybridization effect due to the greater GC content within the coding region. Also, a second build-up is observed over probe 1-20, just before polymerase density falls to the lowest levels detected over probe 1-22, suggesting that transcription termination may be occurring at this position. Figure 3.Transcription run-on analysis of nmt1. The wild-type strain 972h− was grown overnight in minimal medium in the presence or absence of thiamine, and TRO analysis was performed. The signals obtained after the high-stringency wash are shown and are displayed graphically after quantitation. The signals were corrected for the M13 background (M13mp18 without an insert) and for the U content of the nascent RNA hybridizing to each probe, and are expressed relative to probe 1-1. The graph is drawn to scale such that the width of each bar reflects the length of the probe and the height indicates the relative average signal across the probe. The H2A probe is included as an internal control. The profile obtained was similar after an additional RNase A wash (data not shown). Download figure Download PowerPoint A similar analysis for nmt2 is presented in Figure 4. The signals obtained for nmt2 are also sensitive to the presence of thiamine, confirming that the nascent transcription detected is due to polymerases initiating at the nmt2 promoter. For nmt2, a drop in polymerase density is observed consistently over the probe containing the mapped poly(A) sites, suggesting that some polymerases may terminate transcription at this point. The polymerases that continue appear to give an increased signal over probe 2-5 and those immediately downstream, before a gradual decline in signal is observed. As for nmt1, it is notable that polymerases transcribe at least 2.4 kb beyond the poly(A) site and read through the downstream convergent avn2 transcription unit. Figure 4.Transcription run-on analysis of nmt2. Conditions were as for Figure 3. The signals obtained after the high-stringency wash are shown and displayed graphically after quantitation. The signals were corrected as in Figure 3 and expressed relative to probe 2-1. The H2A, leu1 and 18S probes are included as internal controls. The profile obtained was similar after an additional RNase A wash (data not shown). Download figure Download PowerPoint A comparison of the level of signals obtained for nmt1 and nmt2 with that obtained for the other pol II-transcribed genes, leu1 and H2A.1, and the RNA polymerase I-transcribed 18S rRNA gene (Figures 3 and 4), reveals that high levels of nascent transcription are observed for nmt1 and nmt2 which facilitates their analysis. No signal was detected for the sense M13 probes (±T) which would be expected to detect polymerases transcribing gut2 or avn2 (data not shown). Presumably, avn2 and gut2 nascent transcription signals lie below the limits of detection using this TRO assay under the growth conditions employed. Although representative polymerase profiles are shown in Figures 3 and 4, some variation is observed between different experiments (data not shown) in the apparent efficiency of transcription termination [distance beyond the poly(A) site at which polymerase density falls to background levels]. However, the trend in polymerase density remains consistent. For nmt2, polymerase density always falls over the poly(A) signals and increases over probe 2-5, but in some experiments a more pronounced build-up in signal over probe 2-5 appears to be associated with more efficient transcription termination, closer to the poly(A) site. A similar apparent increase in termination efficiency is also observed in some experiments for nmt1, performed in parallel, suggesting that the variation may reflect differences in the growth of the yeast. However, even in these experiments, polymerases still continue for 2.0 kb (nmt2) and 2.6 kb (nmt1) beyond the poly(A) site reading into the downstream transcription unit (data not shown). gut2 and avn2 expression is unaffected by nmt1 or nmt2 nascent transcription Since nascent transcription from both the nmt1 and nmt2 genes extends into the convergent downstream transcription unit, it was of interest to examine whether the high level of nmt1 or nmt2 transcription influenced the expression of the adjacent gut2 or avn2 genes. Therefore, the wild-type strain 972h− was grown on minimal media in the presence or absence of thiamine, with 3% glucose as the sole carbon source or 3% glycerol/0.1% glucose (0.1% glucose was included to facilitate growth; Hoffman and Winston, 1989). These alternative carbon sources were used to test whether gut2 expression was subject to glucose repression as has been found for the S.cerevisiae GUT2 gene (Sprague and Cronan, 1977; Rønnow and Kielland-Brandt, 1993). Total RNA was prepared and subjected to Northern analysis (Figure 5A). The low levels of gut2 expression prevent accurate quantitation. However, based on the leu1 loading control, it is apparent that gut2 expression is repressed by glucose, but not significantly affected by nmt1 transcription. The size of the gut2 transcript was estimated to be 2.6 kb based on RNA markers run alongside. Figure 5.Steady-state analysis of gut2 and avn2 expression. (A) The wild-type strain 972h− was grown overnight in minimal medium containing 3% glucose or 3% glycerol/0.1% glucose as a carbon source, in the presence or absence of thiamine. Total RNA was prepared which was resolved on a 1.6% formaldehyde–agarose gel overnight, transferred to a nylon membrane and sequentially stripped and probed with a random-primed probe of the StyI–NruI leu1-specific fragment, an XhoI–NcoI nmt1-specific fragment and a HindIII–EcoRV gut2-specific fragment. The leu1 signal acts as a loading control. (B) The wild-type strain 972h− and the nmt2 disruption strain Sp92 were grown overnight in minimal medium in the presence and absence of thiamine. Total RNA was prepared and equal amounts resolved on a 1.6% formaldehyde–agarose gel overnight, transferred to a nylon membrane and probed with a random-primed probe of the NdeI–StyI fragment from the avn2 coding region. Download figure Download PowerPoint A similar steady-state Northern analysis of avn2 expression reveals that a doublet (1.56 and 1.8 kb bands) is detected in RNA isolated from the wild-type strain 972h− and the nmt2 disruption strain Sp92, grown both in the presence and absence of thiamine (Figure 5B). The doublet corresponds to 3′ end formation at the two major poly(A) sites mapped by RT–PCR analysis (not shown) as seen in clone 5(−) in Figure 2B. Equal amounts of total RNA were loaded in each lane, therefore expression of avn2 does not appear to be influenced by nmt2 transcription. The lower band visible in the lane containing RNA isolated from 972h− grown in the absence of thiamine represents cross-hybridization to the nmt2-derived transcripts since it is absent in the nmt2 disruption strain Sp92. These data indicate that both the gut2 and avn2 genes are subjected to extensive readthrough transcription from nmt1 and nmt2. However, neither this high level of transcriptional activity on the opposite strand nor the accumulation of antisense transcript appear to affect their steady-state mRNA levels. nmt2 poly(A) signal is required for efficient transcription termination To identify the sequences responsible for controlling the pattern of polymerase density observed at the 3′ end of a gene, it is necessary to reproduce the endogenous gene profile in a plasmid context. Therefore, a 3.67 kb HindIII–HindIII fragment encompassing the nmt2 and avn2 ORFs was cloned into the BamHI site of pIRT2 to create plasmid p2V. Northern analysis confirmed that a thiamine-sensitive transcript was transcribed from plasmid p2V (Figure 6B). When plasmid p2V was used to transform S.pombe nmt2 disruption strain Sp92 and TRO analysis performed, a similar polymerase profile was observed for the plasmid and endogenous genes (compare panels in Figures 6A, and 4), such that the polymerase density fell over the poly(A) site and increased over probe 2-5 before decreasing gradually over subsequent probes to background levels. However, the plasmid-derived TRO profile gave lower signals over the downstream probes covering the 3′ end of avn2 than was observed for the endogenous gene. This may reflect differences in chromatin structure between the chromosomal and plasmid genes which might be expected to influence elongation. The build-up in signal observed over probe 2-5 in both the chromosomal and plasmid nmt2 gen