Transcription and polyadenylation in a short human intergenic region
Transcription and polyadenylation in a short human intergenic region
Simon Brackenridge, Hilary L. Ashe, Mauro Giacca1, Nicholas J. Proudfoot*
Sir William Dunn School of Pathology, University of Oxford, South Parks Road, Oxford OX1 3RE, UK and 1International Centre for Genetic Engineering and Biotechnology, Padriciano 99, I-34012 Trieste, Italy
Received March 18, 1997;Revised and Accepted April 30, 1997
The poly(A) signal of the human Lamin B2 gene was previously shown to lie 600 bp upstream of the cap site of a gene of unknown function (ppv1). However, using RNase protection analysis, we show that ppv1 has two clusters of multiple initiation sites, so that the 5" cap site lies only [sim]280 nt downstream of the Lamin B2 poly(A) signal. We analysed nascent transcription across this unusually short intergenic region using nuclear run-on analysis of both the endogenous locus and of transiently transfected hybrid constructs. Surprisingly, transcription of the Lamin B2 gene does not appear to terminate prior to any of the mapped ppv1 start sites, although pausing of the elongating polymerase complexes is observed downstream of the Lamin B2 poly(A) signal. We suggest that this pausing may be sufficient to protect the downstream gene from transcriptional interference. Finally, we have also investigated the sequences required for efficient recognition of the Lamin B2 poly(A) signal. We show that sequences upstream of the AAUAAA element are required for full activity, which is an unusual feature of mammalian poly(A) signals.
INTRODUCTION
In eukaryotes, the 3" end of a mRNA does not correspond to the point at which the polymerase complex stops transcribing the DNA; rather, the nascent RNA is cleaved at the polyadenylation signal and a poly(A) tail added (for a review, see ref. 1). Poly(A) signals of higher eukaryotes contain the highly conserved AAUAAA motif, first identified 20 years ago (2), that is absolutely required for function both in vivo (3) and in vitro (4). Saturation mutagenesis has demonstrated that very few variants of the canonical AAUAAA sequence support efficient cleavage/polyadenylation (5). As might be expected, the least deleterious change (AUUAAA) represents the most frequent naturally occurring variant. Cleavage appears to occur preferentially at an A residue lying 10-30 nucleotides (nt) downstream of the AAUAAA element (6), with some evidence of a preference for a CA dinucleotide (5). Sequences downstream of the cleavage site have been shown to enhance the efficiency of cleavage at a number of poly(A) signals both in vivo and in vitro (e.g. refs 7-10). In addition, these downstream sequences can influence the position at which cleavage occurs (11,12). Sequences upstream of the AAUAAA hexanucleotide do not appear to be required for efficient 3[prime]-end formation in most vertebrate genes, although examples of upstream sequence elements (USEs) have been reported for several viral systems, including the SV40 late poly(A) signal and HIV-1 (see ref. 1).
Following polyadenylation, the RNA polymerase II complex continues elongating, often terminating several kilobases downstream of the poly(A) signal; for example, in vivo and in vitro labelling studies have shown that transcription of the murine [beta]maj globin gene terminates between 700 and 2000 bp downstream of the poly(A) signal (13). Termination requires an intact polyadenylation signal (14-16) and the strength of a polyadenylation signal correlates with the efficiency of termination (17,18). Two models have been proposed to explain how an RNA processing event influences the ability of a ternary complex to terminate transcription: Darnell and co-workers (15) suggested that the ternary complex carries with it a processivity factor that is lost when a poly(A) signal is encountered, converting the polymerase to a form that is competent to recognise termination signals. A recent study has demonstrated that the CTD tail of RNA polymerase II is required for polyadenylation in vivo and can associate with components of the polyadenylation machinery (19). In the second model, proposed separately by Manley (16) and Proudfoot (20), the uncapped 5" end of the nascent RNA resulting from the cleavage at the poly(A) signal would be accessible to attack by a 5"[rarr]3" exonuclease or helicase. Termination is suggested to result from the ternary complex being disrupted by removal of the nascent RNA. Two lines of circumstantial evidence have emerged that are consistent with this second model: firstly, an exonuclease activity has been described in HeLa nuclear extract that is specific for uncapped RNA (21), and secondly, pause sites appear to be associated with at least some termination sites (22), and it has been suggested that the stalling of the polymerase downstream of a poly(A) signal would increase the efficiency of termination by reducing the time required for the helicase or exonuclease to reach the ternary complex.
It has previously been demonstrated in situations where two closely spaced genes are transcribed in the same direction, that transcription of the upstream gene reading into the promoter of the downstream gene decreases expression of the downstream gene. Examples of this phenomenon, termed promoter occlusion or transcriptional interference, include the retroviral 3" LTR promoter (23) and duplicated human [alpha] globin genes (24). Evidence that promoter occlusion represents a transcriptional event comes from the observation that insertion of polyadenylation and termination signals (individually or in combination) between two synthetic genes results in relief from interference on the downstream promoter (25). For this reason, we have previously studied transcription in the short intergenic region between the human complement C2 and factor B genes (26). These studies led to the identification of the binding site for the zinc-finger transcription factor, MAZ, close to the region of termination of C2 transcription. Indirect assays suggested that binding of MAZ could play a role in directing termination of transcription, and that MAZ could potentially play a role in termination in the human complement g11-C4 intergenic region, as well as in the mouse IgM-D gene (27). We have now turned our attention to another short human intergenic region, that between the Lamin B2 gene and a gene of unknown function, termed ppv1, lying in the sub-telomeric G-negative band p13.3 of human chromosome 19. This region is replicated early during S-phase (28) and considerable evidence has accumulated to suggest the presence of a chromosomal origin of DNA replication lying in the 3" UTR of the Lamin B2 gene (29,30). Approximately 600 bp separates the poly(A) signal of the Lamin B2 gene from the two mapped start sites of ppv1, and expression of the Lamin B2 gene appears to be limited to S-phase, while the downstream gene is not cell cycle regulated (28). We have undertaken a detailed analysis of transcription in this intergenic region, using both steady state and nascent transcript mapping techniques. Our results suggest that the intergenic region is smaller than previously determined, with only [sim]280 nt separating the cleavage site of the Lamin B2 poly(A) signal and the 5"-most cap site of the downstream gene. Surprisingly, there does not appear to be efficient termination of Lamin B2 transcription. We suggest that promoter occlusion may not represent a significant problem for expression of ppv1 due to the limited expression of the Lamin gene. We have also investigated the poly(A) signal of the Lamin B2 gene and have demonstrated that the poly(A) signal depends upon sequences upstream of the AAUAAA hexanucleotide for maximal activity. To date, the only other cellular mammalian gene with such an upstream activatory element present in the poly(A) signal is the human C2 complement gene (31).
MATERIALS AND METHODS
Plasmid constructions
pB48 is a derivative of pBluescribe that contains the 1.1 kb B48 region from human chromosome 19 that encompasses the Lamin B2-ppv1 intergenic region (30). Single stranded DNA probes for this region were produced by cloning the desired restriction fragments from pB48 into the HincII site of M13mp18 or M13mp19 vectors (Table 1). The single stranded probes for the [alpha]2 globin gene have been described previously (22), the two probes for the HIV-1 LTR were constructed by Jan Eggermont (U3) and Mark Ashe (PH), and the mouse Histone H4 and Xenopus 5S rRNA probes by Jan Eggermont. RNase protection constructs were made using pGEM4, with inserts arranged so that transcription with T7 RNA polymerase produces antisense RNAs. pGEMRP1 contains the 290 bp ApaI-PstI fragment from pB48, while pGEMRP2 carries the 347bp PmlI-AvaI fragment from the same plasmid. The [alpha]2W3"PS plasmid used in the poly(A) competition experiments has been described previously (22). The transient expression construct for the HeLa nuclear run-on experiments (pHIV[alpha]B48) was produced by inserting the 1.1 kb B48 region (excised from pB48 as a BamHI-Acc65I fragment, utilising restriction sites in the pBluescribe polylinker) into pSVodL[alpha]2 (25).
Plasmids for use in the poly(A) competition experiments were produced by inserting fragments containing the Lamin B2 poly(A) signal into [alpha]2W3"PS at either the PvuII or HpaI sites. Changes to the upstream sequences were introduced by PCR with mismatched primers as follows (all distances relative to the AAUAAA sequence): m1 alters the 6 T residues of the T10 tract from -55 to -64 to a BamHI site; m2 converts the CAT4 sequences from -31 to -34 to a PvuII site; and m3 changes the T6 tract from -13 to -18 to a BamHI site.
HeLa cell transfection and RNA analysis
Subconfluent HeLa cells were transiently transfected using the calcium phosphate precipitation method and cytoplasmic RNA was isolated after 24-48 h exactly as described previously (32). RNA from untransfected NTera-2/D1 cells was isolated using exactly the same method. S1 probes were made by 3" end-labelling 1 [mu]g of linearised plasmid with 10 [mu]Ci [[alpha]-32P]dGTP (800 Ci/mmol) and Klenow DNA polymerase (Boehringer Mannheim), and purified using a Nick column (Pharmacia). S1 and RNase protection analyses were performed exactly as described previously (25), except RNase digestions were performed at 18°C for 2 h. Quantitation of S1 gels was by phosphorimaging using a Molecular Dynamics Storm 840 system and the results presented are an average of at least three independent experiments.
Figure 1 RNase protection mapping of ppv1. (A) Schematic of the intergenic region with the major features indicated, including the binding sites for a number of transcription factors (bHLH, Sp1, NRF-1 and UBF) and components of the replication origin (including a topoisomerase II binding site, marked Topo II). A number of restriction sites are also marked (filled arrows), as are the two cap sites that have previously been mapped for ppv1, and the poly(A) signal of the Lamin B2 gene (open arrow). (B) RNase protection results using probes 1 (left) and 2 (right): lanes 1 and 4 show molecular weight markers, lanes 2 and 5 cytoplasmic RNA from HeLa cells, and lane 3 cytoplasmic RNA from NT2/D1 cells.
Nuclear run-on analysis
Cells were washed once with PBS, scraped into 10 ml of the same and centrifuged at 1400 r.p.m. for 5 min at 4°C. Cell pellets were resuspended in 4 ml of hypotonic lysis buffer (10 mM Tris-HCl pH 7.5, 10 mM NaCl, 2.5 mM MgCl2)/0.5% NP-40 and incubated on ice for 5 min. Lysates were then underlayered with 1 ml of HLB/0.5% NP-40/10% sucrose, and the nuclei spun through the sucrose cushion (1500 r.p.m. for 5 min.). The nuclei were washed once with HLB to remove residual traces of NP-40, then resuspended in an equal volume (plus 10 [mu]l) of 2× transcription buffer (40 mM Tris-HCl, 1.5 M KCl, 10 mM MgCl2, 40% glycerol, 2 mM DTT) and rATP, rUTP and rGTP were added to a final concentration of 250 [mu]M. Transcription was initiated by the addition of [[alpha]-32P]rCTP (800 Ci/mmol; 20 mCi/ml) to the nuclei (typically 120 [mu]Ci per 50 [mu]l of nuclei), and the reaction incubated at 30°C for 15 min. Nuclei were then pelleted by centrifugation in a microfuge for 30 s, resuspended in 200 [mu]l of high salt buffer (10 mM Tris-HCl pH 7.9, 500 mM NaCl, 10 mM MgCl2) containing 10 U RNase-free DNase I, and incubated at 30°C for 5 min. A volume of 200 [mu]l of 0.5 mg/ml Proteinase K/0.5% w/v SDS was added and the reactions incubated at 37°C for 30 min. Nucleic acid was extracted once against an equal volume of acidic phenol:chloroform (1:1) and ethanol precipitated by the addition of 2.5 vol of ice cold ethanol. Pellets were resuspended in 200 [mu]l of DNase Buffer/10 U RNase-free DNase I, then incubated at 37°C for 30 min. RNA was extracted once against acidic phenol:chloroform, ethanol precipitated following the addition of sodium acetate to 0.3 M, and finally resuspended in 40 [mu]l of water. RNA was partially hydrolysed for 4 min on ice by the addition of NaOH to a final concentration of 200 mM, then neutralised by the addition of 20 [mu]l of 0.2 M Tris/0.2 M HCl and the volume made up to 750 [mu]l in hybridisation solution (6× SSPE, 50% deionised formamide, 5× Denhardt's solution, 50 [mu]g/ml tRNA, 10% SDS), and hybridised overnight at 42°C with Hybond-N (Amersham) slot blots carrying 5 [mu]g of each single- stranded M13 probe. (The slot blots were prehybridised in 2.5 ml of 6× SSPE, 50% deionised formamide, 5× Denhardt's solution, 50 [mu]g/ml tRNA, 10% SDS at 42°C overnight.) Following hybridisation, the slot blots were washed three times in 1.0× SSPE/0.1% w/v SDS at room temperature, then at least once in 0.1× SSPE/0.1% w/v SDS. Membranes were exposed to Kodak AR film at -70°C or to Molecular Dynamics Phosphor storage screens at room temperature (to allow quantitation of the hybridisation signal).
RESULTS
RNase protection mapping of steady state transcripts
As shown in Figure 1A, two different start sites have been reported for ppv1, determined by RNase protection and primer extension (28), one on each side of the NcoI restriction site. In order to confirm that these start sites are the same in the two cell lines employed for the nuclear run-on analysis (NTera-2/D1 and HeLa cells, see below) we have used the RNase protection assay to map the 5" end of the ppv1 gene. The 290 bp ApaI to PstI restriction fragment from the intergenic region, encompassing the two known start sites, was inserted into the HindIII site in the polylinker of pGEM4 to produce pGEMRP1. In vitro transcription of EcoRI-linearised pGEMRP1 with T7 RNA polymerase produces a riboprobe of 356 nt, comprising a 290 nt homologous region flanked by 31 and 35 nt of non-homologous pGEM4 sequence at the 5" and 3" ends, respectively.
The left hand panel of Figure 1Bshows an RNase protection assay using this probe. In lanes 2 and 3, it can be seen that a number of specific bands are produced with both HeLa and NTera-2/D1 cytoplasmic RNA. An identical pattern of protected bands is obtained with cytoplasmic RNA isolated from HepG2 or HL-60 cells, but not tRNA, demonstrating that these bands are not a probe artefact (not shown). The presence of multiple protected probe fragments suggests that ppv1 has considerably more than two start sites in the previously described region, perhaps as a result of the promoter lying in a CpG Island. Also present on the gel is a protected fragment of [sim]290 nt. This fragment apparently represents protection of the entire homologous region of the probe, implying that transcription from upstream of the ApaI site in the intergenic region is being detected.
To map the 5" end of this upstream transcript, a second RNase protection probe has been generated. This probe was produced by transcribing pGEMRP2, a pGEM4 derivative that carries the 347 bp PmlI-AvaI fragment from the intergenic region. This probe does not therefore extend to the multiple start sites mapped using the first probe. As can be seen from lane 5 of Figure 1B (right), HeLa cell RNA protects about three probe fragments of [sim]260-290 nt. This indicates that there is an additional group of start sites for the ppv1 gene lying [sim]280 bp downstream of the Lamin B2 poly(A) site. Thus, the minimum intergenic region that separates the Lamin B2 gene from ppv1 is significantly shorter than previously thought and represents one of the shortest intergenic regions so far described in mammals.
Nuclear run-on transcription analysis of the endogenous intergenic region
The most widely employed method of determining the location of transcription termination is the nuclear run-on procedure (originally described by Weber et al., ref. 33). Nuclei are isolated from cells and incubated in a transcription reaction (in the presence of radiolabelled nucleoside triphosphate), which allows any polymerase complexes engaged in transcription to extend their nascent transcripts for a short distance. The conditions employed are such that de novo initiation and transcript processing are inhibited (34), resulting in labelling of only the nascent transcripts already present. Incorporated label can be detected by isolating nuclear RNA and hybridising to filter-bound single stranded DNA probes containing sequences complementary to the RNA of interest. The intensity of the hybridised signal on each probe is proportional to both the polymerase density across the region encompassed by the probe and the number of labelled nucleotides in that region. Prior to the hybridisation, the nuclear RNA must be partially hydrolysed, ensuring that the incorporated label will be present only in short RNA molecules proximal to the polymerase complex. The locations of the single-stranded DNA probes complementary to the nascent RNA generated by transcription of this region are illustrated in the upper part of Figure 2. Two additional probes have been included in these experiments, one complementary to the histone H4 mRNA and a second complementary to the pol III-transcribed 5S rRNA (denoted by `his' and `5S' on the diagram).
Figure 2Nuclear run-on analysis of the endogenous gene. The upper part shows the Lamin B2-ppv1 intergenic region with the positions of the single-stranded DNA probes marked; the lower part shows the hybridisation pattern obtained with these probes, and quantitation of the signals.
We have employed the embryonal teratocarcinoma cell line NTera-2/D12 for these studies, as these cells appear to express the Lamin B2 gene at a higher than normal level (M.G., personal communication). Because the Lamin B2 gene is expressed only in S-phase, cells were synchronised by the double aphidicolin block method and Figure 2 shows the pattern of hybridisation obtained from nuclear run-on reactions using nuclei isolated 2 h after release from the second aphidicolin block. The graph gives the relative signal for each probe corrected for both the empty M13 hybridisation background and the number of CMP residues in the nascent transcript over the region spanned by the probe. Probes B covers 765 bp of sequence within the Lamin B2 gene, while probe ED covers [sim]1.7 kb in the final exon of the gene. The corrected signal for the probes over the intergenic region has been expressed relative to probe ED. It can be seen from this experiment that there appears to be an increase in the density of polymerase complexes over the region containing the poly(A) signal of the Lamin B2 gene (probe HBy), followed by a further increase downstream (probe PE). The polymerase density over the rest of the intergenic region (probes EBs and BsN) is comparable with that seen over probe ED, and downstream of the multiple start sites of the ppv1 gene there is again a significant increase of polymerase density, as indicated by the 3-fold higher signal for the ppv1 probe HA. Thus it would appear that there is a slight build-up of RNA polymerase downstream of the poly(A) signal of the Lamin B2 gene, perhaps as a result of sequences capable of pausing the elongating polymerase. However, there does not appear to be any appreciable termination of transcription prior to the cap sites of the downstream gene. In view of the steady-state mapping experiments described above, it is possible that initiation of transcription within the intergenic region could result in overlap of the Lamin B2 and ppv1 transcription units, thus obscuring termination of Lamin B2 transcription. Initiation at this site could also contribute to the increased polymerase density observed over probe PE. This point is addressed in the next section.
Figure 3Nuclear run-on analysis using a transient expression construct. (A) Details of the transient expression construct pHIV[alpha]B48, showing the positions of the single stranded probes used to detect transcription over both the hybrid gene and the intergenic region. (B) Nuclear run-on hybridisation profiles and quantitation using HeLa cells transiently transfected with both pHIV[alpha]B48 and pOGS213. (C) Nuclear run-on hybridisation profiles using HeLa cells transiently transfected with pHIV[alpha]B48 alone.
Nuclear run on analysis using transiently transfected constructs
The results presented above suggest that transcription of the Lamin B2 gene does not terminate upstream of ppv1. As mentioned before (and see also the Discussion), the level of signal obtained from the nuclear run-on experiments is very low, making definite conclusions difficult. For this reason, we have performed nuclear run-on analysis using HeLa cells transiently transfected with a hybrid expression construct. The Lamin B2-ppv1 intergenic region was inserted downstream of the HIV-1 LTR promoter/human [alpha]2 globin hybrid in plasmid pSVodL[alpha]2 (25), producing plasmid pHIV[alpha]B48 (Fig. 3A). Co-transfection with plasmid pOGS213 (35), a source of the viral Tat protein, allows trans-activation of the LTR promoter and expression of the hybrid gene in HeLa cells. As the poly(A) signal of the [alpha]2 globin gene has been replaced with that of the Lamin B2 gene in this construct, it was important to demonstrate that the Lamin B2 poly(A) signal is functional in this context. RNase protection mapping using cytoplasmic RNA from HeLa cells transiently transfected with each of these constructs showed that the Lamin B2 poly(A) signal was indeed used both accurately and efficiently (data not shown; see also below). Although the Lamin B2 and ppv1 genes are expressed at low levels in HeLa cells, transient nuclear run on analysis using the HIV promoter construct is possible since the level of signal obtained from such transient constructs is several orders of magnitude greater than the endogenous signals. Thus, we feel that these transient expression constructs are a realistic surrogate for the endogenous gene.EF
Figure 4 The Lamin B2 poly(A) signal is dependent on sequences upstream and downstream of the AAUAAA. (A) The sequence around the polyadenylation signal of the Lamin B2 gene, including 134 nt upstream of the AAUAAA element and 80 nt of sequence beyond the cleavage site. The 5" ends of the various poly(A) signal fragments are marked either as numbers (-131, -56, +1), or restriction sites (HinfI and RsaI) (GenBank accession number M94363). (B)Details of the S1 mapping strategy used in the poly(A) signal competition assay using plasmid [alpha]2W3"PS. Non-homologous probes are used that mis-match at the PvuII site. Use of the [alpha]2 globin poly(A) signal generates mRNA that protects a probe fragment of 212 nt, while use of a poly(A) signal inserted at either the PvuII or HpaI sites gives rise to mRNA that protects a probe fragment of 285 nt. (C) Results of the poly(A) competition assay when the Lamin B2 poly(A) signal fragments are inserted at the PvuII site. Lane 1, molecular weight markers; lane 2, [alpha]WP/L+131; lane 3, [alpha]WP/L+101; lane 4, [alpha]WP/L+56; lane 5, [alpha]WP/L+10; lane 6, [alpha]WP/L+0. Also included S1 reactions was a probe specific for the rabbit [beta]-globin transcript produced by plasmid p[beta]5[prime]SVpBR328 that is included in the transfections as a source of SV40 Large T antigen. The probe is a 3"-end-labelled EcoRI digest of this plasmid, and should protect a fragment of 170 nt. However, several bands are seen as a result of the S1 reaction being performed at 30°C, which allows breathing between the probe and the mRNA. (D) Results of the poly(A) competition assay when the Lamin B2 poly(A) signal fragments are inserted at the HpaI site. Lane 1, molecular weight markers; lane 2, [alpha]WH/L+131; lane 3, [alpha]WH/L+101; lane 4, [alpha]WH/L+56; lane 5, [alpha]WH/L+10; lane 6, [alpha]WH/L+0. The co-transfection control is as (C). (E) Results of the poly(A) competition assay with the gL+10 and [Delta]ds forms of the poly(A) signal. Lane 1, molecular weight markers; lane 2, [alpha]WP/gL+56; lane 3, [alpha]WP/gL+10; lane 4, [alpha]WP/[Delta]ds ; lane 5, [alpha]WH/gL+56; lane 6, [alpha]WH/gL+10; lane 7, [alpha]WH/[Delta]ds. The co-transfection control is as (C).
Figure 3B shows the nuclear run-on analysis of transiently transfected HeLa cells. As with the endogenous Lamin B2-ppv1 profile, a significant build-up of signal occurs over probe PE, again indicating transcriptional pausing in this region. However, transcription continues unabated beyond this pause site, reading through the intergenic region, which again indicates that no detectable termination occurs. The very high signal obtained over the HIV promoter-proximal region (probe PH) may reflect a transcriptional attenuation effect at the beginning of the hybrid gene. One advantage of this transient expression system is that the [EEgr][Igr]V promoter can be effectively turned off by omitting the source of the Tat plasmid and Figure 3C shows the run-on profile obtained for plasmid pHIV[alpha]B48 transfected into HeLa cells in the absence of pOGS213. As can be seen, near background levels of transcription are observed over all of the probes, including PE, relative to the 5S and histone H2B control probes. Thus, it would appear that the signals seen over the intergenic region probes, are due to transcription reading through from the HIV-1 promoter. This would imply that termination is not occurring within the intergenic region. In addition, the strong hybridisation signal seen over probe U3 which lies upstream of the HIV-1 cap site is also dependent upon the presence of Tat protein, indicating that transcription extends around the whole plasmid.
Sequences required for full activity of the Lamin B2 poly(A) signal
The sequence around the Lamin B2 polyadenylation signal is shown in Figure 4A. Because of the link between the efficiencies of a poly(A) signal and the subsequent termination event, we have investigated the sequences required for efficient utilisation of the Lamin B2 poly(A) signal by placing it in competition with a weak reference poly(A) signal (from the human [alpha]2 globin gene). A number of constructs were produced with various fragments containing the Lamin B2 signal inserted downstream of the human [alpha]2 globin gene in plasmid [alpha]2W3"PS (36), as illustrated in Figure 4B. The Lamin poly(A) signal fragments were inserted at the unique PvuII site, which lies [sim]72 bp downstream of the cleavage site of the [alpha] globin poly(A) signal. This plasmid contains the SV40 origin of replication, which allows expression of the [alpha]2 globin gene in HeLa cells in the presence of a source of SV40 Large T antigen [supplied by co-transfecting plasmid p[beta]5[prime]SVpBR328 (37)]. The relative use of the Lamin B2 and [alpha]2 globin poly(A) signals in the [alpha]2W3"PS derivatives is determined by S1 nuclease protection with cytoplasmic RNA harvested from transiently transfected cells. S1 probes are made by labelling [alpha]2W3"PS at the BstEII site in the third exon of the [alpha]2 globin gene (Fig. 4B). Transcripts utilising the [alpha]2 globin poly(A) signal will thus protect a fragment of 212 nt, whilst transcripts that have polyadenylated at the Lamin B2 signal will mismatch at the PvuII site, protecting a fragment of 306 nt. To ensure that correct polyadenylation at the Lamin B2 poly(A) signal is being observed homologous probes have also been employed (data not shown).
Figure 4C shows an autoradiograph of a typical S1 protection assay where the Lamin B2 poly(A) signal has been inserted at the PvuII site. The three longest forms of the poly(A) signal tested (L+131, L+101 and L+56, which carry 131, 101 and 56 nt of upstream sequence) all virtually completely out-competed the upstream [alpha]2 globin poly(A) signal. Surprisingly, when only 10 bp of sequence is included upstream of the AAUAAA of the poly(A) signal (construct [alpha]WP/L+10, which carries the RsaI-BstYI fragment), the ability of the Lamin poly(A) signal to out-compete the [alpha]2 globin signal was greatly reduced: quantitation reveals that 39% of the steady state transcripts now polyadenylate at the [alpha]2 globin signal, despite the fact the Lamin B2 signal has apparently been moved closer to the upstream signal. In the absence of any sequence upstream of the AAUAAA of the Lamin B2 poly(A) signal (construct [alpha]WP/L+0), a further decrease in use of this signal (to 54%) is observed.
The results presented above demonstrate that sequences extending 56 nt upstream of the AAUAAA of the Lamin B2 poly(A) signal are required for efficient utilisation of this signal. It is possible, however, that sequences present further upstream of the poly(A) signal could also be important, but that the assay is not sufficiently sensitive to reveal this fact. Given that the [alpha]2 globin poly(A) signal is relatively weak, it is conceivable that the sequence elements present in the first 55 bp of upstream sequence are sufficient to enhance the efficiency of the basal Lamin B2 polyadenylation signal such that it is strong enough to out-compete the weaker [alpha] poly(A) signal. Indeed, the upstream activatory element(s) may extend beyond the 131 nt arbitrarily chosen as the 5" cut-off in the initial construct.
Since the spacing between two poly(A) signals may influence competition between them, inserting the Lamin B2 poly(A) signal further downstream of the [alpha]2 poly(A) signal cleavage site provides a more stringent test of processing efficiency, and may reveal differences between the L+56, L+101 and L+131 forms. Thus, the various forms of the poly(A) signal were inserted at the unique HpaI site [sim]300 bp downstream of the [alpha]2 poly(A) signal, and the S1 analysis of these constructs is shown in Figure 4D. The L+56 form of the poly(A) signal is now used less frequently than the L+101 form (36% versus 55%), but surprisingly the L+131 form is also less efficient than the L+101 form (being used 47% of the time). This suggests that the USE extends at most 101 nt upstream of the AAUAAA, with the additional 30 nt of sequence present along with the L+131 form accounting for the slight reduction in use. When only 10 nt of upstream sequence is present, use of the Lamin B2 signal again decreases sharply (with only 18% of transcripts having polyadenylated at the inserted signal), while use of the Lamin B2 poly(A) signal falls even further (to 13%) following removal of all of the upstream sequence. Thus, it would appear that the L+101 and L+56 forms are not equivalent, and that the upstream activatory element extends beyond 56 bp upstream of the core poly(A) signal.
Given that the poly(A) signal of the Lamin B2 gene possesses an upstream element, it was important to determine if the sequences downstream of the cleavage site were also required for efficient use of the signal. PCR was used to generate a Lamin B2 poly(A) signal fragment extending from the upstream HinfI site to the A residue at the cleavage site (named [Delta]ds), and its ability to compete against the [alpha]2 globin poly(A) signal was tested at both the PvuII and HpaI sites (Fig. 4E). As can be seen, removal of the 80 bp of downstream sequence has a crippling effect on the poly(A) signal: at the PvuII site, the Lamin signal was only used 13% of the time, whilst at the HpaI site, the Lamin signal was used 17% of the time. The fact that the [Delta]ds form of the poly(A) signal works better when further away from the [alpha]2 poly(A) signal presumably results from the sequence environment immediately downstream of the HpaI site being more favourable than that downstream of the PvuII site. Thus, it would appear that while the signal can function reasonably well in the absence of the upstream sequence, the downstream sequence is absolutely critical for function. Sequences further downstream than the BstYI site do not contribute to the efficiency of the poly(A) signal (data not shown).
Figure 5 Mutagenesis of the poly(A) signal upstream sequence. (A) Schematic diagram showing the locations of the various changes (m1-m3) introduced into the sequences upstream of the AAUAAA (represented by a black rectangle). The cleavage site is shown by the open arrow. (B) Results of the poly(A) competitions using the constructs carrying single clusters of mutations. Lane 1, molecular weight markers; lane 2, [alpha]WP/L+101; lane 3, [alpha]WP/m1; lane 4, [alpha]WP/m2; lane 5, [alpha]WP/m3; lane 6, [alpha]WH/L+101; lane 7, [alpha]WH/m1; lane 8, [alpha]WH/m2; and lane 9, [alpha]WH/m3. Also visible on the gel is an artefact band that results from breathing of the RNA:probe hybrid at the AAUAAA in the [alpha]2 globin poly(A) signal. (C) Results of poly(A) competition with the double and triple clusters of mutations. Lane 1 shows molecular weight markers; lane 2, [alpha]WP/m12; lane 3, [alpha]WP/m23; lane 4, [alpha]WP/m13; lane 5, [alpha]WP/m123; lane 6, [alpha]WH/m12; lane 7, [alpha]WH/m23; lane 8, [alpha]WH/m13; lane 9, [alpha]WH/m123.
Specific sequences are required upstream of the AAUAAA for activation of the poly(A) signal
One possible explanation for the difference in the abilities of the various forms of Lamin B2 poly(A) to compete with the [alpha]2 poly(A) signal is that the upstream sequence merely acts as a `spacer' to protect the Lamin B2 poly(A) signal from inhibitory elements in the sequence downstream of the [alpha]2 globin gene. For example, the presence of a cryptic 5" splice site in the sequence flanking the [alpha]2 globin gene could result in inhibition of the inserted Lamin poly(A) signal (see refs 38,39 for examples of this). Thus, the L+101 and L+131 forms of the signal would compete against the [alpha]2 globin poly(A) signal with the highest efficiency simply as a result of the core Lamin poly(A) signal in these fragment being furthest removed from such an inhibitory element; as the amount of upstream sequence is reduced, the poly(A) signal is moved closer to the [alpha]2 globin flanking sequence. To test this possibility a spacer fragment has been introduced 5" of the L+56 and L+10 forms of the Lamin B2 poly(A) signal inserted at both the PvuII and HpaI sites in [alpha]2W3"PS. This has been achieved by first inserting the L+56 and L+10 poly(A) signal fragments into the HincII site of pGEM4, then inserting the PvuII-EcoRI fragment from these pGEM4 derivatives into [alpha]2W3"PS. These modified forms of the Lamin poly(A) signal (referred to as gL+56 and gL+10) carry an extra 63 bp of sequence upstream and an extra 30 bp downstream of the poly(A) signal fragment, and their ability to compete against the [alpha]2 globin poly(A) signal is shown in Figure 4F. In all cases, the introduction of the spacer decreases the frequency with which the poly(A) signal is used, consistent with the fact that the Lamin B2 poly(A) signal has been moved further from the competing [alpha]2 globin poly(A) signal. The gL+56 form is used 85% of the time when inserted at the PvuII site, but only 34% of the time when inserted at the HpaI site, while the gL+10 form of the poly(A) signal is used [sim]26% of the time when present at the PvuII site but only 4% when present at the HpaI site. If the role of the upstream sequence was to simply act as a spacer element, it would be expected that both the gL+10 and gL+56 forms of the poly(A) signal would compete with efficiencies comparable to that seen for the L+101 form. Thus, specific sequences upstream of the AAUAAA of the Lamin B2 poly(A) signal appear to be required to directly stimulate the efficiency of the core poly(A) signal.
Mutagenesis of the upstream sequence
USEs defined for other poly(A) signals appear to show little sequence homology, other than being somewhat U-rich (1), so mutations were introduced into three of the T-tracts upstream of the AAUAAA to determine if these sequences are important for activation of the Lamin B2 poly(A) signal (Fig. 5A). Individually, these changes appear to have little significant effect on the efficiency of the poly(A) signal when tested at both positions downstream of the [alpha] poly(A) signal (Fig. 5B). When the clusters of mutations are combined, however, effects on the efficiency of the poly(A) signal become apparent (Fig. 5C), suggesting redundancy. Changes m1 and m2 together had no effect on the frequency with which the Lamin B2 poly(A) signal was used, when tested at either site, while the combination of mutations m2 and m3 reduced use of the Lamin poly(A) signal to 81% (at PvuII) or 25% (at HpaI). The combination of mutations m1 and m3 had the greatest effect, with the Lamin poly(A) signal used only 73% of the time when present at the PvuII site, and only 17% of the time at the HpaI site. The triple mutation was found not to reduce the efficiency of the Lamin B2 poly(A) signal any further. Thus, it would appear that these sequences are important for activating the Lamin B2 poly(A) signal, although the relatively modest effects confirm that there is considerable redundancy in the upstream sequence elements present. However, it is apparent that the T-tracts do play an important role.
DISCUSSION
Based on the results of the RNase protection analysis, the intergenic region that separates the Lamin B2 and ppv1 genes is significantly shorter than previously thought, with only [sim]280 bp separating the cleavage site of the Lamin B2 poly(A) signal from the 5[prime]-most cap site of ppv1. This genetic arrangement therefore represents one of the smallest intergenic regions so far described. In addition, while two cap sites were mapped for the downstream gene [sim]600 bp downstream of the Lamin B2 poly(A) addition site, it can be seen that there is a cluster of start sites in the region extending both 5" and 3" of the previously mapped start sites. How the upstream start sites relate to this downstream cluster has yet to be determined. It has been suggested that the promoter for the ppv1 gene lies in a CpG island (30), which may explain the heterogeneous nature of initiation. Some of this heterogeneity may also be contributed by the use of alternative splice sites at the 3" end of the first exon of ppv1; inspection of the sequence in this region reveals only three very poor matches to the 5" splice site consensus, which clearly cannot account for all of the protected products obtained with probe 1.
When considering the results of the nuclear run-on analysis of the endogenous Lamin B2 gene it is important to note that the signals obtained are very low, a fact that is apparent when comparison is made of the relative levels of signal over the intergenic region probes and the histone and 5S rRNA control probes (both of which are heavily over-exposed in Fig. 2). However, probe ED, which lies upstream of the poly(A) signal of the Lamin B2 gene gives a relatively strong run-on signal, and once the signal over the intergenic region probes have been corrected, it is apparent the relative levels of the intergenic region probes are comparable to that of probe ED. Such a correspondence might not be expected if the values obtained for the short intergenic probes were too low to be accurately quantitated. Thus, it seems likely that the endogenous run-on signals are representative of the levels of transcription occurring across the 3" end of the Lamin B2 gene and the intergenic region.
The results of the endogenous nuclear run-on analysis show that there are significant levels of transcription upstream of even the most 5" start site of the ppv1 gene (which would lie in probe PE) relative to those seen downstream of the ppv1 promoter. It follows that transcription of the Lamin B2 gene does not terminate prior to the downstream gene. However, given that transcription of the Lamin B2 and ppv1 genes overlap, it is impossible to determine where transcription of the Lamin B2 gene terminates. For this reason, and also to confirm that the low signals seen over the intergenic region are an accurate reflection of the relative levels of transcription, we have performed nuclear run-on experiments using transiently transfected expression constructs. By using the trans-activatable HIV-1 LTR promoter, it has been shown that there do not appear to be any signals capable of terminating transcription present in the intergenic region downstream of the Lamin B2 poly(A) signal. It must be remembered, however, that these transient experiments have employed a heterologous promoter to drive high levels of transcription, and previous work has demonstrated that the activation domains of different transcription factors can have a significant effect on the processivity of the resulting RNA polymerase complexes initiated (40,41). That two different promoters (from HIV-1 and human [alpha]2 globin, data not shown) have been used with the same results may argue against this, but it is a formal possibility that these promoters both initiate ternary complexes that are capable of reading through any termination signals that may be present downstream of the poly(A) signal of the Lamin B2 gene. It remains entirely possible that efficient termination of Lamin B2 transcription is not required, and that expression of the ppv1 gene is not impaired by transcription from the Lamin B2 gene. We suggest that because the Lamin B2 gene is expressed at such a low level and in a cell-cycle-regulated manner, it may simply be the case that polymerase complexes will read into the ppv1 promoter so infrequently that the level of ppv1 initiation will not be affected significantly.
The high signal observed over probe PE in both the endogenous and transient run-ons does not appear to be a result of initiation at the upstream start sites mapped for ppv1 (signal in the transient run-ons over this probe being dependent upon transcription for the HIV-1 promoter) and may be due to pausing of the transcription complex in the intergenic region in response to certain sequences. Such pausing may provide some degree of protection for the ppv1 promoter, since it has been shown that pause sites can give some relief from interference (25), or may allow the efficient off-loading of the components of the cleavage/polyadenylation machinery known to associate with the CTD tail of RNA polymerase II (19).
The link between the efficiency of a poly(A) signal and the efficiency with which transcription terminates has prompted us to examine the poly(A) signal of the Lamin B2 gene. If this signal functions inefficiently potential termination elements upstream of the ppv1 promoter might also be recognised inefficiently. However, the poly(A) signal competition experiments described here demonstrate that this signal can out compete the weak human [alpha]2 globin poly(A) signal when placed downstream, suggesting that the Lamin B2 signal functions with a fairly high efficiency. Despite the presence of a pause element downstream of the poly(A) site, we do not observe any termination in the transient run-ons. This may result from the close spacing of these two elements, since other termination elements have been shown to function only when considerable distances from the poly(A) signal (42). In addition, we have found that sequences upstream of the AAUAAA element are required for the poly(A) signal to function with full efficiency. To date, only one other non-viral mammalian gene, that encoding the human complement C2 gene (31), has been found to possess a USE. Limited mutagenesis of the upstream sequences has focused on three of the runs of uridines, and it has been found that while the single clusters of mutations had no detectable effect in our poly(A) competition assay, significant decreases in the efficiency of the Lamin B2 poly(A) signal were observed with the double and triple mutations. The magnitude of the effect of these changes suggests that there is considerable redundancy in the upstream sequence, and this is consistent with the observation that the progressive removal of the upstream sequence has no effect until sequences within 56 nt of the AAUAAA element are removed. This USE also corresponds to sequences thought to be important for the chromosomal origin of replication that appears to overlap the 3" UTR of the Lamin B2 gene. It is interesting to speculate that evolution of the origin in this region has fortuitously created sequences that also function to stimulate an RNA processing signal. Both of these non-viral mammalian USEs are present in genes that lie in close proximity to a downstream gene transcribed in the same direction (30,43). It is tempting to speculate that the presence of an upstream element in the poly(A) signal serves to maximise the space available in the intergenic region for the promoter of the downstream gene. Consistent with this, it has been shown for the C2 poly(A) signal that the sequences downstream of the cleavage site do not play a role in the efficiency of the site (31). However, deletion of the sequence downstream of the Lamin B2 poly(A) signal has been found to significantly reduce the efficiency of this poly(A) signal.
Numerous examples of upstream elements have been reported for viral poly(A) signals, including SV40 late (44), ground squirrel hepatitis virus (45,46), adenovirus L3 (47,48), and HIV-1 (49-54). These elements do not appear to possess significant sequence similarity, but appear to be discrete elements that can activate heterologous `core' poly(A) signals (e.g. ref. 45). Our studies on the Lamin B2 poly(A) signal have so far been confined to determination of the sequences required for the poly(A) signal to function. However, the results of preliminary in vitro experiments using nuclear extract as well as partially purified cleavage/polyadenylation factors have demonstrated that the upstream sequence increases the efficiency of the cleavage step, and can apparently stabilise the CPSF-CstF complex in the absence of additional bound factors (data not shown). Interestingly, the USEs of both HIV-1 (55) and Infectious Equine Anaemia virus (54) have been implicated in facilitating CPSF loading onto the RNA, and we predict that the Lamin B2 USE functions in a similar manner.
ACKNOWLEDGEMENTS
We are grateful to members of the N.J.P. laboratory for helpful discussions throughout these studies, with particular thanks to Jan Eggermont for developing the nuclear run-on technology to a reliable level. S.B. was funded by a Wellcome Prize Studentship, and these studies were also supported by a Wellcome Programme Grant to N.J.P. (no. 032773/Z/G5).
REFERENCES
1. Wahle, E. and Keller, W. (1996) Trends Biochem. 21, 247-250.
2. Proudfoot, N.J. and Brownlee, G.G. (1976) Nature263, 211-214.MEDLINE Abstract
3. Fitzgerald, M. and Shenk, M. (1981) Cell24, 251-260.MEDLINE Abstract
4. Zarkower, D., Stephenson, P., Sheets, M. and Wickens, M. (1986) Mol. Cell. Biol.6, 2317-2323.MEDLINE Abstract
M. Dalziel, N. M. Nunes, and A. Furger Two G-Rich Regulatory Elements Located Adjacent to and 440 Nucleotides Downstream of the Core Poly(A) Site of the Intronless Melanocortin Receptor 1 Gene Are Critical for Efficient 3' End Processing
Mol. Cell. Biol.,
March 1, 2007;
27(5):
1568 - 1580.
[Abstract][Full Text][PDF]
M. Sachchithananthan, S. J. Stasinopoulos, J. Wilusz, and R. L. Medcalf The relationship between the prothrombin upstream sequence element and the G20210A polymorphism: the influence of a competitive environment for mRNA 3'-end formation
Nucleic Acids Res.,
February 17, 2005;
33(3):
1010 - 1020.
[Abstract][Full Text][PDF]
B. J. Natalizio, L. C. Muniz, G. K. Arhin, J. Wilusz, and C. S. Lutz Upstream Elements Present in the 3'-Untranslated Region of Collagen Genes Influence the Processing Efficiency of Overlapping Polyadenylation Signals
J. Biol. Chem.,
November 1, 2002;
277(45):
42733 - 42740.
[Abstract][Full Text][PDF]
I. J. Orozco, S. J. Kim, and H. G. Martinson The Poly(A) Signal, without the Assistance of Any Downstream Element, Directs RNA Polymerase II to Pause in Vivo and Then to Release Stochastically from the Template
J. Biol. Chem.,
November 1, 2002;
277(45):
42899 - 42911.
[Abstract][Full Text][PDF]
Y. Aissouni, C. Perez, B. Calmels, and P. D. Benech The Cleavage/Polyadenylation Activity Triggered by a U-rich Motif Sequence Is Differently Required Depending on the Poly(A) Site Location at Either the First or Last 3'-Terminal Exon of the 2'-5' Oligo(A) Synthetase Gene
J. Biol. Chem.,
September 20, 2002;
277(39):
35808 - 35814.
[Abstract][Full Text][PDF]
C. Cooke and J. C. Alwine Characterization of Specific Protein-RNA Complexes Associated with the Coupling of Polyadenylation and Last-Intron Removal
Mol. Cell. Biol.,
July 1, 2002;
22(13):
4579 - 4586.
[Abstract][Full Text][PDF]
D. P. Tran, S. J. Kim, N. J. Park, T. M. Jew, and H. G. Martinson Mechanism of Poly(A) Signal Transduction to RNA Polymerase II In Vitro
Mol. Cell. Biol.,
November 1, 2001;
21(21):
7495 - 7508.
[Abstract][Full Text][PDF]
Y.-J. Kim, S. Noguchi, Y. K. Hayashi, T. Tsukahara, T. Shimizu, and K. Arahata The product of an oculopharyngeal muscular dystrophy gene, poly(A)-binding protein 2, interacts with SKIP and stimulates muscle-specific gene expression
Hum. Mol. Genet.,
May 1, 2001;
10(11):
1129 - 1139.
[Abstract][Full Text][PDF]
S. Brackenridge and N. J. Proudfoot Recruitment of a Basal Polyadenylation Factor by the Upstream Sequence Element of the Human Lamin B2 Polyadenylation Signal
Mol. Cell. Biol.,
April 15, 2000;
20(8):
2660 - 2669.
[Abstract][Full Text]
H. Hans and J. C. Alwine Functionally Significant Secondary Structure of the Simian Virus 40 Late Polyadenylation Signal
Mol. Cell. Biol.,
April 15, 2000;
20(8):
2926 - 2932.
[Abstract][Full Text]
L. C. Chao, A. Jamil, S. J. Kim, L. Huang, and H. G. Martinson Assembly of the Cleavage and Polyadenylation Apparatus Requires About 10 Seconds In Vivo and Is Faster for Strong than for Weak Poly(A) Sites
Mol. Cell. Biol.,
August 1, 1999;
19(8):
5588 - 5600.
[Abstract][Full Text][PDF]
C. Cooke, H. Hans, and J. C. Alwine Utilization of Splicing Elements and Polyadenylation Signal Elements in the Coupling of Polyadenylation and Last-Intron Removal
Mol. Cell. Biol.,
July 1, 1999;
19(7):
4971 - 4979.
[Abstract][Full Text][PDF]
J. Zhao, L. Hyman, and C. Moore Formation of mRNA 3' Ends in Eukaryotes: Mechanism, Regulation, and Interrelationships with Other Steps in mRNA Synthesis
Microbiol. Mol. Biol. Rev.,
June 1, 1999;
63(2):
405 - 445.
[Abstract][Full Text][PDF]
H. Lou, D. M. Helfman, R. F. Gagel, and S. M. Berget Polypyrimidine Tract-Binding Protein Positively Regulates Inclusion of an Alternative 3'-Terminal Exon
Mol. Cell. Biol.,
January 1, 1999;
19(1):
78 - 85.
[Abstract][Full Text][PDF]
C. Williams, L. Xu, and T. Blumenthal SL1 trans Splicing and 3'-End Formation in a Novel Class of Caenorhabditis elegans Operon
Mol. Cell. Biol.,
January 1, 1999;
19(1):
376 - 383.
[Abstract][Full Text][PDF]
A. Moreira, Y. Takagaki, S. Brackenridge, M. Wollerton, J. L. Manley, and N. J. Proudfoot The upstream sequence element of the C2 complement poly(A) signal activates mRNA 3' end formation by two distinct mechanisms
Genes & Dev.,
August 15, 1998;
12(16):
2522 - 2534.
[Abstract][Full Text]
CiteULike 
Connotea 
Del.icio.us What's this?
