| Nucleic Acids Research | Pages |
A simple polypyrimidine repeat acts as an artificial Rho-dependent terminator in vivo and in vitro
Introduction
Materials And Methods
Enzymes and chemicals
Plasmids and bacteria
Isolation and analyses of in vivo RNA transcripts
In vitro transcription assays
Results
Effect of (AG/TC)n repeats on transcription in vivo
Rho protein is involved in the termination process
S1 mapping of the termination endpoints
In vitro Rho-dependent termination
The Rho termination endpoints coincide with RNA polymerase pause sites
Discussion
Acknowledgements
References
A simple polypyrimidine repeat acts as an artificial Rho-dependent terminator in vivo and in vitro
ABSTRACT
INTRODUCTION
Transcription termination in Escherichia coli occurs at two types of terminators that differ by their mechanisms and DNA sequences (1). Whereas transcript release at intrinsic terminators is a spontaneous process and involves only RNA polymerase, termination at Rho-dependent termination sites requires the participation of Rho protein. A variety of Rho-dependent terminators have been identified in prokaryotic genomes. However, only a few examples have been analyzed in detail, mostly due to their complexity. For instance, the well-studied Rho-dependent terminator [lambda]tR1 is composed of two distinct but overlapping parts extending over a region of 200 bp. The upstream part, called rut, encodes a segment of RNA to which Rho binds (2-4). Deletion experiments, as well as oligonucleotide competitions (5), have shown that the rut element, and especially two regions designated rutA and rutB, are required for efficient Rho-dependent termination. The downstream part contains several clusters of sites where termination occurs (6,7). Some of these sites have been correlated with the positions where RNA polymerase pauses extensively during transcript elongation (8,9).
Most of our current understanding of Rho function stems from in vitro investigations of the structural and enzymatic properties of the protein. Rho acts as a hexamer of identical subunits, each subunit having specific domains for binding RNA and ATP (10,11 and references therein). According to a widely accepted view, the Rho-dependent termination process involves three steps (for reviews see 12 and 13). The protein first binds to the RNA transcript at a loading site located well upstream from the termination endpoints (14). This binding activates the RNA-dependent ATPase activity, leading to a preferential 5[prime] to 3[prime] translocation of the protein along the RNA towards the transcription elongation complex stalled at a downstream pause site (15-18). Using its ATP-dependent helicase activity, Rho then dissociates the RNA transcript from the ternary complex (19). The RNA transcript, therefore, plays a key role in Rho function by facilitating a productive interaction between the protein and the transcription elongation complex.
Several features of the RNA have been shown to be important for Rho recognition and action (20 and references cited therein). During the initial step of the termination process, Rho acts as an RNA binding protein. The binding site for the Rho hexamer has been defined as a sequence of ~70-80 nt in length that is rich in cytosines and devoid of secondary structure (20-22). The general preference of Rho for cytosine-rich RNAs led to the postulate that the interactions between cytosine residues and the protein are essential for high affinity binding at the loading site (20,23). However, extensive site directed mutagenesis of the rut elements of [lambda]tR1 and trp t[prime] terminators did not reveal any direct correlation between the efficiency of termination and the content or spacing of the cytosine residues within the Rho binding site (24-26). These results suggested that high affinity binding at the loading site is not the only determinant for efficient termination.
The subsequent translocation of Rho depends upon cycles of RNA binding and release that are coupled to rounds of ATP hydrolysis (11,17,18). Using short RNA oligonucleotides as cofactors, Wang and von Hippel (27) have examined in detail the ATPase activity of Rho as a function of the RNA sequence. These studies have shown that mixed (rUrC) oligomers stimulate the ATPase activity of Rho more than do homo-oligomers of rC. The maximal ATPase activity depends on the position of rU residues within the mixed (rUrC) oligomers. Thus, it was suggested that functional interactions between Rho and the transcript that lead to efficient Rho-dependent termination may rely on a combination of tight binding of RNA segments to individual cofactor binding sites within the Rho hexamer and the rate at which these segments are released following ATP hydrolysis. Apparently, in mixed (rUrC) sequences, the rC residues provide strong binding whereas rU residues facilitate the release of RNA during the translocation process (27,28).
In this study, we have tested in vivo the validity of the current models of Rho-dependent transcription termination. We asked the question whether the presence of an (rUrC)n repeat within a transcript can be sufficient to bring about Rho-dependent RNA release at non-terminator DNA sequences. To this end, we have inserted different lengths of the (AG/TC)n DNA repeat between a promoter and the coding region of the chloramphenicol acetyltransferase (CAT) gene. Our results show that transcription of the insert in the orientation encoding the (rUrC)n repeat on the transcript leads to efficient Rho-dependent termination at a downstream site within the CAT gene. The presence of a stretch of 30 nt of the (rUrC)n sequence on the transcript is sufficient to induce the termination process.
MATERIALS AND METHODS
Enzymes and chemicals
Escherichia coli RNA polymerase was obtained from Sigma (France). Restriction enzymes, the Klenow fragment of DNA polymerase I, T4 DNA ligase, T4 polynucleotide kinase and nuclease S1 were from New England Biolabs (Ozyme, France). The enzymes were used according to the manufacturers' specifications. All the chemicals, including the antibiotics, were from Sigma. The unlabelled NTPs, dNTPs and ddNTPs were bought from Boehringer Mannheim and the 32P-labeled NTPs and dNTPs were from Amersham (France). All oligonucleotides were obtained from Eurogentec Belgium.
Plasmids and bacteria
The plasmid vector pKKTac is a derivative of pKK232-8 (28) in which the Tac promoter was inserted at the SmaI site to drive the transcription of the CAT gene. The Tac promoter was subcloned as a 90 bp HindIII-BamHI fragment from pDR540 (Pharmacia). The pAGn and pTCn plasmid derivatives were constructed by inserting the (AG/TC)n repeats in both orientations at the filled-in SalI site of pKKTac. For pAG44 and pTC44, the (AG/TC)22 repeat was subcloned from an existing plasmid (30). For the other derivatives, the inserts were made by ligating two 10mer complementary oligonucleotides (AG)5 and (TC)5. The ligation products were separated on a 10% polyacrylamide gel and the relevant multimers isolated. The lengths and the orientations of the inserts were determined by dideoxy sequencing. The resulting plasmids were designated according to the sequence of the RNA within the Tac-promoted transcript. The p[lambda]tR1 plasmid was constructed by cloning a 233 bp [lambda]tR1 fragment containing rutA, rutB, tI, tII and tIII sequences into the filled-in SalI site of pKKTac. The terminator fragment was prepared from [lambda] DNA by PCR amplification using the primers 5[prime]-AGCCCTTCCCGAGTAAC-3[prime] and 5[prime]-TAGTGCCTCGTTGCGTT-3[prime]. All of the above constructions and the preparation of plasmid DNAs for in vitro transcription were performed in E.coli HB101 according to standard procedures (31). For the in vivo studies, the plasmids were introduced into HB101 F[prime]Iq from our laboratory collection. The F[prime] episome was maintained by growing the cells in the presence of 5 µg/ml of kanamycine. The Tac promoter was induced by addition of IPTG at the final concentration of 1 mM. For the experiments with the Rho mutant, two Salmonella typhimurium isogenic strains were used: the wild type rho+ strain and the rho-111 mutant which has a mutation in Rho coding region that abolishes Rho-dependent termination at non-permissive temperature (37°C). The two strains were from the collection of J. Roth (32) and they were provided to us by L. Bossi (Gif sur Yvette, France).
Isolation and analyses of in vivo RNA transcripts
Total bacterial RNAs were extracted from the cells harboring the plasmids as previously described (33,34). The induction with 1 mM IPTG was done 20 min prior to the cell harvest. Northern blot analyses were performed by fractionating 10 µg of total RNA on a 2% agarose gel containing 6% formaldehyde. The RNAs were transferred onto a nylon membrane filter and hybridized with 32P-labeled DNA probes as follows. The membrane was prehybridized for 4 h at 40°C in a solution containing 50% formamide, 5× SSPE, 2× Denhardt solution, 1% SDS and 100 µg/ml denatured sonicated calf thymus DNA. After addition of the denatured DNA probe, the hybridization was allowed to proceed for 18 h at 40°C. The membrane was first washed at room temperature in 2× SSC plus 0.5% SDS for 10 min followed by two washes at 30°C in 0.5× SSC plus 0.1% SDS for 15 min. The different probes used, the EcoRI-BamHI Tac fragment, the HindIII-PvuII CAT proximal fragment and the PvuII-StyI CAT distal fragment were labeled by random priming.
S1 mapping was performed as previously described (34) with some modifications. Briefly, a single-stranded HindIII-PvuII DNA probe labeled at its HindIII 3[prime] end was mixed with 15 µg of total RNA in 40 µl hybridization buffer (40 mM PIPES pH 6.5, 0.4 M NaCl, 1 mM EDTA and 80% formamide) at 37°C for 15 h. After addition of 225 µl of H2O, 5 µl of 3 M NaCl and 30 µl of 10× S1 buffer (300 mM sodium acetate pH 4.6, 10 mM zinc acetate, 50% glycerol and 200 µg/ml denatured herring sperm DNA), the sample was digested with 160 U of S1 nuclease for 10 min at 37°C. The protected hybrids were precipitated with ethanol and analyzed on a 5% denaturing polyacrylamide gel.
In vitro transcription assays
The in vitro transcription mixture contained, in a total volume of 25 µl of buffer (40 mM Tris-HCl pH 8; 5 mM MgCl2, 1 mM DTT and 150 mM KCl), 0.15 pmol of supercoiled DNA plasmid, 0.4 pmol of RNA polymerase and, when present, 1 pmol of purified E.coli Rho hexamer (the Rho protein was a generous gift from P. von Hippel; its preparation was previously reported; 27). The sample was incubated for 10 min at 37°C to form open complexes. When desired for competition experiments, 8 pmol of the 18mer oligonucleotides AG18, CT18 or AG18/CT18 duplex were added at this time. The elongation reaction was then carried out for 15 min at 37°C in the presence of three unlabeled NTPs (200 µM each) and one labeled NTP (20 µM non-radiolabeled and 60 nM at 3000 Ci/mmol). The reactions were stopped by the addition of a solution containing 40 µg/ml tRNA, 1 mM EDTA and 0.3 M sodium acetate. After phenol extraction, the samples were ethanol precipitated and resuspended in 80% formamide-dyes solution and analyzed on a 6% denaturing polyacrylamide gel. To study the kinetics of RNA elongation we used the same transcription protocol, except that the KCl concentration was 50 mM, and instead of a supercoiled plasmid we used the EcoRI-PvuII restriction fragment from pTC44 as template. The fragment was cut from the plasmid and purified on a 6% native polyacrylamide gel. The transcription reactions were arrested at the indicated times by addition of the stop solution as above.
RESULTS
Effect of (AG/TC)n repeats on transcription in vivo
Figure
Figure 1. Schematic illustration of the relevant part of the plasmid constructs used in this study. The pAGn and pTCn plasmid series differ by the orientation of the (AG/TC)n insert relative to the direction of transcription initiated at the Tac promoter. In addition to the relevant restriction sites, the translation stop codons (TSC), the ribosome binding sites (SD) and the CAT start codon are indicated. The effect of the (AG/TC)n inserts on the expression of the CAT gene in vivo was first monitored by testing, under inducing conditions, the chloramphenicol resistance of E.coli cells harboring the plasmids (results not shown). As for cells carrying the parent plasmid vector pKKTac, chloramphenicol at 100 µg/ml did not alter the growth of cells harboring pTC20 or any one of the plasmids from the pAGn series (pAG44, pAG30 and pAG20). However, under the same conditions, cells transformed with pTC44 and pTC30 were sensitive to the antibiotic. At lower concentrations of chloramphenicol (10-30 µg/ml), the pTC30-transformed cells grow slowly whereas no growth was observed for cells carrying pTC44. These results suggested a possible interference of the insert at the transcriptional or post-transcriptional levels. To determine if the inhibition of CAT expression in pTC44 and pTC30-transformed cells was exerted at the level of transcription, we performed northern blot analyses of the RNAs produced in vivo. Figure Figure 2. Northern blot analyses of in vivo transcripts initiated at the Tac promoter. (A and B) RNAs extracted from E.coli cells harboring the indicated constructs. (C) RNAs extracted from wild type S.typhimurium (lanes 9 and 10) and from a rho-111 mutant (lane 11) transformed with the indicated plasmids. All the cells were grown at 37°C (non-permissive temperature for rho-111 mutant). The blots in (A and C) were hybridized with the HindIII-PvuII DNA probe whereas the blot in (B) was hybridized with the PvuII-StyI DNA probe. Asterisks indicate slight cross hybridization between the probe and 23S rRNA.
Rho protein is involved in the termination process
The effect of the (AG/TC)n sequence on Tac-promoted transcription depends on the orientation and the length of the insert and therefore on the nature of the transcript. Furthermore, the putative termination endpoints are located downstream from the repeat. These facts prompted us to consider the involvement of Rho protein in the termination process. To test this possibility, the transcriptional behavior of the plasmid pTC44 was analyzed in the S.typhimurium rho-111 mutant. The temperature-sensitive rho-111 allele has been well-characterized in vivo and in vitro (32,35,36). At the non-permissive temperature (37°C), the mutant strain is defective in Rho-dependent termination, presumably due to an altered interaction of the mutant protein with the transcript (36). Figure
S1 mapping of the termination endpoints
As a further step to characterize the Rho-dependent RNA release within pTC44 and pTC30 in vivo, we analyzed the termination endpoints by S1 mapping. Cellular RNAs isolated from E.coli cells harboring the plasmids were hybridized to a 3[prime] end-labeled single-stranded DNA probe (the template strand of the HindIII-PvuII DNA fragment) and digested with S1 nuclease as described in Materials and Methods. The data in Figure
Figure 3. S1 mapping of the Rho-dependent termination end-points. RNAs extracted from E.coli cells containing the indicated construct were hybridized with the template strand of the HindIII-PvuII fragment, digested with S1 nuclease and analyzed on a 6% denaturing polyacrylamide gel. A+G and T+C correspond to the Maxam and Gilbert sequencing ladders obtained from the same DNA fragment.
In vitro Rho-dependent termination
We next investigated the effect of the (AG/TC)n DNA repeats on transcription in vitro on supercoiled plasmids. The pAGn and pTCn plasmid series, at native superhelical density, were transcribed by purified E.coli RNA polymerase in the absence or the presence of Rho protein and the resulting transcripts were analyzed by polyacrylamide gel electrophoresis. As shown in Figure
Figure 4. In vitro transcription of pAGn and pTCn plasmids in the absence (-) or the presence (+) of Rho protein. The Rho-dependent termination products generated from the control plasmid, p[lambda]tR1, are designated by I, II and III and correspond to the preferred subsites within the [lambda]tR1 terminator. MW refers to RNA molecular weight markers that have been produced by run-off transcription using plasmids cut at appropriate restriction sites. To provide additional proof that the (rUrC)n sequence on the transcript is involved in the Rho termination activity, we performed transcription experiments in the presence of oligonucleotide competitors. As shown in Figure Figure 5. Effect of AG18 and CT18 oligonucleotides on in vitro Rho termination of pTC44. p[lambda]tR1 and pTC44 were transcribed in vitro in the presence (+) or in the absence (-) of Rho protein. AG18, CT18 and AG18/CT18 duplex oligonucleotides were added just prior to transcription elongation. The arrow indicates the major Rho-dependent termination product of pTC44 and I, II, III the Rho-terminated RNAs of p[lambda]tR1.
The Rho termination endpoints coincide with RNA polymerase pause sites
To determine whether the sites of Rho-dependent RNA release correlate with positions on the DNA template where RNA polymerase experiences extensive pausing, we analyzed the kinetics of RNA elongation in the absence of Rho. In these experiments, the EcoRI-PvuII restriction fragment from pTC44 was used as template in order to avoid the interference of truncated RNAs originating from other promoters carried on the plasmid. Figure
Figure 6. Determination of in vitro transcription pause sites in pTC44 construct. A linear DNA fragment from pTC44 was transcribed in the absence (-) or presence (+) of Rho protein for the indicated time. The arrow shows the major Rho-dependent termination product generated from the supercoiled pTC44 plasmid.
DISCUSSION
In this work, we present in vivo and in vitro data demonstrating that an artificial Rho-dependent terminator is readily generated by the insertion of (AG/TC)n DNA repeats within a transcription unit. Transcription of the inserts in the orientation encoding the (rUrC)n repetition on the transcript leads to Rho-dependent termination within the downstream non-terminator DNA sequences. The Rho termination activity is observed with the plasmids harboring (AG/TC)n inserts of 30 and 44 bp in length and a nearly 100% termination efficiency is obtained in vivo with the 44 bp repeat.
The molecular mechanism by which Rho brings about transcript release at Rho-dependent termination sites is still not completely understood. It is clear, however, that the interactions between Rho protein and the nascent transcript are at the basis of the termination process (12,13). The protein first binds to a segment of the transcript (loading site) encoded by the upstream part of the terminator (rut element). Following cycles of ATP hydrolysis, Rho acts at a distance by causing transcript release downstream from its initial binding site. Within the framework of this model, our results suggest that the (rUrC)n repetition on the transcript serves as an efficient loading site for Rho protein.
It is generally accepted that the initial interaction of Rho and the RNA transcript is predominantly determined by the proportion of cytosine residues within this stretch of unstructured RNA (20,24). Thus, the Rho-dependent termination activity of our constructs is presumably reflecting a stable association of Rho with the (rUrC)n RNA repeat due to the large content of cytosines (50%). The binding site for hexameric Rho has been shown to involve 70-80 contiguous nucleotides of RNA (20-22). Since the size of the repeats in our constructs is smaller, the contribution of RNA flanking sequences to the initial binding cannot be excluded. However, we observed efficient Rho-dependent termination activity when the repeat was inserted within a different sequence context (in the middle of the CAT gene, in the PvuII site). Also, the RNA sequences (up to 40 nt upstream and downstream) surrounding both insertion sites (SalI and PvuII) do not harbor any clusters of cytosines. Therefore, the (rUrC)n sequence on the transcript seems to play the major role at the initial binding. Additional support to this conclusion is provided by the oligonucleotide competition experiments. In effect, to interfere with Rho termination activity, the heteroduplex must block the initial interaction of the protein at the loading site, otherwise it would be disrupted by the Rho helicase activity (5,37).
The high efficiency of termination obtained in vivo with the 44 bp insert (pTC44) cannot be explained exclusively by a tight binding of Rho at the loading site. Previous studies have shown that there is no direct correlation between the efficiency of termination and the content of cytosine residues within the upstream part of the transcript (24-26). In some instances, an increase of the proportion of cytosines even leads to a reduction of termination efficiency. It is possible that the high efficiency of termination is due to the sequence composition of the repeat. As pointed out by Wang and von Hippel (27,28), the (rUrC) RNA motif may allow faster release of RNA from individual binding sites within the Rho hexamer which could facilitate the translocation of the protein towards the transcription elongation complex. Two models of the translocation process have been proposed. In the simple tracking model, the mechanism involves only reactions of RNA binding and release to individual subunits of the Rho hexamer (17). The tethered tracking model, on the other hand, assumes that certain subunits of the Rho hexamer remain bound to the initial loading site while the enzyme moves towards the 3[prime] end of the RNA (13,18). The experiments reported here cannot distinguish between these two models but they set a lower limit on the size of the putative primary RNA binding site of the tethered tracking model. An (rUrC)n repeat of 30 nt is not sufficient to maximally stimulate the termination activity of Rho, but a 44 nt stretch is fully functional. Since the (rUrC)n repeat clearly dictates the efficiency of Rho-dependent termination of transcription in our system, we can exploit this property in future experiments to assess the influence of the sequences at the entry or tethering site.
The Rho-dependent transcript release in our pTC44 construct occurs predominantly at a site located 130 bp downstream from the repeat, both in vivo and in vitro. The sequence of this site is rich in uridine residues (5[prime]-UUUUG). A similar sequence preference was observed at certain release sites within natural Rho-dependent terminators, such as tI and tII of [lambda]tR1. We have considered the possibility that this release site is related to a neighboring structural motif within the transcript. However, sequence analysis with the Mulfold computer program (38) of the region upstream from the site did not reveal any obvious stem-loop structure. We have also analyzed the kinetics of RNA elongation to determine whether the release site corresponds to a natural pause site of RNA polymerase. Our experiments on linear DNA fragments confirm the well-known relation between RNA polymerase pausing and Rho-dependent RNA release (39-41). However, the results indicate that the efficiency of release at a given site is not directly correlated with the extent of pausing. In particular, we found that the high efficiency of release at the site located 130 bp downstream from the repeat is not the result of a strong transcriptional pausing.
Finally, we note that Hart and Roberts (24,25) have previously reported that efficient Rho-dependent terminators can be obtained by replacing a part of the upstream natural region of [lambda]tR1 terminator with relatively long (70-140 bp) DNA repeats encoding (rArC) rich transcripts. Indeed, our results extend these previous studies by showing that the presence of 30-44 nt of the (rUrC)n repetition on the transcript is sufficient to induce Rho-dependent transcription termination within non-terminator DNA context. Therefore, our findings provide for the first time clear evidence that an adequate loading site is the primary determinant for Rho termination activity. We believe that these artificial terminators will be very useful for future investigations aimed at elucidating the mechanism of Rho-dependent transcription termination.
ACKNOWLEDGEMENTS
We are highly grateful to Dr Marc Leng for continuous encouragement, support and advice. We thank Professor P. von Hippel for the gift of purified Rho protein and Dr L. Bossi for providing the S.typhimurium strains. This work was supported in part by la Ligue contre le Cancer, l'ANRS and E.E.C. (Project no. BMH4-CT97-2485).
REFERENCES
This article has been cited by other articles:
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 17 Oct 1998
Copyright©Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
A. Schwartz, E. Margeat, A. R. Rahmouni, and M. Boudvillain
Transcription Termination Factor Rho Can Displace Streptavidin from Biotinylated RNA
J. Biol. Chem.,
October 26, 2007;
282(43):
31469 - 31476.
[Abstract]
[Full Text]
[PDF]
M. S. Ciampi
Rho-dependent terminators and transcription termination.
Microbiology,
September 1, 2006;
152(Pt 9):
2515 - 2528.
[Abstract]
[Full Text]
[PDF]
J. E. Graham
Sequence-specific Rho-RNA interactions in transcription termination
Nucleic Acids Res.,
June 4, 2004;
32(10):
3093 - 3100.
[Abstract]
[Full Text]
[PDF]
This Article 
Abstract
Print PDF (126K)
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (12)
Request Permissions
Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Articles by Guerin, M.
Articles by Rahmouni, A. R.
Search for Related Content
PubMed
PubMed Citation
Articles by Guerin, M.
Articles by Rahmouni, A. R.
Social Bookmarking
What's this?