Effects of heterologous downstream sequences on the activity of the HIV-1 promoter and its response to Tat
Effects of heterologous downstream sequences on the activity of the HIV-1 promoter and its response to TatMichael E. Greenberg1 and Michael B. Mathews1,2,*
1Cold Spring Harbor Laboratory, PO Box 100, Cold Spring Harbor, NY 11724, USA and 2Department of Biochemistry and Molecular Biology, New Jersey Medical School, UMDNJ, Newark, NJ 07103, USA
Received August 18, 1997; Revised and Accepted October 30, 1997
ABSTRACT
In HIV-1 infection, Tat acts at least in part to control transcriptional elongation by overcoming premature transcriptional termination. In some other genes this process is governed by DNA elements called attenuators in concert with cellular transcription factors. To understand the action of Tat more fully and explore its role as an anti-attenuator, we examined the ability of several natural and synthetic attenuation sequences to modulate transcription initiated at the HIV LTR. Fragments containing these signals were inserted downstream of the TAR element in an HIV-CAT chimera and their effects on transcription were assessed both in vitro and in vivo. Runoff transcription assays in HeLa cell extracts demonstrated that the attenuators give rise to premature termination of transcripts initiated from the heterologous HIV-LTR promoter in vitro. When transiently expressed following transfection into Cos cells, however, premature transcript termination at the attenuation site was not observed. Nevertheless, many of the inserted sequences exerted marked effects on CAT gene expression and on transactivation by Tat at both the RNA and protein levels. The nature and magnitude of the effects depended upon the identity of the attenuator and its orientation but only one of 16 sequences tested met the criteria for a Tat-suppressible attenuator in vivo. One other sequence, in contrast, severely reduced Tat-activated transcription without inhibiting basal transcription These results indicate that sequences downstream of the HIV LTR can influence its function as a promoter and its response to Tat transactivation, but lend little support to their role as attenuators in vivo.
Intense efforts to control the AIDS epidemic have sparked a detailed examination of its etiological agent, the human immunodeficiency virus, HIV-1. This has led to the recognition that HIV encodes novel regulatory proteins which profoundly affect the course of viral gene expression in infected cells. One of these, an essential protein known as Tat, is a potent activator of viral transcription from the viral promoter, located in the long terminal repeat (LTR) (1 -3 ). Early in infection, the basal level of transcription initiating at the HIV LTR is low. The binding of Tat to a structured RNA element termed TAR (the transactivation response element), located between nucleotides +15 and +45 relative to the cap site at +1, brings about a dramatic increase in viral transcription from the LTR. It has been suggested that TAR acts as an RNA enhancer element by recruiting Tat in concert with cellular protein cofactors (3 -5 ). The precise mechanisms remain unclear, but evidence exists suggesting that Tat acts by increasing the rate of transcriptional initiation, by stabilizing transcriptional elongation, or through a combination of the two (6 -13 ). Since the LTR is disposed to produce abortive short transcripts (6 ,14 -16 ), the hypothesis has been advanced that Tat might function to overcome the action of an attenuator element (17 ).
Premature transcriptional termination (or attenuation) is increasingly recognized as a means of gene regulation with far-reaching consequences in eukaryotes (18 ), as well as prokaryotes (19 ). Attenuation contributes to the regulation of several eukaryotic proto-oncogenes and viral genes (for reviews, see 20 -23 ). Nucleotide sequences have been defined which act as specific attenuators able to cause pausing and premature termination when placed in the path of the transcription complex (20 ,22 ). Some of these sequences have been characterized with regard to their predicted RNA secondary structures and the specific elongation factors required to promote efficient readthrough by the polymerase complex (20 ,24 -27 ). For example, transcripts from sequences conferring an elongation block from several viruses are predicted to form stable stem-loop structures followed by a run of uridine residues. Both motifs have been shown to contribute to polymerase stalling, while certain oncogene transcripts contain a uridylate-rich tract with no apparent secondary structure (20 ,24 ,27 ,28 ). The conditional transcription termination imposed by these blocks can be overcome by modifying the arrested RNA polymerase complex to an anti-terminating form, in some cases by recruiting cellular elongation factors. Several cellular factors play an essential role in stabilizing the elongation complex; these include transcription factors TFIIF and TFIIS, elongin or SIII, and ELL (29 -31 ). That uncontrolled transcriptional elongation may contribute to certain forms of human cancer is suggested by the recent finding that elongin is negatively regulated by the product of the tumor suppressor gene VHL (32 ).
Tat increases elongation efficiency both in vivo and in a cell free transcription system (6 -8 ,10 ,11 ). In the absence of Tat, there is an accumulation in vivo and in vitro of short transcripts terminating at the 3' border of the TAR element (6 ,15 ,33 ). TAR adopts a stem-loop conformation (34 ) which resembles some attenuators structurally. Furthermore, it can function as an attenuator in vitro, causing polymerase pausing and termination (35 ). By analogy with the bacteriophage [lambda] anti-terminator N, which also binds a promoter-proximal RNA target via an arginine-rich motif (17 ,36 ,37 ), it was postulated that Tat acts by recruiting specific host factors which modify the transcription complexes to a more processive form (38 ). Evidence consistent with this view has been drawn from several studies, of which two in particular reported attenuator readthrough. First, premature termination was elicited in vitro by insertion of a synthetic attenuator, specifying a stable RNA stem-loop structure followed by a run of Us, between the HIV LTR and downstream viral sequences. The addition of purified Tat protein increased transcription through the attenuator (12 ), suggesting that Tat allows complexes to read through heterologous transcription blocks. Second, Tat was shown to allow readthrough of a natural c-myc attenuator in cells transfected with a plasmid containing the TAR element inserted upstream of the promoter-proximal attenuation site in the c-myc gene (39 ).
As an alternative approach toward defining the activity of Tat, we examined its capacity to function as an anti-terminator when heterologous DNA sequences encoding naturally-occurring attenuation signals were placed downstream of the TAR element in the HIV LTR. Since attenuators differ in their structure, properties and interactions with protein factors, implying that they operate by differing mechanisms (24 ,26 ), Tat might be able to overcome some attenuators but not others, permitting inferences to be drawn about the mechanism of Tat's action. We therefore isolated several DNA fragments encoding attenuators that had been characterized in cell-free systems, inserted them individually in both orientations into a plasmid vector between the TAR element and the CAT reporter gene, and assayed for premature transcription termination in vivo and in vitro. Although the attenuators functioned in vitro when placed downstream of the HIV-1 promoter, technical limitations precluded the simultaneous observation of both the Tat response and attenuation in the cell-free reaction. Transfection experiments in Cos cells showed that several of the inserts affect HIV transcription (either basal or Tat-activated, or both), but suppression of attenuation in vivo was not generally observed.
LTR-CAT constructs were derived from a previously described plasmid, pH2 (15 ). Minimal DNA sequences reported to confer attenuation in vitro were excised from the viral and cellular genes listed in Table 1 and HindIII linkers were attached. To construct the SV40 attenuator-containing plasmid, the SV40 containing vector pSV010 (40 ) was subjected to 10 cycles of PCR using flanking primers, d(CACAAGCTTTCAGGCCATGGTGC) and d(CACAAGCTTGTCAACAGTATCTTCCC), which span the attenuation sequence with HindIII sites (underlined) added near their termini for cloning purposes. The HindIII fragments were ligated into the unique HindIII site at position +79 of pH2. Plasmids carrying inserts in both sense and antisense orientations were characterized by sequence analysis. Plasmids containing the attenuator encoding inserts are named according to the gene from which the insert was subcloned, with the suffix `/r' designating reverse orientation. For synthesis of RNA probes, segments from the pH2 derivatives were subcloned into pGEM-1 or pGEM-3 vectors (Stratagene). The HIV LTR-attenuator-CAT sequence, from the ScaI site at -135 to the EcoRI site at +329, was inserted into the EcoRI-SmaI window of the pGEM polylinker region (Fig. 1 ). pGEM PCNA CAT (41 ) was used to synthesize an RNA probe for protecting downstream CAT sequences.
Calcium phosphate transfections were performed as described previously (9 ). Transfection mixtures contained 3 µg of attenuator plasmid and 0.5 µg of a control plasmid p[alpha]A+C (42 ). Tat was generated by inclusion of 1.0 µg of pCMV-Tat in the transfection mix. The DNA was made up to 20 µg with herring sperm DNA and overlayed onto 80% confluent monolayers of Cos cells in 60 mm tissue culture dishes. At 16 h post-transfection, cells were washed twice with fresh medium then overlayed with 5 ml of medium. Cells were incubated at 37°C.
Total RNA was harvested at 48 h post-transfection by the RNAzol method (43 ). Briefly, cells were washed twice on ice with PBS and scraped into 0.5 ml RNAzol made up of water-equilibrated phenol (USB ultrapure), guanidinium thiocyanate and sarcosyl. RNA was subsequently precipitated in isopropanol, redissolved in water, extracted repeatedly with phenol:chloroform (1:1) and reprecipitated. RNA concentration was determined by A260 measurement. Antisense RNA probes were synthesized with T7 RNA polymerase and [[alpha]-32P]UTP from pGEM vectors linearized with SalI, and purified on sequencing gels. To perform the RNase protection assay, 5 µg RNA was hybridized overnight with 5 × 105 c.p.m. of RNA probe at 40°C in 30 µl 80% formamide, 40 mM PIPES, pH 6.7, 400 mM NaCl and 1 mM EDTA. Single-stranded RNA was then digested with 10 µg/ml RNase T1 in 300 µl 10 mM Tris, pH 6.7, 400 mM NaCl and 1 mM EDTA for 1 h at room temperature, SDS and proteinase K were added to 0.5% and 167 µg/ml, respectively, and incubation was continued for 15 min at 37°C. RNA was extracted with phenol:chloroform, precipitated with ethanol, and analyzed by electrophoresis in 8% polyacrylamide-7 M urea sequencing gel. Protected RNA fragments were detected by autoradiography using an intensifying screen. Quantitation was performed with a Fuji phosphorimager.
Transcription was performed in HeLa cell nuclear extract according to previously described methods (24 ). Briefly, 10 µl of nuclear extract (80-100 µg protein), linearized DNA (0.125 µg) and 4 mM creatine phosphate were incubated for 30 min at 30°C in a final volume of 15 µl. Following pre-initiation, ATP, CTP and GTP were added to a final concentration of 62.5 µM each together with 20 µCi of [[alpha]-32P]UTP in a volume of 5 µl and pulse labeling was conducted for 1.5 min. For the elongation step, 2 µl of each of the four rNTPs (final concentration 0.6 mM) were added and the reaction was allowed to proceed for an additional 30 min at 30°C. Reactions were stopped by addition of 100 µl of 20 mM EDTA, 0.2% SDS, and 200 µg/ml proteinase K and 30 µg tRNA. After incubation at 65°C for 15 min, RNA was extracted with phenol:chloroform and with chloroform, and then was precipitated with ethanol. Transcription products were analyzed in 6% polyacrylamide-7 M urea gels.
aNucleotide number according to numeration in cited reference.
bLength in nucleotides between the HIV start site and the attenuation site published in the literature.
It has been suggested that Tat converts RNA polymerase II transcription complexes initiating at the HIV promoter into a more processive form (3 ), possibly through recruitment of novel host factors. Consistent with such a mode of action, Tat has been shown to augment readthrough of attenuation signals in two cases (12 ,23 ). Several types of attenuation signal have been identified in cellular and viral genes through cell-free transcription studies. Some of these attenuator sequences caused premature transcriptional termination in vitro when placed under the control of a heterologous promoter such as the adenovirus major late (AdML) promoter (24 ,27 ,44 ). We have examined the ability of such sequences, from viral and cellular sources, to attenuate transcription from the HIV LTR in vivo, and have tested their influence on transactivation by Tat.
With the identification of cellular factors recruited by Tat as an objective, we endeavored to extend the finding that Tat can promote transcriptional readthrough of an artificial terminator in vitro (12 ,39 ). To this end, we inserted downstream of the TAR element in the HIV-1 LTR various DNA fragments that have been reported to act as attenuators. The resulting constructs were transfected into Cos-1 cells in the presence or absence of a plasmid expressing the first exon of Tat. Since most of these sequences had previously been characterized as attenuators only in cell-free systems, our initial goals were 2-fold: first, to observe their effect on transcription in vivo; and, second, to test the ability of Tat to impel readthrough of the attenuators. We expected to observe varying degrees of attenuation in vivo with the different attenuators, and hoped that varying degrees of Tat-promoted elongation through the attenuator sequences would throw light on the factors and mechanisms involved.
The sequences studied had all been reported to serve as attenuators of cell-free transcription but most of them had not been tested when placed downstream of the HIV LTR. Cell-free transcription experiments confirmed that several of these subcloned sequences led to the formation of the expected attenuated RNAs initiated from the HIV promoter. Unfortunately, we were unable to determine whether Tat overcame attenuation in vitro because the specialized conditions required to induce the appearance of the attenuated transcripts proved to be incompatible with the Tat response. In transient expression assays, on the other hand, none of the sequences reported to function in vitro behaved as an attenuator although many of the inserts markedly affected transcription as summarized in Figure 5 . Ironically, evidence for attenuation in vivo was obtained only in the case of ODC2/r, in which the sequence is inverted relative to the orientation reported to function as an attenuator in vitro. Moreover, Tat was strikingly effectively in overcoming the inhibitory effect of ODC2/r on RNA readthrough in vivo. It should be noted, however, that short RNA transcripts mapping to the expected site of premature termination were not detected (data not shown), perhaps because they are labile in vivo.
Notwithstanding the disappointing behavior of these sequences as attenuators in vivo, the placement of inserts downstream of TAR exerted remarkably variable effects upon the basal readthrough levels of CAT RNA and on transactivation by Tat (Fig. 5 ). In comparison with the parent pH2 plasmid, the effects of the inserts can be grouped as follows: (i) decreased basal transcription but no effect on transactivated levels (ODC2/r); (ii) decreased basal and transactivated levels (AdMLP, AdMLP/r, myc, fos); (iii) unchanged basal and transactivated levels (GC, MVM, ODC2); (iv) increased basal and transactivated levels (ODC1, ODC1/r); and (v) increased basal levels and decreased transactivated levels (SV40).
The ODC2/r insert was the only one to display attenuator properties: it yielded the greatest degree of inhibition of downstream readthrough RNA (~3-fold), while ODC2 had no significant effect. In the presence of Tat, neither insert had a dramatic effect. Consequently, ODC2/r evinced exceptionally high transactivation whereas ODC2 transactivated to the same extent as pH2. Thus ODC2/r behaved as expected for a true Tat-suppressible attenuator, even though runoff transcription in vitro did not give rise to a detectable attenuated band or to decreased RNA yield (data not shown). The ODC2/r sequence contains a cluster of 30 adenine residues within a 35 nt stretch. Adenine rich sequences have been shown to be intrinsic attenuators in prokaryotes (17 ).
The adenovirus inserts, AdML and AdML/r, brought about an ~2-fold decrease in both basal and Tat-transactivated readthrough transcription (Fig. 2 B). The orientation-independence of the effect suggests that a DNA element embedded within the cloned AdML fragment may function as a general transcriptional downregulator. While transcription repressors are commonly thought to function during initiation, it is conceivable that DNA binding factors may function downstream to inhibit transcription and depress RNA levels. Attenuator sequences have been characterized as RNA elements due to their orientation specificity in vitro and by base substitutions that destabilize the RNA stem-loop structure leading to reduced attenuation activity (20 ). A DNA element, the inducer of short transcripts (IST) directs formation of poorly elongating transcription complexes (16 ) and it is possible that a functionally equivalent sequence is present within the AdML insert.
The myc and fos inserts were mildly inhibitory in the absence of Tat and distinctly inhibitory in its presence. Consequently the fold transactivation by Tat was reduced by these inserts. These effects were not observed when the inserts were reversed, and were largely if not entirely due to the presence of a donor splice site as evidenced by RNase protection and mutagenic data. Our observation that the short RNA species from the myc and fos transfections arise from a splicing event and not premature termination complicates the interpretation of some previous studies which map the attenuation sites in the myc gene adjacent to the exon 1/intron 1 border (27 ,28 ), but other experiments imply that promoter-proximal pausing can occur even when the splice site is deleted (51 ).
Surprisingly, neither the MVM attenuator nor the synthetic terminator sequence GC, both characterized as effective attenuators in vitro, gave substantial effects in our transfection assays. Although the GC insert is identical to the one used by Graeble and co-workers (12 ), the TAR and GC elements are closer to one another in our constructs than in theirs (20 nt downstream from the 3' end of the TAR stem-loop compared to 120 nt in pMAG-10). This difference is unlikely to provide an explanation for the discrepancy, however, since a short RNA corresponding to the 3' end of the termination sequence was observed in the cell-free transcription assay in the presence of high salt (data not shown).
The ODC1 insert enhanced transcription independent of orientation and of the presence of Tat, suggesting that it contains a stimulatory DNA element. The high basal levels obtained for the ODC1 inserts may be explained by the two upstream stimulatory factor (USF) consensus binding sequences located within the ODC1 attenuator sequences (45 ).
The SV40 insert gave rise to a modest increase in basal transcription yet a significant decrease in Tat-activated transcription. Transactivation was reduced to its lowest level (2-fold), implying that Tat-modified transcription complexes may be less able than normal complexes to pass through the SV40 attenuator. Alternatively, the SV40 sequence may render transcription complexes formed at the HIV-1 promoter more refractory to Tat. Earlier studies indicated that the secondary structure of the SV40 RNA attenuator element influences the elongation properties of the polymerase in vitro (47 ,52 ). Therefore it is possible that Tat binding may stabilize the attenuation conformation or block access of a protein required to maintain the readthrough conformation. Use of the GCG program `MFold' to generate alternate RNA conformations of TAR followed by the SV40 attenuator sequence supports this conjecture. Proper folding of the TAR element allows the creation of the SV40 attenuation conformation as described by Hay and Aloni, but the 3' stem of TAR can hybridize with downstream nucleotides to disrupt this attenuation conformation. Tat's interaction with TAR presumably stabilizes the TAR structure, thus allowing for formation of the SV40 attenuator conformation and resulting in a greatly reduced Tat effect.
Why did we not observe attenuation in vivo? Experimentally, attenuation is inferred from the detection of short, prematurely-terminated transcripts or by monitoring a decrease in readthrough transciption levels. We employed both methods to assay for attenuation in our constructs, raising the possibility that the inserted sequences do not function as attenuators in vivo. Before drawing this conclusion, however, several caveats must be considered. First, the failure to detect short transcripts might be related to the kinetics of transcriptional pausing and RNA stability. We analyzed the steady-state levels of RNA accumulated during an extended period after transfection: while RNA polymerase complexes presumably stall throughout this time, it is likely that not all prematurely released transcripts are sufficiently stable to persist in the cell. Although Blair et al. (53 ) were able to detect such truncated RNAs using a large insert specifying a highly organized structural element (the EMCV IRES), other RNA sequences 3' of TAR may be degraded. Indeed, 3' processing leaving a hairpin ending at the 3' base of the TAR structure has been observed both in vivo and in vitro (15 ), and such transcripts were detected in all of the transfection experiments reported here (see Fig. 3 for example). Second, special conditions may be required to enhance the rate of attenuation in the cell, as in cell-free systems where detergent or high salt concentrations are routinely used to induce the appearance of attenuated transcripts. For example, treatment of HL-60 cells with retinoic acid, an inducer of differentiation, results in a downregulation in full-length RNA levels and a 10-fold increase in attenuated RNAs (21 ). Appropriate conditions might be missing in the transfection assays used here. Third, the nature of the promoter may have an influence. Although several of the inserts gave rise to short RNAs of the expected sizes in vitro when placed downstream of the HIV promoter (data not shown), attenuation may be largely promoter dependent in the cell (51 ). A comparison of our findings with those of Wright et al. (39 ) is consistent with this possibility. When these authors placed the TAR element downstream of the myc P2 promoter, they detected an accumulation of polymerase complexes upstream of the myc attenuator sequence that was alleviated by transfection of Tat. When the myc attenuator sequence was installed downstream of the HIV promoter, on the other hand, we failed to detect attenuation.
Finally, our observations underscore the role of downstream sequences in determining transcription levels. Both basal and Tat transactivated levels varied over a >20-fold range depending on the nature of the inserted downstream sequences. Certain DNA binding proteins function as architectural components to load or recruit transcription factors and basal components. The DNA sequences within the subcloned fragments may also block access of elongation factors, such that RNA levels are altered. Recently, local chromatin structure has been identified as an important factor affecting transcription levels (54 ). The variable transcription levels of our chimeric templates are worthy of further investigation from these perspectives.
This paper is dedicated to the memory of Yosef Aloni, in whose laboratory at the Weizmann Institute of Science, Israel, this study was begun. We thank Naomi Ragimov, Tsafrira Pe'ery and Ruhama Leizerowitz for help in the initial stages of this project. M.B.M. held a Meyerhoff Visiting Professorship at the Weizmann Institute and was a Fellow of the J.S. Guggenheim Foundation. The work was supported by NIH grant AI31802 and ACS grant SG-199.
55 Resnekov,O. and Aloni,Y. (1989) Proc. Natl. Acad. Sci. USA,86,12-16.MEDLINE Abstract
56 Tooze,J. (1981) in Tooze,J. (ed.), Molecular Biology of Tumor Viruses. Cold Spring Harbor, 2nd Ed., p. 948.
*To whom correspondence should be addressed at: Department of Biochemistry and Molecular Biology, New Jersey Medical School, UMDNJ, 185 South Orange Avenue, Newark, NJ 07103, USA. Tel: +1 973 972 4411; Fax: +1 973 972 5594; Email: mathews{at}umdnj.edu
C. Mischiati, G. Feriotto, M. Borgatti, P. Giacomini, and R. Gambari Characterization of a Major Histocompatibility Complex Class II X-Box-Binding Protein Enhancing Tat-Induced Transcription Directed by the Human Immunodeficiency Virus Type 1 Long Terminal Repeat
J. Virol.,
October 1, 2000;
74(19):
8989 - 9001.
[Abstract][Full Text]
CiteULike 
Connotea 
Del.icio.us What's this?
