Skip to main content
NCBI home page
Nucleic Acids Research logoLink to Nucleic Acids Research
. 2000 Feb 15;28(4):901–910. doi: 10.1093/nar/28.4.901

Multiple copies of the Mason–Pfizer monkey virus constitutive RNA transport element lead to enhanced HIV-1 Gag expression in a context-dependent manner

Harald Wodrich 1, Axel Schambach 1, Hans-Georg Kräusslich 1,a
PMCID: PMC102582  PMID: 10648781

Abstract

Retroviral gene expression requires nuclear export and translation of incompletely spliced RNA. In the case of human immunodeficiency virus (HIV), this is facilitated by the viral Rev protein binding to its cognate RNA response element (RRE), while other retroviruses contain constitutive transport elements (CTE) binding to cellular factors. These CTE can substitute for the HIV-1 Rev/RRE system, albeit with reduced efficiency. Here, we show that multimeric copies of the CTE restore HIV-1 protein expression to levels comparable to or higher than Rev/RRE in various cell lines from different species. We suggest that multimerization of export factors is important for CTE function, as reported for Rev. CTE function was not affected when the element was displaced from its natural position close to the poly(A) signal, while insertion of an intron into the 3′-untranslated region (3′-UTR) severely reduced CTE activity. In this case, cytoplasmic RNA degradation was observed, which may be mediated by nonsense-mediated RNA decay. In contrast, Rev-dependent gene expression was insensitive to an intron in the 3′-UTR. Finally, we show that the putative CTE-binding protein RNA helicase A is not specifically translocated into the cytoplasm upon overexpression of CTE-containing RNA.

INTRODUCTION

Most retroviruses synthesize only a single primary transcript from their integrated proviral genome. Balanced expression of retroviral gene products therefore requires regulated post-transcriptional modification of this pre-mRNA. A key step in this process is the export of intron-containing RNAs, which would normally be retained in the nucleus. All retroviruses, however, use their unspliced primary transcript as genomic RNA and as mRNA for the inner structural proteins of the virion (Gag proteins) and for the replication enzymes. Consequently, retroviral replication requires the balanced nuclear export of spliced and unspliced RNA, while cellular RNAs are normally not exported unless all intronic sequences have been removed (1,2). Retroviruses have evolved different mechanisms to direct the nuclear export of intron-containing viral RNAs. These consist of cis-acting RNA elements which interact either with cellular proteins or with viral trans-acting export factors.

The best understood RNA export system is that of the complex retrovirus human immunodeficiency virus (HIV-1). This system consists of a trans-acting viral protein (Rev) which facilitates export through binding to its RNA recognition sequence (Rev response element, RRE; summarized in 3,4). The primary transcript of the HIV-1 genome is a 9 kb RNA, which is differentially spliced to a class of partially spliced 4 kb transcripts and a class of 2 kb transcripts (5,6). The latter contain no removable intronic sequences and are thought to be processed by the normal cellular splicing and export machinery. All transcripts which retain introns, on the other hand, contain the RRE within the env open reading frame and are exported through binding to the Rev protein (summarized in 3,4). Rev is a 116 amino acid phosphoprotein which is produced from a fully spliced transcript and whose expression is therefore not dependent on its own function. Rev contains a nucleic acid-binding domain, which also serves as a nuclear import signal (NLS) (7,8) and a leucine-rich nuclear export signal (NES) (9). This signal was the first of a class of similar NES sequences observed in various viral and cellular proteins (summarized in 4). The NES binds to the cellular export factor Crm1, a member of the superfamily of importin-β-like proteins (10–12; summarized in 13). Crm1 binding targets the RRE-containing RNAs to the cellular export machinery and leads to efficient nuclear export of incompletely spliced HIV-1 RNAs. Rev therefore appears to function as a bridging factor between viral RNAs and a cellular export system.

Rev/RRE or related systems are only found in complex retroviruses, while simple retroviruses encode only the viral structural proteins and replication enzymes, but no regulatory factors like Rev. These viruses also need to export incompletely spliced RNAs from the nucleus to permit viral replication, but the mechanism of nuclear export is less well understood. A subgroup of simple retroviruses, the simian D-type retroviruses Mason–Pfizer monkey virus (M-PMV) and simian retroviruses 1 and 2 (SRV-1/2), contain a cis-acting RNA export element in the 3′-part of their genome, located between the env region and the 3′ long terminal repeat (LTR) (Fig. 1; 14,15). These elements have been termed constitutive transport elements (CTE) (14). The presence of the CTE within the unspliced genomic transcript of M-PMV is necessary for cytoplasmic accumulation and expression of this RNA, but is dispensable for expression of the spliced env message (16). Displacement of the CTE from its natural position at the 3′-end of the viral genome abrogated its function and suggested a role for the polyadenylation signal in CTE function (17). The CTE corresponds to an incomplete inverted direct repeat sequence which folds into a stable stem–loop structure. This structure contains two internal loop regions and the nucleotide sequence of these loops is essential for CTE function, making them likely candidates as docking sites for cellular export factors (16,18–20).

Figure 1.

Figure 1

Open in a new tab

Analysis of Rev-independent HIV-1 Gag expression. (A) Schematic diagram of plasmids pK-R-gpII, pK-gpII, pK-gpII-C and pK-gpII-CCCC. All plasmids contain part of the 5′-UTR and the coding region for the Gag and PR domains of HIV-1. Relevant RNA transport elements, either the RRE of HIV-1 or the CTE of M-PMV, are depicted as black boxes. All plasmids contain the SV40 large T antigen splice and poly(A) signal (pA). At the top, the position of the gag/PR region and of the RRE as well as the coding exons of Rev are highlighted in the HIV-1 genome. At the bottom, the relative position of the CTE is highlighted in the M-PMV genome. (B) Immunoblot analysis of transfected HeLa cells using antiserum against HIV-1 CA. Cells were transfected with plasmids pK-gpII (lane 1), pK-R-gpII (in the absence or presence of Rev; lanes 2 and 3), pK-gpII-C [CTE in the antisense (a.s.) or sense orientation; lanes 4 and 5] or pK-gpII-CCCC (lane 6). The Gag polyprotein (Pr55) and the CA protein are identified on the right; molecular mass standards (in kDa) are given on the left. The two additional proteins migrating between Pr55 and CA in lane 3 correspond to the Gag intermediate cleavage products MA-CA-NC and MA-CA.

It has been shown that the CTE can efficiently promote the nuclear export of intron-containing RNAs encoding reporter proteins and of HIV-1-derived RNAs, normally dependent on Rev/RRE (14,18,19,21–24). When inserted into an intronic sequence, the CTE was able to direct the export of the intron itself without inhibiting the splice reaction in microinjected Xenopus oocytes (25). Furthermore, the Rev/RRE system of HIV-1 or SIV could be functionally replaced by the CTE in the context of a complete proviral clone, leading to the production of infectious virus, albeit at a reduced titer (14,15). Although the CTE can functionally substitute for Rev/RRE, the two elements make use of different nucleocytoplasmic transport pathways. The Rev/NES-mediated export competes with the nuclear export of 5S rRNA and U snRNA, whereas the CTE competes with mRNA export (20,25). This competition is most likely due to the binding of a cellular factor(s) present at limited concentration. To date, two cellular proteins have been suggested to be specific CTE-binding factors and to play a role in CTE-mediated nucleocytoplasmic transport. In vitro crosslinking experiments identified RNA helicase A (RHA) as a specific CTE-binding protein (26), although this result could not be reproduced in a second study (20). RHA was shown to be a shuttling protein and overexpression of CTE-containing RNAs resulted in its translocation into the cytoplasm, suggesting that RHA is part of the functional CTE export complex (26). The second CTE-binding protein is the Tip-associated protein Tap (27,28). Tap has been shown to bind to the wild-type CTE, but not to non-functional mutated elements, and to have nucleocytoplasmic shuttling capacity. Tap promotes CTE-dependent export in microinjected Xenopus oocytes (28–30). Furthermore, overexpression of Tap can rescue CTE function in the non-permissive quail cell line QCL-3, suggesting that Tap is indeed the limiting cellular factor for CTE export (31). Besides RHA and Tap, the CTE appears to bind additional proteins of ~190 and 85 kDa, which have not been identified to date (20).

Although the CTE can functionally substitute for Rev/RRE, CTE-mediated expression was consistantly lower than that directed by Rev (14,15,18,22,23,32). Significantly lower expression levels were also observed in studies where the HIV-1 Gag-Pol proteins were produced in a CTE-dependent way (24). Furthermore, CTE function has been reported to be influenced by the position of the element on the RNA (17). In this study, we report that multimer copies of the CTE can direct the expression of HIV-1 Gag proteins at a level equivalent to or higher than those achieved by the Rev/RRE system. We also show that CTE function is not position dependent, while the presence of an intron in the 3′-untranslated region (3′-UTR) of the RNA abrogates CTE but not Rev function. Finally, we provide evidence that RHA is not specifically translocated into the cytoplasm upon overexpression of CTE-containing RNA, making RHA an unlikely candidate as a CTE export factor.

MATERIALS AND METHODS

Plasmids

Construction of the eukaryotic expression vector pK-R-gpII for the HIV-1 Gag polyprotein has been described before (33; Fig. 1). Briefly, this plasmid contains part of the 5′-untranslated region (5′-UTR), the complete gag gene and the protease (PR) coding region of the pol gene of HIV-1 strain BH-10 (34), as well as the RRE under the control of the strong cytomegalovirus (CMV) promoter/enhancer and SV40 splice and polyadenylation sequences. The HIV-1-specific sequences were excised using flanking EcoRI sites and cloned into pBluescript SK+ to give pBSG (Bluescript gpII). The M-PMV CTE sequence (nt 8006–8175) was PCR amplified from the molecular clone SIVMPCG (GenBank accession no. M12349), kindly provided by E. Hunter. Primers were: 5′-CTE (GCTCTAGAGCCAGATAGGCCC) introducing a XbaI site (underlined) and 3′-CTE (AAAGCTTGCTAGCTGATCAACACATCCCTCGGAGGC) introducing NheI and BclI sites (underlined). The amplified sequence was cloned as a XbaI–NheI fragment into pBSG, which yielded plasmids containing a single CTE in the sense or antisense orientation as well as plasmids containing multimer copies of the CTE. The relevant regions were excised as SalI–NotI fragments and cloned into pKex (35), resulting in the CTE-containing expression vectors pK-gpII-C (sense and antisense) and pK-gpII-CCCC (containing four copies of the CTE) (Fig. 1).

To exchange the 3′-UTR and polyadenylation signal, the complete expression cassette including the respective RNA element was excised as a ScaI–NotI fragment from plasmid pK-gpII and derivatives and introduced into the expression vector pcDNA3 (Invitrogen), which contains the polyadenylation signal of bovine growth hormone (BGH pA) (Fig. 2A). The resulting plasmids were termed 3-(–) (without element), 3-RRE, 3-C, 3-Ca.s. and 3-CCCC (Fig. 2A). To introduce an intron into the 3′-UTR of plasmid 3-C, the rabbit β-globin intron of the expression vector pGRE5 (Amersham) was PCR amplified using the following primers: 5′-BG Intron (GGCTCGAGCG-CGCTGAGAACTTCAGGGTGAGTTTGGGG) introducing XhoI and BssHII sites (underlined) and 3′-BG Intron (GGCTCGAGCCATGGCGCGCTCTTTGCCAAAATGATGAGACAGC) introducing XhoI, NcoI and BssHII sites (underlined). The PCR fragment was digested with XhoI and cloned into plasmid 3-C, linearized with the same enzyme. The resulting plasmids contained the intron in the sense and antisense orientations, respectively.

Figure 2.

Figure 2

Figure 2

Open in a new tab

Multimer copies of the M-PMV CTE lead to enhanced Gag expression. (A) A schematic diagram of plasmids 3-RRE, 3-(–), 3-C and 3-CCCC is shown at the top. These plasmids contain the bovine growth hormone poly(A) signal (BGH pA) instead of the SV40 splice and poly(A) signal. Immunoblot analysis of HeLa cells transfected with the specified plasmids and analyzed with an antiserum against CA is shown below. HIV-1-specific products and marker proteins are identified as in Figure 1. (B) Northern blot analysis of transfected HeLa cells. Cells were mock-transfected or transfected with plasmids 3-RRE (±Rev), 3-C and 3-CCCC as indicated above the lanes. Nuclear (N) and cytoplasmic (C) RNA fractions were prepared from the cell lysate and subjected to northern blot analysis. Gag-specific RNA was detected with a radiolabeled probe corresponding to the complete gag/PR coding region (upper panel). To validate the quality of the fractionation, blots were rehybridized with a probe against U6 snRNA which is restricted to the nucleus (lower panel). The main gag-specific product is identified on the right; a slower migrating gag-specific RNA product observed in lanes 7–10 is marked with an asterisk. This RNA is most likely produced by read-through of the poly(A) signal, giving a longer transcript also containing the neomycin resistance gene as shown at the bottom. The migration of RNA standards is depicted on the left.

To generate retroviral vectors for HIV-1 Gag expression, the HIV-1-specific fragment was cloned into the multiple cloning site of plasmid MP 110 to give MP 110 gpII. Plasmid MP 110 is a derivative of the retroviral vector SF 10 (36), but contains the LTR regions of the murine myeloproliferative sarcoma retrovirus (MPSV). The M-PMV CTE was cloned into MP 110 gpII as a BamHI–HindIII fragment to give MP 110 gpII-C. MP 110 gpII-CCCC was made by excising the four CTE copies from 3-CCCC with NotI and EcoRI, filling the NotI site by the Klenow reaction and ligating the fragment into MP 110. Subsequently the HIV-1-specific sequence was introduced as an EcoRI fragment.

Cells and transfection

HeLa P4 cells (37) stably expressing the CCR5 co-receptor (kindly provided by M. Alizon, Institut Cochin de Genetique Moleculaire, Paris, France) were maintained in Dulbecco’s modified Eagle’s medium supplemented with 10% heat-inactivated fetal calf serum, 100 U/ml penicillin, 100 µg/ml streptomycin and 2 mM glutamine. For transient transfection, 5 × 105 cells were plated on a 10 cm dish and grown overnight. On the day of transfection, the medium was exchanged and 10 µg of vector DNA as well as 1 µg of a reporter plasmid encoding the enhanced green fluorescent protein (eGFP) were transfected using the modified calcium phosphate co-precipitation technique (38). To determine the transfection efficiency, coverslips were included in the transfection and were fixed for 10 min with 4% paraformaldehyde in phosphate-buffered saline (PBS) at 48 h post-transfection. Transfection efficiency was determined by counting GFP-positive cells. Alternatively, unfixed cells were analyzed for GFP fluorescence by FACS.

For retroviral transduction experiments, vector particles were generated by transient transfection of the amphotropic Phoenix packaging cell line (39) with 8 µg of the retroviral vector MP 110 and its derivatives or of the GFP-expressing vector MP 110 eGFP. Prior to transfection, chloroquin was added to the medium to a final concentration of 25 µM. Twenty-four hours post-transfection the medium was replaced. Forty-eight hours post-transfection, the medium was collected, filtered through a 0.45 µm filter and used to infect 5–10 × 104 target cells. Transduction was performed in the presence of 8 µg/ml polybrene to improve transduction efficiency. Transduction experiments were performed with human SW 480 cells, murine NIH 3T3 cells, canine Cf2th cells and the monkey cell line Cos7, all cultured in the medium described above. The vector titer was determined to be ~3 × 105/ml using MP 110-Neo vector particles produced from the Phoenix packaging cell line and analyzed by colony formation on NIH 3T3 cells.

Immunofluorescence staining and microscopy

Cells grown on coverslips were fixed with methanol/acetone (1:1) at –20°C for 10 min and air dried. Staining of HIV-1 Gag proteins was performed with undiluted supernatant from hybridoma 183, recognizing the HIV-1 capsid (CA) protein (40), followed by FITC-labeled anti-mouse IgG (DAKO), diluted 1:100 in PBS containing 10% heat-inactivated calf serum. Antiserum against RHA was kindly provided by F. Grosse and was used at a dilution of 1:100, followed by Texas red-labeled anti-rabbit IgG (1:100) (Dianova). Immunostaining was performed for 1 h in a 37°C moist chamber for each antibody and cells were analyzed on a Zeiss Axioscop fluorescence microscope.

Western blot analysis and ELISA

Cells were harvested at 48 h post-transfection. Cell lysates normalized for percentage of GFP-positive cells were analyzed on SDS–polyacrylamide gels containing 17.5% polyacrylamide (200:1 ratio of acrylamide to N,N-methylene bisacrylamide). Following electrophoresis, the proteins were transferred for 2 h at room temperature to nitrocellulose membrane (0.45 µm; Schleicher & Schuell). The membrane was blocked with 10% dry milk in PBS for 1 h and stained with antiserum against HIV-1 CA (diluted 1:2500 in 5% dry milk in PBS containing 0.5% Triton X-100) overnight, followed by an additional 1 h blocking step and incubation with peroxidase-conjugated anti-rabbit secondary antibody (Dianova; diluted 1:10.000 in 2.5% dry milk in PBS containing 0.5% Triton X-100) for 2 h. Detection was performed by enhanced chemiluminescence (Amersham) according to the manufacturer’s protocol. Blots were exposed to Kodak X-Omat AR films. Release of virus-like particles into the supernatant of transfected cells was determined using a quantitative capture ELISA for the HIV-1 CA protein (41).

RNA preparation and analysis

For preparation of nuclear and cytoplasmic RNA, cells were collected at 48 h post-transfection, resuspended in 400 µl hypotonic buffer (10 mM HEPES–KOH pH 7.9, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM dithiothreitol) and incubated on ice for 10 min. Subsequently, cells were disrupted by vortexing for 10 s and nuclei were pelleted by centrifugation at 2000 r.p.m. for 1 min at 4°C. The supernatant corresponding to the cytoplasmic fraction was subjected to RNA extraction by the RNAzol B method (WAK Chemicals) according to the manufacturer’s protocol. The nuclei were washed twice in 500 µl hypotonic buffer prior to RNA extraction. Integrity of the nuclei was monitored by light microscopy. The RNA content was determined photometrically or by gel analysis.

For northern blot analysis, 10 µg of each RNA were loaded on denaturing formaldehyde gels containing 1% agarose. Separation was performed at 0.6 V/cm2 for 4 h and the RNA was transferred to Biodyne B membrane (0.45 µm; Pall) overnight by capillary blot. Subsequently, RNAs were UV crosslinked (Stratalinker) and heat fixed for 2 h at 80°C. Transfer efficiency was determined by staining the blot for 5 min with methylene blue (0.04% in 0.5 M NaOAc pH 5.2). The blot was pre-hybridized for 1 h at 42°C in 50% formamide, 6× SSC, 5× Denhardt’s, 0.5% SDS, containing 100 µg/ml denatured salmon sperm DNA and hybridized overnight at 42°C against the specific probes (activity >107 c.p.m.). The probes (25 ng) were labeled with [α-32P]dATP by random prime labeling using the Prime It II kit (Stratagene) and separated from unincorporated nucleotides on spin columms (Mobitec). Filters were washed, sealed and exposed to X-ray films (Kodak X-Omat AR) or quantified by phosphorimager analysis. For validation of the RNA fractionation, the filters were rehybridized with a probe specific for the nuclear U6 snRNA (kindly provided by R. Lührmann, University of Marburg).

RESULTS

Multimer copies of the M-PMV CTE lead to enhanced Rev-independent HIV-1 Gag production in a context-dependent way

Expression of the HIV-1 structural proteins is dependent on the Rev/RRE system, which can be functionally replaced by the CTE of D-type retroviruses. To analyze CTE function, we constructed plasmids carrying the HIV-1 Gag/PR coding region under the control of a CMV promotor and SV40 splice and poly(A) signals (Fig. 1). These plasmids were derived from the previously described expression vector pK-R-gpII (33) and contain either no element, the RRE, one copy of the M-PMV CTE in the sense or antisense orientation or four copies of the CTE in the sense orientation, respectively (Fig. 1). HeLa cells transfected with these plasmids in the presence or absence of a Rev expression vector were analyzed by immunoblotting using antiserum specific for the HIV-1 CA protein. An expression vector for GFP was co-transfected in all cases and cell lysates were normalized for percentage of GFP-positive cells. A strong Rev-dependent signal for the Pr55gag polyprotein, the proteolytically cleaved CA protein as well as two intermediate cleavage products was observed for pK-R-gpII transfected cells as described previously (Fig. 1, lanes 2 and 3) (33). No Gag-specific products were detected in the absence of RRE and CTE (Fig. 1, lane 1), while transfection of pK-gpII-C yielded some Rev-independent Gag expression (Fig. 1, lane 5), not observed for the CTE in the antisense orientation (Fig. 1, lane 4). However, only a weak signal for Pr55gag was detected in these samples and virtually no cleavage products were observed (activation of HIV-1 PR depends on dimerization of polyproteins and cleavage products are only observed at higher expression levels). Furthermore, no enhanced expression was achieved upon multimerization of the CTE (Fig. 1, lane 6).

The weak CTE-mediated expression was unexpected because the M-PMV-derived region used had been reported to be fully functional in conferring Rev independence (16). We therefore analyzed whether other segments of the transcript could interfere with CTE function and replaced the 3′-UTR carrying SV40 splice and poly(A) sequences by the poly(A) signal of bovine growth hormone (Fig. 2A). The resulting plasmids were transfected into HeLa cells in the presence or absence of Rev. Efficient Rev-dependent HIV-1 Gag expression was observed for the RRE-containing plasmid (Fig. 2A, lane 4), while transfection of a plasmid lacking any element yielded only a weak signal (lane 2), which was even lower for a plasmid containing the RRE in the absence of Rev (lane 3). Transfection of plasmids 3-C and 3-CCCC in this case led to efficient Rev-independent Pr55gag expression which was comparable to or even higher than that observed for the Rev/RRE system (Fig. 2A, lanes 5 and 7) and depended on CTE orientation (lane 6). Co-transfection of a Rev expression vector with the CTE-containing plasmids did not alter Pr55gag expression significantly (data not shown). Similar results were observed when the 3′ poly(A) signal was replaced by the LTR of a murine retrovirus (data not shown).

The results illustrated in Figure 2A suggested that multimerization of the CTE significantly enhanced Rev-independent Gag expression. Production of Pr55gag is sufficient to direct the release of virus-like particles from transfected cells, which can be measured in the culture medium. We therefore determined the amount of extracellular HIV-1 capsid protein (a domain of Pr55gag) using a quantitative ELISA. Relative expression levels were calculated as a percentage of the amount of extracellular HIV-1 antigen in the case of 3-RRE (+Rev) transfected cells, which was set as 100% (Table 1). It should be noted, however, that production of extracellular antigen requires a threshold concentration of polyproteins for assembly and processing to occur and the baseline expression may therefore be somewhat underestimated. Data from four independent experiments gave an average baseline expression for the 3-(–) vector of 3%, which was reduced to undetectable levels for 3-RRE in the absence of Rev. CTE-mediated Rev-independent expression yielded ~3-fold less antigen compared to Rev/RRE, consistent with previous reports that CTE is a weaker export element than Rev/RRE (18,24). Transfection of 3-CCCC, on the other hand, yielded a 2-fold higher expression level than observed for Rev/RRE (Table 1). From these data, we calculated approximate stimulation factors for the respective expression elements by dividing the amount of extracellular HIV-1 antigen in each sample by that observed in the case of 3-(–). This gave a stimulation factor of 10 for the CTE, 30 for Rev/RRE and 60 for the multimerized CTE (Table 1).

Table 1. Comparative analysis of Rev/RRE-mediated and CTE-mediated Gag expression.

Element Relative expression compared to Rev/RRE (=100%) Stimulation factor [relative to 3-(–)]
(–) 3.3 ± 1.5% 1×
RRE <1% <1×
RRE + Rev 100% 30×
C 31.8 ± 8% 10×
CCCC 206 ± 71% 60×
Open in a new tab

HeLa or Cos7 cells were transfected with plasmid 3-(–) and derivatives containing either the RRE or CTE. At 48 h post-transfection, cleared media were analyzed for release of virus-like particles using a quantitative ELISA detecting HIV-1 CA. The amount of antigen observed in the RRE plus Rev sample was set at 100% and relative expression levels were calculated for the other transfections. The mean of extracellular antigen in the case of 3-RRE (+Rev) was 315 ± 91 ng/ml in HeLa and 19 ± 2 ng/ml in Cos cells. Data from four independent experiments were used to calculate relative expression levels. In the right column, stimulation factors were calculated for each element by dividing the antigen level observed in the medium of transfected cells by that observed in a parallel sample of 3-(–) transfected cells (lacking any transport element).

To analyze whether the enhanced Gag expression correlated with increased nucleocytoplasmic transport of the respective mRNAs, we performed northern blot analysis on nuclear and cytoplasmic fractions of transfected HeLa cells. The HIV-1-specific RNA was detected with a radiolabeled probe against the Gag/PR coding region present in all RNAs, while the quality of the fractionation was validated by reprobing the blot with a U6 snRNA probe, which should be exclusively nuclear (Fig. 2B). Because initial experiments showed some U6 snRNA in the cytoplasmic fraction when NP-40 was used (data not shown), fractionations were performed in the absence of detergent. Under these conditions, no nuclear contamination of the cytoplasmic fraction was detected (Fig. 2B, lower panel), while the presence of cytoplasmic RNAs in the nuclear fraction cannot be excluded. Gag-specific RNAs of the expected sizes were observed in all transfections, while no HIV-1-specific signal was detected in mock-transfected cells (Fig. 2B, lanes 1 and 2). The RRE-containing RNA migrated slightly faster upon co-transfection with Rev (Fig. 2B, lanes 5 and 6). Furthermore, slower migrating specific RNAs were observed in the case of transcripts lacking or containing the CTE, but not RRE-containing transcripts (Fig. 2B, lanes 7 and 9, marked by an asterisk, and data not shown). These longer RNAs probably correspond to read-through transcripts produced by skipping the poly(A) signal of the gag/PR transcription unit and terminating at the downstream poly(A) signal of the neomycin cassette (Fig. 2B, bottom), but are not specific for the CTE-containing RNA. Virtually no gag-specific RNA was detected in the cytoplasm of 3-RRE transfected cells in the absence of Rev (Fig. 2B, lane 4), while it was readily observed in the presence of Rev (lane 6) and for the CTE- and multimerized CTE-containing transcripts (lanes 8 and 10). The signal was significantly stronger in case of the multimerized CTE, indicating that enhanced expression was directly correlated with increased nuclear export of this mRNA.

An intron in the 3′-UTR virtually abolishes CTE function

The described CTE-containing plasmids (Figs 1 and 2) gave very different levels of Rev-independent Pr55gag expression although they contained exactly the same M-PMV-derived sequence. This difference could be due to either the presence of an intron in the 3′-UTR of the pK plasmids or the distance of the CTE from the poly(A) signal, as reported previously (17). We therefore constructed plasmids carrying an intron in either the sense or antisense orientation (Fig. 3A). The intron in reverse orientation served as a spacer sequence, separating the CTE by 600 nt from the 3′-end of the RNA. Immunoblot analysis of transiently transfected HeLa cells showed virtually unaltered Rev-independent Pr55gag expression when the spacer was present (Fig. 3A, lanes 2 and 3). The presence of an intron in the sense orientation, on the other hand, led to a marked reduction in Gag expression, most easily demonstrated by the virtual absence of the CA cleavage product (Fig. 3A, lane 4). Quantitative analysis of released HIV-1-specific antigen showed a 2-fold reduction for the spacer-containing plasmid (54 ± 13% of the amount detected in the case of 3-C), while the intron in the 3′-UTR caused a 50-fold reduction (1.9 ± 0.8% of the amount detected in the case of 3-C), similar to the results observed for the pK plasmids (Fig. 1). From these data we conclude that the presence of an intron in the 3′-UTR strongly interferes with CTE function, while the distance to the poly(A) signal has only a marginal influence.

Figure 3.

Figure 3

Figure 3

Open in a new tab

An intron in the 3′-UTR leads to decreased CTE-mediated Gag expression. (A) A schematic diagram of plasmid 3-C and derivatives containing an intron in the antisense (serving as spacer) or sense orientation is shown in the upper panel. The lower panel shows immunoblot analysis of HeLa cells transfected with the plasmids specified and analyzed with antiserum against CA. (B) Northern blot analysis of nuclear (lanes 1–3) and cytoplasmic (lanes 4–6) RNA preparations from HeLa cells transfected with plasmids 3-C (lanes 1 and 4), 3-C(intron) (lanes 2 and 5) and 3-C(spacer) (lanes 3 and 6) and hybridized with a radiolabeled probe against the gag/PR region. Arrows point to the unspliced and spliced RNA in the preparation of 3-C(intron) transfected cells. The dotted arrow indicates a gag-specific degradation product observed mainly in the cytoplasmic fraction of 3-C(intron) transfected cells. The additional bands of lower mobility are read-through products, as explained in Figure 2B. A methylene blue staining of the blot is shown below to indicate the integrity and quality of the RNA preparation.

To analyze whether this loss of function correlates with impaired nuclear export of the respective mRNA, we performed northern blot analysis of nuclear and cytoplasmic fractions from transfected HeLa cells. Transcripts carrying the intron (Fig. 3B, lanes 2 and 5) or spacer (lanes 3 and 6) sequence migrated slower than the 3-C-derived transcript (lanes 1 and 4) and splicing of the intron-containing transcript was readily observed (Fig. 3B, lane 2). Cytoplasmic fractions of 3-C and 3-C(spacer) transfected cells revealed equivalent amounts of HIV-1-specific RNA (Fig. 3B, lanes 4 and 6). In contrast, significantly lower amounts of specific RNA were detected both in the nucleus (lane 2) and cytoplasm (lane 5) of 3-C(intron) transfected cells. The spliced RNA was only found in the nuclear fraction, while low amounts of the unspliced RNA were also observed in the cytoplasm. Furthermore, an additional signal which had a significantly faster and more diffuse mobility and specifically hybridized with the HIV-1 gag probe was detected in the cytoplasmic and (to a lesser extent) in the nuclear fraction of 3-C(intron) transfected cells (Fig. 2B, lanes 2 and 5, marked with a dotted arrow). Methylene blue staining of the blot showed that there was no difference in the integrity and quality of the RNA preparations (Fig. 2B, lower panel) and it appears likely that the smaller RNA species correspond to specific degradation products of the gag RNA.

The multimeric CTE yields efficient HIV-1 Gag expression in a variety of cell lines after retroviral transduction

Most analyses of CTE function have relied on transient transfection of expression plasmids which leads to multiple extrachromosomal transcription units in the nucleus of transfected cells. Retroviral replication, on the other hand, involves chromosomal integration of one or few copies of the viral genome. In order to analyze whether efficient Rev-independent HIV-1 Gag expression can also be mediated by multimeric copies of the CTE in this setting, we constructed MuLV-based retroviral vectors (MP 110) (36), carrying the gag/PR transcription unit with or without one or four copies of the CTE. Vector particles were produced by transfection of the Phoenix packaging cell line (39) with plasmid MP 110 gpII, MP 110 gpII-C, MP 110 gpII-CCCC or MP 110 eGFP as a control. Subsequently, human epithelial SW 480 target cells were transduced and Gag expression was analyzed by indirect immunofluorescence and immunoblot after 7 days. Figure 4C shows strong Rev-independent Gag expression for cells transduced with MP 110 gpII-CCCC, while the empty control vector (MP 110, Fig. 4A) and MP 110 gpII lacking the CTE (Fig. 4B) gave no detectable expression. A significantly weaker signal was also seen in cells transduced with MP 110 gpII-C carrying a single copy of the CTE and these results were confirmed by immunoblot analysis (data not shown). Northern blots revealed a strong Gag-specific signal only in the case of MP 110 gpII-CCCC transduced cells (data not shown). CTE function in this case is required both for nuclear export of the genomic RNA in the packaging cell line and for export of the mRNA in the transduced cells, which may explain the stronger requirement for CTE function in the retroviral vector system.

Figure 4.

Figure 4

Open in a new tab

Analysis of CTE-mediated Gag expression in various cell lines from different species. The gag/PR expression cassette lacking (B) or containing four copies of the CTE (C–F) was inserted into a retroviral vector and vector particles produced from an amphotropic packaging cell line were used for transduction of human SW 480 (A–C), murine NIH 3T3 (D), canine Cf2th (E) and simian Cos7 (F) cells. (A) SW 480 cells transduced with the empty retroviral vector. Cells were fixed 7 days after transduction and HIV-1-specific products were detected by indirect immunofluorescence using antiserum against CA.

Next we tried to transduce a variety of cell lines with MP 110 gpII-CCCC to look for CTE-mediated Rev-independent expression in cells from different species. Efficient transduction was achieved in cell lines of human origin (e.g. SW 480, Fig. 4C) as well as in murine (NIH 3T3, Fig. 4D), canine (Cf2th, Fig. 4E) and simian (Cos7, Fig. 4F) cell lines. No major differences were observed in the relative expression levels. To analyze whether the presence of the CTE also enhances expression of a Rev-independent mRNA, we performed experiments with GFP-carrying retroviral vectors containing or lacking one or four copies of the CTE. No significant difference in GFP expression was observed by FACS analysis of transduced cells (data not shown), indicating that the presence of this retroviral export element does not enhance expression from a heterologous intron-less mRNA.

RHA is not specifically translocated to the cytoplasm of cells expressing CTE-containing RNAs

The CTE of D-type retroviruses are cis-acting RNA signals presumed to bind to cellular factors facilitating the export of intron-containing RNAs. RHA has been described as a CTE-interacting protein which is specifically translocated from the nucleus to the cytoplasm in the presence of a transcription inhibitor or in cells overexpressing CTE-containing RNA (26). To analyze the role of RHA in CTE function, we determined the localization of RHA in HeLa cells transfected with the CTE-containing plasmids 3-C and 3-CCCC or the Rev-dependent plasmid 3-RRE. Cells grown on coverslips were stained with a HIV-1 CA-specific monoclonal antibody and polyclonal antiserum against RHA 48 h after transfection. No specific translocation of RHA into the cytoplasm of CTE-expressing cells was observed. Figure 5C and D shows double fluorescence analysis of 3-CCCC transfected HeLa cells indicating that RHA was nuclear in both HIV-1 Gag-positive and Gag-negative cells. Similar results were observed for all Gag expression vectors tested (data not shown). A cytoplasmic localization of RHA was observed in a few cells in all experiments, independent of the transfected plasmid, and most of these cells were in different stages of cell division. Figure 5A shows a cell with cytoplasmic RHA localization in a mock-transfected sample (marked by an asterisk). The relative number of cells with cytoplasmically localized RHA was similar irrespective of the transfected plasmid (Table 2, left column). Furthermore, it was also similar in cells productively transfected with CTE-carrying plasmids and in HIV-negative neighboring cells (Table 2, right column). Taken together, these results show that overexpression of the CTE does not alter the subcellular localization of RHA in HeLa cells. Addition of the transcription inhibitor actinomycin D, on the other hand, completely shifted RHA from the nucleus to the cytoplasm (Fig. 5B), as described previously (26).

Figure 5.

Figure 5

Open in a new tab

CTE-mediated RNA export does not lead to specific translocation of RNA helicase A into the cytoplasm. Untransfected HeLa cells (A and B) or HeLa cells transfected with plasmid 3-CCCC (C and D) were either left untreated (A, C and D) or treated with cycloheximide (50 µg/ml) and actinomycin D (5 µg/ml) for 3 h (B). Subsequently, cells were fixed and subjected to indirect immunofluorescence using antiserum against RHA. (C and D) Double immunofluorescence using a monoclonal antibody against HIV-1 CA in addition was performed and cells were analyzed for CA expression in (C) and for RHA expression in (D). A single cell exhibiting predominantly cytoplasmic localization of RHA in untransfected cells is marked with an asterisk in (A).

Table 2. Analysis of the subcellular localization of RHA.

Element Cells (total)
 Cells (Gag-positive)
  Nuclear Cytoplasmic Nuclear Cytoplasmic
Mock 366 17 n.a. n.a.
RRE 282 12 n.a. n.a.
RRE + Rev 354 18 140 3
C 289 7 31 0
CCCC 412 16 77 0
Open in a new tab

HeLa cells were either mock-transfected or transfected with plasmid 3-RRE (± Rev), 3-C or 3-CCCC and subjected to indirect immunofluorescence using double staining with antisera against HIV-1 CA and RHA. Cells in representative fields were analyzed for nuclear or cytoplasmic localization of RHA. In the two right-most columns, only Gag-positive cells were analyzed for RHA localization. n.a. not applicable.

DISCUSSION

In this study, we have shown that multimeric copies of the M-PMV CTE allow Rev-independent expression of the HIV-1 Gag polyprotein at a level comparable to or higher than observed for authentic Rev-dependent production. Previous investigations have measured the stimulatory activity of the CTE using either a reporter construct containing the chloramphenicol acetyltransferase (CAT) coding sequence within the intron of a pre-mRNA (42) or expression vectors for the Rev-dependent HIV-1 structural proteins. Depending on the type of vector, the presence of the CTE stimulated Gag expression ~10- to 20-fold (18,24) and led to a 5-fold enhancement of CAT activity (22,23). Gag or CAT expression was stimulated by a factor of 20–100, on the other hand, if the Rev/RRE system was used instead of the CTE (23,43,44). Comparable results were obtained in this study (CTE, 10-fold; Rev, 30-fold stimulation), confirming that CTE activity is significantly weaker than Rev activity. By introducing four copies of the CTE, however, we could reproducibly achieve Gag expression levels higher than those observed for Rev/RRE, corresponding to a 6-fold enhanced activity compared to a single CTE.

These results indicate that CTE activity may be dependent on binding of several molecules of a transport factor or of multiple transport factors to a single RNA molecule. This may correspond to the described requirement for multiple Rev molecules to be recruited to the target RNA response element (45). Experimental determination of the CTE structure has revealed a stable RNA helix with two internal loop sequences, where the integrity of the helix and the primary sequence of the single-stranded loop regions are both essential for activity (16,18,19). A single loop has been shown to be sufficient for binding of the export factor Tap in vivo (31) and to promote RNA export in Xenopus oocytes (25,28), while two internal loops are required for CTE-dependent expression in mammalian cells (16,20). Accordingly, we observed that a shorter element, unable to form both internal loops, was poorly functional while two copies partially restored activity (data not shown). Computer analysis of the RNA folding potential using the MulFold program (46,47) predicted that at least six single-stranded CTE loops were formed when four copies of the CTE were present on the RNA. Each loop can serve as a binding site for the RNA export factor Tap and other putative export proteins and binding of multiple factors may lead to cooperative enhancement of RNA export. We suggest that CTE function, similar to Rev function, requires binding of multiple molecules of export factors and may be further enhanced by creating additional binding sites. Enhancement of activity has recently also been described when segments of the hepatitis B virus post-transcriptional regulatory element were used in multiple copies (43).

If several binding sites for export factors are present on the multimeric CTE, one would expect that translocation of these factors to the cytoplasm occurs to a higher degree than in the case of a single CTE. Previously, RHA has been reported to be quantitatively translocated to the cytoplasm of transfected cells in a CTE-dependent way (26). In contrast to that study, we were unable to detect any CTE-mediated translocation of RHA to the cytoplasm using constructs with either a single or four copies of the CTE. A small number of cells contained RHA in the cytoplasm in all cases and this number was not different in transfected and untransfected cells. Although this discrepancy may be due to technical differences, we would like to point out that we used the same cell line (HeLa) and the same basic expression vector (pcDNA3) as in the previous study by Tang et al. (26), suggesting that the different results cannot be explained simply by different expression levels. In a recent report, RHA was shown to be associated with poly(A)-containing mRNA in the nucleus and to a much lesser extent in the cytoplasm of HeLa cells (48). These authors concluded that RHA is a pre-mRNA- and mRNA-binding protein which may have a transient role in DNA binding during the process of transcript formation. Similar to the previous report and in accordance with our results, they also observed translocation of RHA to the cytoplasm upon inhibition of RNA transcription. Given that RHA appears to be a general mRNA-binding protein (48) which is not specifically translocated to the cytoplasm by the CTE (this study), that RHA was not detected among CTE-binding factors obtained by affinity purification (20) and that it is not needed for CTE-mediated RNA export in Xenopus oocytes (28), we suggest that RHA is most likely associated with CTE-containing as well as other mRNAs, but is probably not a specific CTE export factor.

Our results indicate that besides the requirement for multiple factor binding sites, CTE function is also dependent on the context. Rizvi et al. (17) previously reported that inserting a 750 nt spacer sequence between the CTE and the adjacent poly(A) site of M-PMV led to a reduction in the titer of the resulting CTE-dependent vectors close to background levels. Based on these results, the authors suggested that a functional interaction between the CTE and the poly(A) signal may be necessary for CTE function. We did not observe a reduction in CTE function when a 600 nt spacer sequence was inserted between the CTE and the poly(A) signal. A possible explanation for this discrepancy would be that the M-PMV poly(A) signal may be rather weak and can be influenced by the upstream CTE sequences, thus arguing for a CTE-mediated enhancement of polyadenylation rather than for a poly(A) signal-mediated enhancement of CTE function. This effect would not be without precedent since the hepatitis B virus post-transcriptional regulatory element as well as the post-transcriptional control elements on the herpes simplex thymidine kinase and mouse histone H2a RNAs have been shown to enhance polyadenylation (49). Furthermore, we observed that the specific RRE-containing RNA was shorter in the presence of Rev than in its absence, which is most likely due to an influence of Rev on poly(A) tail length.

While the distance between the CTE and poly(A) site appeared not to be critical, our results clearly showed that the presence of an intron in the 3′-UTR of the mRNA severely reduced CTE function. Northern blot analyses revealed a small amount of the unspliced RNA in the cytoplasm of transfected cells, while the spliced RNA was reduced in the nuclear and virtually absent in the cytoplasmic fraction and appeared to be specifically degraded (a significant signal was observed for cytoplasmic degradation products). In contrast, no loss of the specific RNA was observed for a RRE-containing RNA with an intron in the 3′-UTR in the presence of Rev (data not shown). Specific degradation of RNAs that had been spliced in the 3′-UTR has recently been reported as a cellular control mechanism referred to as nonsense-mediated mRNA decay (NMD) (50,51). NMD has been defined as a mechanism to prevent the translation of aberrant polypeptides from mRNAs containing premature stop codons and has been most extensively studied in the case of β-globin. It is based on the hypothesis that exon junctions are marked with specific factors during the splicing reaction and these factors remain on the RNA and are recognized during translation. If the RNA contains a premature stop codon, there are such factors 3′ to the stop codon which direct specific degradation of this RNA (50,51; reviewed in 52). Based on this concept, we suggest that an intron in the 3′-UTR of our Gag expression vectors did not primarily interfere with CTE-mediated nuclear export but led to NMD of the spliced transcript. Accordingly, only some unspliced RNA (which would not be a substrate for NMD) was detected in the cytoplasmic fraction of cells transfected with the intron-containing vector. In contrast, the Rev/RRE system allowed export of the spliced RNA as well and yielded efficient Gag expression even in the presence of a 3′ intron. It is interesting to note that the CTE-containing simple retroviruses harbor no introns 3′ to the stop codon of the first open reading frame (Gag), while there are many 3′ introns in the case of HIV and other complex retroviruses. Conceivably, Rev may circumvent NMD by inhibiting the splicing reaction, by interfering with marking of the exon junction or by rapidly shifting the RNA to the export pathway, and this function cannot be substituted for by the CTE.

Acknowledgments

ACKNOWLEDGEMENTS

We thank F. Grosse, M. Alizon and R. Lührmann for antisera, the HeLa P4 cell line and probes. We are grateful to C. Baum and M. Hildinger for supplying retroviral vectors and to G. Nolan for the Phoenix packaging cells. We thank T. Kock for help with tissue culture and M. Dittmar for critically reading the manuscript. This work was supported in part by grants from the Deutsche Forschungsgemeinschaft (SFB545 and Ri192/21-1-1).

REFERENCES

Articles from Nucleic Acids Research are provided here courtesy of Oxford University Press

RESOURCES