• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of narLink to Publisher's site
Nucleic Acids Res. Apr 1, 2000; 28(7): 1625–1634.
PMCID: PMC102795

Initiation of translation by non-AUG codons in human T-cell lymphotropic virus type I mRNA encoding both Rex and Tax regulatory proteins

Abstract

Human T-cell lymphotropic virus type I (HTLV-I) double-spliced mRNA exhibits two GUG and two CUG codons upstream to, and in frame with, the sequences encoding Rex and Tax regulatory proteins, respectively. To verify whether these GUG and CUG codons could be used as additional initiation codons of translation, two chimeric constructs were built for directing the synthesis of either Rex–CAT or Tax–CAT fusion proteins. In both cases, the CAT reporter sequence was inserted after the Tax AUG codon and in frame with either the Rex or Tax AUG codon. Under transient expression of these constructs, other proteins of higher molecular mass were synthesized in addition to the expected Rex–CAT and Tax–CAT proteins. The potential non-AUG initiation codons were exchanged for either an AUG codon or a non-initiation codon. This allowed us to demonstrate that the two GUG codons in frame with the Rex coding sequence, and only the second CUG in frame with the Tax coding sequence, were used as additional initiation codons. In HTLV-I infected cells, two Rex and one Tax additional proteins were detected that exhibited molecular mass compatible with the use of the two GUG and the second CUG as additional initiation codons of translation. Comparison of the HTLV-I proviral DNA sequence with that of other HTLV-related retroviruses revealed a striking conservation of the three non-AUG initiation codons, strongly suggesting their use for the synthesis of additional Rex and Tax proteins.

INTRODUCTION

After transcription of human T-cell lymphotropic virus type I (HTLV-I) integrated proviral DNA, different mRNAs are produced by complex mechanisms of differential splicing. In the absence of viral proteins, a unique viral primary transcript is synthesized in small amounts by transcription of the full-length integrated proviral DNA. After double splicing of this primary transcript, the main mRNA species reaches the cytoplasm. This 2.1 kb mRNA does not code for structural viral proteins (Fig. (Fig.1A).1A). It contains two overlapping coding sequences starting with two AUG initiation codons of translation in different frames, both AUG being located in the second exon. Translation initiating from the first AUG codon directs Rex protein synthesis, whereas that initiating from the second AUG codon directs Tax protein synthesis (1). The 27 kDa Rex protein (p27rex) is a post-transcriptional regulator of gene expression that acts primarily by binding to a responsive element (XRE) located at the 3′ end of all HTLV-I RNA species (2). The first 19 amino acids of the NH2-terminal sequence of Rex protein encompass both nuclear and nucleolar localization signals and also the XRE binding domain (3,4). Complex molecular events, involving Rex protein binding to the XRE in the nucleus, followed by its oligomerization and association to nucleoporin-like proteins, lead to the nucleocytoplasmic transport of single- or un-spliced mRNAs encoding structural viral proteins (57). The 40 kDa Tax protein is a transcriptional transactivator of viral and cellular gene expression (810). This last characteristic confers on Tax protein the status of an oncogenic protein (11,12).

Figure 1
Structure of HTLV-I double- and single-spliced mRNA. (A) HTLV-I double-spliced mRNA results from addition of exon I, the sequence of which is entirely in the 5′ UTR, then exon II bearing two GUG, two CUG and the Rex and Tax AUG initiation codons, ...

In another mRNA species resulting from alternative splicing leading to removal of the second exon, translation initiates by another AUG codon located further downstream in the Rex coding sequence. This mRNA is devoid of the Rex and Tax AUG initiation codons described above and was supposed to direct only the synthesis of a 21 kDa Rex NH2-terminally-truncated protein (p21rex) of unknown function (13,14). Similarly, an mRNA coding exclusively for Tax protein seems to be produced, at least in some cases, by use of a putative internal promoter located just upstream to the ATG triplet specifying the beginning of Tax protein synthesis (15,16). Other proteins, namely Rof, Tof, p13II and p12I, probably involved in regulatory functions, are potentially synthesized by translation of other mRNA species resulting from still other alternative splicing of the primary transcript (14,17). However, the 2.1 kb mRNA species produced by double splicing of the full-length primary transcript encodes both 27 kDa Rex and 40 kDa Tax proteins, the synthesis of which results from translation initiating by two different non-in frame AUG codons, located in the second exon. Therefore, the presence of two partly overlapping coding sequences in the same mRNA molecule implies a coordinated regulation of the translation initiation processes that should lead to the regulation of HTLV-I gene expression, and subsequently to either a productive infection or a latency, eventually followed by T-cell transformation.

HTLV-I single-spliced 4.2 kb mRNA exhibits an organization similar to that of the 2.1 kb double-spliced mRNA species with an identical 5′ UTR and the same two AUG codons (Fig. (Fig.1B).1B). However, translation initiated by the first AUG codon of this 4.2 kb mRNA potentially directs the synthesis of a putative 26 amino acid long peptide, the first 19 being identical to the NH2-terminal sequence of the 27 kDa Rex protein. Such a putative peptide of 26 amino acids has never been described in HTLV-I infected cells. Translation initiated by the second AUG codon directs the synthesis of HTLV-I envelope glycoprotein (Env) (for more details see 14).

The 5′ UTR of HTLV-I single- and double-spliced mRNA species exhibits two GUG and two CUG codons 5′ to, and in frame with, Rex and Tax/Env AUG initiation codons, respectively. Some eukaryotic cell and viral mRNA encoding mostly regulatory proteins exhibit non-AUG codons such as GUG, CUG and more rarely ACG or AUU or even CUU codons which can be used for initiating translation (1822 and references therein). Among them, some of these non-canonical initiation codons are the only possible sites for initiating translation. This is the case for the synthesis of EIAV Tat protein and MuLV cell surface antigen which result from translation initiated by a CUG codon (23,24). However, when it exists, translation initiated by non-AUG codons usually occurs in addition to that initiated mainly by a regular AUG codon located further downstream. These non-canonical initiations of translation result generally in the synthesis of proteins exhibiting subcellular distributions and/or functions somewhat different to that of a major protein synthesized by translation initiated mainly by an AUG codon (2427). Therefore, initiation of translation by non-AUG codons seems to be a rare event that is mRNA species-dependent. Indeed, specific mRNA features are required for such initiations by non-AUG codons (28). Efficiency of most translations initiated by non-AUG codons appears to be increased when the sequence ancillary to the initiation codon is similar to that described in the so-called scanning model, i.e. a purine at –3 and a G at +4 of the initiation codon (29,30). Finally, translation initiated by either AUG or non-AUG codons can occur also, at least in some cases, by direct, cap-independent, internal ribosome entry (for a review see 31). This last mechanism requires a specific mRNA structure, called internal ribosome entry site or segment.

Considering the particular features of HTLV-I double- and single-spliced mRNA described above, we hypothesized that the two GUG and the two CUG codons located upstream to, and in frame with, the Rex and Tax/Env AUG initiation codons of translation, respectively, could be used as additional translation initiation codons for directing the synthesis of additional forms of Rex, Tax and Env proteins. By using the chloramphenicol acetyl transferase (CAT) reporter coding sequence replacing that of Rex and Tax or Env, we showed that both GUG codons located upstream to and in frame with Rex AUG initiation codon can be used for initiating translation. Similarly we showed that only the second CUG codon, located upstream to and in frame with the Tax/Env AUG initiation codon, can also be used for initiating translation. Moreover, an additional Tax protein was detected in HeLa cells after transfection with a construct directing the synthesis of an mRNA identical to HTLV-I double-spliced mRNA. Such additional Tax protein was detected also in C91PL and MT-2 cells which are permanently infected with HTLV-I. Two additional forms of Rex were also detected in HTLV-I-infected MT-2 cells only.

MATERIALS AND METHODS

Construction of expression vectors

pHTL refers to plasmids in which the sequences to transcribe were placed under the control of the HTLV-I promoter. For efficient plasmid amplification, and simplification of subcloning steps, all pHTL were inserted in a pUHD10-3 vector (32). Nucleotides were numbered with respect to the HTLV-I double-spliced mRNA cap site (position +1). Each construct used in this study was verified by DNA sequencing.

pHTL-Rex–CAT and pHTL-Tax–CAT. For constructing pHTL-Rex–CAT and pHTL-Tax–CAT, the Env coding sequence in pHLX-I was exchanged for the CAT coding sequence which was inserted in frame with either Rex or Tax ATG codons specifying initiations of translation (33). For removing the Env coding sequence, the entire sequence coding for exon II up to the Tax/Env initiation codon, on one hand, and that coding for exon III starting after the Env termination codon, on the other hand, were first amplified by PCR. The first one was amplified using HindIII no. 7 (5′-TCCTGAACTGCGTCCGCCGT-3′) and BssHII no. 2 (5′-TCCAAACCCTGGGAAGCGCGCCA-TGGTGTTGGAGG-3′) as 5′ and 3′ primers, respectively. The second one was amplified using BssHII no. 1 (5′-CCTCCAA-CACCATGGCGCGCTTCCCAGGGTTTGGA-3′) and MluI no. 6 (5′-GGAACTGTAGAGCTGAGCCG-3′) as 5′ and 3′ primers, respectively. The first 15 nt of the 5′ primer BssHII no. 1 are identical to that coding for the end of exon II while its last 16 nt are identical to that coding for the beginning of exon III. The four additional nucleotides CGCG located in the middle of the primer permit creation of a BssHII restriction site (underlined) immediately after the Tax ATG initiation codon. This BssHII restriction site was created for subsequent cloning of CAT coding sequence in frame with either Rex or Tax ATG initiation codons (see below). Primers no. 1 and no. 2 are complementary. The resulting fragments which partly hybridized to each other were then mixed and amplified using primers HindIII no. 7 and MluI no. 6. The new fragment was then digested by HindIII and by MluI for subsequent insertion between the same restriction sites in pHLX-I, in which the HindIII site located at the end of the sequence coding for exon I was made unique. This new construct deleted from the Env coding sequence was called pHTL-2-BssHII.

The CAT coding sequence was amplified by PCR using pSV2–CAT as template and either 5′-GGGGGACGCGTGGA-GAAAAAAATCACTGGATATACCACCGT-3′ (primer no. 4) or 5′-GGGGGACGCGTGAGAAAAAAATCACTGGATAT-ACCACCGT-3′ (primer no. 5) as 5′ primer, and 5′-GGGGGGCG-CGCACTTATTCAGGCGTAGCACCA-3′ (primer no. 3) as 3′ primer. Primer no. 4 differs from primer no. 5 in that it contains an additional G (in bold). Use of either primer no. 4 or primer no. 5 allowed us to place the CAT coding sequence in frame with either Rex or Tax ATG initiation codon, respectively. Both 5′ primers and the 3′ primer permit generation of MluI and a BssHII restriction sites (underlined), respectively. After digestion by MluI and BssHII, the amplified fragments were cloned in the unique BssHII restriction site of pHTL-2-BssHII giving rise to pHTL-Rex–CAT +i and pHTL-Tax–CAT +i. These two last constructs contained a part of the sequence coding for intron I, going from the splicing donor site to the HindIII site, which was subsequently removed. This was done by PCR using a strategy similar to that described above for removing the Env coding sequence from pHLX-I. The two constructs which were thus obtained were called pHTL-Rex–CAT and pHTL-Tax–CAT (Fig. (Fig.2).2). Transcription of these last two constructs leads to the synthesis of two different mRNAs with an identical 5′ UTR in which HTLV-I intron I has been removed from nt 119–4641 (see table 1 in 17).

Figure 2
Chimeric mRNA synthesized by transcription of pHTL-Rex–CAT and pHTL-Tax–CAT. The CAT coding sequence devoid of its initiation codon was inserted in a plasmid derived from pHLX-I (33), downstream from AUG306 and in frame with either Rex ...

pHTL-2 and pCMV-2. To permit synthesis of an mRNA identical to the HTLV-I double-spliced mRNA, another plasmid was derived from pHLX-I (33). First, reverse transcription was performed on total RNA purified from HeLa cells transfected with pHLX-I using primer MluI no. 6. The cDNA were then double-stranded and amplified by PCR using HindIII no. 7 and MluI no. 6 as 5′ and 3′ primers, respectively. Two cDNAs were thus obtained, a long one corresponding to a part of the unspliced primary transcript, therefore containing the Env coding sequence, and a short one deleted from the Env coding sequence. The short cDNA was digested by SphI and MluI. The resulting fragment was then inserted between the same restriction sites in pHLX-I, in which the SphI restriction site located in the sequence coding for exon II was made unique. This new construct containing the Rex and Tax coding sequences was called pHTL-2 +i. It contained a part of the sequence coding for intron I which was removed as for obtaining pHTL-Rex–CAT and pHTL-Tax–CAT. This new construct was called pHTL-2.

Rex and Tax coding sequences were also placed under the control of the CMV promoter by changing the HTLV-I 5′-LTR of pHTL-2 for the CMV promoter. The CMV promoter sequence was first amplified using CMV-βgal (Clontech) as template and 5′-CTAGTCTAGAGCTTCCCATTGCATACG-TTGT-3′ and 5′-GGGGGTCGACCGGTTCACTAAACGAG-CTCTG-3′ as 5′ and 3′ primers, respectively. The amplified fragment was successively digested by SalI (underlined site in the 3′ primer), then by mung bean nuclease and finally by XbaI (underlined site in the 5′ primer). The sequence in pHTL-2 going from the cap site to the MluI site was amplified using 5′-GGGGGTCGACTCGCATCTCTCCTTCAC-3′ and MluI no. 6 as 5′ and 3′ primers, respectively. The amplified fragment was successively digested by SalI at the underlined site, then treated by mung bean nuclease and finally digested by MluI. This last fragment, and that containing the CMV promoter, were then inserted together between the XbaI and MluI sites of pHTL-2, giving rise to pCMV-2 coding for both Rex and Tax proteins.

pLTR-βgal. To place β-galactosidase coding sequence under the control of HTLV-I promoter, the LTR sequence of Rous sarcoma virus (RSV) was first removed from RSV-βgal by digestion with HindIII restriction enzyme. The sequence containing the RSV promoter was then replaced by that of HTLV-I which was obtained from pHTL-2 by digestion with BamHI. The linearized plasmid containing the β-galactosidase coding sequence and the sequence containing HTLV-I promoter were then ligated after blunt ending both DNA molecules by the Klenow fragment of DNA polymerase I.

Site-directed mutagenesis of Rex and Tax ATG initiation codons of translation

The ATG triplet specifying initiation of Rex protein synthesis (position 250) was exchanged for ATT using pHTL-Rex–CAT as template and 5′-GGGGCATGCATTCCCAAGACCCGTC-GGA-3′ and MluI no. 6 as 5′ and 3′ primers, respectively. The amplified fragment was successively digested by SphI at the underlined site, then treated with T4 DNA polymerase to remove overhanging ends and finally digested by MluI restriction enzyme. The DNA fragment was subsequently ligated in pHTL-Rex–CAT between blunt ended SphI restriction site and MluI restriction site, for obtaining pHTL-Rex–CAT(ATT250). To change ATG250 for ATT in pHTL-Tax–CAT, the XbaI–BstXI fragment of pHTL-Tax–CAT was exchanged for that of pHTL-Rex–CAT(ATT250).

ATG250 was exchanged for TAG using pHTL-Rex–CAT as template and 5′-GAGCTGCTAGCCCAAGACCCGTC-3′ and MluI no. 6 as 5′ and 3′ primers, respectively. The amplified fragment was successively digested by NheI at the underlined site, then treated with T4 DNA polymerase and finally digested by MluI. The DNA fragment was subsequently ligated in pHTL-Rex–CAT between blunt ended SphI restriction site and MluI restriction site for obtaining pHTL-Rex–CAT(TAG250). To change ATG250 for TAG in pHTL-Tax–CAT, the XbaI–BstXI fragment of pHTL-Tax–CAT was exchanged for that of pHTL-Rex–CAT(TAG250).

The ATG triplet specifying initiation of Tax protein synthesis (position 306) was exchanged for TTG using pHTL-Tax–CAT as template and 5′-AGACCTCCAACACCTTGGCGCGTG-3′ and MluI no. 6 as 5′ and 3′ primers, respectively. The amplified fragment was then digested by BstXI at the underlined site and by MluI for subsequent insertion between the same restriction sites in pHTL-Tax–CAT, giving rise to pHTL-Tax–CAT(TTG306).

Site-directed mutagenesis of GTG and CTG codons

GTG217 was exchanged for either ATG or GTA using pHTL-Rex–CAT as template and either 5′-GGGGAGCGCTAGTTCTGCC-CAATGGATCCCGTGGAGAC-3′ or 5′-GGGGAGC-GCTAG-TTCTGCCCAGTAGATCCCGTGGAGAC-3′ as 5′ primer and MluI no. 6 as 3′ primer. Each amplified fragment was then digested by Eco47III at the underlined site and by BstXI for subsequent insertion between the same restriction sites in pHTL-Rex–CAT, giving rise to pHTL-Rex–CAT(ATG217) and pHTL-Rex–CAT(GTA217).

GTG226 was exchanged for either ATG or GTA using pHTL-Rex–CAT as template and either 5′-GGCGCGGATCCCATGGAGACTCCTCAAGC-3′ or 5′-GGCGCGGATCC-CGTAGAGACTCCTCAAGC-3′ as 5′ primer and MluI no. 6 as 3′ primer. Each amplified fragment was then digested with BamHI at the underlined site and by MluI for subsequent insertion between the same restriction sites in a plasmid containing a unique BamHI restriction site that differed from pHTL-Rex–CAT in that part of the vector was derived from pG3. The XbaI–BstXI fragment of each construct was then inserted between the same restriction sites in pHTL-Rex–CAT, giving rise to pHTL-Rex–CAT(ATG226) and pHTL-Rex–CAT(GTA226).

CTG210 was exchanged for either ATG or CTT using pHTL-Rex–CAT as template and either 5′-GGGGAGC-GCTAGTTATGCCCAGTGGATC-3′ or 5′-GGGGAG-CGCTAGTTCTTCCCAGTGGATC-3′ as 5′ primer and MluI no. 6 as 3′ primer. Each amplified fragment was then digested by Eco47III (underlined site) and MluI for subsequent insertion between the same restriction sites in pHTL-Rex–CAT, giving rise to pHTL-Rex–CAT(ATG210) and pHTL-Rex–CAT(CTT210). To exchange CTG210 for either ATG or CTT in pHTL-Tax–CAT, the XbaI–BstXI fragment of pHTL-Tax–CAT was exchanged for that of either pHTL-Rex–CAT(ATG210) or pHTL-Rex–CAT(CTT210).

CTG246 was exchanged for either ATG or CCG using pHTL-Rex–CAT as template and XbaI no. 8 as 5′ primer and either 5′-GGGGGGCATGCATCTCGCTTGAGGAGTC-3′ or 5′-TCTTGGGCATGCGGCTCGCTTGAGGAGTC-3′ as 3′ primer. The amplified fragments were then digested by XbaI and SphI at the underlined site for subsequent insertion between the same restriction sites in pHTL-Tax–CAT, giving rise to pHTL-Tax–CAT(ATG246) and pHTL-Tax–CAT(CCG246).

Cell culture and DNA transfection

HeLa cells were grown as monolayers in Eagle’s minimum essential medium (EMEM) containing 5% fetal calf serum, at 37°C in the presence of 5% CO2. For DNA transfection, cells were seeded at 7.5 × 105 cells per 100 mm diameter plate, 6 h before transfection. Cells were transfected by the calcium phosphate precipitation procedure (34) modified as follows. One hour before adding the DNA-calcium phosphate precipitate, the culture medium was replaced by a transfection medium made of EMEM containing 5% fetal calf serum and complemented with 15 mM HEPES, pH 7.8. Cells were then incubated in the presence of 1.5% CO2. Transfections were performed in duplicate with 12 µg of total DNA containing 4 µg of wild-type or modified pHTL-Rex–CAT or pHTL-Tax–CAT, 1 µg of pcTax and 1 µg of pLTR-βgal. After overnight incubation, the medium was removed. After washing the cells once with phosphate-buffered saline (PBS) containing 4 mM EDTA, a fresh transfection medium was added. Forty-eight hours after transfection, cells were harvested, lysed in 100 mM Tris–HCl pH 7.8 containing 0.7% NP-40 and sonicated. Cell lysates were used for western blots as well as for β-galactosidase and CAT assays (35).

Western immunoblotting

Total cell proteins were separated by SDS–PAGE then transferred onto a nitrocellulose membrane (see 36 for a complete description of the electrophoresis and immunoblotting procedures). CAT proteins were detected with polyclonal anti-CAT antibodies (5 Prime-3 Prime, USA). Tax proteins were detected with polyclonal anti-Tax antibodies which were obtained through the AIDS Research and Reference Reagent Program (Division of AIDS, NIAID, NIH). The antiserum to HTLV-I Tax was from Dr Kuan-Teh Jeang (37,38). HTLV-I proteins were detected with a serum obtained from a patient seropositive for HTLV-I. Polyclonal monospecific antibodies directed against each of the three main proteins detected in MT-2 cell extracts were purified from this serum. For such purification, total proteins of MT-2 cell extracts were first separated by SDS–PAGE after loading them in a well 12 cm wide, then transferred onto a nitrocellulose membrane. Three horizontal strips, each containing one of the three main proteins detected in MT-2 cell extracts, were cut out of the nitrocellulose membrane, then incubated with HTLV-I positive serum. Monospecific antibodies recognizing each of the three main proteins detected in MT-2 cells were eluted by treatment of the strips with a 3.5 M solution of potassium thiocyanate containing 1% BSA, then dialysed against PBS containing 0.1% sodium azide and used immediately for western immunoblotting. HTLV-I infected C91PL and MT-2 cells and Jurkat cells were provided by L. Gazzolo (Lyon, France).

RESULTS

Experimental strategy

HTLV-I double-spliced mRNA displays in its 5′ UTR four putative non-AUG initiation codons of translation (Fig. (Fig.1A).1A). Two GUG codons are located at positions 217 and 226, in frame with Rex AUG initiation codon (AUG250), and two CUG are located at positions 210 and 246, in frame with the Tax AUG initiation codon (AUG306). No other non-canonical initiation codon is present in the 5′ UTR, in frame or not with the Rex and Tax coding sequences. HTLV-I single-spliced mRNA exhibits a 5′ UTR strictly identical to that of the double-spliced species, up to a G in position 309 (Fig. (Fig.1B).1B). For detecting whether other initiations of translation may occur in the HTLV-I double-spliced mRNA, the CAT coding sequence was inserted in frame with either the Rex or the Tax AUG initiation codon, giving rise to two CAT constructs called pHTL-Rex–CAT and pHTL-Tax–CAT, respectively (Fig. (Fig.2,2, see details in Materials and Methods). pHTL-Rex–CAT was used for detecting potential initiations of translation in the Rex coding frame, while pHTL-Tax–CAT was used for identification of similar events, but in the Tax coding frame. pHTL-Rex–CAT and pHTL-Tax–CAT differ only by one nucleotide (see Materials and Methods). They were equally efficient for directing transcription of the CAT reporter gene (data not shown). In addition, as shown by S1 mapping, no transcription was detected in HeLa cells by use of the internal promoter supposedly used to direct the synthesis of a short mRNA encoding Tax protein only (not shown) (7,15). After transfection of pHTL-Rex–CAT and pHTL-Tax–CAT in HeLa cells, the CAT activity was measured (35). Both Rex–CAT and Tax–CAT proteins were easily detected by western blot with the same anti-CAT antibody. They exhibited similar half-lives as demonstrated by measure of CAT activity after treatment with cycloheximide (not shown). To identify translation initiation codons, different constructs were derived from pHTL-Rex–CAT and pHTL-Tax–CAT by point mutations. First, Rex and Tax AUG initiation codons were exchanged for other codons non-suitable for initiating translation, allowing us to characterize their use as bona fide initiation codons. Second, GUG and CUG codons present 5′ to and in frame with the Rex and Tax AUG initiation codons, respectively, were exchanged for either a non-initiation codon or AUG. This allowed us to characterize the GUG and CUG that were used as additional initiation codons of translation. Such codon usage has been characterized at least four times in completely different experiments.

GUG and CUG codons are used to synthesize additional Rex–CAT and Tax–CAT proteins

In transient expression, pHTL-Rex–CAT led to the synthesis of three CAT-fusion proteins, a main 29 kDa protein and two less abundant proteins of 30 and 31 kDa, while pHTL-Tax–CAT directed the synthesis of two proteins, a main 26 kDa protein and a supplementary 29.5 kDa Tax–CAT protein (Fig. (Fig.33).

Figure 3
Detection of several CAT-fusion proteins in HeLa cells transiently expressing either pHTL-Rex–CAT or pHTL-Tax–CAT. After transfection of HeLa cells with either pHTL-Rex–CAT or pHTL-Tax–CAT, total proteins were separated ...

In pHTL-Rex–CAT mRNA, when AUG250 was exchanged for AUU, the 29 kDa protein was no longer detected (Fig. (Fig.4A),4A), thus indicating that the main Rex–CAT protein was synthesized by translation initiating at AUG250. In this case, the 30 and 31 kDa additional Rex–CAT proteins were no longer detected either. When either GUG217 (Fig. (Fig.4B,4B, lane 2) or GUG226 (Fig. (Fig.4C,4C, lane 2) was exchanged for GUA in pHTL-Rex–CAT mRNA, the 31 (Fig. (Fig.4B)4B) or 30 kDa (Fig. (Fig.4C)4C) proteins were no longer detected. However, such changes did not impair the synthesis of the main 29 kDa Rex–CAT protein, neither did they impair that of the other additional Rex–CAT protein. Exchange of the first or second GUG for AUG led to an increased synthesis of the 31 and 30 kDa proteins, respectively, and almost to the absence of synthesis of the other Rex–CAT protein (Fig. (Fig.4B4B and C, lane 3). Altogether, these results demonstrated that GUG217 and GUG226 codons were used as translation initiation codons, leading to the synthesis of the 31 and 30 kDa additional Rex–CAT proteins.

Figure 4
Identification of three initiation codons directing the synthesis of Rex–CAT proteins. HeLa cells were transfected with either wild-type or modified pHTL-Rex–CAT in which putative AUG and GUG initiation codons were submitted to site-directed ...

In pHTL-Tax–CAT mRNA, when AUG306 was exchanged for UUG the 26 kDa protein was no longer detected (Fig. (Fig.5A),5A), thus indicating that the main Tax–CAT protein was synthesized by translation initiated by AUG306. In this case, the additional 29.5 kDa Tax–CAT protein was still weakly detected. Synthesis of the 29.5 kDa Tax–CAT protein was unmodified by exchange of CUG210 for CUU in pHTL-Tax–CAT mRNA (Fig. (Fig.5B,5B, lane 2), while exchange of CUG210 for AUG led to the synthesis of a CAT-fusion protein of even higher molecular mass than the 29.5 kDa protein (Fig. (Fig.5B,5B, lane 3). In this case, the main 26 kDa Tax–CAT protein was still synthesized, but apparently to a much lower extent (Fig. (Fig.5B,5B, lane 3). In contrast, when CUG246 was exchanged for CCG, the 29.5 kDa protein was no longer detected (Fig. (Fig.5C,5C, lane 2), while when exchanged for AUG, the 29.5 kDa protein was the only one to be detected, the main 26 kDa Tax–CAT protein being no longer detected (Fig. (Fig.5C,5C, lane 3). These results demonstrated that CUG246, but not CUG210, was recognized as a translation initiation codon for directing synthesis of the additional 29.5 kDa Tax–CAT protein.

Figure 5
Identification of two initiation codons directing the synthesis of Tax–CAT proteins. HeLa cells were transfected with either wild-type or modified pHTL-Tax–CAT in which putative AUG and CUG initiation codons were submitted to site-directed ...

AUG250 does not impair selection of CUG246 as an initiation codon

CUG246 and AUG250 are separated by only one nucleotide, a C, and the three nucleotides of AUG250 are located 5–7 bases downstream from CUG246. In order to verify whether AUG250, acting as an initiation codon, was interfering with translation initiated by CUG246, AUG250 was exchanged for either AUU or UAG. Obviously UAG stop codon cannot be used for initiating translation and, as described above, AUU250 could not be used in pHTL-Rex–CAT mRNA (Fig. (Fig.4A).4A). However, AUU but not UAG permits maintaining A and U +5 and +6 to CUG246. In pHTL-Tax–CAT mRNA, exchange of AUG250 for AUU did not significantly change the synthesis of the additional 29.5 kDa Tax–CAT protein (Fig. (Fig.6,6, lane 2). However, when AUG250 was exchanged for UAG instead of AUU, synthesis of the 29.5 kDa Tax–CAT protein was almost completely abolished (Fig. (Fig.6,6, lane 3). These results indicated that translation initiated by CUG246 was not significantly impaired by the presence of downstream AUG250, although it can also be used for initiating translation in another frame. They suggested also that A and U located +5 and +6 to CUG might be favorable features for use of CUG as an initiation codon, at least when this CUG is not immediately followed by a G (see Discussion). Table Table11 summarizes the nucleotide contexts of all non-AUG and AUG initiation codons present in the 5′ UTR of HTLV-I double-spliced mRNA.

Figure 6
Influence of AUG250 and CUG nucleotide context on translation initiated by CUG246. HeLa cells were transfected with either wild-type pHTL-Tax–CAT (lane 1) or the same construct in which AUG250 was exchanged for either AUU (lane 2) or UAG (lane ...
Table 1.
Nucleotide contexts of potential initiation codons located in the 5′ UTR of HTLV-I double-spliced mRNA

Detection of additional Rex and Tax proteins

Synthesis of additional Rex and Tax proteins was investigated in C91PL and MT-2 cells in which HTLV-I is permanently expressed, and in HeLa cells transfected with a construct (pCMV-2) directing the synthesis of an mRNA identical to HTLV-I double-spliced mRNA.

Use of three different sera from patients infected with HTLV-I allowed us to detect reproducibly several proteins in C91PL and MT-2 cell extracts, two main ones of 21 and 26 kDa, as well as another of 28–29 kDa in MT-2 cell extracts only. A typical analysis with one of these HTLV-I positive sera is shown in Figure Figure7.7. The 28–29 kDa protein was strongly detected as a doublet in the MT-2 cell extract, but never in that of C91PL. To investigate whether such proteins could correspond to Rex proteins, antibodies directed against each of these three proteins used as specific antigens were purified from the most reactive HTLV-I positive serum (see Materials and Methods). As expected, these antibodies recognized proteins by which they were purified. They also recognized the two other proteins, as well as the Rex protein produced as a GST–Rex (glutathione S-transferase) fusion protein but not GST alone (not shown). These results suggested that the 21 and 26 kDa proteins were very likely p21rex and p27rex, respectively, and that the 28–29 kDa proteins were two additional forms of Rex protein. Consistent with their apparent molecular mass, these additional Rex proteins were very likely synthesized by translation initiated by GUG217 and GUG226.

Figure 7
Detection of different Rex proteins in C91PL and MT-2 cells. Total proteins from C91PL and MT-2 cells, as well as from Jurkat and HeLa cells used as controls, were separated by SDS–PAGE (12%) then transferred onto a nitrocellulose membrane. ...

In addition to the expected 40 kDa Tax protein, which constantly appeared as a doublet in C91PL cells (Fig. (Fig.8A)8A) but as a single band in MT-2 cells and in pCMV-2 transfected HeLa cells (Fig. (Fig.8B8B and C), anti-Tax antibodies revealed an additional protein of ~43 kDa. This 43 kDa protein was revealed in pCMV-2 transfected HeLa cells, as well as in C91PL and MT-2 cells, permanently infected with HTLV-I (Fig. (Fig.8).8). The size of this additional Tax protein was compatible with translation initiated by CUG246. In HeLa cells transiently expressing the pCMV-2 construct, both Tax proteins of 40 and 43 kDa appeared to accumulate between 24 and 36 h after transfection (Fig. (Fig.88C).

Figure 8
Detection of an additional Tax protein. Total proteins from C91PL and MT-2 cells as well as from pCMV-2 transfected HeLa cells were separated by SDS–PAGE (10%), then transferred onto a nitrocellulose membrane. Immunodetection of Tax proteins ...

Conservation of non-AUG initiation codons in HTLV, STLV and BLV

If non-AUG initiation codons are conserved in other HTLV-I-related retroviruses at positions corresponding to GUG217, GUG226 and CUG246 codons in HTLV-I double-spliced mRNA, this would indicate that such codons might also be used as translation initiation codons, therefore reinforcing the evidence reported here for their use as initiation codons. Therefore, the HTLV-I sequence bearing the non-AUG initiation codons was compared to that of HTLV-2b, three strains of STLV and BLV (Table (Table2).2). This analysis revealed that non-canonical translation initiation codons are conserved, in almost all cases, at positions corresponding to HTLV-I GUG and CUG codons. Both GUG and the second CUG are conserved in HTLV-2b, STLV-1 and STLV-2PP1664. A CUG codon is found in STLV-PH969 instead of the first GUG and in BLV instead of the second GUG. Similarly, a GUG codon is found in STLV-PH969 instead of the second CUG. These viruses do not exhibit stop codons in frame with the three non-AUG codons upstream to Rex AUG. In BLV only, the second CUG was replaced by a UCA codon, not suitable for initiating translation. In all cases, a pyrimidine and a G are present at –3 and +4 of both GUG, respectively. With the exception of BLV, which does not exhibit a CUG matching HTLV-I CUG246, all CUG matching HTLV-I CUG246 codon exhibit a purine at position –3 and are followed by a pyrimidine at position +4. Even more remarkable is that a pyrimidine and a C are always found at positions –3 and +4 respective to all AUG matching HTLV-I AUG250, specifying the beginning of Rex protein synthesis. Finally, the mRNA of all these retroviruses, except STLV-1, contains a codon that cannot be used for initiating translation, instead of the corresponding HTLV-I first CUG codon, which is never used for initiating translation in the experiments reported here.

Table 2.
Comparison of HTLV-I sequence to that of HTLV-I-related retroviruses

DISCUSSION

The strategy used in the experiments described above allowed us to characterize the use of AUG and of non-canonical initiation codons of translation to direct the synthesis of Rex–CAT and Tax–CAT proteins. This was carried out by evaluating independently translation initiated by codons used for synthesis of either HTLV-I Rex or Tax and Env proteins in a living cell. This evaluation was made possible in transient expression of two different constructs that direct the synthesis of two mRNA species differing only by one nucleotide. In addition, the 5′ UTR of both mRNA species is identical to that of HTLV-I double-spliced mRNA. Translation of these two almost identical mRNA, instead of a single one encoding both Rex and Tax proteins, directs the synthesis of very similar CAT proteins differing only by a few amino acids at the N-terminus. Moreover, this strategy allowed us to overcome the expected changes in the amount and behavior of HTLV-I double-spliced mRNA encoding both Rex and Tax, inevitably occurring after their synthesis (1,2). Therefore, following this strategy, we were able for the first time to demonstrate unambiguously by site-directed mutagenesis that the same double-spliced mRNA molecule potentially directs the synthesis of three Rex and two Tax proteins by translation initiated by AUG and non-AUG codons.

Synthesis of all Rex–CAT and Tax–CAT proteins was shown by western blot, using the same anti-CAT antibody. This allowed us to estimate roughly that the two additional Rex–CAT as well as the additional Tax–CAT proteins were synthesized reproducibly in significant amounts. The complete loss as well as the recovery of their synthesis after exchange of GUG and CUG for either codons non-suitable for initiating translation or AUG codons, was direct evidence that the two GUG and the second CUG codons were used to initiate translation. Similarly, the complete vanishing of the main Rex–CAT and Tax–CAT proteins when AUG250 and AUG306 were exchanged for AUU and UUG, respectively, confirmed that these two AUG were used to initiate translation. These last results show also that AUU cannot be used to initiate HTLV-I mRNA translation when replacing AUG250, although AUU can be used sometimes as an initiation codon (see 20 for more details on the last point). In addition, change of downstream AUG250 for AUU interferes with the use of upstream GUG217 and GUG226 as initiation codons. On the contrary, removal of downstream AUG306 does not interfere with the use of upstream CUG246. These results suggest a mechanism of translation initiation different for Rex and Tax coding frames. In HTLV-I double-spliced mRNA, five codons are potentially used to initiate translation, three of them for the synthesis of Rex protein, the two others for that of Tax protein. Therefore, it is difficult to conceive that a simple scanning of the 5′ UTR by the 43S pre-initiation complex would be sufficient to explain such an intricate process for initiating translation (39). Indeed, the scanning has to be partially leaky many times before initiation of translation could ultimately proceed for the synthesis of Tax protein. Obviously, further work is needed to clarify how initiation of translation is regulated for ensuring Rex and Tax protein synthesis in appropriate amounts.

Insertion of additional AUG upstream to AUG250 by replacement of GUG217 and GUG226 in the Rex coding frame, as well as upstream to AUG306 by replacement of CUG246 in the Tax coding frame, led to the vanishing of the main Rex–CAT and Tax–CAT proteins, suggesting a scanning mechanism to initiate translation specifying synthesis of both Rex and Tax proteins. However, AUG250 that specifies the beginning of Rex protein synthesis did not impair the use of CUG246 as translation initiation codon in the Tax coding frame, although both codons are separated by only one nucleotide, even if this nucleotide is a C and not a G at position +4 of CUG246. Indeed, as clearly demonstrated, the main generalizable feature of the ancillary sequence immediately downstream from AUG and non-AUG initiation codons is a G at position +4, rather than AU at +5 and +6 (30). However, change of AUG250 for UAG abolished the use of CUG246 as initiation codon, while its change for AUU did not. At first glance, and because AU of AUG250 are at +5 and +6 of CUG246, these results would refute the demonstration that the optimal context for translation initiation does not extend beyond the nucleotide at position +4 (30). It would suggest also that AU at +5 and +6 are favorable features for use of CUG246 as initiation codon, as already suggested (19,40). However, the nature of nucleotides at –3 and +4 of both CUG246 and AUG250 (Table (Table2)2) may explain this result, at least partially.

The use of two GUG and one CUG codons to initiate translation was demonstrated for the synthesis of Rex–CAT and Tax–CAT proteins. Therefore, if these non-AUG codons were also used to initiate translation of HTLV-I double-spliced mRNA, additional Rex and Tax proteins of expected molecular mass should be found in permanently HTLV-I infected cells. They should also be expected in HeLa cells transfected with pCMV-2 which directs the synthesis of an mRNA identical to that of the HTLV-I double-spliced species. Synthesis of an additional Tax protein by C91PL and MT-2 cells, as well as by pCMV-2 transfected HeLa cells, was demonstrated by western blot using polyclonal anti-Tax antibodies. The weak signals given by the anti-Tax antibodies suggested a poor reactivity of these antibodies and/or a low level of Tax protein. However, both 40 and 43 kDa Tax proteins were reproducibly obtained several times. If these results are not direct evidence for the use of CUG246 as an additional initiation codon, they suggest it very strongly. Finally, if CUG246 may be used to initiate translation of the Tax coding sequence in the double-spliced mRNA, in addition to the main translation initiated by AUG306, CUG246 could also potentially be used to initiate translation of the Env coding sequence in the single-spliced mRNA, in addition to the main translation initiated by the same AUG306.

Demonstrating the bona fide synthesis of additional Rex proteins was much more difficult. The main reason was that despite all our efforts we were unable to find good available anti-Rex antibodies, as we were unable to obtain them by immunization of rabbits with purified GST–Rex fusion protein constructed for other purposes (41). To overcome this difficulty, we purified in sufficient amounts, what turned out to be anti-Rex antibodies, from the serum of an HTLV-I-infected patient. Because these antibodies independently obtained using the 21, 27 and 28–29 kDa proteins as purified antigens, were reactive against all 21, 27 and 28–29 kDa proteins as well as against GST–Rex fusion protein, we assumed that all these proteins exhibited Rex protein epitopes. Such indirect results strongly suggested that MT-2 cells synthesized in significant amounts all the different expected forms of Rex protein, i.e. p27rex and the two additional forms synthesized by translation initiating at GUG217 and GUG226, as well as p21rex (13,14).

Translation initiated by non-AUG codons has been demonstrated for only a few mRNAs, although the number of these mRNAs is growing. Such mRNAs are not only of viral origin, but they are also of cellular origin. When not used as sole translation initiation codons, non-AUG codons generally direct the synthesis of N-terminus-extended proteins which sometimes exhibit different localizations and/or functions (1822). Additional Rex–CAT and Tax–CAT proteins exhibited localizations identical to that of the main Rex–CAT and Tax–CAT proteins. This was demonstrated by indirect immuno-fluorescence using anti-CAT antibodies and vectors directing the synthesis of mRNA in which GUG and CUG codons were exchanged for AUG making the N-terminus-extended Rex–CAT and Tax–CAT proteins the only ones to be synthesized (not reported here). However, this does not preclude that the additional Rex and Tax proteins should exhibit somewhat different functions than the main Rex and Tax proteins in the regulation of HTLV-I gene expression, the pattern of which appears much more complex than initially thought.

ACKNOWLEDGEMENTS

We thank L. Gazzolo for the culture and preparation of C91PL, MT-2 and Jurkat cells, K. Kindbeiter for preparation of purified GST–Rex, and R. Grantham and J.-J. Diaz for critical reading of the manuscript. This work was supported by the Centre National de la Recherche Scientifique (UMR 5537), and by grants from the Association Nationale de Recherches sur le SIDA and Association pour la Recherche sur le Cancer. S.C. and T.M. were supported by fellowships from the Ministère de l’Enseignement Supérieur et de la Recherche and Association Nationale de Recherches sur le SIDA, respectivley.

REFERENCES

1. Nagashima K., Yoshida,M. and Seiki,M. (1986) J. Virol., 60, 394–399. [PMC free article] [PubMed]
2. Cullen B.R. (1992) Microbiol. Rev., 56, 375–394. [PMC free article] [PubMed]
3. Grassmann R., Berchtold,S., Aepinus,C., Ballaun,C., Boehnlein,E. and Fleckenstein,B. (1991) J. Virol., 65, 3721–3727. [PMC free article] [PubMed]
4. Siomi H., Shida,H., Nam,S.H., Nosaka,T., Maki,M. and Hatanaka,M. (1988) Cell, 55, 197–209. [PubMed]
5. Cullen B.R. (1998) Virology, 249, 203–210. [PubMed]
6. Hidaka M., Inoue,J., Yoshida,M. and Seiki,M. (1988) EMBO J., 7, 519–523. [PMC free article] [PubMed]
7. Nosaka T., Ariumi,Y., Sakurai,M., Takeuchi,R. and Hatanaka,M. (1993) Nucleic Acids Res., 21, 5124–5129. [PMC free article] [PubMed]
8. Bex F. and Gaynor,R.B. (1998) Methods, 16, 83–94. [PubMed]
9. Felber B.K., Paskalis,H., Kleinman-Ewing,C., Wong-Staal,F. and Pavlakis,G.N. (1985) Science, 229, 675–679. [PubMed]
10. Seiki M., Inoue,J.-I., Takeda,T. and Yoshida,M. (1986) EMBO J., 5, 561–565. [PMC free article] [PubMed]
11. Schmitt I., Rosin,O., Rohwer,P., Gossen,M. and Grassmann,R. (1998) J. Virol., 72, 633–640. [PMC free article] [PubMed]
12. Yoshida M., Suzuki,T., Fujisawa,J. and Hirai,H. (1995) Curr. Top. Microbiol. Immunol., 193, 79–89. [PubMed]
13. Berneman Z.N., Gartenhaus,R.B., Reitz,M.S., Blattner,W.A., Manns,A., Hanchard,B., Ikehara,O., Gallo,R.C. and Klotman,M.E. (1992) Proc. Natl Acad. Sci. USA, 89, 3005–3009. [PMC free article] [PubMed]
14. Koralnik I.J., Gessain,A., Klotman,M.E., Lo Monico,A., Berneman,Z.N. and Franchini,G. (1992) Proc. Natl Acad. Sci. USA, 89, 8813–8817. [PMC free article] [PubMed]
15. Ariumi Y., Shimotohno,K., Noda,M. and Hatanaka,M. (1998) FEBS Lett., 423, 25–30. [PubMed]
16. Nosaka T., Siomi,H., Adachi,Y., Ishibashi,M., Kubota,S., Maki,M. and Hatanaka,M. (1989) Proc. Natl Acad. Sci. USA, 86, 9798–9802. [PMC free article] [PubMed]
17. Ciminale V., Pavlakis,G.N., Derse,D., Cunningham,C.P. and Felber,B.K. (1992) J. Virol., 66, 1737–1745. [PMC free article] [PubMed]
18. Arnaud E., Touriol,C., Boutonnet,C., Gensac,M.C., Vagner,S., Prats,H. and Prats,A.C. (1999) Mol. Cell. Biol., 19, 505–514. [PMC free article] [PubMed]
19. Boeck R. and Kolakofsky,D. (1994) EMBO J., 13, 3608–3617. [PMC free article] [PubMed]
20. Hann S.R. (1994) Biochimie, 76, 880–886. [PubMed]
21. Jackson R.J. and Wickens,M. (1997) Curr. Opin. Genet. Dev., 7, 233–241. [PubMed]
22. Sasaki J. and Nakashima,N. (1999) J. Virol., 73, 1219–1226. [PMC free article] [PubMed]
23. Carroll R. and Derse,D. (1993) J. Virol., 67, 1433–1440. [PMC free article] [PubMed]
24. Prats A.-C., De Billy,G., Wang,P. and Darlix,J.-L. (1989) J. Mol. Biol., 205, 363–372. [PubMed]
25. Bugler B., Amalric,F. and Prats,H. (1991) Mol. Cell. Biol., 11, 573–577. [PMC free article] [PubMed]
26. Florkiewicz R.Z. and Sommer,A. (1989) Proc. Natl Acad. Sci. USA, 86, 3978–3981. [PMC free article] [PubMed]
27. Prats H., Kaghad,M., Prats,A.-C., Klagsbrun,M., Lelias,J.M., Liauzun,P., Chalon,P., Tauber,J.P., Amalric,F., Smith,J.A. and Caput,D. (1989) Proc. Natl Acad. Sci. USA, 86, 1836–1840. [PMC free article] [PubMed]
28. Gupta K.C., Ono,E., Ariztia,E.V. and Inaba,M. (1994) Biochem. Biophys. Res. Commun., 201, 567–573. [PubMed]
29. Kozak M. (1989) J. Cell Biol., 108, 229–241. [PMC free article] [PubMed]
30. Kozak M. (1997) EMBO J., 16, 2482–2492. [PMC free article] [PubMed]
31. Jackson R.J. and Kaminski,A. (1995) RNA, 1, 985–1000. [PMC free article] [PubMed]
32. Gossen M. and Bujard,H. (1992) Proc. Natl Acad. Sci. USA, 89, 5547–5551. [PMC free article] [PubMed]
33. Iwakura Y., Tosu,M., Yoshida,E., Takiguchi,M., Sato,K., Kitajima,I., Nishioka,K., Yamamoto,K., Takeda,T., Hatakana,M., Yamamot,H. and Sekiguchi,T. (1991) Science, 253, 1026–1028. [PubMed]
34. Chen C. and Okayama,H. (1987) Mol. Cell. Biol., 8, 2745–2752. [PMC free article] [PubMed]
35. Desbois C., Massé,T. and Madjar,J.-J. (1992) Trends Genet., 8, 300–301. [PubMed]
36. Diaz J.-J., Simonin,D., Massé,T., Deviller,P., Kindbeiter,K., Denoroy,L. and Madjar,J.-J. (1993) J. Gen. Virol., 74, 397–406. [PubMed]
37. Jeang K.-T., Giam,C.-Z., Neremberg,M. and Khoury,G. (1987) J. Virol., 61, 708–716. [PMC free article] [PubMed]
38. Jeang K.-T., Widen,S.G., Semmes,O.J.I. and Wilson,S.H. (1990) Science, 247, 1082–1084. [PubMed]
39. Kozak M. (1992) Crit. Rev. Biochem. Mol. Biol., 27, 385–402. [PubMed]
40. Grünert S. and Jackson,R.J. (1994) EMBO J., 13, 3618–3630. [PMC free article] [PubMed]
41. Diaz J.-J., Duc Dodon,M., Schaerer-Uthurralt,N., Simonin,D., Kindbeiter,K., Gazzolo,L. and Madjar,J.-J. (1996) Nature, 379, 273–277. [PubMed]

Articles from Nucleic Acids Research are provided here courtesy of Oxford University Press
PubReader format: click here to try

Formats:

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...