Transfer of Primer Binding Site-Mutated Simian Immunodeficiency Virus Vectors by Genetically Engineered Artificial and Hybrid tRNA-Like Primers

Abstract

Simian immunodeficiency viruses (SIV) harbor primer binding sites (PBS) matching tRNA or tRNA. To study determinants of primer usage in SIV, a SIVmac239-based vector was impaired by mutating the PBS to a sequence (PBS-X2) with no match to any tRNA. By cotransfection of a synthetic gene encoding a tRNA^Pro-like RNA with a match to PBS-X2, the activity of this vector could be restored to a transduction efficiency slightly lower than that of the wild-type vector. A vector with a PBS matching tRNA^Pro was functional at a level slightly below that of the wild-type vector, but higher transduction efficiency could be obtained by cotransfection of a gene for an engineered tRNA^Pro-tRNA hybrid with a match to PBS-Pro. The importance of tRNA backbone identity was further analyzed by complementing the PBS-X2 vector with a gene for a matching x2 primer with a tRNA backbone, which led to three- to fourfold-higher titers than those observed for the x2 primer with the tRNA^Pro backbone. In summary, our results demonstrate flexibility in PBS and primer usage for SIVmac239, with PBS-primer complementarity being the major determinant, in analogy with previous findings for murine leukemia viruses and human immunodeficiency virus type 1.

Retroviruses replicate through reverse transcription of the viral RNA genome into a double-stranded DNA form that is integrated into the genome of the host. The virus-encoded reverse transcriptase (RT) enzyme mediates this process through the use of a host-encoded tRNA. The 3′-terminal 18-nucleotide (nt) segment of the tRNA primer molecule anneals to the RNA genome at a complementary sequence known as the primer binding site (PBS). In order to synthesize a double-stranded DNA copy, two template shifts are involved. In plus-strand synthesis, the 3′ 18-nt segment of the tRNA primer is reverse transcribed into a DNA copy, which mediates the second template shift by annealing to the minus-strand DNA copy of the PBS. Accordingly, the PBS sequence of the transduced provirus may originate from a copy of the genomic PBS RNA sequence or from copying of the 3′ 18 nt of the tRNA primer.

Different tRNA species are used as primers among retroviruses; however, for each group of viruses, the PBS is stably maintained during replication. Murine leukemia viruses (MLV) and human T-cell leukemia virus use a tRNA^Pro, the avian sarcoma-leukosis virus (ASLV) group uses a tRNA^Trp, and a tRNA is used as a primer for reverse transcription in mouse mammary tumor virus and in human and simian immunodeficiency viruses (HIV and SIV, respectively) (reviewed in references 33 and 35).

Mechanisms of PBS maintenance during replication have been studied in a variety of viruses. MLV mutated in the PBS to match the 3′ end of other tRNAs are able to use noncognate primers in multiple rounds of replication (6, 30, 36). In contrast, PBS-mutated HIV-1 and ASLV may initially replicate, but in time revert to their cognate tRNA during serial transfer of PBS-mutated viruses (7, 25, 47, 49). The selective reversion and maintenance of the cognate primer in these viruses may be accomplished by non-PBS determinants.

The lentiviruses of the SIV group have attracted attention due to the development of animal models for the study of AIDS (9) and the development of lentivirus vectors for gene therapeutic purposes (34, 43). Although SIVs and HIV-1 share common features, the SIVmac239 isolate used in this study differs from HIV-1 in lacking an adenosine-rich loop situated in the U5 hairpin (3, 4), which has been implicated in determining primer specificity in HIV-1 (19–24, 27, 46, 52). To study the determinants for primer specificity in SIV, we have established a system in which the viral RNA genome as well as the tRNA primer may be manipulated. By using synthetic genes encoding tRNA-like molecules complementary to PBS-mutated SIV vectors, it is demonstrated that the PBS is the primary determinant for primer specificity in initiation of first-strand synthesis and plus-strand transfer resulting in the formation of a provirus. In this experimental setup, it is possible to study the effect on the transduction titers exerted by sequences outside the PBS-complementary 3′ 18 nt of the tRNA-like primer, because no endogenous competitor tRNA molecule is able to compete for binding to the mutant PBS.

Design of PBS-mutated vectors and engineered tRNA-like primers.

The PBS of a SIV-based green fluorescent protein (GFP) vector (SIVGFP) was mutated, and constructs encoding putative tRNA-like primers were designed. SIVGFP-Pro contains a PBS-Pro for which the corresponding tRNA proline is found in the cell. SIVGFP-X2, however, harbors a PBS-X2 for which no matching endogenous tRNA exists (Fig. 1). In the case of SIVGFP-Pro, a PBS for the tRNA proline was introduced into the PBS of SIVGFP by PCR-mediated site-directed mutagenesis (29) in which primer set A (primers 1 and 2) (Table 1) and primer set B (primers 3 and 4) generated two PCR products with SIVGFP as a template. The two PCR products were combined in an overlap extension reaction with primers 2 and 4. The resulting fragment was cloned into SIVGFP. For SIVGFP-X2, the PBS was mutated at 11 nt positions that were originally selected to give a minimal match to any known murine tRNA (28) with primer set A (primers 5 and 2) and primer set B (primers 6 and 4). The SIVGFP vector is an env deletion mutant of SIVmac239, in which the nef gene has been replaced by the enhanced GFP (EGFP) reporter gene (Fig. 1) (Δenvgfp in reference 43).

FIG. 1.

Vector structure. The SIVmac239-based vector SIVGFP is shown at the top. The sequences of the introduced PBS mutations are shown below. The mismatch marker mutation relative to the engineered corresponding tRNA-like primer is shown in boldface.

TABLE 1.

Oligonucleotides used in this study

Primer	Sequence (5′→3′)
1	p-TTGGGGGCTCGTYCGGGATTTGAAGGAGAGTGAGAGACTCCTGA
2	CGGGCTTAATGGCAGGTGGACA
3	p-GTACTCAGGAGTCTCTCACTCTCCTTCAAATCCCGRACGAGGGGC
4	CGCTGAAACAGCAGGGACTT
5	p-TTGGTCAGYTGCAGGGGATTTGAAGGAGAGTGAGAGACTTCTGA
6	p-TTCAAATCCCCTGCARCTGACCAATCTGCTAGGGATTTTC
7	GGAAAGCTTAAAAAAGACATGC
8	CCGGAATTCAAAAAGTAAAGCTCTCGTGAAGACAGCTGCTAGCTCAGTCGGTAGAGCATCAGACTTTTAATCTGAGGGTCAGGGGTTCAAATCCCCTGCAGCTGAGCATGTCTTTTTTAAGCTTTCC
9	CCGGAATTCAAAAAGTAAAGCTCTCTCGTGAAGAGGCTTAGCTCAGTCGGTAGAGCATCAGACTTTTAATCTGAGGGTCCCGGGTTCAAATCCCGGACGAGCCCGCATGTCTTTTTTAAGCTTTCC
10	TGGTCAGCTGCAGGGGAT
11	Biotin-AAAAATTGGTCAGCTGCAGGGGAT
12	GGGGAATTCTGGAAGGGATTTATAAGAG
13	CCCCTTCCTGGATAAAAGACA

The constructs designed to encode the synthetic tRNA-like primer molecules ptRNA^x2pro, ptRNA^x2lys-3, and ptRNA^prolys-3 were constructed from deoxyoligonucleotides and cloned as described in reference 28. ptRNA^x2pro is identical to ptRNA^x2 in reference 28, whereas ptRNA^x2lys-3 and ptRNA^prolys-3 were constructed by annealing the 22-nt elongation primer 7 to the 127-nt primers 8 and 9, respectively. The sequences of primers 8 and 9 were based on that of tRNA (GenBank accession no. K01797) (41). The predicted structures of ptRNA^x2pro and ptRNA^prolys-3 are shown in Fig. 2A. The 3′ 18-nt sequence of ptRNA^x2pro matches that of PBS-X2, but otherwise it resembles the sequence of tRNA^Pro, whereas ptRNA^prolys-3 matches PBS-Pro at the 3′ end, but resembles tRNA. The synthetic tRNA-like primers were processed correctly with regard to trimming of the tRNA precursor at the 5′ and 3′ termini, because the transcript sizes correspond to the expected 75 and 76 nt for ptRNA^x2pro and ptRNA^x2lys-3, respectively (Fig. 2B), as shown by Northern hybridization after transfection into human BOSC 23 packaging cells (37), as described in reference 28. The addition of the nonencoded CCA sequence to the trimmed 3′ end of the synthetic tRNAs was verified by a tRNA tagging assay (31), in which the presence of the CCA tail was a designed prerequisite (Fig. 2C) for the synthesis of the radioactively labeled product seen in Fig. 2D.

FIG. 2.

(A) Structure of the murine tRNA^Pro (16) and tRNA (38) and putative structures of synthetic tRNAs. Mutations relative to tRNA^Pro and tRNA are boxed. (B) Northern hybridization. Ten micrograms of total RNA from human BOSC 23 cells transfected with ptRNA^x2lys-3 (lane 2) or ptRNA^x2pro (lane 3) was separated on an 8% polyacrylamide–urea gel, blotted, and hybridized to an X2-specific probe (primer 10). The DNA markers are shown in lane 1. (C) tRNA tagging assay. Total RNA from BOSC 23 cells was incubated with a biotinylated oligonucleotide complementary to the 3′ 18 nt of tRNA^Pro (31) or tRNA^x2lys-3 (primer 11), respectively. Following magnetic separation, the annealed tRNA was extended in the presence of [³²P]dATP and separated on an 8% polyacrylamide–urea gel. (D) tRNA tagging gel showing total RNA from BOSC 23 cells incubated with the tRNA^Pro-specific oligonucleotide (lane 2). In lane 3, total RNA from BOSC 23 cells transfected with ptRNA^x2lys-3 was incubated with primer 11. Lane 1 shows DNA size markers.

A mismatch between vector PBS and primer was introduced by altering 1 nt in the PBS of SIVGFP-X2 and SIVGFP-Pro, yielding SIVGFP-X2m and SIVGFP-Prom, respectively (Fig. 1), to distinguish the PBS sequences of the primer and PBS origin during replication.

Complementation of PBS mutations by designed tRNA-like primers.

The effect of PBS mutations on the infectivity of SIVGFP was determined by transient transfections of 293T cells (293ts/A1609) (10) with the different mutants of SIVGFP together with the vesicular stomatitis virus glycoprotein G (VSV-G)-encoding pHIT-G plasmid (14) following calcium phosphate coprecipitation (44). Table 2 shows the SIVGFP vector titers in the supernatant of the transfected cells determined on 293A cells (Quantum Biotechnologies, Laval, Canada). 293A cells, a subclone of 293 cells, were used for titration since they show stronger adherence to plastic dishes than 293T cells. The transfection efficiency was monitored by determination of capsid antigen levels in the supernatant of the transfected cells by employing an HIV-1 capsid enzyme-linked immunosorbent assay kit (Innogenetics, Ghent, Belgium). The 293A target cells were seeded in 24-well plates at a density of 5 × 10⁴ cells per well. The following day, the medium was removed and cells were incubated for 2 to 4 h with serial dilutions of 200 μl (per well) of the supernatant from the transfected cells. Two days after infection, the number of GFP-positive cells was counted. The infectivity of the vectors was calculated by dividing the vector titer by the amount of capsid antigen and is expressed relative to the infectivity of wild-type SIVGFP.

TABLE 2.

Vector transduction efficiencies

Plasmid cotransfected with VSV-G		Results for:
PBS mutants	tRNA/mock	Expt 1		Expt 2
		Titer (GFU)/ng of p27CAa	% Infectivityb	Titer (GFU)/ng of p27CAa	% Infectivityb
SIVGFP	pBluescript	5.5 × 105/2.7	100	1.5 × 105/9.2	100
SIVGFP-X2	pBluescript	<5/3.1	<0.0008	5/11.4	0.0027
SIVGFP-X2	ptRNAx2pro	8.5 × 104/2.6	16	2.4 × 104/6	25.9
SIVGFP-X2m	pBluescript	<5/5.7	<0.0004	20/14	0.0088
SIVGFP-X2m	ptRNAx2pro	9.8 × 104/2.7	18	8.7 × 104/6.1	87.5
SIVGFP-Pro	pBluescript	7.7 × 104/8.7	4.3	2.4 × 104/14.6	10
SIVGFP-Pro	ptRNAprolys-3	2.8 × 105/8.7	16	9.2 × 104/15.8	36
SIVGFP-Prom	pBluescript	1.9 × 104/6.4	1.5	3.1 × 103/14.9	1.2
SIVGFP-Prom	ptRNAprolys-3	1.8 × 105/7.7	11.5	2.1 × 104/11	11.9

^aValues are in green fluorescence-forming units (GFU) and nanograms of p27CA antigen per milliliter of supernatant of 293T cells transfected with the indicated plasmids. 

^bPercentage of infectivity (titer of p27CA antigen) relative to wild-type SIVGFP. Of each plasmid DNA, 2.5 μg was used for transfection of a six-well plate. 

Mutation of the PBS of SIVGFP to the X2 sequence reduced the titer by at least 4 orders of magnitude with essentially no background (Table 2). By the addition of a synthetic tRNA-like primer, ptRNA^x2pro, matching PBS-X2, the infectivity of SIVGFP-X2 was increased by at least 4 orders of magnitude and was only fivefold lower than the infectivity of SIVGFP wild-type particles. The PBS-mutated vector could therefore be complemented by the synthetic ptRNA-like primer matching the 18 nt of the PBS-mutated vector.

Direct sequence evidence for ptRNA^x2pro usage was obtained from amplification with primers 12 and 13 and sequencing the PBS region of a mixed pool of cell lysates originating from transductions with the supernatant of SIVGFP-X2m transfected cells. In the left panel of Fig. 3, the PBS sequence of the PBS-X2m origin is shown, whereas the panel to the right shows the PBS sequence, originating from a copy of ptRNA^x2pro, thus providing the evidence for primer usage. According to the mechanism of reverse transcription, the PBS sequence of a transduced provirus originates from copying the genomic PBS RNA or the 3′ 18 nt of the tRNA primer. Consequently, it would be predicted that 50% of the clones would contain the primer sequence. It was demonstrated that 5 subclones out of 13 analyzed contained the artificial tRNA sequence.

FIG. 3.

Genetic evidence for tRNA^x2pro usage. PCR products spanning the PBS region from lysates prepared from a mixed pool of transduced cells were cloned and sequenced. In the left and right panels are shown the PBS sequences resulting from a copy of the vector SIVGFP-X2m and primer tRNA^x2pro, respectively. The mismatch positions are indicated by arrows.

Effect of using primers with a tRNA-like backbone.

A comparison of the infectivity of SIVGFP-Pro in the presence or absence, respectively, of a hybrid primer, tRNA^prolys-3, which had an acceptor stem matching PBS-Pro but otherwise resembled the tRNA molecule (Fig. 2A), showed that cotransfection of ptRNA^prolys-3 increased the infectivity fourfold (Table 2). This suggests that sequences or structures in the backbone of the cognate tRNA compared to the endogenous tRNA^Pro are important for optimal initiation of reverse transcription or packaging. In these experiments, accurate quantification of the expression levels is difficult due to abundant expression of endogenous tRNAs closely homologous to the transfected tRNAs. However, primer selectivity by the retroviral machinery during primer packaging, placement, and utilization may be the determining factor rather than abundance of the tRNA. To compare the effects exerted by the tRNA sequences outside the PBS complementary region, ptRNA^x2lys-3 was designed, containing a tRNA backbone, but matching the PBS-X2 (Fig. 2A), and used in transient transfections. The infectivity of the SIVGFP-X2 obtained by cotransfection with ptRNA^x2lys-3 resulted in three- to fourfold-higher transduction titers relative to cotransfection with ptRNA^x2pro (Table 3).

TABLE 3.

Effect of tRNA backbone identity on vector transduction efficiencies

Plasmid cotransfected with VSV-G		Results for:
PBS mutants	tRNA mock	Expt 1		Expt 2		Expt 3
		Titer (GFU)/ng of p27CAa	Infectivityb	Titer (GFU)/ng of p27CAa	Infectivityb	Titer (GFU)/ng of p27CAa	Infectivityb
SIVGFP-X2m	pBluescript	<5/131		10/95		<5/86
SIVGFP-X2m	ptRNAx2pro	1.4 × 105/132	1 × 103	5.9 × 104/45	1.3 × 103	4.9 × 104/44	1.1 × 103
SIVGFP-X2m	ptRNAx2lys-3	4.1 × 105/106	3.9 × 103	2.4 × 105/54	4.4 × 103	3.1 × 105/109	2.8 × 103
x2lys3/x2pro ratio		2.9	4	3.9	3.4	6.3	2.5

^aValues are in green fluorescence-forming units (GFU) and nanograms of p27CA antigen per milliliter of supernatant of 293T cells transfected with the indicated plasmids on 293A cells. 

^bInfectivity is given as the titer of p27CA antigen. ptRNA expression plasmid DNA (1.25 μg) was cotransfected with 2.5 μg of SIVGFP-X2 and 2.5 μg of pHITG. The total amount of DNA transfected was adjusted to 10 μg for a six-well plate. 

While normalization of transfection efficiency by p27CA antigen levels within each experiment gave consistent results, normalization of transfection efficiency between different experiments was not found to be very useful. Although CA antigen levels differed largely between different experiments (compare Tables 2 and 3), vector titers did not seem to increase in a linear way with increasing p27CA antigen levels. Therefore, CA antigen levels might not be the limiting factor for vector titer, or increased toxicities at higher transfection rates might lead to stronger inhibition of formation of infectious vector particles than protein expression. To confirm that the vector titer depends on the amount and type of tRNA, two independent titration experiments were performed with cotransfection of increasing amounts of ptRNA^x2pro and ptRNA^x2lys-3, respectively, in cotransfections with SIVGFP-X2 (Fig. 4A). For amounts of tRNA expression plasmid up to 1 μg, the transduction titer increases for a given amount of vector DNA. Increasing the amount of ptRNA above the saturating level seems to have an inhibitory effect on transduction efficiency. Throughout the entire range of DNA concentrations, the cotransfection of ptRNA^x2lys-3 results in higher transduction titers than those of ptRNA^x2pro, with the exception of the experiment using 2.5 μg of ptRNA^x2pro. Such a preference for ptRNA^x2lys-3 was not observed in an MLV system using PBS-mutated Akv-based vectors in which the addition of tRNA^x2pro results in slightly higher titers than those used for addition of the ptRNA^x2lys-3 (Fig. 4B). Construction of MLV-based vectors and the titer assay was described previously in reference 28. Thus, it seems that SIV shows a weak preference for synthetic tRNA-like primers with a tRNA backbone.

FIG. 4.

Titration of engineered tRNA-like DNAs differing in their tRNA backbone compositions. Dose-response curves of the effects of increasing amounts of engineered tRNA-like primer DNA ptRNA^x2pro and ptRNA^x2lys-3 on the infectivity of an SIV-based vector (A) and the titer of an MLV-based vector (B) are shown. (A) Engineered tRNA-like primers were cotransfected in the indicated amounts with SIVGFP-X2 and pHITG into 293T cells. The infectivity (green fluorescence-forming units [GFU] per nanogram of p27CA) in the supernatant of the transfected cells was determined on 293A cells. (B) Engineered tRNA-like primers were cotransfected in the indicated amounts with the MLV-based vector pPBS-X2 (28) into the packaging cell line BOSC 23. Vector titers in the supernatant of the transfected cells were determined on NIH 3T3 cells. Black and grey bars represent cotransfections with ptRNA^x2pro and ptRNA^x2lys-3, respectively. Replica experiments are shown for each amount of ptRNA added.

Our data demonstrate that complementarity between the PBS and a matching synthetic tRNA-like primer is the primary determinant for selection and use of the tRNA species in initiation of reverse transcription in SIV. However, we note that the transduction titers are lower than for the wild-type vector. An MLV Akv-based vector carrying PBS-X2 was greatly impaired in its ability to replicate in single-round transfections, but could be rescued by the addition of the designed complementary tRNA^x2pro, resulting in a complete provirus (28). Similarly, an HIV-1 provirus with a PBS complementary to the yeast tRNA^Phe relied on the transfection of a yeast tRNA^Phe, using a single-round transfection system (51). Thus, SIVmac239 seems to be similar to MLV, ASLV, and HIV-1, in which the PBS and tRNA complementarity is the primary determinant for selection of the tRNA primer for initiation of reverse transcription and the ability to proceed through a complete replication cycle.

Northern analysis of the expression levels of the synthetic tRNAs relative to a control U2 small nuclear RNA probe (48) has shown that ptRNA^x2lys-3 is expressed at higher levels than ptRNA^x2pro in the producer cells used (data not shown). However, it is not clear whether the level of tRNA abundance per se has any effect on transduction efficiency, because the amount of synthetic tRNA DNA added in the transfections seems to be saturating (Fig. 4). Secondary determinants such as non-PBS interactions may be responsible for the apparent preference for the tRNA backbone, as has been demonstrated for HIV-1. Viral genomic RNA sequences other than the A-rich loop may interact with the primer in SIVmac239, although interactions between tRNA^Pro and the Moloney MLV RNA have been found to be restricted to the PBS (13). In HIV-1, primer placement on the viral genome and primer packaging may depend on destabilization of the secondary structure of the tRNA and the viral RNA template by the p55^gag precursor (1, 5, 8, 12, 17, 18, 26, 32, 39, 40), which also seems to be the case for avian leukosis virus, but not for MLV (15). Primer and RT interaction may also be necessary for initiation of reverse transcription (1, 2, 11, 42, 50), although RT may recognize the overall L-shaped tRNA rather than specific features of its cognate tRNA primer (45). Thus, we cannot exclude that one or several of the possible non-PBS interactions may contribute to the observed higher titers in SIV for the tRNA backbone in this study; however, the primary determinant for primer specificity is the complementarity between the viral PBS and the tRNA primer in SIVmac239.