Substitution of the Rev-response element in an HIV-1-based gene delivery system with that of SIVmac239 allows efficient delivery of Rev M10 into T-lymphocytes

Background Human immunodeficiency virus type 1 (HIV-1)-based gene delivery systems are popular due to their superior efficiency of transduction of primary cells. However, these systems cannot be readily used for delivery of anti-HIV-1 genes that target constituents of the packaging system itself due to inimical effects on vector titer. Here we describe HIV-1-based packaging systems containing the Rev-response element (RRE), of simian immunodeficiency virus (SIV) in place of the HIV-1 RRE. The SIV RRE-containing packaging systems were used to deliver the anti-Rev gene, Rev M10, into HIV-1 susceptible target cells. Results An HIV-1 based packaging system was created using either a 272- or 1045-nucleotide long RRE derived from the molecular clone SIVmac239. The 1045-nucleotide SIV RRE-containing HIV-1 packaging system provided titers comparable to that of the HIV-1 RRE-based one. Moreover, despite the use of HIV-1 Rev for production of vector stocks, this packaging system was found to be relatively refractory to the inhibitory effects of Rev M10. Correspondingly, the SIV RRE-based packaging system provided 34- to 130-fold higher titers than the HIV-1 RRE one when used for packaging a gene transfer vector encoding Rev-M10. Jurkat T-cells, gene modified with Rev M10 encoding HIV-1 vectors, upon challenge with replication defective HIV-1 in single-round infection experiments, showed diminished production of virus particles. Conclusion A simple modification of an HIV-1 gene delivery system, namely, replacement of HIV-1 RRE with that of SIV, allowed efficient delivery of Rev M10 transgene into T-cell lines for intracellular immunization against HIV-1 replication.


Background
Lentivirus-based gene delivery systems have been used extensively for gene transfer into a variety of different target cells, both ex vivo and in vivo [1]. A possible application of lentivirus-based packaging systems based on human immunodeficiency virus type 1 (HIV-1) is for the delivery of anti-HIV-1 genes, such as siRNAs or genes that encode transdominant proteins, to HIV-1 susceptible cells for intracellular immunization [2]. However, the delivery of such genes using a packaging system based on HIV-1 is hampered by the inhibitory effect of the anti-HIV-1 genes, such as Rev M10, on the expression of either the helper or gene-transfer vector RNAs in the producer cells, resulting in low vector titers. Thus, HIV-1-based packaging systems are most useful if the anti-HIV-1 genes target those regions or products of the viral genome not present in the helper or gene-transfer vector constructs or target host genes, such as the gene for the CCR5 coreceptor [3][4][5].
All lentivirus-based gene delivery systems contain packaging or helper constructs for expression of viral Gag/Pol and gene transfer vectors that encode the transgene of interest. The expression of RNAs from both the Gag/Pol helper and the gene transfer vector constructs in HIV-1based packaging systems requires the coexpression of viral trans-acting regulatory protein Rev and its target sequence in the viral envelope coding region, the Rev response element or RRE [6,7].
It was previously shown that HIV-1 Rev could function with the RRE from HIV-2 or simian immunodeficiency virus (SIV), but the Rev proteins from HIV-2 or SIV were unable to function with HIV-1 RRE [8,9]. It should therefore be feasible to replace the HIV-1 RRE with the RRE from SIV in an HIV-1-based packaging system. In the present study, HIV-1 packaging systems containing the SIV RRE from SIVmac239 were created and found to provide titers equivalent to those obtained with HIV-1 Rev/ RRE-based system. Additionally, despite the use of HIV-1 Rev for vector stock production, the SIV RRE-based HIV-1 packaging system was found to be relatively refractory to the inhibitory effects Rev M10, a transdominant mutant of Rev [10]. The SIV RRE containing HIV-1 packaging system was used for the delivery of Rev M10 to Jurkat T-cells, which, upon challenge with HIV-1 in single-round infection assays, produced fewer virus particles than untransduced control cells.

Effect of homologous and heterologous transport proteins on vector production by HIV-1 and SIV RRE-based HIV-1 packaging systems
The SIVmac239 RRE exhibits about 87% homology with HIV-2 RRE. Lewis and coworkers mapped the RRE within HIV-2 env [9] and showed that the RRE activity was localized within a 1045 bp fragment. The activity could be narrowed down to a smaller fragment of 272 bp. Since SIV RRE had not been tested in an HIV-1 packaging system, several packaging and gene transfer vectors containing HIV-1 or SIV RREs were created ( Figure 1). The packaging constructs contained either a 1045-or a 272-nt putative minimal RRE. The gene transfer vectors were modified with the 1045 nt SIV RRE to more closely mimic the remnant HIV-1 RRE containing env sequence present in the control vector. Thus, both test and control gene transfer vectors included the 3'tat/rev splice acceptor site upstream of the transgene expression cassette.
As a first step, we wished to determine the effect of different Rev-like 'transport' proteins on vector stock production. To this end, vector stocks were produced in 293T cells using various combinations of packaging and genetransfer vectors encoding EGFP. All transfections received a vesicular stomatitis virus G glycoprotein expression construct (pMD.G), a Tat expression construct (pCMVtat) and a plasmid encoding secreted alkaline phosphatase (SEAP). Each transfection also received a Rev (HIV-1 Rev or SIV Rev) or HTLV-1 Rex expression construct. Virus titers in the supernatants of transfected cells were determined by infection of naïve 293T cells followed by flow cytometry to enumerate GFP+ cells [11]. The titers were adjusted for transfection efficiency by normalizing to the SEAP levels in the vector containing supernatant.
The results of vector titer determinations are shown in Figure 2(A-F). The control packaging system (Figure 2A) that used HIV-1 RRE in both packaging and gene transfer vector constructs provided SEAP-adjusted titers of 9.9 ± 0.45 × 10 6 infectious units per ml (I.U/ml) in the presence of HIV-1 Rev. Lower titers (1.7 ± 0.07 × 10 4 IU/ml) were achieved with HTLV-1 Rex. The SIV Rev was unable to function with the HIV-1 RRE as deduced from the basal vector titers obtained. When the 1045 bp SIV RRE was used in both the packaging and gene transfer vector constructs ( Figure 2D), as anticipated, viral titers significantly above basal were obtained with all three 'transport' proteins. Again, highest titers (1.1 ± 0.08 × 10 7 ) that were comparable to titers obtained with the control HIV-1 RRE based packaging system were obtained with the HIV-1 Rev. Titers were similar for SIV Rev (7.4 ± 0.2 × 10 5 ) and HTLV-1 Rex (9.0 ± 0.7 × 10 5 ), but the titers achieved were about an order of magnitude lower. The results were similar for a packaging system that used SIV RRE of 272nucleotide length in the packaging construct ( Figure 2F); however, the titers (3.6 ± 0.09 × 10 6 IU/ml) with HIV-1 Rev were lower than for the 1045 nt SIV RRE-based packaging system. When a combination or mixed packaging system was used, i.e. the packaging and gene transfer vectors used HIV-1 RRE for expression of one construct and the SIV RRE for the other construct ( Figure 2B, 2C and 2E), higher than basal viral titers were obtained in the presence of HIV-1 Rev or HTLV-1 Rex. The SIV Rev achieved only a marginal increase in titer over that obtained in the absence of any 'transport' protein expression construct. These results suggest that both packaging and gene transfer vector constructs must contain SIV RRE to provide useful titers with SIV Rev. The results also demonstrated that the HIV-1 packaging system with the1045 nt RRE provided titers higher than one with the 272 nt RRE. Finally, the results showed that a packaging system with 1045 nt SIV RRE achieved titers equal to that of the HIV-1 RRE-based one. The results, demonstrating the non-reciprocal nature of interaction of HIV-1 and HIV-2 or SIV Revs with the homologous and heterologous RREs, are consistent with the previous observations of Lewis, et al. [9] and Berchtold et al. [8].

RRE Used
To correlate vector titers to particle production, the supernatants used for infection were tested for HIV-1 p24 by ELISA. The SEAP-adjusted p24 levels are depicted in Fig

Determination of optimal concentrations of HIV-1 and SIV Rev expression plasmids for use with SIV RRE containing packaging system
HIV-1 Rev, in the previous experiment, achieved approximately 10-fold higher titers with the SIV RRE containing packaging system than SIV Rev. One possible interpretation of this result would be that HIV-1 Rev was more efficient than SIV Rev with SIV RRE. An alternative explanation could be that the steady state levels of HIV-1 Rev protein produced were higher than that of SIV Rev. In this case it should be possible to overcome the titer differences with a titration experiment to determine the optimal amounts of each of the Rev expression constructs required with the SIV RRE based packaging system. To this end, the SIV RRE containing packaging system was tested with increasing amounts (0.05 to 1.0 μg) of pCI-HIV Rev or pCI-SIV Rev constructs. The total amount of the 'transport' plasmid used in each transfection was kept constant by using pCI-Neo as a 'filler.' The titers of the resultant vector stocks shown in Figure 3 indicate that pCI-HIV Rev achieved higher titers with the SIV-RRE based packaging system than pCI-SIV Rev with its cognate RRE at all input amounts of each of the Rev expression constructs. To determine if these results could be explained by the steady state levels of the proteins, the lysates of 293T cells transfected with different amounts of pCI-HIV Rev and pCI-SIV Rev were subjected to an immunoblot assay procedure using anti-HA antibody. For the same input amount of Rev expression construct, pCI-HIV Rev showed approximately two-fold higher steady state levels of protein than pCI-SIV Rev (see Additional File 1). At the 0.1 μg amount, pCI-HIV Rev with the SIV RRE containing packaging system provided titers equivalent to that achieved by 1.0 μg of pCI-SIV Rev (indicated by a dashed line in Figure 3). The steady state levels of HIV-1 Rev protein at 0.1 μg was considerably lower than that of SIV Rev at 1.0 μg (see Additional File 1). These data suggest that the increased efficiency of HIV-1 Rev could be partly explained by better Rev expression levels and partly attributed to increased efficiency with SIV RRE. Clearly, additional work is necessary to further probe the reasons for the apparent increased efficiency of HIV-1 Rev over SIV Rev with SIV RRE.

The SIV RRE-based HIV-1 packaging system is relatively refractory to inhibitory effects of Rev M10
A previous study [8], using a luciferase-based reporter system, showed that the SIV RRE rendered the reporter less susceptible to inhibition by Rev M10. To determine the validity of this observation in the context of gene delivery systems, different amounts (0 μg to 1.0 μg) of an M10 expression construct, pCI-Rev M10, were added during production of vector stock with either the HIV-1 RRE or the SIV RRE-based packaging systems. The total amount of plasmid added was kept constant by using pCI-Neo as a 'filler' plasmid. The pCI-HIV-1 Rev was used for production of vector stock from the HIV-1 RRE-based packaging system. For the SIV RRE-based packaging system, in one set of transfections 0.1 μg of pCI-HIV Rev was used while Determination of optimal amounts of pCI-HIV-1 Rev for pro-duction of vector stocks with the SIV RRE-based HIV-1 pack-aging system Interestingly, the SIV RRE-based packaging system, when used with HIV-1 Rev, also proved to be less susceptible to inhibition than the HIV-1 RRE-based packaging system but more susceptible than the system that used SIV Rev, particularly at high input amounts of pCI-M10. This occurred despite the use of lower amounts of pCI-HIV Rev. A second experiment, using a different HIV-1 vector (pN-GIT72) [7] with the control HIV-1 RRE-based packaging system, provided comparable results (see Additional File 2).

Jurkat T-cells transduced with Rev M10 encoding HIV-1 vectors containing HIV-1 or SIV RRE produce fewer virus particles than cells transduced with control vectors upon challenge with replication defective HIV-1
Jurkat The differences between the different vector transduced cell populations could not be attributed to differences in infection levels since flow cytometry using PE-conjugated antibody to mouse CD24 (heat stable antigen) present in the challenge virus showed comparable levels of infection (see Additional File 5).
Virus particle production in Jurkat T-cells transduced with different HIV-1 vectors upon challenge with a replication defective HIV-1

Figure 5 Virus particle production in Jurkat T-cells transduced with different HIV-1 vectors upon challenge with a replication defective HIV-1.
Jurkat T-cells were separately transduced with each of the indicated vectors (X-axis) and sorted to greater than 95% purity. Each population was either mock-infected or infected with VSV-G pseudotyped pNL4-3.HSA. R -E -. The supernatants from mock or virusinfected cells were obtained on days 1, 4 and 7 and assayed for HIV-1 p24 capsid protein using a commercial ELISA kit. The Y-axis shows mean p24 levels produced by each of the different cell populations on days 4 and 7 normalized to the p24 produced by infected but untransduced Jurkat cells (which was set at 100%). The results shown are from three independent experiments. Error bar = 1 SD.

Discussion
There has been a resurgence of interest in evaluating Rev M10 for intracellular immunization in HIV-1 infected patients [2]. However, the usage of HIV-1-based packaging systems to deliver Rev M10 has been particularly problematic due to the inhibitory effect of Rev M10 on vector stock production (Figure 4). Modifications to the HIV-1 packaging system to render it resistant to Rev M10, such as the use of the constitutive transport element of Mason-Pfizer monkey virus [12], would enable its use for anti-HIV-1 gene therapy.
In this study, we describe the use of SIV RRE to replace the HIV-1 RRE in a HIV-1 based-packaging system to achieve the same ends. The results showed that the SIV RRE was able to substitute for the HIV-1 RRE in both packaging as well as the gene transfer vector constructs. The SIV RREbased packaging systems were found to be not only as efficient as the HIV-1 RRE-based one for production of vector stocks ( Figure 2), but also relatively refractory to Rev M10 (Figure 4), despite the use of HIV-1 Rev for production of vector stocks. Our study confirms and extends the earlier study by Berchtold and coworkers [8] who, using a different reporter construct based on expression of luciferase, also showed resistance to Rev M10 of SIV RRE containing construct. To the best of our knowledge, our study is the first to evaluate SIV RRE in the context of an HIV-1-based gene delivery system.
The ability of SIV RRE to render the packaging system relatively resistant to Rev M10 allowed the production of high-titered stocks of vectors encoding Rev M10 (Table 1). When Jurkat T-cells transduced with M10 encoding SIV-RRE containing vectors were challenged with a replication defective HIV-1, in single round infection assays, cells transduced with the Rev M10 encoding vectors produced lower amounts of virus particles than cells transduced with vectors encoding EGFP alone ( Figure 5). The differences observed between the Rev M10 expressing cells and control cells, unmodified or expressing EGFP alone, could not be attributed to different levels of infection of the cells since flow cytometry using anti-mouse CD24 antibodies to heat-stable antigen revealed that the percentage of cells infected in the M10 expressing population was similar to the control EGFP expressing cells (Additional File 5). The differences between the different vectors could also not be assigned to variations in the level of Rev M10 or EGFP expression since the sorting was carried out using a narrow window to ensure that comparable levels of EGFP expression was present in the different populations. Moreover, both vectors achieved similar levels of expression of Rev M10 as deduced from EGFP levels since EGFP and Rev M10 expression was linked at the translational level.
Despite the ability of SIV RRE to mitigate the inhibitory effects of Rev M10 during vector stock production, the observation that the SIV RRE containing vector encoding Rev M10 was found to be as efficient as the control Rev M10 expressing vector containing HIV-1 RRE in decreasing HIV-1 particle production ( Figure 4) in transduced target cells can be explained as follows. The presence of constitutively expressed Rev M10 in the target cell, due to gene modification, would ensure interference with the function of wild-type Rev produced from the challenge virus, even at the earliest time points. This would then prevent significant accumulation of full-length viral RNA that encodes Gag/Pol or the vector RNA containing SIV RRE. Thus, the concentration of SIV RRE containing transcript in the gene-modified Jurkat T-cell is likely to be too low to obtund Rev M10 function.
In addition to the use of the SIV RRE based packaging system for delivery of Rev M10, one could possibly use such a system for targeting the HIV-1 envelope sequence employing RNAi approaches. Employing distinct RNA transport elements for expression of helper and gene transfer vector RNA can reduce the risk of recombination between the packaging and gene transfer vector constructs during vector stock production [13,14]. Such packaging systems can be used for delivery of any transgene of interest. Here we have demonstrated that the SIV RRE can replace HIV-1 RRE in either the packaging or gene transfer vector with no loss of titer.
An alternative approach to decreasing recombination frequency between components of packaging systems is by using hybrid packaging systems consisting of helper and gene transfer constructs derived in their entirety from viruses with low sequence homology, such as SIV (or HIV-2) and HIV-1 [15,16]. The major concern in the case of the hybrid packaging systems is the low efficiency of encapsidation of the heterologus vector RNA [17,18] in comparison to the homologus vector RNA. In contrast to those studies, HIV-1 packaging systems that utilize only the SIV RRE of the different viruses are not likely to have such drawbacks. However, a direct comparison of the different packaging systems is necessary to determine the suitability of different packaging systems for specific therapeutic applications.
It was previously hypothesized that the Rev M10 protein inhibits wild-type Rev function by formation of mixedmultimers with wild-type Rev protein [19,20]. The findings in this study, and that of Berchold and coworkers [8], appear to challenge that hypothesis since the mere presence of SIV RRE in the producer cell seemed to obtund the inhibitory effect of Rev M10 on wild type Rev.

Conclusion
The present study demonstrated that an HIV-1-based packaging system containing only the RRE sequence from SIV can be used for efficient delivery of Rev M10 into HIV-1 susceptible cells to achieve intracellular immunization. Furthermore, the studies showed that SIV RRE could be used in the context of a reciprocal or combination packaging system to improve its safety without compromising vector titers.

Gene-transfer vectors
The gene-transfer vectors ( Figure 1B) are similar to the previously described pN-EF1α-MGMT-WPRE vector [22]. The vector pN-EF1α-EGFP/HIV-1 RRE was derived from the molecular clone pNL4-3 and has a deletion between proximal (nt 1247) and distal (nt 6738) NsiI sites of pNL4-3. The remnant portion of the HIV-1 env contains the RRE. The vector has an engineered frame-shift (FS) mutation in gag [6] and the central polypurine tract and central termination sequences (CPPT/CTS) to improve gene-transfer efficiency [23][24][25]. The transgene expression cassette, positioned between the BamHI site in the second coding exon of Rev that overlaps the 3' end of env and the XhoI site in nef, consists of human elongation factor 1 alpha (EF1α) promoter driving enhanced green fluorescent protein (EGFP). The woodchuck post-transcriptional regulatory element (WPRE) [26] was placed downstream of the EGFP coding sequence. To create the SIV RRE containing vector pN-EF1α-EGFP/SIV RRE, a 1045 nt SIV RRE (described above for pGP/SIV 1045 RRE) was inserted between BsaBI and EcoRI sites of pN-EF1α-EGFP/HIV-1 RRE, effectively replacing the HIV-1 RRE with that of SIV. The vectors pN-EF1α-EGFP-2A-M10/HIV-1 RRE and p-EF1α-EGFP-2A-M10/SIV RRE are identical to the abovedescribed vectors but express both EGFP and Rev M10 instead of EGFP alone. The EGFP and Rev M10 coding sequences were linked in-frame by the foot and mouth disease virus 2A cleavage factor sequence. Inclusion of the 2A sequence in-frame results in the cleavage and release of EGFP-2A and Rev M10 proteins from the engineered polyprotein and ensures equimolar expression of both transgenes [27,28].

pCI-HIV-Rev and pCI-Rev M10
These contain the HIV-1 Rev coding sequence amplified from pCMVRev (corresponds to nt 970 to nt 1320 in HIVPCV12; [GenBank:M11840]) with an added hemagglutinin (HA) epitope tag (MYPYDVPDYA) at the N-terminus and inserted into pCI-Neo (Promega Corp., Madison, WI) between the human cytomegalovirus immediate early promoter and polyadenylylation signal of SV40 virus. A synthetic intron is present upstream of the Rev coding sequence. pCI-Rev M10 is identical to pCI-HIV-Rev but contains the classic mutation in the nuclear export sequence (LQLPPLERLTLD) of HIV-1 Rev in which residues LE (CTTGAG) were changed to DL (GATCTC) [10].

pCI-SIV Rev
This plasmid contains the Rev coding sequence amplified from p239SpE3' [29] which contains the 3' half of SIVmac239. The SIV Rev corresponds to nt 6784 to nt 6853 (first coding exon) and nt 9062 to nt 9315 (second coding exon) of SIVmac239 joined in-frame using splicing by overlap extension (SOE) PCR [30,31]. An N-terminal HA epitope tag was engineered in the same manner as for pCI-HIV-Rev. The amplified sequence was inserted into pCI-Neo as described above for pCI-HIV-Rev.