table of contents
The authors describe a novel lentiviral packaging system in which not only …
Biology Articles » Virology » Design of a trans protease lentiviral packaging system that produces high titer virus » Results
- Design of a trans protease lentiviral packaging system that produces high titer virus
Delivering the Pol proteins in trans to the viral particles
During the viral life cycle, the Gag (Pr55Gag) and Gag-Pol (Pr160Gag-Pol) precursor polyproteins are targeted to the cell membrane for assembly via the membrane-binding domain (M), which consists of a N-terminal myristylic acid group and a highly basic stretch of amino-acids at the N terminus of MAp17 protein [30-33]. The first step in designing a split Gag-Pol packaging system is to consider how to deliver the Pol proteins, which are normally incorporated via the Gag-Pol precursor polyprotein, to the viral assembly site. Since Vpr can be efficiently incorporated into viral particles (approximately 200 molecules per virion) by an independent mechanism, that is, through an interaction with the C-terminal of P6 on the Gag precursor polyprotein [34-36], we chose to use Vpr to supply the Pol proteins (PR and RT/IN) independently. A truncated form of Vpr (1–88) was selected since it has the ability to be packaged in HIV particles as efficiently as wild type Vpr but is strongly defective in its ability to induce a G2 cell cycle arrest . A representation of Vpr tethering of the Pol components supplied in trans to the viral assembly site is shown in (Fig. 1B), while the packaging plasmids for each of the 3 lentiviral systems presented here are shown in (Fig. 2).
Figure 1 Schematic of the components involved in the 5, 6, and 7 plasmid systems. (A) Diagram of the 4 plasmids used in common for all three packaging systems for the production of virus, followed by a brief description of the packaging plasmids used for each of the corresponding systems (more detail is shown in Fig. 2). (B) Schematic depicting the assembly site of the viral proteins as it takes place in the 5 plasmid system, here the Gag and Gag-Pol precursor proteins are targeted to the cell membrane through the membrane-binding domain located at the N-terminus of MAp17, and the assembly sites of the 6 and 7 plasmid systems where the Gag proteins are targeted to cell membrane by MAp17, and the Pol proteins (PR and RT/IN) are targeted through tethering of Vpr to P6.
Figure 2 Schematic showing the packaging plasmids used in the 5, 6, and 7 plasmid systems. Gag proteins are represented in blue, the Pol proteins in green, Vpr in orange, and Vif in pink. The packaging plasmid in the 5 plasmid system is located at the top of the diagram, only one plasmid is used to express both Gag and Gag-Pol (after frameshifting). The packaging plasmids in the 6 plasmid system are located in the middle of the diagram, two plasmids are used, one that expresses Gag and Gag-PR (after frameshifting) and the other expressing Vpr-RT/IN-Vif (reverse transcriptase, integrase, and Vif). The packaging plasmids in the 7 plasmid system are located at the bottom of the diagram, three plasmids are used, the first expressing Gag alone (there is no frame shift), and the second and third plasmids expressing the Pol components, Vpr-PR (protease alone) and Vpr-RT/IN-Vif (reverse transcriptase, integrase, and Vif), respectively.
Structure of the three lentiviral packaging systems
The data presented here compares 3 different lentiviral packaging systems. The first, referred to as the "5 plasmid system", is a conventional lentiviral packaging system where Gag-Pol is supplied from a single expression plasmid. In addition to the packaging plasmid, which contains both Gag-Pol and Vif (Vpr, Vpu, Tat, Rev, ENV, and Nef were all deleted), four other expression plasmids are used to generate virus: the first contains the lentiviral vector that encodes GFP, the second expresses Tat, the third Rev, and the fourth VSV-G. The second system, referred to as the "6 plasmid system", is a split-packaging system in which the Gag-Pol functions are expressed by two separate plasmids, one for Gag-PR and the other for RT-IN. The Gag-PR expression plasmid was derived from the aforementioned Gag-Pol plasmid in which all the RT, IN, and Vif sequences were deleted. The second packaging plasmid consists of Vpr fused to RT/IN-Vif, a splice donor site to allow for the proper splicing and expression of Vif, and the natural PR cleavage site for RT (33 bases before the start of RT) to allow for proper PR processing of the RT and IN proteins. The third system, referred to as the "7 plasmid system", is a "super-split" packaging system in which the functional components of the Gag-Pol are expressed from three separate plasmids. The first plasmid contains only the Gag gene from which the frameshift has been mutated and all the regions that encode the Pol proteins deleted. The second plasmid contains PR fused to Vpr along with the natural PR cleavage site (15 bases before the start of PR). The third plasmid is the same Vpr-RT/IN-Vif fusion plasmid used for the 6 plasmid system. Diagrams of the plasmids used for all three packaging systems are shown in (Figs. 1 and 2).
Titer analysis of the 5, 6, and 7 plasmid systems
Optimizing parameters, such as molar ratios of one plasmid to another, as well as comparing one system to another, were performed by means of a wt-LTR lentiviral vector that expresses GFP driven by an EF1α promoter. Since the 6 and 7 plasmid systems described here are not conventional, we suspected that p24 and RT assays may not accurately reflect viral titers. The p24 assay gives information about the amount of CAp24 present but does not discriminate infectious from non-infectious particles. In the same respect, the RT assay gives information on RT activity, but it may be difficult to interpret as the 6 and 7 plasmid systems supply RT in trans. We thus chose instead to measure functional infectious viral titers by scoring stable GFP expression in target cells upon chromosomal integration of the provirus. These titers were determined by transfecting 293T cells with 5, 6, or 7 plasmids, collecting the supernatants 48 h later, transducing NIH 3T3 and Jurkat cells with varying amounts of these viral supernatants, and then monitoring the transduced NIH 3T3 and Jurkat cells for the production of GFP by FACS.
Results from the split-packaging 6 plasmid system
The initial question in constructing the 6 plasmid system was how to best separate the Gag-Pol polyprotein precursor without affecting the processing of the viral particles. We decided that the safest location to separate the Gag-Pol was likely to be between PR and RT. There were two reasons for choosing this location. The first was to preserve the frameshift in order to minimize disturbing PR expression by maintaining the 20:1 ratio with Gag. The second was to avoid the 208 nucleotide overlap that occurs between the end of Gag and the start of Pol. To determine if the viral particles produced by this system would be infectious, 293T cells were transfected with either the 5 plasmid or 6 plasmid system and the resulting supernatants were used to transduce NIH 3T3 cells. Titers were then determined by FACS analysis for the expression of GFP. As shown in (Fig. 3A), titers obtained with the 6 plasmid system averaged 2.4 × 105 TU/ml whereas the titers obtained with the 5 plasmid system averaged 2.2 × 106 TU/ml. While these results indicate that the 6 plasmid produces infectious particles at respectable titers, the titers generated were consistently 9 times lower than those of the conventional 5 plasmid system. We hypothesized that the lower titers generated by the 6 plasmid system may be caused by less efficient processing of the precursor polyproteins as a result of splitting RT/IN from Gag-Pol. In order to determine if there was defective processing of viral polyproteins by the 6 plasmid system, we pelleted viral particles from culture supernatants and analyzed virion-associated protein products by immunoprecipitation using serum from an HIV positive patient. Results in Fig. 4 show that RT and IN are efficiently packaged into virions for the 6 plasmid system, with the levels of RT and IN to Gag (Pr55Gag and CAp24) comparable to those found in the conventional 5 plasmid system. This indicated that the fusion of Vpr with RT/IN was successful in delivering RT and IN to the virions. Next, we looked at the processing of the precursor polyproteins. We found that there was efficient processing of the virions produced by the 5 plasmid system with little accumulation of the Gag (Pr55Gag) or Gag-Pol (Pr160Gag-Pol) whereas virions produced by the 6 plasmid system showed a processing defect indicated by an accumulation of both Pr55Gag and Vpr-RT/IN (Fig. 4). Quantitative analysis by laser densitometry scanning of the CAp24 and Pr55Gag bands showed that the ratio of CAp24 (from processed Gag) to Pr55Gag (unprocessed precursor Gag) was 3-fold lower in the 6 plasmid system than in the 5 plasmid system (CAp24/Pr55Gag; 5 plasmid system 6.1 and 4.3, 6 plasmid system 1.6 and 1.8, without and with Vif respectively). Taken together, these results indicate that activation and release of PR were inefficient, and that the titers of the 6 plasmid system could possibly be rescued by increasing expression of PR.
Figure 3 The 7 plasmid system: functional titers and modifications. (A) Functional titers were obtained using a wt-LTR lentiviral vector containing green fluorescent protein (GFP) driven by an EF1α promoter. NIH 3T3 cells were infected with serial dilutions of viral supernatants produced by the 5, 6, or 7 plasmid systems as mentioned in the Methods. The number of transducing units (TU) was determined by multiplying the number of cells plated by the percentage of GFP positive cells (determined by FACS) by the dilution factor. The mean titer for the 5 plasmid system, shown in blue, was 2.2 × 106 TU/ml, for the 6 plasmid system, shown in green, 2.4 × 105 TU/ml, for the 7 plasmid system with the optimized Gag, shown in light pink, 4.4 × 105 TU/ml, and for the 7 plasmid system, shown in dark pink, 7.4 × 105 TU/ml. Error bars represent SEM, 5 independent experiments are represented (N = 5), * p = 0.03 (6P versus 7P-Opt), ** p = 0.003 (7P-Opt versus 7P), *** p < or = 0.0002 (6P versus 7P, 5P versus 6P, and 5P versus 7P-Opt, 5P versus 7P) as determined by unpaired t-test using Prism 4 software. (B) Schematic showing the safety modifications incorporated into the 7 plasmid system. Gag proteins are represented in blue and Pol proteins in green. (1) The Gag to Gag-Pol frameshift was eliminated (AAT TT TTA GGG became AAC TTC TTA GGG). (2) PR was expressed independently of both Gag and Pol. In addition, the active site of PR was changed from DTG to DSG to create the T26S mutant PR. (3) The sequences that overlapped between the packaging signal (Ψ) and Gag, and between Gag and Pol (at P6) were greatly reduced.
Figure 4 Protein analysis of viral particles generated by the 5, 6, and 7 plasmid systems. 293T cells were transfected with the 5 plasmid system (lanes 1 and 2), the 6 plasmid system (lanes 3 and 4), or the non-optimized 7 plasmid system (lanes 5 and 6), with (lanes 2, 4, 6) or without (lanes 1, 3, 5) Vif. Transfected cells were labeled with [35S] methionine for 12 h, 48 h post transfection. Radiolabeled viral particles were pelleted, lysed, immunoprecipitated with anti-HIV serum and analyzed on 12.5% SDS-PAGE. The position of viral proteins are indicated, the boxed region shows the location of Vpr-RT/IN in which less accumulation of unprocessed Vpr-RT/IN can be seen for the 7 plasmid system as compared to the 6 plasmid system. Quantitative analysis of the CAp24 and Pr55Gag bands showed a 3 fold decrease in ratio of CAp24 to Pr55Gag for the 6 plasmid system and a 2 fold decrease for the 7 plasmid system (Cap24/Pr55Gag; 5 plasmid system 6.1 and 4.3, 6 plasmid system 1.6 and 1.8, 7 plasmid system 2.6 and 2.1, without and with Vif respectively). Lanes are not loaded equally. Mock, uninfected.
Titer rescue of the 6 plasmid system by supplying PR in trans
To correct the processing problem detected with the 6 plasmid system, we decided to express PR separately from Gag, resulting in the development of a "super-split" 7 plasmid system. Before constructing this new system, there were three areas of concern that needed to be addressed: (i) What to do with the frameshift, (ii) How to deliver PR in trans without cytotoxicity or loss of infectivity, and (iii) How to minimize the sequence overlap between the packaging signal and Gag, and between Gag and Pol, see (Fig. 3B).
In confronting the first concern, we decided to remove the frameshift in order to completely separate Gag from Pol. This was performed using PCR to generate a fragment, between the Nsi I site (found in the CAp24) and the Bgl II site (just after NCp7), which encompasses the area of frameshift at the end of NCp7. This frameshift sequence was changed from AAT TTT TTA GGG to AAC TTC TTA GGG. A second PCR was performed from the Bgl II site (just after NCp7) to the stop codon of P6 in order to eliminate PR. The result was a Gag expression plasmid in which both the frameshift and PR had been eliminated.
The next step was to determine how to express PR optimally. This was problematic in that PR is central to the processing of the precursor polyproteins and as a consequence to the maturation of the viral particles . The main concern was that too much PR may be expressed resulting in premature processing and cytoxicity [9,14,17]. To address this concern, we expressed a less active PR mutant as an alternative to the wt PR. In searching the literature, we chose a PR mutant with an altered active site in which Asp-Thr-Gly was changed to Asp-Ser-Gly (T26S) [9,38,39]. This mutant was shown to have a slightly reduced protease activity (4–10 fold), with very little effect on viral assembly or maturation, and a markedly reduced cytotoxicity that may result from a shift in the pH needed for its activation [38,39]. The T26S mutation was included in the construction of the PR expression vector in which PR was fused to Vpr, leaving only 15 bases before PR for protease processing. To test whether there was an advantage in using the mutant form of PR, pilot studies were preformed to optimize viral titers by varying the concentrations of the Gag, PR, and RT/IN expression plasmids in order to compensate for molar differences of the plasmids used, as well as for differences in the activity of wt versus mutant protease. One of these pilot studies is shown in (Fig. 5). In this study 293T cells were transfected with the 7 plasmid system, in which the lentiviral GFP vector, Rev, Tat, VSV-G, Gag, and Vpr-RT/IN DNA amounts remained constant, while the concentrations of either the wt protease (Vpr-wt PR, shown in blue) or mutant protease (Vpr-T26S PR, shown in red) expression plasmids varied. As can be seen in (Fig. 5), the titers obtained when mutant or wt PR was delivered independently of Gag ranged from 0.4 × 105 TU/ml to 3.0 × 105 TU/ml, indicating that PR can be supplied in trans to produce infectious particles. In addition, when the T26S mutant PR was used, replacing the wt PR, we were able to obtain equivalent or higher (3 fold) titers than those obtained with the wt PR. We consequently continued to optimize DNA concentrations further improving viral titers produced with the 7 plasmid system, using the T26S mutant PR in place of the wt PR (Fig. 3A).
Figure 5 Comparison of titers produced by PR expression plasmids: wt versus T26S mutant. NIH 3T3 cells were infected with serial dilutions of viral supernatants produced by the 7 plasmid system with either the wt (blue) or T26S mutant (red) PR. Titers were determined by monitoring transduced NIH 3T3 cells for the production of GFP by FACS. In this study, the lentiviral vector (wt-LTR expressing GFP), Rev, Tat, VSV-G, Gag (non-optimized), and Vpr-RT/IN DNA amounts remained constant, while the DNA amounts of Vpr-wt PR (wt protease, shown in blue) and Vpr-T26S PR (mutant protease, shown in red) varied. Experiments were performed using two concentrations for Gag and Vpr-RT/IN: (1×) using 1.3 μg Gag and 2.3 μg Vpr-RT/IN DNA with varying amounts of PR DNA (0.7 μg, 1.0 μg, 1.3 μg, and 1.6 μg), indicated on the graph by circles (●), and (2×) using 2.6 μg Gag and 4.5 μg Vpr-RT/IN DNA along with varying amounts of PR DNA (0.8 μg, 1.4 μg, 2.0 μg 2.6 μg, 3.2 μg), indicated on the graph by triangles (▲). In these initial studies the Vpr-RT/IN plasmid did not contain Vif, the functional titers ranged from 0.4 × 105 TU/ml to 3.0 × 105 TU/ml. N = 1.
The third goal in constructing the 7 plasmid system, as seen in (Fig. 3B), was to minimize the sequence overlap between packaging signal and Gag and between Gag and Pol. The first overlap consisted of 542 bases and was minimized (from 542 to 55 bases) by optimizing the codons at the start of Gag, that is, by using alternate nucleotides for the codons while maintaining the originally encoded Gag amino acid sequence. The second overlap, located at the junction of Gag and Pol, was minimized in the two previous steps by removing the frameshift and separating Gag from PR (208 bases reduced to 54 bases). To determine whether the use of this optimized Gag had an impact on titers generated by the 7 plasmid system, we compared functional titers obtained with the original versus the Gag-optimized expression plasmids. Titer results for the 5 plasmid system, the 6 plasmid system, the 7 plasmid system after optimizing Gag, and the 7 plasmid system where Gag is not optimized, are shown in (Fig. 3A). These titers were obtained after optimizing transfections for variations in total DNA concentration and for molar differences in plasmids used to generate virus for the 6 (Gag-PR and Vpr-RT/IN-Vif) and 7 (Gag, Vpr-T26S PR, and Vpr-RT/IN-Vif) plasmid systems. As can be seen in (Fig. 3A), the 7 plasmid system in which PR is supplied independently of Gag and RT/IN generated titers that were about 2–3 fold higher than those obtained with the 6 plasmid system. Titers achieved with the 6 plasmid system averaged 2.4 × 105 TU/ml and were 9 fold lower than titers obtained with the 5 plasmid system, whereas titers obtained with the 7 plasmid system averaged 4.4 × 105 TU/ml with the optimized Gag, and 7.4 × 105 TU/ml with the non-optimized Gag, that is only about 3 to 5 fold lower than with the 5 plasmid system.
In addition to looking at the functional titers, we analyzed the viral particles generated by the 7 plasmid system to determine whether protein processing had improved by supplying PR independently of Gag. The results shown in (Fig. 4) demonstrate that the Vpr fusions are effective in supplying the Pol components in trans for both mutant PR and RT/IN. The virions produced by the 7 plasmid system, in which PR is delivered independently, showed more processed proteins (CAp24, RT, and IN) with less accumulation of both Pr55Gag and Vpr-RT/IN. Quantitative analysis of the CAp24 and Pr55Gag bands revealed that the ratio of CAp24 (CA from processed Gag) to Pr55Gag (unprocessed Gag precursor) had improved compared to the 6 plasmid system and was now just 2 fold lower than with the 5 plasmid system (Cap24/Pr55Gag; 5 plasmid system 6.1 and 4.3, 6 plasmid system 1.6 and 1.8, 7 plasmid system 2.6 and 2.1, without and with Vif respectively). In addition, because we generally saw a slight increase in titers in the presence of Vif (data not shown) we also looked at the processing in relation to the presence of Vif for all three systems. We were unable to establish conclusively that a change had occurred in the processing of the Gag precursor in the presence of Vif, although we detected an improvement in the processing of Vpr-RT/IN with the 6 plasmid system, as can be seen in Figure 4 by the concurrent reduction in Vpr-RT/IN and increase in RT (lanes 3 and 4).
Self-inactivating (SIN) vector improves viral titers
In addition to modifying the packaging system, we also constructed a SIN lentiviral vector to improve further the safety of the system by decreasing the risk of provirus mobilization and RCL formation. This SIN vector was constructed by modifying the U3 and U5 regions of the 3' LTR, as follows: a 400 bp deletion was created in the U3 region between the EcoRV to the Pvu II restriction sites to remove viral enhancer and promoter, and the U5 region was entirely eliminated and replaced by an "ideal" termination/polyadenylation sequence (ATG TGT GTG TTG GTT TTT TGT GT). In addition, two stop codons were also introduced within the region where the packaging signal and Gag overlap, so that Gag could not be reconstituted if a recombination occurred and to prevent the translation of a residual Gag peptide. The remaining portions of this vector are identical to those of the wt-LTR lentiviral vector, that is, they both contain an unmodified 5' LTR (so that the lentiviral vector remains Tat dependent), the central polypurine tract, RRE, and an Ef1α promoter driving GFP expression (Fig. 6A). In conjunction with the SIN vector we chose to continue to supply Tat in trans due to safety concerns, that is to say, since the 5' LTR in our vectors do not contain a strong promoter (such as CMV or RSV) and still require Tat to properly activate their HIV-1 promoter, than the Tat transactivation of the promoter acts as a safeguard preventing the production of full length packagable transcripts by the integrated vector. To determine if this SIN vector would significantly affect titers, 293T cells were transfected with plasmids composing each of the three packaging systems in conjunction with the SIN vector (Fig. 6B), the resulting supernatants were then used to transduce NIH 3T3 cells. Titers were determined by FACS analysis for the expression of GFP. For all three systems, titers improved by 1.4 fold when the SIN vector was used. The 5 plasmid system increased from 2.2 × 106 TU/ml to 3.1 × 106 TU/ml, the 6 plasmid system from 2.4 × 105 TU/ml to 3.4 × 105 TU/ml, the 7 plasmid system with the optimized Gag from 4.4 × 105 TU/ml to 6.0 × 105 TU/ml, and the 7 plasmid system with the non-optimized Gag from 7.4 × 105 TU/ml to 1.0 × 106 TU/ml. The increase in viral titers when the SIN vector was used was such that the 7 plasmid system provided functional titers (1.0 × 106 TU/ml) that were just 2 fold lower than those obtained when the wt lentiviral vector was used with the conventional 5 plasmid system (2.2 × 106 TU/ml).
Figure 6 Titer results for the 5, 6, and 7 plasmid systems with the SIN lentiviral vector. (A) Diagram showing the structure of the SIN lentiviral vector which contains the following safety features: a 400 bp deletion in the U3 region of the 3' LTR, a complete deletion of the 3' LTR U5 region replaced by an ideal termination/polyadenylation sequence, and two stops placed within the packaging signal (Ψ) to prevent the production of unwanted transcripts. This vector also contains an unmodified 5' LTR, the central polypurine tract, RRE, and GFP driven by an EF1α promoter. (B) NIH3T3 and Jurkat cells were infected with serial dilutions of viral supernatants produced using a SIN lentiviral vector along with the 5, 6, or 7 plasmid packaging systems. Titers were determined by monitoring transduced cells for the production of GFP by FACS. For NIH3T3 cells the mean titer with the 5 plasmid system, shown in blue, was 3.1 × 106 TU/ml, the 6 plasmid system, shown in green, 3.4 × 105 TU/ml, and the 7 plasmid system with and without the optimized Gag, shown in light and dark pink, 6.0 × 105 TU/ml, and 1.0 × 106 TU/ml, respectively. ** p = 0.009 (6P versus 7P-Opt), *** p < or = 0.0002 (6P versus 7P, 7P-Opt versus 7P, 5P versus 6P, and 5P versus 7P-Opt, 5P versus 7P). For Jurkat cells the mean titer for the 5 plasmid system, shown in blue, was 2.5 × 107 TU/ml, the 6 plasmid system, shown in green, 2.7 × 106 TU/ml, the 7 plasmid system with and without the optimized Gag, shown in light and dark pink, 5.5 × 106 and 6.9 × 106 TU/ml, respectively. * p = 0.01 (6P versus 7P-Opt), *** p < 0.0001 (6P versus 7P, 5P versus 6P, and 5P versus 7P-Opt, 5P versus 7P). Error bars represent SEM, data represents 6 independent experiments (N = 6), statistical analysis was determined by unpaired t-test using Prism 4 software.
To demonstrate that the 7 plasmid system is capable of efficiently transducing other cell types, such as human T cells, we also transduced Jurkat cells using the GFP SIN vector along with the 5, 6, and 7 plasmid systems. As shown in (Fig. 6B), titers obtained with the 6 plasmid system averaged 2.7 × 106 TU/ml, once again 9 fold lower than titers obtained with the 5 plasmid system, titers obtained for the 7 plasmid system averaged 5.5 × 106 TU/ml with the optimized Gag and 6.9 × 106 TU/ml with the non-optimized Gag, these titers were 2–3 times higher than those obtained with 6 plasmid system and just 4 times lower than those obtained using the 5 plasmid system.
rating: 5.00 from 1 votes | updated on: 14 May 2008 | views: 8410 |