|Home | About | Journals | Submit | Contact Us | Français|
Hammerhead ribozymes have been shown to silence human immunodeficiency virus-1 (HIV-1) gene expression by site-specific cleavage of viral mRNA. The two major factors that determine whether ribozymes will be effective for post-transcriptional gene silencing are colocalization of the ribozyme and the target RNAs, and the choice of an appropriate target site on the mRNA. An effective screening strategy for potential targets on the viral genome is the use of ribozyme libraries in cell culture. Capitalizing on previous findings that HIV-1 and ribozymes can be colocalized in the nucleolus, we created a novel hammerhead ribozyme library by inserting hammerhead ribozymes with fully randomized stems 1 and 2 into the body of the U16 small nucleolar RNA (snoRNA). Following three rounds of cotransfection with an HIV-1 proviral DNA harboring the herpes simplex virus thymidine kinase (HSV-TK) gene, we selected for gancyclovir-resistant cells and identified a ribozyme sequence that could potentially target both the U5 and gag genes of HIV-1 regions on the HIV-1 genome through partial homologies with these targets. When the ribozymes were converted to full complementarity with the targets, they provided potent inhibition of HIV-1 replication in cell culture. These results provide a novel approach for identifying ribozyme targets in HIV-1.
The nucleolus is a distinctive subnuclear compartment, thought of primarily as the site of ribosomal RNA biogenesis and ribosome assembly. However, a number of studies have now shown diverse roles for the nucleolus in mRNA production and even viral replication.1,2 For example, several mRNAs have been shown to pass through the nucleolus (e.g., c-myc, N myc, and myoD3) and at least one microRNA localizes to the nucleolus.4 Transcription and replication of Borna disease virus (a negative strand RNA virus) occurs in the nucleolus5 while human T-lymphotropic virus env RNAs6 and the human immunodeficiency virus-1 (HIV-1) immediate-early proteins Tat and Rev have been shown to localize to the nucleolus.7–9 The fact that these proteins interact with HIV-1 RNA, and the finding that a nucleolar-localized hammerhead ribozyme provided strong inhibition of HIV-1 (ref. 10) constitute strong evidence to suggest that HIV-1 RNA traffics through the nucleolus as part of its replication cycle. In support of this notion, Cantó-Nogués et al. have demonstrated, using electron microscopy, that HIV-1 transcripts can localize in the nucleolus and in peripheral dense chromatin in peripheral blood mononuclear cells.11 We have previously shown that efficient inhibition of HIV-1 can be achieved with nucleolar-localizing ribozymes and decoys for Tat and Rev.10,12,13
Ribozymes are small RNA molecules with distinct catalytic motifs that can be adapted for trans-target cleavage by modifying the hybridizing arms to bind to target RNAs through Watson–Crick base pairing. In the presence of magnesium, their intrinsic enzymatic activities cleave the phosphodiester backbone of targeted mRNAs.14 Specifically, any RNA can be cleaved as long as the ribozyme can pair with the target RNA and the target contains an NUC triplet where N = A, G, C, or U. This property, combined with the ability of ribozymes to undergo multiple turnover reactions, makes them attractive agents for modulating gene expression.14 The two most common forms of ribozymes employed for mRNA cleavage are the hairpin and the hammerhead motifs. Recently, a number of strategies involving libraries of RNA-based antivirals have been employed in forward genetic screens for identifying genes on the basis of their functions. For ribozymes, merely randomizing the hybridizing arms can generate a library of 4n distinct ribozyme sequences, depending on the length of the hybridizing arms (n). The application of ribozymes in forward genetic screens involves delivery of these libraries to target cells, either through transient transfection or transduction, followed by screening or selection of a particular phenotype. Using this approach, ribozymes that target several important genes involved in viral and oncogenic pathways have been identified.15–21 Importantly, the library approach selects only for target inhibition, and does not require 100% matched base pairing to the target.21 In general, therefore, this approach allows only the identification of ribozymes that achieve the desired phenotype, and not necessarily of ribozymes with the greatest cleavage activities.
We have taken advantage of a hammerhead ribozyme library approach to identify targets in the HIV-1 genome that are highly susceptible to ribozyme-mediated downregulation. We have also taken into consideration our previous finding that a nucleolar-localized ribozyme can provide potent inhibition of HIV-1 replication in human T lymphocytes.10 The nucleolar localization is accomplished by insertion of the hammerhead ribozyme into the structure of the small nucleolar RNA (snoRNA) U16 (Figure 1). The approach we utilized is rapid and can be carried out with transient transfection of the ribozyme expression units, bypassing the steps of cloning and packaging the libraries in viral vectors. Moreover, stably transduced libraries often suffer from a potent disadvantage in that the site of vector integration can determine the level of expression of the ribozyme or small interfering RNA, oftentimes leading to heterogeneous levels of ribozyme expression in cell populations. Our approach also takes advantage of a potent survival assay using the herpes simplex virus thymidine kinase (HSV-TK) gene embedded in infectious HIV-1 proviral DNA. Ribozymes that inhibit HIV-1 replication or gene expression will render cells resistant to gancyclovir, while unprotected cells are killed following phosphorylation of gancyclovir, converting it from a prodrug to a toxic nucleoside.
The use of a ribozyme library embedded in the U16 snoRNA stem provides two major advantages. First, it provides a mechanism for colocalizing the ribozyme with full-length and singly spliced viral RNA targets in the nucleolus.10 Second, the snoRNA portion provides a PCR primer-binding site for retrieving the active ribozyme species from treated cells. We demonstrate here the selection of three ribozyme sequences from a pool of 4 (ref. 14) U16 ribozyme chimeras that resulted in gancyclovir resistance of treated cells. Of these, one sequence was isolated from three different colonies, suggesting a selective enrichment of this ribozyme sequence. Sequence alignment with HIV-1 showed that the ribozyme could potentially target two sites on the viral genome, at positions 550 and 4870 of HIV-1 NL4-3. Both these sites were found to be very efficient for ribozyme-mediated inhibition, when tested with perfectly matched ribozyme sequences. Inhibition of HIV-1 by the selected ribozyme and its perfectly matched variants was further confirmed in CEM cells stably transduced with these ribozymes.
The ribozyme library was generated as described in the Materials and Methods section. Four pools of ribozymes were generated for the screens. We used both positive and negative selection approaches in our quest to isolate ribozymes inhibiting HIV-1 gene expression. The selection approach used in our study involved cotransfecting HEK293 cells with infectious proviral DNA pNL-TK (which has the HSV-TK inserted in the nef region of the HIV-1 genome) and the ribozyme library. Twenty-four hours after transfection, the cells were treated with gancyclovir and incubated for another 48 hours. The cells were then washed two times with phosphate-buffered saline to remove dead cells in suspension and the total RNA was extracted using STAT-60. The rationale for this approach is that, in the absence of any protective ribozymes, HSV-TK would be expressed, resulting in cell killing and loss of ribozyme library components in those cells, whereas survivors would contain the effective ribozymes. The ribozymes were rescued from the total RNA with reverse transcriptase PCR using primers complementary to the U16 stem. The PCR products containing the amplified ribozymes were then recloned under the U6 promoter in the pTzU6 + 1 vector. This cycle of cotransfection and selection was repeated three times. After each cycle, colonies were randomly selected and sequenced to check for enrichment of particular sequences.
After three rounds of selection, 50 colonies from pool I were randomly selected and sequenced. As can be seen from Figure 2b, two ribozymes appeared to have been selectively enriched during the pNL-TK-mediated screening process. Ten ribozymes including the two selectively enriched ribozymes were directly tested in a pilot cotransfection assay with wild-type proviral DNA pNL4-3. The U16 snoRNA under the U6 promoter was used as a negative control. The results of this pilot experiment are depicted in Figure 2a. One of the sequences, isolated from four different colonies, did not show any inhibition in p24 output, whereas RzC36 showed good inhibition, as did RzC10, when compared with the U16 control-transfected cells. An unexpected observation was that the ribozyme sequence C9 (RzC9) gave a higher p24 readout as compared to the U16-transfected control.
The three selected ribozymes that showed anti-HIV-1 activity were retested in a cotransfection assay with HIV-1 pNL4-3 using a higher ratio of ribozymes to HIV-1 vector. As can be seen from Figure 2c, RzC36 yielded almost 70% inhibition of HIV-1 gene expression while RzC10 was moderately effective. RzC9 continued to show higher p24 levels as compared to the U16 controls. In order to better understand why RzC9 came through the selection process, we cotransfected HEK293 cells with the selected ribozymes and HIV-1 pNL-TK. Forty-eight hours after treatment of the cells with gancyclovir, total cells were collected, stained with trypan blue, and the results expressed as a ratio of live/dead cells. U16-transfected cells were used as a control. As can be seen from Figure 2d, RzC10-expressing cells had a moderate survival advantage whereas both RzC9 and RzC36 conferred strong survival advantages. At this point we have not been able to explain this paradox; however, it is quite possible that RzC9 could actually be targeting a cellular factor involved in either viral RNA export or sorting of viral RNA between translation and packaging.
Using cDNA microarray experiments we sought to analyze gene expression patterns in HEK293 cells as a function of HIV-1 pNL4-3 transfection and gene expression, and also investigated whether RzC36 altered this profile. We focused on genes that showed a plus or minus log2 change of ~1.0. As can be seen in Figure 3, distinct changes in the expression patterns of several genes are observed during viral replication. These changes could be a direct or indirect result of viral gene expression in cells. We next compared this pattern with that observed in cells cotransfected with HIV-1 pNL4-3 and RzC36. If the ribozyme’s target site was on the viral RNA, the patterns of altered gene expression might be expected to resemble those found in non-HIV-1 pNL4-3-transfected cells. Interestingly, these analyses demonstrated that all the virus-mediated changes observed with HIV-1 pNL4-3-transfected cells were reversed in the presence of RzC36. Moreover, the extent of change (multiplying factor) in gene expression produced by the RzC36-transfected cells was equal and opposite to that produced by HIV-1 pNL4-3 transfection alone. Apart from the reversal of virus-mediated changes, no other changes in the pattern of gene expression were observed in cells cotransfected with HIV-1 pNL4-3 and RzC36.
Having observed that RzC36 reversed the pattern of HIV-1-mediated changes in host gene expression, we next sought to determine the potential ribozyme cleavage sites in the HIV-1 genome. Aligning the substrate recognition sequence of RzC36 with the HIV-1 NL4-3 sequence did not turn up any sites with perfect Watson–Crick base pairing. This is not an unusual finding, in that Waninger et al. found similar mismatched pairings with their targets, using a hairpin ribozyme library.21 These results can be attributed to the fact that the assays used require a phenotypic change and do not necessarily require complete inhibition of the target. We next devised an approach for target site determination on HIV-1 RNA that would allow us to consider mismatches, bulges of 1 or 2 nucleotides (nt), and wobble base pairing. We searched for targets containing an NU(A, C, or U) triplet flanked by at least 5 nt complementary to helices 1 and 3 of RzC36 with mismatches in between. As can be seen in Figure 4, we could identify two potential targets for RzC36 on the NL4-3 sequence. One site was within the HIV-1-LTR in the R/U5 region at position 105. This site is unique, in that it is very close to a previously reported, highly conserved site in the HIV-1 genome that was also used by us in our earlier studies with a nucleolar-localized hammerhead ribozyme.10 A number of studies including our own have shown efficient inhibition of HIV-1 gene expression by targeting this region, thereby suggesting that the region is highly accessible for ribozyme or antisense interaction.
The other site lies in the gag/pol open reading frame in the pol region at position 4870. The U5 site would be present in spliced, un-spliced, and full-length forms of the transcript, whereas the pol site would be present only in the full-length form. It is quite possible that the full-length or partially spliced forms of the HIV-1 transcript, which are known to involve rev-mediated export and thus, possibly, traffic through the nucleolus, could become targets for RzC36 during this transit.
In order to determine which of the two sites was more highly accessible and therefore the true target for RzC36, we designed variants of RzC36 that had perfect complementarity to each of the two sites. Catalytically inactive variants were generated by mutating the C to a G in helix II (Figure 4a and b). Both these ribozymes and their corresponding catalytic mutants were tested in transient cotransfection assays with HIV-1 pNL4-3 into HEK293 cells. U16 cotransfected with HIV-1 pNL4-3 was used as a control. As can be seen from Figure 5, both ribozyme variants produced pronounced inhibition of HIV-1 gene expression. Approximately 95% inhibition of HIV-1 p24 was observed with both the U5 and the Pol variants as compared to the U16-transfected controls. Although we did observe some inhibition with the catalytically inactive mutants, the inhibition with active ribozymes was much more pronounced, thereby suggesting that the major reduction in viral output was caused by ribozyme-mediated catalytic activity. The partial inhibition by the catalytically inactive variants could have been on account of an antisense effect, given that these were also tightly colocalized with the HIV-1 targets. In order to check whether the decrease in p24 output was indeed caused by a ribozyme-mediated cleavage rather than by a nonspecific block in viral translation, HEK293 cells were cotransfected with HIV-1 pNL4-3 and either the Pol or U5 variants or the parent RzC36. Culture supernatants were collected and analyzed using a B-DNA assay to check for the presence of viral genomic RNA in the supernatant. As can be seen, nearly complete inhibition of pNL4-3 RNA levels was observed in cells transfected with RzC36 or its perfectly matched variants.
In order to check the efficiency of these ribozymes in a potential gene therapy setting, the U16 ribozyme cassettes (RzC36, U5 variant, and the Pol variant), were cloned in the lentiviral vector pHIV-7-GFP and packaged as described earlier.22 CEM cells were transduced with either vector alone, the perfectly matched ribozyme variants, or the parent RzC36, and challenged with ~150 ng p24 equivalent of HIV-1 NL4-3 virus. At designated time-points the culture supernatant was collected and analyzed for p24 levels as an index of viral replication. As can be seen from Figure 4c, an average ~65–70% inhibition of HIV-1 gene expression was observed. All three ribozymes showed a comparable level of inhibition, thereby suggesting that both the U5 and the Pol target sites selected by our approach were efficient targets. It is therefore quite possible that RzC36 was chosen for its ability to actually target both sites on the viral RNA. This is not surprising, given that the potent TK-gancyclovir suicide axis would require considerable knockdown of the TK gene to allow the cells to survive in the presence of gancyclovir. This would ensure the selection and enrichment of only the most potent ribozyme molecules from the library.
Anti-HIV-1 therapy using ribozymes has focused mainly on either targeting the viral RNA or mRNA of cellular genes that may have a role to play in the viral life cycle. In order to achieve a near complete knockdown of viral gene expression, one needs to take into account several factors that may influence the cleavage efficiency of these antiviral agents. The most important considerations are target site accessibility and colocalization of the ribozyme and target. Although a number of guiding parameters are available to aid the selection of efficient target sites, most of these parameters are established in vitro and then tested in cells. Most approaches to ribozyme design begin with choosing an appropriate NU(A, C, or U) target site in the message of interest. These approaches do not take into account factors that can alter the efficiency of the ribozymes in an intracellular environment, such as local structural features in the target mRNA and RNA–protein interactions. By using an in vivo or intracellular library approach these concerns are minimized. Moreover, by modifying the selection process, one can enrich for ribozymes that achieve the desired phenotype without necessarily requiring complete target knockdown.
In this report, we have exploited the potent killing effect of the TK-gancyclovir axis in an HIV-1 proviral DNA setting. The hammerhead ribozymes were incorporated in the U16 snoRNA stem to localize them to the nucleolus, an intracellular compartment that we found to be effective for ribozyme targeting of singly spliced and unspliced HIV-1.10 We reasoned that this approach would result in a large number of ineffective ribozymes being eliminated in the first round of selection. Indeed, we saw almost ~80% of the cells detached from the surface during the first round of selection. On the other hand, cells harboring a protective ribozyme would inhibit expression of TK and survive. The ribozyme molecules from these cells could then be rescued by reverse transcriptase-PCR of total RNA using primers specific to the U16 stem. After three rounds of selection with this approach, random sampling of the colonies revealed two sequences that had been enriched during the process. We also randomly tested other library ribozymes along with these two for their activity against HIV-1. We were surprised to find one that routinely resulted in a higher p24 output as compared to the U16-transfected control. When we tested this ribozyme in cell survival assays using HIV-1 pNL-TK we found that the ribozyme promoted cell survival at a level comparable to that observed with RzC36. We are currently investigating the mode of inhibition and the potential target site for this ribozyme sequence. Another ribozyme sequence that we had obtained in four different clones did not show any inhibition of HIV-1 p24 levels. However, further analysis showed that the ribozyme had a marked complementarity with the TK open reading frame and could promote cell survival when cotransfected with pNL-TK in the presence of gancyclovir (Supplementary Figure S1). This could explain why the ribozyme had been selectively enriched during screening but was unable to show inhibition of viral gene expression with HIV-1 pNL4-3, because HIV-1 pNL4-3 lacks the target site for RzC43. However, the target site is not a canonical NUH target and is, in fact, a GUG. Earlier studies have presented conflicting observations regarding cleavage of GUG targets. Koizumi et al.23 found that a target site with this sequence could not be cleaved, whereas Sheldon and Symons24 observed considerable cleavage of a GUG substrate in a unimolecular ribozyme substrate reaction. Perriman et al.25 have demonstrated that a GUG triplet could be induced to cleave by modifying the helix II. Indeed, there were minor differences among the catalytic motifs used in these studies. In our work, we have used a minimal helix II reported by Persson et al.26 It is quite possible that the selection process we employed transcends the guidelines of cleavage established in vitro using perfectly matched Watson–Crick base pairing, and takes into account interactions between the bulges and mismatches in helix I and III with the helix II. Using in vitro selection procedures, Kore et al. were able to isolate hammerhead ribozymes with a mutation at position 7 in the catalytic motif,27 which is capable of cleaving GAC triplets, another anomaly to the established NUX rule.
With the goal of determining effective target sites in the HIV-1 genome, we decided to proceed with the investigation of RzC36 which, in addition to surviving gancyclovir, showed potent inhibition of HIV-1 gene expression. RzC36 did not share perfect complementarity with any region of the HIV-1 genome. However, a cDNA array analysis of cells cotransfected with RzC36 and HIV-1 pNL4-3 showed a complete reversal of the cDNA array pattern observed with HIV-1 pNL4-3 cotransfection with the control U16 backbone vector. These results suggested that RzC36 was targeting viral RNA, which could explain the complete reversal of virus-mediated changes observed when this ribozyme was cotransfected with HIV-1 pNL4-3.
In order to identify potential target sites for RzC36, we relaxed the rules of target site determination to a minimum of 6 nt of base pairing in the hybridizing arms of the ribozyme with the target sequence. Using this approach we were able to identify two potential target sites in HIV-1 NL4-3 in the U5 region and in pol. The U5 site is present in all forms of viral transcripts, whereas the site in pol at position 4850 would not be present in the partially or fully spliced forms of the viral RNAs. Both sites turned out to be very good ribozyme sites in HIV-1, resulting in ~75 and 80% inhibition of p24 levels in transient transfection assays with pNL4-3, respectively. Moreover, these ribozymes were protective in HIV-1 challenge assays when they were stably expressed in CEM cells, thereby suggesting that RzC36 could have been selected for its ability to target both sites. This is not surprising, given that a significant level of knockdown of TK expression would be required for the cells to survive the gancyclovir treatment.
In summary, the use of nucleolar-localizing hammerhead ribozymes has proven to provide potent inhibition of HIV-1 replication in cell culture.10 The results presented in this study provide further support for our model that HIV-1 traffics through the nucleolus, and also provides a rationale for the using nucleolar-localizing ribozymes to inhibit HIV-1 infection.
The construction and characteristics of the lentiviral vector pHIV-7 have been described earlier.28 Plasmids HIV-1 pNL4-3 and HIV-1 pNL-TK were obtained from the NIH AIDS Research and Reference reagent program (Rockville, MD).
The U16Rz library was prepared synthetically with overlapping PCR using the primers A, B, C, D, under the conditions described earlier29:
A: ctt gct atg atg tcg taa ttt gcg tct tac tct gtt ctc agc gac agt tga aa
B: ttt cga aaa ctc atc agn nnn nnn ttt tca act gtc gct gag aac
C: ctg atg agt ttt cga aan nnn nnn aaa acc tgc tgt cag taa gct ggt aca gaa g
D: ttt ctt gct cag taa gaa ttt tcg tca acc ttc tgt acc agc tta ctg ac
The U16Rz library was reamplified from the PCR mix using primers E and F which include appended SalI and XbaI restriction endonuclease sites respectively.
E: ccc ccc cgt cga cct tgc tat gat gtc gta att tg
F: ccc ctc tag aaa aaa ttt ctt gct cag taa gaa ttt
The resulting ribozyme library flanked by SalI and XbaI sites was digested and ligated in a 20 μl ligation reaction using T4 DNA ligase with a similarly digested pTz/U6 + 1 expression cassette.30 This would place the U16 ribozyme library under the transcriptional control of the U6 promoter. Six thymidines were added at the 3′-end of the ribozyme coding sequence to create a Pol III transcriptional termination signal.
In order to maximize the number of library members, several transformations were carried out using the above ligation mix, and the colonies were pooled as four mini-libraries. The U16 control plasmid was created by amplifying the endogenous U16 snoRNA using primers E and F, digesting with SalI and XbaI, and ligating this into pTz/U6 + 1.
The perfectly matched variants of RzC36 were cloned similarly. Primers B and C were modified such that the hybridizing arms were complementary to the putative sites in the U5 and Pol region.
For the U5 version, the primer sequences used were oligo B: tttc gaa aac tca tca gaa agt agt ttt caa ctg tcg ctg aga ac, and oligo C: ctg atg agt ttt cga aaa gca ctc aaa acc tgc tgt cag taa gct ggt aca gaa g.
For the pol version, the primer sequences used were oligo B: ttt cga aaa ctc atc aga aag ggc ttt tca act gtc gct gag aac, and oligo C: ctg atg agt ttt cga aaa gct tga aaa acc tgc tgt cag taa gct ggt aca gaa g.
For cloning in the lentiviral vector pHIV-7-GFP, the entire U16 Rz cassette (RzC36 and the U5 or pol variant) was amplified using primer E appended to a NotI site and primer F appended to a XbaI site.
HT1080 and HEK293 cells were maintained in Dulbecco’s modified Eagle’s medium 20% fetal bovine serum. Twenty-four hours before transfection, cells were replated in 6-well plates at 50–70% confluency with fresh media, without antibiotics. The human T-cell line CEM was maintained in Rosewell Park Memorial Institute 1640 medium supplemented with 10% fetal bovine serum. For all transfection experiments, lipofectamine plus reagent (GibcoBRL, Bethesda, MD) was used in accordance with the manufacturer’s instructions.
Thymidine kinase–mediated selection and rescue of effective ribozyme molecules: HEK293 Cells were plated at ~1 × 106 cells per 100-mm dish 1 day before the transfection, and then transiently cotransfected with HIV-1 pNL-TK and the mini-libraries (1:2 wt/wt ratio respectively). The total DNA used for transfection was 450 ng. Twenty-four hours after transfection, gancyclovir was added at a final concentration of 5 nmol/l, and the cells were incubated further. Eight hours after gancyclovir addition, the plates were washed to remove detached cells, and total RNA was extracted from adherent cells using RNA STAT-60 (TEL-TEST “B”). The ribozyme sequences were rescued using primer pairs E and F following electrophoresis in a 4% agarose gel. A doublet was observed, with the lower band being the endogenous U16 and the upper band being the library. The band corresponding to the library was purified, digested with SalI and XbaI, and recloned in similarly digested pTz/U6 + 1. The process was repeated an additional two times for each mini-library. After three rounds of selection, 50 colonies were randomly selected, sequenced, and analyzed for anti-HIV-1 activity in a pilot cotransfection assay with HIV-1 pNL4-3, using pNL4-3:ribozyme in a ratio of 1:2 by weight]. Subsequent co-transfection assays with HIV-1 pNL4-3 involved using pNL 4-3: ribozyme in a ratio of 1:4 by weight with 500 ng of total DNA for each transfection. Positive control involved cotransfecting cells with U16 cloned in pTzU6 + 1 in a 1:4 ratio by weight.
HEK293 cells were cotransfected with HIV-1 pNL4-3 and RzC36 or the U16 control, in duplicate experiments. Seventy-two hours after transfection, total RNA was extracted from cells and used for cDNA array analyses. Experiments were carried out as biological replicates.
Arrays were prepared and processed by the Microarray Core Facility of City of Hope. cDNA clones were amplified from the 40K Research Genetics Human Genome Collection (Invitrogen, Carlsbad, CA) and printed on in-house-made poly-L-lysine slides using a DeRisi Version II Microarrayer. The labeling protocol was modified from the Vanderbilt Microarray Shared Resource Labs (http://array.mc.vanderbilt.edu/cDNA/downloads/Coupling_protocol_0903.pdf). Fifteen microgram of total RNA were combined with 2 μg of anchored oligo (dT)12–28 primer and denatured for 5 minutes at 65 °C, followed by a snap cooling in ice for 2 minutes. Sufficient master mix of the first strand was made for all samples, 18 μl per sample (6 μl 5× first strand buffer, 3 μl of 0.1 mol/l dithiothreitol, 0.5 μl of 40 U/μl Rnasin, 0.6 μl of deoxyribonucleotide triphosphates at the concentrations: 25 mmol/l deoxyadenosine triphosphate, 25 mmol/l deoxycytidine triphosphate, 25 mmol/l deoxyguanosine triphosphate, 10 mmol/l deoxythymidine triphosphate; 4.5 μl of 2 mmol/l aminoallyl deoxyuridine triphosphate, 2 μl of 200 U/μl SuperScriptII and 1.4 μl of nuclease-free water per sample). The samples were incubated at 42 °C for 2 hours, followed by enzyme inactivation at 95 °C for 5 minutes. The RNA was digested with 12.9 μl of 1 mol/l NaOH and heated at 65 °C for 15 minutes. The samples were cooled to room temperature before being neutralized with 12.9 μl of 1 mol/l HCl, followed by 6.2 μl of 1 mol/l Tris–HCl (pH 7.0). The cDNAs were purified using the QIAquick PCR Purification Kit (Qiagen, Valencia, CA) as recommended by the manufacturer, with modifications: wash buffer was replaced with a phosphate wash (0.5 ml of 1 mol/l phosphate buffer*, 15.25 ml nuclease-free water, 84.25 ml 95% ethanol) and the elution buffer was replaced with 4 mmol/l phosphate buffer, (*9.5 ml of 1 mol/l K2HPO4, 0.5 ml of 1 mol/l KH2PO4). The purified cDNA was concentrated using speed vacuum to 5 μl, and fluorescent dye coupling was initiated by adding 3 μl of 25 mg/ml sodium bicarbonate plus the respective Alexa Fluor reactive dye, 555 for Cy3 or 647 for Cy5 (Invitrogen), suspended in 2 μl of dimethyl sulfoxide. Coupling was carried out at room temperature in the dark with mixing every 15 minutes. The reaction was stopped with 10 μl 3 mol/l sodium acetate (pH 5.2). The volumes of the samples were brought up to 100 μl, and the samples were purified using the QIAquick PCR Purification Kit in accordance with the manufacturer’s recommendations, with modifications: 75% ethanol was used as the wash buffer and water as the elution buffer. The prehybridized array (washed for 30 seconds with 0.2% sodium dodecyl sulfate, and incubated at 55 °C for 45 minutes in 1% bovine serum albumin, 1% sodium dodecyl sulfate, 5× sodium chloride and sodium citrate (SSC); followed by five 30-second washes in double-distilled water, and then dipped in isopropanol for 30 seconds, and spun dry) was loaded onto a caliper hybridization cassette (Caliper LifeSciences, Hopkinton, MA). The purified cDNA was paired with its respective sample with a formamide-based hybridization buffer (50% di formamide, 10× SSC, and 0.2% sodium dodecyl sulfate), denatured, and loaded onto the array. The array was hybridized at 42 °C for 16 hours. Posthybridization washings were carried out in an ArrayIt wash station (TeleChem international, Sunnyvale, CA) with agitation: 2× SSC, 0.1% sodium dodecyl sulfate heated to 55 °C for 10 minutes; 1× SSC 10 minutes; and 0.1× SSC 10 minutes, and then spun dry. An Axon 4000B (Molecular Devices, Sunnyvale, CA) scanner was used for scanning the images, and GenePix 5.0 was used for the data acquisition. Genes showing a change with a P value of <0.01 were considered significant.
CEM cells, (2 × 105) were placed in a 15-ml centrifuge tube with 1 ml culture medium in the presence of lentiviral vector at a multiplicity of infection of 10, and 8 μg/ml polybrene. Following centrifugation at 2,000 rpm for 1 hour, the cells were transferred into a 24-well culture plate, and after 24 hours the culture medium was replaced. Seventy-two hours after transduction, the cells were sorted for enhanced green fluorescent protein expression and used for HIV-1 challenge experiments.
Seventy-two hours after being sorted, 1 × 106 CEM T cells transduced with vector alone, and equal numbers of cells transduced with RzC36 and each of its variants, were infected with 150 ng p24 equivalent of HIV-1 NL4-3 virus. After overnight incubation, the cells were washed three times with Hanks’ balanced salts solution and cultured in medium with R10 (Rosewell Park Memorial Institute 1640 + 10% fetal bovine serum). At each of the designated time-points from day 5 to 15, 500 μl of culture supernatant was withdrawn for cell counting. The cells were then pelleted and returned to the infection. The culture supernatants were used for p24 assays and replaced with 500 μl of fresh medium.
Culture supernatants were collected at designated time-points and analyzed for HIV-1 p24 antigen using an enzyme-linked immunosorbent assay (Beckman Coulter, Fullerton, CA). The p24 values were calculated using a Dynatech MR5000 enzyme-linked immunosorbent assay plate reader (Dynatech Lab, Chantilly, VA). The B-DNA assay for detection of viral RNA was performed using 90 μl of the culture supernatant. These assays utilized the QuantiGene branched DNA reagent system (Panomics, Redwood City, CA) according to the manufacturer’s protocol. The HIV complementary probe targets 671 nt within the p24 segment of the HIV gag gene.
Figure S1. RzC43 targets thymidine kinase.
This research was supported by National Institutes of Health (NIH) grant AI29329 to J.J.R. and an AMFAR grant 106690-39 RFNT to H.J.U. We acknowledge the NIH AIDS Reagent and Repository for supplying several of the viral strains and cell lines used in this study.