Real-Time PCR Assay of Individual Human Immunodeficiency Virus Type 1 Variants in Coinfected Human Lymphoid Tissues

Abstract

In the course of infection, human immunodeficiency virus type 1 (HIV-1) mutates, diverging into a “swarm” of viral quasispecies, and the predominance of CCR5- or CXCR4-utilizing quasispecies is strongly associated with the pattern of disease progression. Quantification of CCR5- and CXCR4-utilizing viruses in viral swarms is important in the investigation of the mechanisms of this phenomenon. Here, we report on a new real-time PCR-based methology for the evaluation of replication of individual CCR5- and CXCR4-utilizing variants. The assay is highly reproducible, with a coefficient of variation of <3%, and it accurately estimates the numbers of virus-specific RNA copies even when their difference in the mixture is 2 orders of magnitude. We demonstrate that replications of CCR5- and CXCR4-utilizing variants can be evaluated and distinguished in experimentally coinfected human lymphoid tissue. The assay we developed may facilitate study of the mechanisms of the R5-to-X4 switch in viral swarms in human tissues infected with HIV-1.

In the course of infection, RNA viruses mutate, diverging into a “swarm” of viral quasispecies, whose dominant forms change in the course of disease (10). This divergence is best documented in human immunodeficiency virus (HIV) infection, since the extremely high mutation rate results in rapid swarm development and the dominance of certain quasispecies is strongly associated with acceleration of disease progression (3, 4, 7, 12-14, 24, 25, 29, 30, 32). It is firmly established that HIV type 1 (HIV-1) uses two surface molecules to infect cells, CD4 and one of several chemokine receptors (“coreceptors”). In vivo, HIV-1 utilizes predominantly one of two coreceptors: CXCR4 and CCR5. CCR5-utilizing (R5) HIV-1 variants generally dominate early HIV-1 infection, whereas CXCR4-utilizing variants (X4) evolve later and may become dominant. The switch from R5 to X4 is associated with acceleration of disease progression (reviewed in reference 2). The mechanisms of the R5-to-X4 switch are not known and, for them to be studied experimentally, methods for quantification of R5 and X4 viruses in viral swarms must be developed. Although quantification of R5 and X4 viruses in viral swarms is important, until recently the presence of R5 and X4 HIV-1 variants has been determined by testing whether these viruses infect CXCR4- or CCR5-transfected cell lines or by viral sequencing (21, 26, 28, 30). Although such methods can demonstrate the presence of viral quasispecies in viral swarms, they are at most semiquantitative. Recently, more quantitative methods of analysis of viral swarms, based on heteroduplex mobility, have been used (5, 6). However, additional quantitative assays are necessary to differentiate replication of individual HIV-1 variants in viral swarms.

We report here on a new real-time reverse transcription-PCR (RT-PCR)-based methology to evaluate replication of individual HIV-1 variants, and we demonstrate that replications of R5 and X4 HIV-1 variants can be detected and distinguished in experimentally coinfected human lymphoid tissue.

MATERIALS AND METHODS

Preparation of HIV-1 variant standards.

We subcloned the gp120 gene from each of the SF162, LAV.04, 89-v345.SF, and 89.6 HIV-1 variants into the pCR2.1-TOPO plasmid by using the TOPO TA cloning kit (Invitrogen Corp., Carlsbad, Calif.) according to the manufacturer's instructions. Also, we constructed a plasmid that provides a reproducible source of reference DNA, containing gp120 genes from LAV.04 and SF162. The LAV.04 gp120 was subcloned into pCR2.1-TOPO, which already contained SF162 gp120. Similarly, we constructed a plasmid that contained both 89-v345.SF and 89.6 gp120 genes. By using these dual viral plasmids, we avoided spectrophotometric quantification errors, which are inevitable when two separate plasmids with a single gp120 are used.

Viral RNA extraction.

We extracted viral RNA from 100 μl of culture medium by using a PURESCRIPT RNA isolation kit (Gentra Systems, Minneapolis, Minn.) according to the manufacturer's instructions. RNA was eluted in 25 μl of RNA hydration solution.

RT reaction.

RT was performed in 100-μl reaction mixtures containing 1× RT buffer; 5.5 mM MgCl₂; dATP, dCTP, dGTP, and dTTP (each at 500 μM); 2.5 μM random hexamers; 40 U of RNase inhibitor; 125 U of MultiScribe reverse transcriptase (Applied Biosystems, Foster City, Calif.); and 10 μl of the RNA solution. Each reaction was performed for 10 min at 25°C, 30 min at 48°C, and 5 min at 95°C.

Primer design.

Primers for the separate quantification of each HIV-1 variant were chosen within the V3-V5 region of the gp120 gene. Primer sequences were selected according to Applied Biosystems guidelines. However, some of the sequences we used do not satisfy all of the criteria recommended in these guidelines because the primers should be designed for gp120 sequences that are different for different viruses, and these sequences are small and scattered throughout the gene. The specific primer sets used to detect SF162, LAV.04, 89-v345.SF, and 89.6 were as follows: SF162 forward, 5′-TATGCAACAGGAGACATAATAGGAGATATAAG-3′; SF162 reverse, 5′-TAGTTCCATTAGTGTTATTTGGCCCTATAG-3′; LAV.04 forward, 5′-GTTACAATAGGAAAAATAGGAAATATGAGACA-3′; LAV.04 reverse, 5′-TTTGACCCTTCAGTACTCCAAGTACTATTAA-3′; 89-v345.SF forward, 5′-TATAGGGCCAAATAACACTAATGGAACTA-3′; 89-v345.SF reverse, 5′-CGGTGGTGTTACTGATCTCTTTACCA-3′; 89.6 forward, 5′-GGAGGGACAAATGGCACTGAA-3′; and 89.6 reverse, 5′-AGATCTCAGTCTCAGTCTCAGTACTATTACCT-3′ (Fig. 1).

FIG. 1.

Primer designs for R5 and X4 HIV-1 variant-specific real-time PCR assay. (A) Primer design for sequences for the SF162 and LAV.04 gp120 gene based on the V3-V4 region. (B) Primer design for sequences for the 89-v345.SF and 89.6 gp120 gene based on the V4-V5 region. The sequences of four primers for each HIV-1 variant are aligned. Primer sequences are shown in boxes. The percentages of homology between each of the R5/X4 variant pairs are given in brackets.

Three of these 8 primers, SF162 forward, SF162 reverse, and LAV.04 forward, have been reported previously (19). Primers for the sequences common to all HIV-1 variants were chosen within the V5 region of the gp120 gene according to Applied Biosystems guidelines. The sequences for this primer set were as follows: V5 forward, 5′-AATGTATGCCCCTCCCATCA; and V5 reverse, 5′-TCACTTCTCCAATTGTCCCTCAT.

The lengths of the PCR amplicons were as follows: SF162, 272 bp; LAV.04, 266 bp; 89-v345.SF, 188 bp; 89.6, 199 bp; and V5, 145 bp.

Real-time quantitative RT-PCR assay.

We evaluated the viral replication of each of the HIV-1 variants (SF162, LAV.04, 89-v345.SF, and 89.6) by means of quantitative real-time PCR by using the ABI Prism 7700 sequence detector (Applied Biosystems). The reaction mixture contained SYBR Green PCR master mix (Applied Biosystems), each primer at 300 nM, and 10 μl of cDNA. After activation of the AmpliTaq Gold for 10 min at 95°C, we carried out 45 cycles, with each cycle consisting of 15 s at 95°C, followed by 1 min at 60°C. A reference standard curve was obtained from serially diluted plasmids containing the target genes. Using this system, we measured each viral load separately in SF162/LAV.04- and in 89-v345.SF/89.6-coinfected tissues. We analyzed the dissociation curve for each amplification to confirm that there were no nonspecific PCR products.

Virus stocks.

We used two prototypic monotropic viruses: an X4 isolate (LAV.04) and an R5 isolate (SF162). We obtained both through the National Institutes of Health (NIH) AIDS Research and Reference Program. Also, we used the isolate 89.6 (11) and its viral chimera, 89-v345.SF, in which the V3-through-V5 env region of 89.6 was replaced with that of SF162 (31), kindly donated by R. Collmann. Both 89.6 and 89-v345.SF are dual-tropic R5X4 variants, as evidenced by their ability to mediate fusion and to infect cells transfected with CD4 and either CCR5 or CXCR4 (18, 31), but they behaved like monotropic X4 and R5, respectively, in lymphoid tissue (18, 27).

Tissue culture and viral infection.

Human tonsils surgically removed during routine tonsillectomy and not required for clinical purposes were received within 5 h of excision, dissected, cultured, and infected as described earlier (16, 20). Briefly, the tonsils were washed thoroughly with medium containing antibiotics, sectioned into 2- to 3-mm blocks with an average weight of 5 mg, and placed on top of collagen sponge gels in culture medium at the air-liquid interface. In a typical experiment, 3 to 5 μl of clarified medium containing ca. 1 ng of p24 was applied to the top of each tissue block. The amount of inoculated virus was chosen to produce comparable amounts of HIV-1 for each pair of the viral variants in tissue blocks from a given donor (17). The culture medium, bathing 54 tissue blocks in six wells, was collected every 3 days after viral inoculation, and the viral load was measured.

p24 ELISA.

We assessed productive HIV-1 infection by measuring p24 in the culture medium with an HIV-1 p24 antigen enzyme-linked immunosorbent assay (ELISA; Beckman-Coulter, Miami, Fla.).

RESULTS

Optimization of primer concentrations.

We used different primer concentrations to evaluate the magnitude of the PCR products' signals from plasmids that contained the gp120 gene for each of the viruses used. We tested each designed primer at concentrations of 50, 300, and 900 nM (data not shown). We found the same signal magnitude for all primer sets at concentrations of 300 and 900 nM, and thus we used the 300 nM concentration of primer for all further experiments.

Assessment of standard curves.

To evaluate standard curves for HIV-1 quantification, we made five 10-fold serial dilutions of plasmids. The standard curves obtained were linear, with R ≥ 0.99 for any specific primer set (Fig. 2), with slopes between −3.46 and −3.91, and with y-axis intercept points between 38.8 and 42.6.

FIG. 2.

Standard curves for HIV-1 variant-specific real-time PCR assay. Representative standard curves for one of nine experiments are shown. Tenfold serial dilutions ranging from 10¹ to 10⁵ copies of plasmid were tested in duplicate.

Reproducibility of the assay.

We tested the reproducibility of this assay by comparing the threshold cycle (C_T) for each of five 10-fold dilutions of a plasmid run in duplicate in the same plate (intra-assay variability), as well as by comparing the C_T for each plasmid evaluated in a different plate (interassay variability). These experiments demonstrated a high level of reproducibility for the assay. The variation between duplicate measurements within the same assay for the SF162 primer set was between 0.33 and 2.16%; for LAV.04, it was between 0.36 and 1.56%; for 89-v345.SF, it was between 0.08 and 2.65%; for 89.6, it was between 0.08 and 1.17%. The coefficients of variation for interassay variability for these viral variants were between 5.82 and 9.64%, 5.14 and 7.12%, 2.21 and 6.21%, and 1.83 and 5.14% (n = 7), respectively (Table 1).

TABLE 1.

Reproducibility of real-time PCR assay for HIV-1 variants

Variant	No. of DNA copies	Intra-assay variabilitya (%)	Interassay CVb (%)
SF162	105	0.42	9.48
	104	0.33	9.64
	103	0.79	7.06
	102	1.00	6.28
	101	2.16	5.82
LAV.04	105	0.36	6.56
	104	0.45	7.12
	103	0.65	5.93
	102	1.56	5.62
	101	1.32	5.14
89-v345.SF	105	0.08	6.21
	104	0.19	5.16
	103	0.64	4.71
	102	0.92	2.21
	101	2.65	3.85
89.6	105	0.65	5.14
	104	0.38	4.14
	103	0.08	4.82
	102	1.17	1.83
	101	1.09	1.86

^aThe values were calculated as follows: (A − B)/A × 100, where A and B are the C_T values of two measurements in one assay.

^bCoefficients of variation (CVs) were calculated for each dilution of each plasmid in seven experiments.

Comparison with p24 ELISA.

Since the ELISA of p24 antigen is currently used as the conventional assay to evaluate HIV-1 replication, we determined whether the results obtained with the developed real-time PCR assay correlate with those obtained with this conventional assay. We infected tissue blocks of human lymphoid tissue from seven donors with LAV.04 or SF162 and collected culture medium samples on days 3, 6, 9, and 12 postinfection. In these 28 samples, we evaluated the numbers of HIV-1 RNA copies by using a real-time PCR assay of V3-V5 region-specific primer sets for LAV.04 and SF162 and then evaluated the concentrations of HIV-1 p24 by ELISA. The results obtained with the real-time PCR were highly consistent with the p24 ELISA data. Figure 3 shows the correlation curve between these two assays for one experiment. The coefficients of correlation (R) for SF162 and LAV.04 were between 0.87 and 1.0 (P < 0.01, n = 7). The differences between different samples may be due to different amounts of free RNA and free p24 from disintegrated viral particles.

FIG. 3.

Comparison of real-time PCR assay and p24 ELISA. The correlation between the numbers of RNA copies obtained with the real-time PCR assay and the concentrations of p24 in the same samples is shown. Representative data for two of seven experiments for each viral variant are shown.

Comparison of the real-time PCR assay of V3-V5-specific primer sets and of a primer set common to all HIV-1 variants (“universal primer”).

To verify the amount of HIV-1 RNA with different primer sets, we estimated the numbers of copies of HIV RNA in the same specimens with primer sets based on variant-specific gp120 sequences of the V3-V5 regions and a universal primer set based on conserved gp120 sequences of the V5 region. Also, the latter allowed us to estimate the total amount of HIV-1 RNA by using the same real-time PCR assay. We analyzed culture medium samples from tissue blocks obtained from seven donors. Each set of tissue blocks from each donor was infected with LAV.04 or SF162. Similarly, tissues from four donors were infected with 89.6 and 89-v345.SF. For each infection, culture medium was collected at different times postinfection. The results of these experiments demonstrate the consistency of measurements with two different primer sets. The coefficients of variation for the numbers of viral RNA copies for SF162, LAV.04, 89-v345.SF, and 89.6 evaluated with variant-specific and universal primers were between 0.96 and 1.0.

Amplification of individual plasmids in their mixture.

We tested whether the designed primer sets can interact with a noncomplementary target gene. First, we used the LAV.04-specific primer to amplify the signal in a specimen containing 10⁵ copies of the SF162-specific plasmid and no LAV.04-specific plasmid. No signal was detected in this case, whereas a strong signal was detected with the SF162-specific primer. Converse experiments were done with a sample of 10⁵ copies of the LAV.04-specific plasmid which did not contain the SF162-specific plasmid. No signal was detected when the SF162-specific primer sets were used, whereas a strong signal was detected when the LAV.04-specific primer sets were used.

Next, we mixed two HIV-1 plasmids and estimated the amount of one of them in the presence of an excessive amount of the other. We compared the amplification plot for 10³ copies of the SF162 plasmid with that for a mixture comprising 10³ copies of the SF162 and 10⁵ copies of the LAV.04 plasmid. As shown in Fig. 4A, the two amplification plots were almost identical. We performed complementary experiments by mixing 10³ copies of LAV.04 plasmid with 10⁵ copies of SF162, and we compared the amplification plot for LAV.04 in this mixture with that for 10³ copies of LAV.04 plasmid alone. Again, the amplification plots were almost identical (Fig. 4B). Thus, specific primers do not react with noncomplementary targets, and the presence of the second plasmid does not interfere with the amplification of the first plasmid with its specific primers.

FIG. 4.

Amplification plots for mixtures of viral plasmids. (A) Comparison of duplicated amplification plots of 10³ SF162 plasmids with duplicated amplification plots of the same plasmid in a mixture with 10⁵ copies of LAV.04 plasmid. (B) Comparison of duplicated amplification plots of 10³ LAV.04 plasmids with duplicated amplification plots of the same plasmid in a mixture with 10⁵ copies of SF162 plasmid.

HIV-1 variant replication in coinfected lymphoid tissue.

Finally, we applied the assay we developed to measure the amounts of R5 and X4 HIV-1 variants produced in human lymphoid tissue coinfected ex vivo with a mixture of two viruses. Tissue blocks from 11 donors were coinfected with SF162 and LAV.04 or infected with SF162 and LAV.04 separately. Similarly, blocks from three donors were coinfected with 89-v345.SF and 89.6 or else infected with these viruses separately. The culture medium was changed 3 days after inoculation. On day 6, the culture medium bathing 54 tissue blocks was collected for each condition, and the amount of each viral variant was determined by means of quantitative real-time PCR (Fig. 5). The viral loads of HIV-1 variants in culture medium on day 6 were not significantly different in singly infected and coinfected tissues. The results of these measurements demonstrate that the above-described assay allows measurement of the number of RNA copies of individual HIV-1 variants replicated in human lymphoid tissue infected with viral mixtures.

FIG. 5.

Replication of R5 and X4 HIV-1 variants in coinfected human lymphoid tissue ex vivo. Shown are numbers of RNA copies of LAV.04 and of SF162 HIV-1 variants (A) and of 89.6 and of 89-v345.SF HIV-1 variants (B) on day 6 postinfection in matched tissues infected separately or coinfected with two viruses. For each condition, RNA that was specific for each HIV-1 variant was quantified on day 6 postinoculation. Each datum point represents a mean (± the range) measurement of two pooled medium samples from six wells containing 54 tissue blocks.

DISCUSSION

HIV-1 viral load is one of the important correlates of the status of the HIV-infected patient and of disease progression (8, 22). A variety of methods to measure HIV-1 viremia have been developed, including evaluation of p24 antigen with ELISA and evaluation of viral RNA with various modifications of PCR and other molecular biology assays (9, 21). These methods determine the total amounts of viral components in a swarm or identify drug-resistant variants. In the present study, we describe a real-time PCR assay for evaluating the contribution of individual viral variants in a model of ex vivo tissues coinfected with R5 and X4 HIV-1 variants. Earlier, we used a similar method to study the influence of herpesvirus 6 on HIV-1 replication (19). Here, having optimized the assay, we report on new variant-specific primers and prove the accuracy and reproducibility of results obtained with different primer sets.

We used two pairs of HIV-1 variants. The first pair consisted of two isolates: a prototypic R5 virus, SF162, and a prototypic X4 virus, LAV.04. The second pair consisted of an R5X4 isolate, 89.6, and a genetic construct, 89-v345.SF, which is also an R5X4 variant (31). Whereas the first pair consisted of two genetically diverse HIV-1 isolates, the second pair consisted of viruses that were isogenic to each other except for limited sequences in V3-V5 loops of gp120. In tissues, these two dual-tropic viruses behave monotropically: whereas 89.6 uses CXCR4 preferentially, 89-v345.SF uses CCR5 (18). Our goal was to quantitatively assess replication of individual viruses in tissues coinfected with these viral pairs.

It might have been useful to design a real-time RT-PCR assay based on two primers common to all viruses and one probe that discriminates amplicons from different viral variants. However, it turned out that this was not feasible because in the case of an excess of one virus, its amplicon inhibits amplification of a target gene sequence of another virus, as happens in a competitive PCR assay (15). This is why we chose a more straightforward strategy and designed our assay on the basis of virus-specific primers and SYBR Green I dye to visualize the amplification product. The gp120 sequence differences between HIV-1 variants that determine their preferential CCR5 or CXCR4 utilization and on which we based the primers' design were restricted to a few small regions scattered throughout gp120.

Because of this, we were not able to design primers that fit all of the standard requirements for real-time PCR (1). Thus, we have had to ignore one or more of these requirements in our design strategies. Although trials and errors are inevitable, our results show that if two 30-bp regions differing at least by 23% and separated by up to 300 bp can be located, appropriate primers can be designed. How much difference in sequences between HIV-1 variants must there be in order for this assay to discriminate between them? The sequence differences that we were able to discriminate with our assay in the present study were on average 44% over the length of two 30-bp segments. In general, the position of the sequence difference within the targeted segment may be important for discrimination between closely related viruses. For example, it has been reported (23) that a sequence difference as small as a few base pairs but located next to the 3′ end reduces PCR efficiency and thus potentially can be used to design primers that discriminate between closely related virus variants.

We have shown here that (i) the designed primers recognize only their target genes and nontarget genes are not amplified, (ii) the assay variations were small, and (iii) the standard curves were linear, allowing a reliable determination of the numbers of RNA copies in experimental samples. We verified the accuracy of the assay by comparing the results obtained with virus-specific primers and those obtained with a primer common to all of the viruses. Moreover, the amounts of HIV-1 RNA copies in various experimental conditions correlated with the concentrations of p24, another conventional assay to determine virus concentration. We successfully applied the newly developed assay to an ex vivo model consisting of blocks of human lymphoid tissue infected with mixtures of R5 and X4 HIV-1 variants.

The above-described real-time quantitative PCR assay to measure replication of individual HIV-1 variants in coinfected tissues may be important for understanding the mechanism of the R5-to-X4 switch and can be used in other experimental systems in which such mechanisms are studied (cell lines, peripheral blood mononuclear cells, experimental animals, etc). Moreover, if HIV-1 in blood or tissue biopsy samples of an infected individual is sequenced, the method described above may allow study of whether the changes in the relative presence of individual variants in a viral swarm in the course of HIV infection are of prognostic value for disease progression.