Fluorometric assay for phenotypic differentiation of drug-resistant HIV mutants

Convenient drug-resistance testing of viral mutants is indispensable to effective treatment of viral infection. We developed a novel fluorometric assay for phenotypic differentiation of drug-resistant mutants of human immunodeficiency virus-I protease (HIV-PR) which uses enzymatic and peptide-specific fluorescence (FL) reactions and high-performance liquid chromatography (HPLC) of three HIV-PR substrates. This assay protocol enables use of non-purified enzyme sources and multiple substrates for the enzymatic reaction. In this study, susceptibility of HIV mutations to drugs was evaluated by selective formation of three FL products after the enzymatic HIV-PR reaction. This proof-of-concept study indicates that the present HPLC-FL method could be an alternative to current phenotypic assays for the evaluation of HIV drug resistance.

The present study aims to develop a high-performance liquid chromatography (HPLC)-based FL assay method that uses multiple acetyl peptide substrates for phenotypic differentiation of drug-resistant HIV-PR mutants in non-purified HIV-PR samples, utilizing the catechol, NaIO 4 and borate reagents 33 for the FL detection of N-terminal Phe-, Leu-or Val-containing peptides. Consequently, we found that the developed HFA method offers great potential for facile, informative and reliable phenotypic differentiation of HIV-PR mutations that engender drug resistance.

Results
HFA protocol for the phenotypic differentiation of drug resistant HIV-PR mutants. We developed HPLC-based FL assay (HFA) for phenotypic differentiation of drug-resistant HIV-PR mutants by using the enzymatic and peptide-specific FL reactions and HPLC analysis of three HIV-PR substrates. Figure 1 shows the schematic HFA protocol. The HIV-PR gene from an AIDS patient was first cloned into a prokaryotic expression vector and expressed in E. coli cells. The cell lysate containing HIV-PR was then directly used for enzymatic reactions with three N-terminal acetyl substrates in the absence or presence of a pharmaceutical HIV-PR inhibitor. Subsequently, the N-terminal free peptides enzymatically produced from the acetyl peptides were converted with catechol to FL compounds, which were analyzed by the following HPLC. The susceptibility of the patient's HIV-PR was evaluated by the formation ratio of three cleaved peptides. In addition, drug resistance of HIV-PR mutants towards inhibitors was analyzed on the basis of the change of IC 50 , comparing with that of wild-type HIV-PR.
Multiple substrates and HPLC-FL detection of enzymatic products. Three peptide substrates, SGIFLETSLE, ARVLFEAM and KSGVFVQNGL, were selected from several HIV-PR substrates [41][42][43][44][45] reported for wild-type HIV-PR, and used as acetyl peptides to prevent the FL reaction between substrate and catechol. To confirm suitability of the three HIV-PR substrates, their enzymatic products were reacted with catechol in the presence of NaIO 4 and borate 33 , and then determined by reversed-phase HPLC with a fluorescence (FL) detector (Fig. 2).
Three FL peaks corresponding to LETSLE, FEAM and VQNGL were separated and then detected within 15 min by the HPLC as shown in Fig. 2A. It means that the three N-terminal free peptides of LETSLE, FEAM and VQNGL were enzymatically produced at pH 5.5 from their acetyl substrates and successfully reacted with the catechol reagent in a buffered borate aqueous solution (pH 7.0) to generate corresponding FL compounds, which could be sensitively detected by the HPLC with the FL detector. Additionally, low background noises in the chromatogram indicate that most other components in the enzymatic reaction (e.g., buffers, proteins, nucleic acids, free amino acids and lipids) did not disturb the FL detection for the three peptide products.
The quantitative determination of the LETSLE, FEAM and VQNGL products was performed by the HPLC analysis using their standard curves (Fig. 2B). Areas of their FL peaks in the chromatogram were calculated and plotted against those of peptides. Proportional relationship between peptide concentrations and FL signals could be obtained over a wide concentration range (1-3000 pmol per injection) for each peptide (R 2 = 0.97-0.99). Detection limits were 0.5-1.0 pmol per injection at a signal-to-noise ratio of 3. These results indicate that the present HFA can provide sensitive and simultaneous determination for peptides produced by the enzymatic reaction with HIV-PR.
Activity assay of HIV-PR and its mutants. The wild-type HIV-PR gene was derived from the pNL4-3 HIV-1 clone and expressed in E. coli cells. We determined the Wt HIV-PR concentration in cell lysate with a quantitative immunoblotting method against a standard curve of a commercially available purified HIV-PR. The cell lysate was directly used and measured for the HIV-PR activity without further purification.
The cell lysates containing different concentrations of the wild-type HIV-PR were incubated with the three substrates for 4 h, followed by the FL reaction and HPLC analysis (Fig. 3A). Their FL peaks corresponding to the peptide products of LETSLE, FEAM and VQNGL were proportionally increased with the amounts (1.7-6.6 pmol) of the HIV-PR in the enzymatic reaction mixture. This result means that the present HFA can measure the proportional activity of HIV-PR depending on the enzyme concentration. The cell lysates that contained 5 pmol of the wild-type HIV-PR were incubated with the three acetyl substrates at 37 °C for prolonged periods (Fig. 3B). The HIV-PR activity was constant for at least 6 h, indicating that all three substrates were enzymatically cleaved by HIV-PR. Thus, the present assay can be used to measure specific HIV-PR activities in cell lysates. The present assay format could differentiate wild-type (Wt) HIV-PR and its mutants. Mutants, Ma and Mb were from drug-binding sites of G48V and V32I, respectively 24 . The cell lysate of Ma or Mb was incubated with three substrates under the same enzymatic conditions as for Wt HIV-PR, followed by the FL reaction and HPLC analysis. Enzymatically produced peptide fragments of LETSLE, FEAM and VQNGL from each substrate were determined, and their enzymatic productions differed slightly between Wt HIV-PR and its mutants (Ma and Mb) as shown in Fig. 4A. Therefore, the activities of Wt HIV-PR and its mutants were individually calculated for the three substrates. Figure 4B shows their activity ratios against the substrate of [Ac]-SGIFLETSLE for each enzyme source. The results explain that each activity ratio for the three substrates reflects different patterns for the mutants, and suggest that HIV-PR affinity varies by substrate with a different Km value.
Evaluation of drug resistance. The proposed assay format was further used to analyze the phenotypic drug resistance of the HIV-PR mutants for HIV-PR inhibitors of saquinavir, indinavir, lopinavir and ritovavir, which are clinically used. The HIV-PR mutants of Ma (G48V) and Mb (V32I) are reportedly resistant to saquinavir [46][47][48] and indinavir 48,49 , respectively. Thus, different levels of the drug resistance of the mutants and Wt HIV-PR were first studied by using saquinavir and indinavir as HIV-PR inhibitors. In this experiment, the inhibition rates by these drugs on HIV-PR activity were measured by the decrease in the FL peak area corresponding to enzymatic peptide products. Figure 5A-C show the inhibition rate (%) of the enzyme activity for the substrates, [Ac]-SGIFLETSLE, [Ac]-ARVLFEAM and [Ac]-KSGVFVQNGL, respectively, depending on the saquinavir concentration in the enzymatic reaction mixture. As the results, the Ma mutant showed hysteretic inhibition curves with saquinavir for all the substrates, but not for indinavir. However, mutant Mb showed a hysteretic inhibition curve only with indinavir for the [Ac]-SGIFLETSLE substrate (Fig. 5D).  To more readily assess the drug-resistance profile of each HIV-PR mutant, we determined each fold resistance value that was corresponded to the ratio of the inhibition rate at a single concentration of an inhibitor for each mutant (Ma and Mb) against that for Wt HIV-PR. Figure 6 depict the drug resistance profiles for the HIV-PR mutants. In this experiment, we used four inhibitor drugs (saquinavir, indinavir, lopinavir and ritonavir) towards the three substrates and calculated each value by the formula shown in the caption of Fig. 6. The results explain that Ma showed weak resistance to saquinavir, and not to other drugs; whereas Mb showed strong resistance to indinavir; however, neither Ma nor Mb were resistant to lopinavir or ritonavir.
The above data for IC 50 doses and fold change in HIV-PR inhibition especially with sauinavir and indinavir were listed in Table 1. The data represent that the mutants, Ma and Mb clearly gave higher IC 50 and fold change values with sauinavir and indinavir, respectively than Wt HIV-PR. It means that Ma is resistant to saquinavir, while Mb is resistant to indinavir.

Discussion
By taking advantage of the simplicity and sensitivity of the FL reaction of N-terminal free oligopeptides with catechol 33 , we have developed a novel method for phenotypic differentiation of drug-resistant HIV-PR mutants. Using this method, we obtained IC 50 values of several HIV-PR inhibitors by the simultaneous detection of three FL peaks that corresponded to enzymatically produced peptides from three substrates, and thus could evaluate drug resistance of two representative mutants (Ma and Mb) of HIV-PR according to the IC 50 doses of inhibitor drugs (Fig. 5). We determined the fold changes for the drug resistance profiles using several HIV-PR inhibitors towards the three substrates for the HIV-PR mutants (Fig. 6). The values of the IC 50 dose and fold change of two inhibitors, saquinavir and indinavir that are clinically used, were also listed in Table 1. Fold changes in resistance evaluation could help quickly produce extensive drug-resistance profiles of HIV-PR in patients. Those results were consistent with their reported phenotypes 48 , which support that the present HFA method can allow the phenotypic assay of drug-resistant HIV-PR correctly, although more mutant samples are needed to determine cutoff values for different drugs.
The present HFA method uses three different substrates simultaneously, and thus makes the outcome more informative and reliable. Resistance mutations in HIV-PR reduce the binding affinity of PR to its inhibitors and other substrates to varying degrees 14 , so using more than one substrate can avoid the possible influence of substrate preference of HIV-PR mutants. The HFA method could be performed within 1-2 weeks, which is much quicker than clinically used recombinant virus-based methods that need 3-4 weeks 30 . Turnaround time for HFA would be further decreased by using advanced HPLC that allows automatic sampling, automatic measurement and high-throughput analysis. The present HFA permitted direct use of cell lysate and highly sensitive detection of enzymatic peptide products. It suggests that the method might directly use specimens such as patients' blood and lymph fluid. Further studies with patient's blood specimens should be needed for the practical application of HFA, though we could not get AIDS patient's specimens.
In principle, development of similar assays for phenotypic differentiation of drug-resistant HIV-reverse transcriptase, and a combination assay for both drug-resistant HIV-PR and HIV-reverse transcriptase is possible. Taken together, the present HFA method shows great potential to enable inexpensive, rapid, informative and reliable phenotypic differentiation of drug-resistant HIV-PR mutants.

Preparation of Wt HIV-PR and its drug-resistant mutants. A Wt HIV protease gene (Gene
Bank accession no. M19921) was cloned into the pMAL-c2x plasmid (New England Biolabs, Ipswich, UK) to make a Wt HIV-PR-expressing vector (pMAL-c2x-PR) using conventional cloning techniques In this formula, A i is the FL peak area of each enzymatic product in the presence of each drug, and A 0 is the peak area of the product in the absence of the drug. Wt is wild-type HIV-PR, and M is its mutant. and primers: 5′ -GAGGGAAGGATTTCATCCTTCAGCTTCCCTCAGATCACTCTTTGG-3′ and 5′ -AGGAAGCTTTTAAAAATTTAAAGTGCAGCC-3′ . As two drug-resistant mutations (G48V 47 and V32I 48 ) for the HIV-1 protease gene were reported, their genes were introduced into the pMAL-c2x-PR to generate two HIV-PR mutants, the Ma-and Mb-expressing vectors, respectively. The mutations were carried using a QuickChange II Site-Directed Mutagenesis kit (Agilent Technologies, CA, USA) and primers: 5′ -GATGGAAACCAAAAATGATAGTGGGAATTGGA GG-3′ and 5′ -CAGGAGCAGATGATACA ATTTTAGAAGAAATGAATTTGCCAGG-3′ . The vectors were confirmed by DNA sequencing. Each vector containing the Wt or mutant HIV-PR gene was transformed into E. coli strain DH5α cells. The transformed E. coli cells were cultured in LB broth until they attained OD 600 of 0.6-0.8; 1.0 mM ITPG was then added to induce HIV-PR expression over 2 h. Cells were centrifuged at 13,400 × g for 3 min at 4 °C and then washed with 1 × PBS. After re-suspension in water, the cells were sonicated three times in ice water at power 40 with 10-sec intervals (Model 300 V/T, BioLogics Inc. NC, USA). The lysate was centrifuged at 13,400 × g for 3 min at 4 °C; its supernatant was stored at − 40 °C. The HIV-PR concentration in the lysate was determined by western blotting 50  FL reaction. Conditions for the FL reaction of peptides were as previously reported 33,39 , but reaction conditions for the enzymatically produced peptides were slightly modified as follows: a portion (50 μ l) of the enzyme reaction mixture or an aqueous solution of synthetic standard peptide was successively mixed with 2.5 μ l of 0.1 N NaOH, 10 μ l of PBS, 25 μ l of 0.3 M H 3 BO 3 -NaOH (pH 7.0), 50 μ l of 2.5 mM catechol and 25 μ l of 2 mM NaIO 4 . The mixture (162.5 μ l) was then heated to 100 °C for 10 min. After the reaction, the mixture was immediately cooled in ice water to stabilize the FL compounds. Finally, the mixture was centrifuged at 13,400 × g for 3 min at 4 °C and the supernatant containing the FL compounds was subjected to HPLC.

HIV-PR
HPLC analysis. The FL compound quantities were analyzed by an HPLC system, which consisted of a pump with degasser (PU-2089 plus, JASCO, Japan), UV detector (UV-2070, JASCO), FL detector (FP-920, JASCO) and chart recorder (Ross, USA). The analytical column was TSK gel ODS-80Ts (Tosoh, Japan). The FL reaction mixture (5-20 μ l) was injected into the HPLC system. Gradient elution of the mobile phase composed of eluent A (methanol), B (0.25 M H 3 BO 3 -NaOH, pH 7.0) and C (H 2 O) was performed for the separation as follows: 0-35% of A, 5% of B, and 95-60% of C for 0-25 min; 35-80% of A, 5% of B, and 60-15% of C for 25-26 min; 80% of A, 5% of B, and 15% of C for 26-33 min; 80-0% of A, 5% of B, and 15-95% of C for 33-34 min; followed by at least 10 min for the reconstitution of the column before the next analysis. The flow rate was 1.0 ml/min. The FL detector was set at 400 nm for excitation and 490 nm for emission. The enzymatically produced peptides were quantified by measuring their FL peak areas in the chromatogram. The inhibition rate of an inhibitor on HIV-PR activity was calculated by the relative decrease in product peptides. The concentration of the inhibitor that inhibited 50% the HIV-PR activity (IC 50 ) on each substrate was calculated with the Forecast function of Excel software.
Evaluation of drug resistance. Cell lysates containing HIV-PR (< 5 pmol) were incubated with the three substrates as described above, with varying concentrations (0-2.5 μ M) of therapeutic HIV-PR-inhibitors: saquinavir mesylate (Sigma, China), indinavir sulfate salt hydrate (Sigma, China), lopinavir (Santa Cruz Biotechnology, CA, USA) or ritonavir (Toronto Research Chemicals Inc., Canada). The enzymatically formed peptides were analyzed as described above.