WO2000066774A2

WO2000066774A2 - Improved assay and reagents therefor

Info

Publication number: WO2000066774A2
Application number: PCT/GB2000/001639
Authority: WO
Inventors: Edward Duncan Blair; Laurence Henry Robinson; Barbara Wendy Snowden; Sylvia Margaret Tisdale
Original assignee: Glaxo Group Limited
Priority date: 1999-04-28
Filing date: 2000-04-28
Publication date: 2000-11-09
Also published as: GB9909793D0; AU4586700A; WO2000066774A3; JP2002542804A; EP1226270A2

Abstract

The present invention relates to methods for generating recombinant viruses from samples such as uncharacterised virus samples or clinical specimens, to the use of the viruses so generated in assays, predominantly for the purpose of detecting altered viral susceptibility to anti-viral drugs, and to reagents, more particularly deoxyribonucleic acid (DNA) vector constructs, for use in such methods and assays. The assays are adapted to detect resistant virus more accurately and sensitively than known assays by taking into account compensatory mutations arising in nucleotide sequences other than those encoding the anti-viral drug target.

Description

Improved Assay And Reagents Therefor

Resistance of human immunodeficiency virus type 1 (HIV-1 ) to anti-retroviral drugs presents a major challenge in the chemotherapeutic prevention of progression to Acquired Immunodeficiency Syndrome (AIDS). The acquisition of viral resistance is accentuated by the rapid turnover rate of the virus in vivo combined with the high error rate of the viral reverse transcriptase (RT) enzyme (Coffin, 1995; Ho et al., 1995; Mansky and Temin, 1995; Wei et al., 1995). Thus, the current goal of anti-HIV therapy is maximally to suppress replication of the virus so as to delay the appearance of drug-resistant variants and maintain healthy levels of CD4⁺ immune cells for as long as possible (Vandamme et al., 1998). To this end, the more recently introduced therapies of HIV infection usually involve combinations of three more anti-retroviral drugs.

The use simultaneously of reverse transcriptase (RT) inhibitors (RTIs) and inhibitors of the viral protease (Pro or PR) enzyme (protease inhibitors or Pis) has led to the emergence of HIV-1 viruses with resistance mutations in either or both drug of the targets RT and Pro (Hertogs et al., 1998). These mutations alter the structure and/or chemical affinities of the target enzymes such that their ability to interact with the drugs is altered or reduced and the drugs show diminished activity against the mutated virus. Drug resistance mutations can potentially occur in the drug target molecules of any other viruses. For example, although in other viruses antiviral drugs may act on the viral protease or, if the virus is an RNA virus, at the reverse transcriptase, other targets may also be employed. Indeed any viral nucleic acid or protein critical to viral reproduction or infectivity may be a potential drug target, for example several anti-herpes nucleoside analogue drugs e.g. aciclovir act against HSV 1 or 2, Varicella Zoster and other herpes family viruses through phosphorylation of the drug molecule by the viral thymidine kinase and further processing by host cellular enzymes, including incorporation into viral DNA by host DNA polymerase activity. Mutations associated with reduced susceptibility to nucleoside analogues have been demonstrated in the thymidine kinases or DNA polymerases of herpes viruses (Kimberlin and Whitley, 1996; Balfour, 1999). Other viral drug targets include the DNA maturation factors and DNA polymerases of certain DNA viruses.

It has recently been reported that mutations can occur in drug-resistant viruses at sites other than the drug targets themselves, for example at the cleavage sites (CSs) at which Pro cleaves precursor proteins to create the structural and functional viral proteins. In HIV-1 , the p7 protein is cleaved from the Gag and Gag-Pol precursor polyproteins at the p7/p1 cleavage site (CS); viral proteins p1 and p6 are cleaved from the Gag polyprotein at the p1/p6 CS; Pro, RT and integrase (INT) are also produced from cleavage of Gag-Pol. Mutations at cleavage sites are thought to compensate for impaired polypeptide cleavage activity of the mutant Pro which would otherwise lead to loss of viral fitness and are hereafter referred to as compensatory mutations. Compensatory mutations have been documented at the p7/p1 and the p1/p6 CSs (Doyon et al., 1996; Maschera et al., 1996a, b; Zhang et al., 1997; Carrillo et al., 1998; Mammano et al., 1998; Zennou et al., 1998). Thus, mutations associated with reduced susceptibility to currently marketed anti-retroviral agents occur in at least three distinct regions of the HIV-1 genome: RT, Pro and the CSs. It can be valuable to monitor the emergence of drug-resistant virus as part of the management of infection and disease. For example, assays of the resistance phenotype of HIV-1 isolates play a major role in the management of disease by identifying the development of reduced susceptibility to therapy, cross- resistance and re-sensitisation. The co-culture of peripheral blood mononuclear cells (PBMC) from infected subjects with uninfected donor PBMC is one method which has been used to isolate HIV-1 from patients for phenotypic assays (Japour et al., 1993). Co-culture of PBMC is not ideal for large scale regular application because it involves the isolation of fresh PBMC and the long culture times have been shown to select for minority or less drug-resistant variants (Kusumi et al., 1992; Mayers et al., 1998). The recombinant virus assay (RVA) enables the rapid and reproducible determination of phenotypic susceptibility of HIV-1 from plasma (Kellam and Larder, 1994; Maschera ef al., 1995; Hertogs et al., 1998) and has thus been instrumental in directing the choice of drugs used in HIV-therapy. In the RVA, HIV RT or Pro sequences are amplified from plasma by reverse transcription-polymerase chain reaction (RT-PCR) and co- electroporated into CD4⁺ cultured cells (MT4) with a molecular clone of an RT or Pro-deleted "provirus" of the HIV-1 standard laboratory strain wild type HXB2, termed the 'vector'. "Provirus" is the term used to describe the DNA copy of an HIV virus genome which integrates into the host cell's DNA. The plasma-derived

RT or PR sequences insert into the corresponding deletion site of the wild type vector by homologous recombination, and the resulting recombinant vector DNA inserts into the cellular DNA to yield full length infectious proviruses. Gene expression from the proviruses yields a population of viruses with a drug susceptibility phenotype representative of the subject's plasma viruses. The resulting virus stocks can be used for drug sensitivity tests against PR or RT inhibitors, depending on which part of the genetic information in the recombinant virus is derived from the clinical isolate. The RVA has several advantages over the co-culture of PBMC in addition to its rapidity and reproducibility. The production of virus in a common backbone enables a more accurate comparison of the drug sensitivities and growth characteristics of the viruses produced, and the shorter culture times minimise the outgrowth of minor variants.

Similarly, RVA-type assays can be directed to the detection of drug-resistant mutants of viruses other than HIV, by the recombination of sequences corresponding to the anti-viral drug target, derived by PCR or RT-PCR from a patient tissue sample, with a DNA vector including a wild type (or other standard laboratory strain) viral genome carrying a deletion corresponding to the sequence encoding the drug target.

According to a first aspect, the present invention provides an assay for the detection of virus resistant to an anti-viral drug by recombination of a DNA vector with a nucleotide sequence derived from a sample of virus suspected of including drug-resistant virus to produce viable recombinant virus having the resistance profile of the virus sample, wherein the DNA vector comprises a wild- type laboratory virus strain genome carrying a deletion of at least a portion of the sequence encoding the antiviral drug target and a further deletion of sequences comprising a potential site of compensatory mutation. The nucleotide sequence derived from the sample of virus, which may be obtained by PCR, RT-PCR, or related methods, comprises a region of the viral genome substantially corresponding to the sequences deleted from the DNA vector.

Accordingly, the invention provides an assay for the detection of virus resistant to an anti-viral drug comprising the step of generating recombinant virus having a drug resistance profile of the virus in a viral sample by recombination of a DNA vector with a nucleotide sequence derived from the sample, wherein the DNA vector comprises a wild-type laboratory virus strain genome carrying a deletion of at least a portion of the sequence encoding the antiviral drug target and a further deletion of sequences comprising a potential site of compensatory mutation. The assay may further comprise the steps of infecting cells with recombinant virus so produced, incubating infected cells with an antiviral drug and detecting the viability of virus or of infected cells after incubation to determine the sensitivity of the recombinant virus to the drug.

In preferred embodiments, the drug target is a viral protease, polymerase or reverse transcriptase enzyme. Desirably, the virus is a herpes family virus or an

HIV virus, preferably an HIV virus, more preferably HIV-1 , in which case the DNA vector comprises the sequence of an HIV provirus carrying a deletion in a gene encoding a drug target and in a site of compensatory mutation. In certain preferred embodiments, the DNA vector carries a deletion in the sequence encoding the viral protease and a further deletion of a sequence encoding one or more protease cleavage sites. Desirably, the DNA vector carries a deletion of one or both of the HIV-1 p7/p1 and p1/p6 protease cleavage sites in addition to a deletion of the sequence encoding the viral protease, or a fragment thereof.

In other embodiments, the assay employs a DNA vector carrying a deletion of at least a fragment of the sequence encoding the viral reverse transcriptase. In certain embodiments, sequences encoding the viral reverse transcriptase and the viral protease, or fragments thereof, are deleted in addition to a deletion of one or more protease cleavage sites. For example, the DNA vector may carry a deletion of one or both of the HIV-1 p7/p1 and p1/p6 protease cleavage sites, of the entire sequence encoding the viral protease, of the entire sequence encoding the viral reverse transcriptase and, optionally, of the sequence encoding the viral polymerase, or a fragment thereof.

In a second aspect, the present invention provides DNA vectors comprising a wild-type laboratory virus strain genome carrying a deletion of at least a portion of the sequence encoding an antiviral drug target and a further deletion of sequences comprising a site of compensatory mutation. In preferred embodiments, the drug target is a viral protease, polymerase or reverse transcriptase enzyme. Desirably, the virus is a herpes family virus or an HIV virus, preferably an HIV virus, more preferably HIV-1 , in which case the DNA vector comprises the sequence of an HIV-1 provirus carrying a deletion in a sequence encoding a drug target and in a site of compensatory mutation. In certain preferred embodiments, the vector carries a deletion in the sequence encoding the viral protease and a further deletion of a sequence encoding one or more protease cleavage sites. Desirably, the vector carries a deletion of one or both of the HIV-1 p7/p1 and p1/p6 protease cleavage sites in addition to a deletion of the sequence encoding the viral protease, or a fragment thereof. In other embodiments, the DNA vector carries a deletion of at least a fragment of the sequence encoding the viral reverse transcriptase. In certain embodiments, sequences encoding the viral reverse transcriptase and the viral protease, or fragments thereof, are deleted in addition to a deletion of one or more protease cleavage sites. For example, the DNA vector may carry a deletion of one or both of the HIV-1 p7/p1 and p1/p6 protease cleavage sites, of the entire sequence encoding the viral protease, of the entire sequence encoding the viral reverse transcriptase and, optionally, of the sequence encoding the viral polymerase, or a fragment thereof.

In a further aspect, the present provides DNA vectors such as set out above for use in an assay according to the first aspect of the present invention.

It can also be envisaged that, in HIV-1 or other viruses, compensatory mutations may occur in viral proteins associated with the RT enzyme in its functional role (accessory proteins), or in the binding, activation or initiation sites in the RNA genome from which the RT enzyme commences reverse transcription of the genome in the first step of viral replication. Similarly, in viral chemotherapy of certain infections, the viral polymerase enzyme is the drug target, in which circumstances resistance mutations may occur in the viral sequences encoding the polymerase and compensatory mutations might be envisaged in accessory proteins or in polymerase binding, activation, initiation etc. sequences of the viral genome. Other possible mechanisms by which viruses could overcome drug induced mutations which are deleterious to viral growth include the modulation of expression of those drug targets by compensatory mutations in regulatory nucleic acid sequences or by mutations affecting viral factors which control levels, timings or patterns of expression. Compensatory mutations in regulatory sequences may or may not cause amino acid changes in any proteins they encode. The skilled man will be able to modify the teachings of the present invention to create DNA vectors for use in RVA assays in which the sites of such compensatory mutations are deleted, in addition to deletion of RT, Pro, polymerase or other drug target sequence (or a region thereof). Such vectors are included within the present invention.

Where the assay of the present invention is applied to HIV, sample (RT-)PCR products derived from infected subjects may be co-transfected with a vector derived from any previously characterised strain of HIV-1 to produce viable recombinant virus, for example strains other than the standard HXB2 laboratory strain may be used. In this way, the assay can easily be adapted to generate variants in different 'vector' strains, which might be useful for examining the effects of the vector on fitness, drug resistance or pathogenesis. The strain chosen to provide the vector sequence should preferably be well characterised in order that variations in growth and sensitivity to antivirals in the recombinant virus produced can be assigned to the vector strain or to the sequences derived from the viral (e.g. patient) sample. A wild-type strain is generally suitable as it contains no pre-existing resistance mutations. The vector strain also needs to be chosen with regard to the cell culture conditions to be used in the generation of recombinant virus and the subsequent drug resistance. The skilled man will be able to select a vector strain which displays sufficiently strong replication characteristics in the cell culture of interest. Thus, throughout this specification and the claims which follow, reference to a "laboratory strain", "laboratory virus strain", "wild type strain" or to a "wild-type laboratory virus strain" should be understood to mean any previously characterised viral strain. In a further aspect, the present invention provides a kit for the performance of an assay according to the first aspect of the invention, the kit comprising a DNA vector construct according to the second aspect of the invention. Also provided is the use of a DNA vector construct according to the present invention in an assay for the detection of virus resistant to an anti-viral drug.

The present invention is further described and illustrated in the following non- limiting Examples and by reference to the accompanying drawings in which:

FIG. 1. illustrates diagrammatically the construction of three CS-deleted plasmids for the RVA;

(A) shows the plBI20 vector containing the Pro-deleted HIV-1 sequence pHXBΔPro including the Apa\ site in the vector cloning site;

(B) shows pHXBΔProA made by removing the Apa\ site from pHXBΔPro; (C) shows pHXBΔProA made by removing the Apa\ site from pHXBΔPro;

(D) shows plasmid pHXBΔCSPro made by removal of a 200 base pair fragment encompassing the p7/p1 and p1/p6 CSs from pHXBΔProA;

(E) shows plasmid pHXBΔCSPRTA made by extension of the deletion in pHXBΔCSPro to include RT up to codon 232; (F) shows plasmid pHXBΔCSPRTC in which the deletion in pHXBΔCSPro is extended up to RT codon 483;

FIG. 2

(A) shows diagrammatically regions of the HIV-1 Gag and Gag-Pol polyproteins, with the locations of the Pro CSs indicated with a Δ (TF:- transframe protein;

NC:- nucleocapsid protein);

(B) shows the locations of the primers used in sequencing and in the generation of PCR products for assay, with their 5'->3' orientations indicated by arrows;

(C) shows diagrammatically by double pointed arrows in parts a) and c)-e) the regions of the HIV-1 genome that are deleted in the RVA plasmids shown in

Fig. 1 A and C-E; FIG. 3. shows a diagrammatic summary of the competitive RVA experiments of Example 3.

FIG. 4. shows the relative growth kinetics of recombinant drug resistant viruses generated from RVA vectors in which the CS region is included from the plasma isolate or from the laboratory strain.

Examples

In these studies, we have extended the utility of existing RT or Pro-deleted RVA constructs (Kellam and Larder, 1994; Maschera et al., 1995; Goulden) to create new plasmids with contiguous deletions of: 1) the p7/p1 and p1/p6 CSs and the entire Pro gene; 2) the CSs, Pro and RT up to codon 232; or 3) the CSs, Pro and RT up to codon 483. These constructs enable a rapid method for assessing the drug-susceptibility phenotypes of plasma virus in subjects receiving any of the currently available anti-HIV drugs. Furthermore, by performing competitive RVAs with vector/PCR fragment combinations that include or omit CS changes, we have demonstrated the importance of including the CSs in the RVA.

Example 1 - Construction of CS-deleted HIV-1 provirus clones for use in the RVA.

In order to perform phenotypic analysis of HIV-1 from patients receiving Pro and RT inhibitors, we constructed RVA plasmids with deletions in Pro, RT and the p7/p1 and p1/p6 CSs by extending the deletion in the existing Pro-deleted HIV-1 proviral clone (Maschera et al., 1995). The Pro deletion (Fig. 1A) was extended to include the CSs (Fig. 1 C) and this CS and Pro-deleted construct was then modified using current RT-deleted constructs to include deletions in RT (Figs. 1 D and 1 E). The CSs are located in a 200 bp region between the Apa\ site in gag and the BstEW site at the deletion in pHXBΔPro. We deleted the CSs by removing this fragment. The proviral clone pHXBΔPro comprises Pro-deleted WT HIV-1 virus HXB2 cloned into the plBI20 vector. As can be seen from Fig. 1A, in addition to the Apa\ site in gag, the provirus has a further Apa\ site at the cloning region of the plBI20 vector. Before the 200bp Apa\-BstE\\ fragment containing the CSs could be removed, first it was necessary to remove this additional Apa\ site. Plasmid pHXBΔPro was digested with Mlu\ and Xba\ to remove a 29 bp fragment encompassing the /\pal site and blunt-ended with T4 DNA polymerase (New England Biolabs). The blunt ends were ligated together with an Xba\ linker

(Stratagene) to yield plasmid pHXBΔProA. Ligations were carried out using the rapid DNA ligation kit from Roche Diagnostics. Ligated DNA was transformed into Eschehchia coli strain XL-1 Blue supercompetent bacteria (Stratagene) and positive colonies were retransformed into strain JM109 (Stratagene) to maximise preparation yields. The CSs were removed from pHXBΔProA by digestion with

/.pal and Ss-EII and ligation with a custom-made Hpal linker (Applied Biosystems) which restored the Apa\ and BsfEII sites and introduced a Hpal site, yielding pHXBΔCSPro (Fig. 1 B). The sequence of the Hpal linker is:

Apa\ Hpal BstEW

5'-CGCGTTAACGCG-3' 3'-CCGGGCGCAATTGCGCCACTG-5'

Plasmid pHXBΔCSPRTA (Fig. 1 D) contains a deletion in the CSs, Pro and RT up to codon 232 and was constructed by replacing the 5.9 kbp Hpal / BamHI fragment from pHXBΔCSPro with the 5.2 kbp Bstl 1071 / BamHI fragment from the plasmid pHIVDRTBs-11071 which has a deletion in RT from codon 39 to 232 (Goulden). Plasmid pHXBΔCSPRTC (Fig. 1E) contains a deletion in the CSs, Pro and RT up to codon 483 and was constructed by replacing the 5.9 kbp

BstEW I BamHI fragment from pHXBΔCSPro with the 4.4 kbp Bs.EII / BamHI fragment from the plasmid pHIVΔBs , which has a deletion in RT from codon 41 to 483 (Kellam and Larder, 1994). Plasmids were verified by restriction endonuclease analysis and sequencing through the modified region with a Perkin Elmer ABI Prism 377 DNA sequencer and Big Dye terminator cycle sequencing. The primer used for sequencing was CS1 (table 1 ; Fig. 2). Sequences were analysed by the Factura feature identification software and aligned in Sequence Navigator (Perkin Elmer Applied Biosystems).

Experiment 2 - Detection of Drug Resistance

The three new RVA constructs were used for co-transfection experiments to ensure that they would allow for detection of drug resistant virus in a sample by production of viruses with drug sensitivity phenotypes consistent with the input PCR products and the regions deleted in the plasmids. Co-electroporations were performed with PCR products derived from an amprenavir-resistant mutant with three Pro mutations, M46I, I47V and I50V, created by site-directed mutagenesis (see below); a lamivudine-resistant mutant with a single M184V mutation in RT created by site-directed mutagenesis (Tisdale et al., 1993); and HXB2 WT virus. The effects of including or excluding CS mutations from two cloned CS mutants (see below) was also investigated. These experiments were analogous to the RVA of plasma virus, but with the crucial difference that the templates used to obtain the PCR products were clonal rather than the heterogeneous populations of viruses found in clinical specimens, thus the possibility of fitter viruses in a mixed plasma virus population out-growing the mutants was negated.

One of the cloned CS mutants was from a subject designated Subject A, who had received indinavir therapy but viral load data had indicated therapy failure. Sequencing of plasma virus from Subject A revealed an A to V mutation at the P2 position (AP2V) of the p7/p1 CS, a mutation observed previously in subjects receiving indinavir therapy (Zhang et al., 1997). Clone A1 had the AP2V p7/p1 CS mutation and also 115V, I54V, R57K, I62V, L63P, H69Y, A71T, I72E, V82A and I85V differences from the consensus subtype B Pro sequence. Clones were also isolated from a subject, Subject B, who had failed amprenavir therapy and acquired an L to F mutation at the P1' position (LP1'F) of the p1/p6 CS, which is a mutation also observed after in vitro selection of resistance with the protease inhibitors BILA 1906 BS or BILA 2185 BS (Doyon et al., 1996), during indinavir or saquinavir therapy (Zhang et al., 1997; Mammano et al., 1998) and with ABT- 378 in vitro (Carrillo et al., 1998). Clone B1 had 115V, E34G, M36I, S37E, I50V and L63P amino acid differences from the consensus subtype B Pro sequence, in addition to the LP1'F p1/p6 CS mutation.

The amprenavir-resistant M46I/I47V/I50V Pro mutant was created by mutagenesis of the M13 clone mpRT1/H (Larder et al. 1989) with a single synthetic oligonucleotide as described (Zoller et al., 1982; Kunkel, 1985), followed by co-electroporation into MT4 cells with an RT-deleted cloned HXB2 provirus.

Infected MT4 cell DNA was used as the template to generate PCR products of the amprenavir-resistant M46I/I47V/I50V Pro mutant, the lamivudine-resistant M184V RT mutant and WT HXB2. Cellular DNA was purified from infected MT4 cell pellets by incubating for 16 hours at 37° in 25 mM Tris pH 7.5, 5 mM EDTA,

0.5% (w/v) SDS and 50 μg/ml proteinase K, followed by phenol/chloroform extraction, chloroform extraction and ethanol precipitation. Plasma viral RNA was prepared using the Roche Amplicor HIV-1 Monitor test kit, according to the manufacturer's instructions (Mulder et al., 1994).

Primers used to generate RT-PCR products for cloning were RVA5' and RVA3' for the first round, and CS1 and dRTC3' for the nested reaction (table 1 and Fig 2B). Products were cloned using the TOPO TA cloning kit (Invitrogen). PCR products that cover the CSs and Pro gene, for co-transfection with pHXBΔCSPro, were generated with primers CS2 and CP2 (table land Fig 2B).

RVAs using pHXBΔPro and a PCR product from clones A1 or B1 which covered the Pro gene but not the CSs (generated with -CS [outer] and CP2) were also carried out to create viruses with identical Pro mutations to that derived from pHXBΔCSPro but without the CS mutations. Primers for the generation of PCR products for RVA with pHXBΔCSPRTA were CS2 and dRTA3'. For pHXBΔCSPRTC, the primers used were CS2 and IN3'.

Table 1.

RT-PCR of RNA from plasma was carried out as follows: The first round PCR consisted of two layers of reagents in a single tube, separated by a wax layer (Ampliwax PCR Gem 100, Perkin Elmer). The upper layer contained the reagents for reverse transcription in a total volume of 50 μl and consisted of 50 mM Tris-HCI (pH 8.3), 75 mM KCI, 3 mM MgCI₂, 10 mM DTT, 5% (v/v) DMSO, 500 μM each dNTP, 20 μg/ml BSA, 250 ng 3^' primer, 40 u Rnasin ribonuclease inhibitor (Promega), 200 u Superscript II RT (Gibco BRL) and 25 μl RNA template. The lower layer contained (in a total volume of 50 μl) 1 mM Tris (pH 8.0), 0.1 mM EDTA, 5% (v/v) DMSO, 250 ng 5' primer and 2.5 u AmpliTaq DNA Polymerase (Perkin Elmer). Second round PCR reactions and amplifications from infected MT4-cell DNA or plasmid clones consisted of (in a total volume of 100 μl) 20 mM Tris (pH 8.8), 25 mM KCI, 1.5 mM MgCI₂₎ 5% (v/v) DMSO, 200 μM each dNTP, 250 ng of each primer and 5 μl template DNA. Thermal cycling was carried out in a Perkin Elmer GeneAmp PCR system 9600 with the following cycles: 45° for 45 minutes (first round only); 95° for 20 seconds, 55° for 10 seconds, 72° for 60 seconds (5 cycles); 90° for 10 seconds, 55° for 10 seconds, 72° for 60 seconds + 5 extra seconds for each consecutive cycle (30 cycles).

Before transfection, the RVA plasmids were linearised by digestion with a restriction enzyme that cut at the site of the deletion. Plasmids pHXBΔPro, pHXBΔCSPro and pHXBΔCSPRTC were cut with BstEW, pHXBΔCSPRTA was cut with >Apal. Transfection of MT4-cells was based on the method described by Kellam and Larder (1994). Briefly, 10 μg of linearised plasmid was electroporated into MT4 cells with approximately 5 μg of the PCR product, which had been purified with the QIAquick Spin PCR Purification Kit (Qiagen). Cultures were fed every 2 or 3 days and harvested when extensive cytopathic effect (CPE) was apparent, after 11-14 days. Drug susceptibility was determined by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT)-reduction colourimetric cell viability assay described by Pauwels et al. (1988). Two-fold serial dilutions of drug were made in 96-well tissue culture plates and 4 x 10⁴ MT4 cells infected with 4 TCID50 units of virus were added to each well. After incubation at 37° for 5 days, MTT was added and the amount of conversion to the blue formazan product after incubation at 37° for 2 hours was assessed by reading the optical absorbance at 540nm. The concentration of drug required to increase the absorbency to levels 50% of those in uninfected control cells (IC50) was calculated by regression analysis of a plot of absorbency against drug concentration. Resistance was expressed as the fold-increase in IC50 compared to a wild type HXB2 control (fold resistance, FR). Plasmids pHXBΔCSPR and pHXBΔCSPRTA produced APV-resistant virus when co-transfected with PCR products derived from the M46I/I47V/I50V PR mutant (25 FR); whilst virus derived from the M184V RT mutant showed APV sensitivity comparable to HXB2. The constructs therefore appear to produce viruses with a PR phenotype expected from the input PCR product.

The RT phenotype of viruses produced in the RVA were also consistent with the properties of the input PCR products and the region of deletion in the plasmids, as demonstrated by the acquisition of lamivudine resistance when pHXBΔCSPRTA or pHXBΔCSPRTC (>18 FR), but not pHXBΔCSPR was co- transfected with PCR products from the M184V RT mutant.

It was unclear whether the inclusion or exclusion of CS mutations would have any effects on the detection of phenotypic resistance, either during growth in the RVA, or in the susceptibility assay itself, which was a 5-day MT4 cell-killing assay (see above). Therefore the effects of including or excluding the CS in the RVA on the determination of drug susceptibility was investigated using cloned RT-PCR products derived from Pl-resistant plasma viruses containing CS mutations (Table 2). These experiments were analogous to the RVA of plasma virus, with the crucial difference that the templates used to generate the PCR products used in the RVA were clones, thus the possibility of fitter viruses in a mixed plasma virus population out-growing the mutants and giving false data was negated. One of the cloned CS mutants was from a subject designated Subject A, who had received indinavir (IDV) therapy but viral load data had indicated therapy failure (previous therapy had also included zidovudine, lamivudine and stavudine). Sequencing of plasma virus from Subject A revealed an A to V mutation at the P₂ position (AP V) of the p7/p1 CS, a mutation observed previously in subjects receiving IDV therapy (Zhang et al., 1997). Clone A1 had the AP₂V p7/p1 CS mutation and also 115V, I54V, R57K, I62V, L63P, H69Y, A71T, I72E, V82A and I85V differences from the consensus

Subtype B PR sequence. Clones were isolated from a second subject, Subject B, who had failed amprenavir (APV; VX-478; 141W94) therapy (therapy also included zidovudine and lamivudine). Plasma virus from Subject B had acquired an L to F mutation at the P-T position (LP-i'F) of the p1/p6 CS, which is a mutation also observed after in vitro selection of resistance with the protease inhibitors BILA 1906 BS or BILA 2185 BS (Doyon et al., 1996), during IDV or saquinavir therapy (Zhang et al., 1997; Mammano et al., 1998) and with ABT-378 in vitro (Carrillo et al., 1998). Clone B1 had 115V, E34G, M36I, S37E, I50V and L63P amino acid differences from the consensus Subtype B PR sequence, in addition to the LPi'F p1/p6 CS mutation.

Constructs pHXBΔPR and pHXBΔCSPR produced viruses with similar IDV susceptibility (8.1 and 8.6 FR respectively) with PCR products derived from clone A1 (Table 2). Similar results were seen with viruses derived from clone B1 when they were assessed for APV resistance with the two constructs (2.5 and 2.6 FR respectively). Thus the presence or absence of the CS mutations did not significantly affect the drug susceptibility phenotype of RVA products generated from a homogeneous template. This is consistent with the original data of Doyon et al. (1996), which suggested an effect of CS mutations on viral fitness rather than directly on drug susceptibility. Interestingly, virus B1 appeared to have increased susceptibility to IDV (>2.4 fold increased susceptibility in both recombinants, the actual values were out of the range of this experiment). Small increases in cross-susceptibility of APV-resistant viruses with other Pis have been documented previously (Tisdale, 1996).

Table 2.

Experiment 3 - Competitive RVAs of Pl-resistant viruses with WT.

When the CSs are not included in the PCR product used for the RVA, one would predict that when CS mutations were present in the source plasma virus population, growth of the resistant viruses may be impaired and viruses with fewer mutations and hence less resistance could have a growth advantage. Selection of less resistant viruses in the RVA as a result of a growth advantage may lead to an inaccurate determination of resistance and cross resistance to Pis in subsequent analyses, possibly impacting decisions made about therapy regimens. The selection process that might occur in the RVA during the growth of a heterogeneous population of recombinant viruses derived from plasma was simulated by mixing known proportions of input PCR products from a CS mutation harbouring Pl-resistant virus with those from WT virus (Fig. 3). Molecular clones were used as PCR templates to enable the ratio of mutant : WT to be precisely controlled.

Molecular clones of the CSs and Pro gene were made from the plasma virus of 2 subjects infected with viruses containing CS mutations. The Pro gene, or CSs and Pro gene were amplified from the clones by PCR and mixed with WT products in ratios of mutant to WT of 4:1 or 9:1. The Pro-only products were co- electroporated into MT4 cells with pHXBΔPro and the CS+Pro products were co- electroporated with pHXBΔCSPro. Relative proportions of mutant and WT in the resulting recombinant virus populations were determined by assessing their susceptibility to indinavir or amprenavir and sequencing cloned PCR products derived from the infected cell pellets.

PCR products that covered the CSs and Pro gene were generated from clone A1 using primers CS2 and CP2. Products covering only the Pro gene were generated with primers -CS (outer) and CP2. Corresponding WT products were generated using plasmid pHIVΔBs-EII (Kellam and Larder, 1994) as the template

DNA. The concentration of the purified PCR products were determined from the absorbency at 260 nm or by visualising in agarose gels, and the products from clone A1 were mixed with WT in ratios of 4:1 or 9:1 (mutant:WT). Approximately 5μg of the mixtures were coelectroporated into MT4 cells with 5μg of linearised pHXBΔCSPro (for products which included the CSs) or pHXBΔPro (products which did not include the CSs). Co-electroporations of A1 PCR DNA alone with the RVA constructs were also performed to generate A1 -derived viruses with or without the CS mutation. Extensive viral CPE occurred over an 11-19 day period, after which the susceptibilities to indinavir of the virus in the tissue culture supematants were determined. In addition, DNA was extracted from the cell pellets and used as templates to generate viral PCR products with primers

RVA5' and Comb3. The PCR products were cloned and 11 or 12 clones from each RVA were sequenced. The resulting susceptibility and sequence data are presented in table 3.

As noted before, the virus reconstructed from clone A1 without the CS mutation (A1-CS) had indinavir sensitivity similar to the reconstructed virus with the CS mutation (A1 +CS) when they were generated from a homogeneous PCR product (table 3, experiment A). When A1-CS PCR products were mixed with WT-CS products in ratios of 4:1 or 9:1 , the resulting RVA virus mixtures were only marginally more resistant to IDV than WT virus. Furthermore, all 11 or 12 cloned PCR products derived from the cell DNA were WT, thus WT virus had out-competed A1-CS and given an RVA product unrepresentative of the input PCR products. When A1 +CS PCR products were mixed with WT+CS products in a ratio of mutant to WT of 4:1 , the resulting virus population had drug susceptibility comparable to that of A1+CS. In addition, 11 out of 11 cloned PCR products derived from the infected cell DNA were A1 +CS. Clearly A1 +CS grew in the mixed RVA at least as efficiently as WT virus, therefore we conclude that the CS mutation compensated sufficiently for the loss of fitness in A1-CS to fully restore the detection of resistance. The reconstructed viruses B1 +CS and B1-CS showed 4.2 and 2.4-fold resistance to amprenavir respectively (table 4). Similarly to experiments with clone A1 , competitive RVAs of B1-CS DNA with WT DNA led to the loss of detection of resistance and the predomination of WT products being amplified from the infected cell DNA. The inclusion of the CS mutation in the RVA led to a near total restoration of mutant growth in a mixture with WT of 4:1 , and a partial restoration from an input ratio of 9:1.

Experiment 4 - Virus growth comparisons

The growth rates of the virus clones were compared with that of a WT virus that was reconstructed from pHXBΔCSPro. Duplicate cultures of 1 x 10⁶ MT4 cells were infected with an amount of virus stock containing 10 ng of p24 for each of the PI resistant viruses or with a clonal WT virus derived from pHXBΔCSPR. The infected cells were incubated in 10 ml of growth medium and 0.5 ml aliquots of the supematants were taken at 2, 4, 6 and 8 days post-infection- The p24 concentrations in each aliquot were then determined in order to assess the relative virus growth rates. Concentrations of HIV-1 p24 in tissue culture supematants or virus stocks were determined with the Murex HIV Antigen Mab kit (Abbott Diagnostics). The data is presented in figure 4.

The results were consistent with the effects of CS mutations observed by others and with the competitive RVA experiments (Doyon et al., 1996; Zhang et al., 1997; Carrillo et al., 1998; Mammano et al., 1998). Growth of A1-CS was impaired compared to the WT virus, but the defect was almost completely restored when the CS mutation was included (A1 +CS). Growth of B1-CS was the most severely affected virus; levels of p24 were 86-fold lower than that of WT at day 4. However, inclusion of the mutant CS region only partially improved B1 growth since B1 +CS remained significantly defective compared to WT, p24 levels being 7.4-fold lower than WT at day 4. Thus the reduction in virus fitness induced by resistance mutations can be almost completely restored by compensatory mutations in the PR and CS, but in the case of virus B1 , compensation at these sites was insufficient to fully restore fitness. Table 3.

N

N-t

Table 4

Discussion

We have constructed plasmids containing a cloned HXB2 provirus deleted in all of the three known sites of resistance to current anti-HIV-1 drugs; the RT, Pro and the p7/p1 and p1/p6 Pro CSs. By RVA with PCR products derived from virus in the plasma of subjects failing therapies, we have a rapid method for detecting resistance and cross-resistance with all of the current drugs simultaneously. The new constructs should play a crucial role in subject management by facilitating the phenotypic analysis of drug resistant mutants arising from the complex multiple drug combinations used in HIV-1 therapy.

When the CSs were not included in competitive RVAs of a heterogeneous PCR population from a CS mutation-harbouring Pro mutant and WT, the WT virus out-grew the mutant, disproportionately producing a drug-sensitive virus population and no detectable mutant when 11 or 12 clones were sequenced. This outgrowth of WT virus occurred even when the input mutant PCR products outnumbered the WT by 9 to 1. The loss of viral fitness was restored upon inclusion of the CSs from the indinavir escape mutant A1 , enabling the detection of resistance, and partially restored upon inclusion of the CSs from the amprenavir-resistant viral clone B1. We conclude that the loss of viral fitness associated with inefficient polypeptide cleavage by mutant Pros had an important influence on the RVA, and that in the clones tested inclusion of the p7/p1 or p1/p6 CS mutations in the input PCR products was crucial to minimise the outgrowth of a fitter virus.

Pl-resistance mutations lead to multiple Gag and Gag-Pol cleavage defects (Doyon et al., 1996; Maschera et al., 1996a, b; Mammano et al., 1998; Zennou et al., 1998). Thus the possibility that mutations occur at CSs other than p7/p1 and p1/p6 should be considered when examining resistance to Pro inhibitors in the RVA, since sites outside the input PCR product would select against resistant virus growth. Nevertheless the p7/p1 and p1/p6 sites appear to be the preferential sites of mutation in response to the reduction of viral fitness. No mutations were observed at any other CSs in gag and pol when HIV-1 strain IIIB was exposed to the inhibitor BILA 2185 BS for 58 passages and had developed 8 mutations in the Pro gene, conferring 1500-fold resistance (Doyon et al., 1996). Similar results were seen with BILA 1906 BS. In another study, when all of the CSs in gag and pol from 3 clones obtained from each of 3 different late viral passages with the PI ABT-378 were sequenced, mutations were limited to the p7/p1 and p1/p6 junctions (Carrillo et al., 1998). Substitutions in the MA/CA CS were observed in viruses from subjects receiving ritonavir or saquinavir, and a CA/p2 substitution was observed in a subject receiving ritonavir, indicating that CS mutations can occur outside p7/p1 and p1/p6 (Mammano et al., 1998). The MA/CA and CA/p2 substitutions have not yet been shown to play a role in improving viral fitness, however. The observation that the p7/p1 and p1/p6 CSs are the rate limiting sites of cleavage in Pl-resistant viruses and hence preferentially undergo mutation, might be related to the inefficiency of cleavage at these sites in WT viruses (Darke et al., 1988; Tozser et al., 1991 ; Wonrak et al., 1993). It is important to note however, that different PR mutants may behave differently, and cleavage may be affected by varying degrees at the different CS. Since multiple Gag and Gag-Pol cleavage defects occur in Pl-resistant viruses it may not be possible to be completely eliminate selection against Pl-resistant viruses can in culture-based assays of susceptibility. However, our new RVA constructs and assays overcome the most important deleterious effects identified so far, and our competitive RVA experiments demonstrate that they significantly improve the detection of resistance in the RVA. References

We thank Matthew Goulden for plasmid pHIVΔRTBsf1107l and Sharon Kemp for the M46I/I47V/I50V Pro mutant.

Carrillo, A., et al. 1998. J. Virol. 72:7532-7541.

Coffin, J. M. 1995. Science 267:483-489. Darke, P. L., et al. 1988. Biochem. Biophys. Res. Commun. 156:297-303.

Doyon, L, et al. 1996. J. Virol. 70:3763-3769.

Goulden, M. G. Unpublished data.

Hertogs, K., et al. 1998. Antimicrob. Agents Chemother. 42:269-276.

Ho, D. D., et al. 1995. Nature 373:123-126. Japour, A. J., et al., 1993. Antimicrob. Agents Chemother. 37:1095-1101.

Kellam, P., and B. A. Larder. 1994. J. Virol. 38:23-30.

Kunkel, T.A. 1985. Proc. Natl. Acad. Sci. USA 82:488-492.

Kusumi, K., et al. 1992. J. Virol. 66:875-885.

Larder, B. A., et al. 1989. Proc. Natl. Acad. Sci. USA 86:4803-4807. Mammano, F., et al. 1998. J. Virol. 72:7632-7637.

Mansky, L. M., and H. M. Temin. 1995. J. Virol. 69:5087-5094.

Maschera, B et al. 1996. J. Biol. Chem. 271 :33231-33235.

Maschera, B., et al. 1996. abstr., p. 85. In, Programme and Abstracts of the 5 th International Workshop on HIV Drug Resistance 1996.

Maschera, B., et al. 1995. J. Virol. 69:5431-5436.

Mayers, D. L., et al. 1998. abstr. 62, p. 42. In Programme and Abstracts of the 2^nd International Workshop on HIV Drug Resistance and Treatment Strategies 1998. Mulder, J., et al. 1994. Journal of Clinical Microbiology 32:292-300. Pauwels, R., et al. 1988. Journal of Virological Methods 20:309-321.

Tisdale, M., et al. 1993. Proc. Natl. Acad. Sci. USA 90:5653-5656. Tisdale, M. 1996. HIV protease inhibitors - resistance issues. Antiviral news 4:41-43.

Tozser, J., et al. 1991. FEBS Lett. 281 :77-80. Vandamme, A.-M., et al. 1998. Antiviral Chemistry and Chemotherapy. 9:187- 203.

Wei, X., et al. 1995. Nature 373:117-122. Wondrak, E. M._f et al. 1993. FEBS Lett. 333:21-24.

Zennou, V., et al. 1998. J. Virol. 72:3300-3306.

Zhang, Y. -M., et al. 1997. J. Virol. 71 :6662-6670.

Zoller M. J., and M. Smith. 1982. Nucl. Acids Res. 10:6487-6500.

Claims

Claims:

1. An assay for the detection of virus resistant to an anti-viral drug by recombination of a DNA vector with a nucleotide sequence derived from a sample of virus suspected of including drug-resistant virus to produce viable recombinant virus having the resistance profile of the virus sample, wherein the DNA vector comprises a wild type laboratory virus strain genome carrying a deletion of at least a portion of the sequence encoding the antiviral drug target and a further deletion of sequences comprising a site of compensatory mutation and the nucleotide sequence derived from the sample of virus comprises a sequence substantially corresponding to the sequences deleted from the DNA vector.

2. An assay according to claim 1 wherein the drug target is a viral protease, polymerase or reverse transcriptase enzyme.

3. An assay according to claim 1 or claim 2 wherein the virus is HIV-1.

4. An assay according to any preceding claim wherein the DNA vector carries a deletion of at least a fragment of the sequence encoding the viral protease and a further deletion of a sequence encoding one or more protease cleavage sites.

5. An assay according to claim 3 and claim 4 wherein the DNA vector is derived from wild type HIV-1 provirus and carries a deletion of the sequence encoding the HIV-1 p7/p1 protease cleavage site.

6. An assay according to claims 3 and 4 or claim 5 wherein the DNA vector is derived from wild type HIV-1 provirus and carries a deletion of the sequence encoding the HIV-1 p1/p6 protease cleavage site.

7. An assay according to any preceding claim wherein the DNA vector carries a deletion of at least a fragment of the sequence encoding the viral polymerase.

8. An assay according to claim 7 wherein the vector carries a deletion of at least a fragment of one or more sequences encoding a polymerase accessory protein.

9. An assay according to claim 7 or claim 8 wherein the vector carries a deletion of at least a fragment of one or more viral polymerase binding or activation sites.

10. An assay according to any preceding claim wherein the DNA vector carries a deletion of at least a fragment of the sequence encoding the viral reverse transcriptase.

11. An assay according to claim 10 wherein the vector carries a deletion of at least a fragment of one or more sequences encoding reverse transcriptase accessory proteins.

12. An assay according to claim 10 or claim 11 wherein the vector carries a deletion of at least a fragment of one or more reverse transcriptase binding or activation sequences.

13. An assay according to any preceding claim wherein the DNA vector carries a deletion of one or both of the HIV-1 p7/p1 and p1/p6 protease cleavage sites, of the entire sequence encoding the viral protease and of the entire sequence encoding the viral reverse transcriptase.

14. An assay for the detection in a viral sample of virus resistant to an anti-viral drug comprising the steps of; (i) generating recombinant virus having the drug resistance profile of the viral sample by recombination of a DNA vector with a nucleotide sequence derived from the sample, wherein the DNA vector comprises a wild-type laboratory virus strain genome carrying a deletion of at least a portion of the sequence encoding the antiviral drug target and a further deletion of sequences comprising a potential site of compensatory mutation;

(ii) infecting cells with recombinant virus so produced; (iii) incubating the infected cells with an antiviral drug; and

(iv) detecting the viability of virus or of the infected cells after incubation, to determine the sensitivity of the recombinant virus to the drug.

15. A DNA vector for use in an assay for the detection of virus resistant to an anti-viral drug by recombination of the vector with a nucleotide sequence derived from a sample of virus suspected of including drug-resistant virus to produce viable recombinant virus having the resistance profile of the virus sample, characterised in that the vector comprises the genome of a wild type laboratory virus strain carrying a deletion of at least a fragment of the sequence encoding the antiviral drug target and a further deletion of a sequence comprising a site of compensatory mutation.

16. A DNA vector according to claim 15 wherein the drug target is a viral protease, polymerase or reverse transcriptase enzyme.

17. A DNA vector according to claim 15 or claim 16 wherein the virus is HIV-1 and the viral sequence is derived from wild type HIV-1 provirus.

18. A DNA vector according to any one of claims 15 to 17 which carries a deletion of at least a fragment of the viral protease and a further deletion of one or more protease cleavage sites.

19. A DNA vector according to claims 17 and 18 which is derived from wild type

HIV-1 provirus and carries a deletion of one or both of the HIV-1 p7/p1 and p1/p6 protease cleavage sites.

20. A DNA vector according to any one of claims 15 to 19 which carries a deletion of at least a fragment of the sequence encoding the viral reverse transcriptase.

21. A DNA vector according to any one of claims 15 to 20 which carries a deletion of one or both of the HIV-1 p7/p1 and p1/p6 protease cleavage sites, of the entire sequence encoding the viral protease and of the entire sequence encoding the viral reverse transcriptase.

22. A DNA vector according to any one of claims 15 to 21 which is pHXBΔCSPro.

23. A DNA vector according to claim 19 which is pHXBΔCSPRTA.

24. A DNA vector according to claim 19 which is pHXBΔCSPRTC.

25. A kit for the performance of an assay according to claim 1 or claim 14 comprising a DNA vector according to any one of claims 15 to 24.

26. The use of a DNA vector according to any one of claims 15 to 24 in an assay for the detection of virus resistant to an anti-viral drug.

27. A DNA vector according to any one of claims 15 to 24 for use in an assay according to claim 1 or claim 14.