Attenuated retroviruses are currently the most widely used vectors in clinical gene therapy because of their potential to effect stable and permanent gene transfer. Since gene delivery is accompanied by random insertion of foreign genetic material into the recipient chromosomal DNA, the potential for insertional mutagenesis exists. In this study, we used a defective retrovirus vector containing a selectable marker, the hygromycin phosphotransferase gene, to investigate the mutagenic effects of vector integration on the mammalian genome. V79 Chinese hamster cells were infected with virus supernatants or by coculture with virus producer cells, and provirus insertion events occurred at low and high frequencies, respectively. The frequency of hprt mutagenesis was increased by a factor of 2.3 over the spontaneous hprt mutation frequency only following multiple provirus insertions/cell genome. Multiple provirus insertions (>3/genome) resulted in instability at the hprt locus in 63% of the virally induced hprt mutants, as indicated by rearrangements at the molecular level, whereas no rearrangements were found when the provirus copy number was 1–2/genome. To demonstrate direct proviral involvement in mutagenesis, the defective MLV vector was retrieved along with flanking genomic hprt sequences from one mutant, and localized within intron 5 of the hprt gene. These data suggest that provirus copy number is a key factor when considering the potential hazards of using retrovirus vectors for gene therapy.
Retrovirus vectors have significant importance in gene therapy because they can deliver therapeutic genes into host chromosomal DNA by integration, leading to permanent gene transfer. An important consideration is that the integrating vector should not lead to adverse effects on the recipient host. Although mutagenesis has been induced experimentally using replication-competent retroviruses, little is known about the effects of multiple integrations of attenuated vectors following in vivo or ex vivo gene delivery. Recent evidence reported by Miller et al1 described the integration of an AAV vector into chromosome 19 within gene sequences leading to chromosomal rearrangements at the sites of vector integration. Retroviruses and lentiviruses integrate essentially at random into the mammalian genome suggesting that vectors derived from these viruses may also be able to cause deleterious effects on normal cellular function following integration near or within a variety of important host genes.
Several reports have demonstrated that wild-type retroviruses can cause animal tumours following integration, and tumorigenesis is known to arise by activation of oncogene sequences occurring as a result of promoter/enhancer insertion, and/or tumour suppressor gene (TSG) inactivation.2 Alteration of gene expression was also demonstrated in the Mov 13 mouse, where provirus insertion into the first intron of the collagen α-11 gene was associated with embryonic lethality and the dilute mutation where proviral insertion alters coat colour.3,4 Provirus LTRs can also alter gene expression after integration at some distance from the affected locus5 or homologously recombine producing grossly altered chromosomal structure.6 These events have led to proto-oncogene activation of the c-myc, vin-1,7 pim-18,9 and Sic 110 genes, and also to inactivation of the p53 TSG.11,12 Recently, Li et al13 identified insertional oncogenesis induced by a defective retrovirus vector resulting from a combination of the activation of a proto-oncogene with signal interference induced by a truncated cell-surface receptor gene carried on the vector.
Although the risk of inactivating both alleles of the same TSG by an integrating vector should in theory be remote, evidence exists that functional inactivation of a single copy of a TSG may in some cases provide growth advantage due to haploinsufficiency and thereby predispose to further neoplastic clonal evolution via additional genetic or epigenetic events.14,15,16,17 Single-copy TSG loci located on the X chromosome may also be of considerable importance in somatic cell growth control, and consequently such genes and genes involved in genetic diseases present on the X chromosome should not be overlooked when considering the risk of insertional mutagenesis by attenuated vectors used in gene therapy protocols.
The frequency of provirus-mediated mutagenesis of the mammalian genome at a single-copy locus (in this case the X-linked hprt gene) by a single provirus insertion has previously been estimated by Goff to occur at a frequency of about one inactivating mutation in 106 virally exposed cells.18 The hprt locus, which is approximately 30 kb in size, contains about 2 kb of exon sequence. In practice, however, a frequency of one in 108 provirus insertions has been shown to promote hprt mutagenesis using replication-competent retroviruses.19 Naturally, this risk would be expected to increase when multiple integrations occur per genome.
V79 cells were used in the present study because, being male in origin, they harbour a single copy of the hprt gene. They have a low frequency of spontaneous hprt mutagenesis and have been used extensively to study the mutagenic effects of chemical and physical agents at this locus. These cells also have a stable karyotype. The hprt model is well suited to study insertional mutagenesis because existing mutants can be purged from culture populations using HAT treatment, and then virally induced mutants can be isolated following selection of hprt-positive cells using 6-thioguanine (6TG). Here, the V79 cell line was used to determine if mutagenesis could be induced by a defective MLV-derived retrovirus vector, LHL (carrying a hygromycin B resistance marker). We show that by infecting V79 cells using high-titre virus supernatants, resulting in only 1–2 provirus integrations per genome, no increase in hprt mutagenesis over background levels can be detected. However, following coculture procedures that produce more than three provirus integrations per cell, we show that an increase in the frequency of mutagenesis could be convincingly demonstrated. Southern analysis of hprt mutants revealed gross alterations in hprt architecture. Moreover, retrieval of proviral sequences with flanking genomic DNA suggested vector involvement in mutagenesis as confirmed by the presence of the attenuated retrovirus directly within the hprt gene.
Spontaneous mutagenesis of V79 and V79E cells
The spontaneous frequency of hprt mutagenesis of V79 and V79E (V79 cells expressing the ecotropic receptor) cells was investigated prior to insertional mutagenesis. To eliminate all pre-existing hprt mutants from the initial cell populations, cells were exposed to HAT medium for 3 days. Based on an expected spontaneous mutation frequency of 10−6, 107, V79 and V79E cells were grown in the presence of 6TG for 2 weeks to kill all hprt-expressing cells. The spontaneous frequency of hprt mutagenesis of both V79 and V79E cells was found to be 2.9 × 10−6 (±0.79, n=6) after scoring methylene blue stained colonies. Six spontaneous hprt-negative clones were isolated for Southern analysis of their respective hprt profile using a Chinese hamster hprt cDNA probe. Since this probe is unable to detect the exon 1 sequence, PCR was used to produce an exon 1 probe to identify rearrangements of exon 1 for each clone. Exon 1 PCR was then optimized using primers designed to span 5.7 kb of the hprt genome sequence following amplification. Positive bands were obtained up to this fragment size across this region in all spontaneous hprt-negative clones.
The band pattern of exons 2–9 of these mutant clones observed by Southern analysis using the cDNA probe was identical to that of wild-type DNA indicating that the spontaneous inactivation is most likely due to point mutations or minor alterations of the hprt locus (data not shown).
Retrovirus insertional mutagenesis of V79E cells
To test the effects of the LHL retrovirus vector on the frequency of insertional mutagenesis, V79 cells were first made permissive to infection by ecotropic MLV pseudotyped LHL by expression of the ecotropic receptor gene. These V79E cells were rendered resistant to hygromycin B following infection with the LHL virus.
To examine the mutagenic effects of LHL provirus infection at low and high copy numbers, supernatant and coculture virus infection, respectively, was applied to V79E cells followed by simultaneous selection in hygromycin B and 6TG. Control mock-infected cells were treated, in parallel, in the same manner with using producer cells not transfected with the LHL retrovirus backbone, and colonies resistant to both selective agents were scored in parallel after methylene blue staining. For V79E cells, the frequency of spontaneous hprt mutagenesis was not found to be significantly different (P=0.42) for either cells mock infected with supernatants (2.9 × 10−6±0.79, n=3) or when cocultured with producer cells (3.3 × 10−6±0.24, n=3).
No increase in hprt mutagenesis was observed following supernatant infection at an MOI of 10. However, following coculture infection, a highly significant 2.3-fold increase (P=0.004) in the number of hprt mutants in infected cells (7.4 × 10−6±0.73, n=3) over the mock infection treated cells (3.3 × 10−6±0.24, n=3) was observed. This suggested provirus involvement in hprt mutagenesis with this infection procedure.
Supernatant infection produces cells with low provirus copy number
Clones of hprt-negative cells also resistant to hygromycin B following infection were isolated and their genomic DNA examined by Southern analysis after HindIII digestion to determine provirus copy number using an hyg probe hybridizing to the hyg gene on the provirus genome. Clones obtained from cells exposed to virus supernatant contained 1–2 proviruses per genome (Figure 1a).
In addition to direct exon 1 PCR to identify the presence of this exon in each of the mutant clones obtained after virus infection, long-distance PCR was used to find provirus integrations in the region surrounding this exon using exon 1 forward and reverse primers in combination with primers designed to amplify outwardly from the LHL provirus. The PCR products obtained were then hybridized with the exon 1 PCR probe and hyg cDNA probe. Exon 1 was found in all Hygr6TGr mutant clones infected by virus supernatants. No positive insertions of the LHL virus in this region were identified for any of these clones (data not shown).
Southern analysis of the hprt profile after HindIII digestion (HindIII is not present in the LHL vector) of genomic DNA for the presence of the remaining exons 2–9 using the hprt cDNA probe for each clone is shown in Figure 1b and illustrated diagrammatically in Figure 1c. No deviations from the wild-type hprt banding patterns were identified in clones subjected to supernatant infection or genomic DNA of V79 cells subjected to mock infection supernatant (data not shown).
Provirus copy number increases following coculture infection
Southern analysis after BamHI digestion (BamHI is absent from the LHL vector) of genomic DNA of coculture-infected mutant clones showed between 1 and 15 provirus insertion events per genome using the hyg probe (Figure 2a). PCR of exon 1 for the presence or absence of this exon for each of these mutants is shown in Table 1. Long-distance PCR using exon 1 forward and reverse primers in combination with LHL outwardly facing primers were hybridized with the exon 1 PCR and hyg cDNA probes and gave no positive provirus insertions in this region for any of the Hygr6TGr clones (data not shown).
The results of Southern analysis of the hprt profile for the presence of the remaining exons 2–9 using the hprt cDNA probe are shown in Figure 2b and illustrated diagrammatically in Figure 2c. In contrast to the normal hprt bands observed following supernatant infection, of the 19 clones examined, 63% of these showed hprt rearrangements from the wild-type profile using BamHI-digested genomic DNA (Figure 2b). Table 1 summarizes the characteristics of each of the coculture-infected clones analysed by Southern blotting and PCR with respect to these rearrangements. A wild-type hprt profile was found in clones 6, 8, 11, 15, 16, 17 and 18. Rearrangements found in the remaining clones include loss of exons 2–9 but exon 1 present (clones 1 and 19), loss of exons 4–9 but exon 1 present (clone 3), loss of exons 4–9 and 1 (clones 5 and 14) and loss of exon 1 only (clones 7 and 13). A 4.5 kb band of unknown origin was observed for clones 2, 4, 9, 10 and 12. Clone 9 also appeared to show an increased fragment size representing exons 4–9 and was subjected to further investigation at the molecular level. No deviations from the wild-type hprt banding patterns were identified in clones subjected to mock coculture infection (data not shown).
Retrieval of the LHL provirus from the hprt locus
Since clone 9 appeared to show an increase in fragment size from exons 4 to 9 (Figure 2b), using Southern analysis this clone was analysed in more detail. A unique PstI restriction site within the provirus situated 1.1 kb downstream to the 5′ viral LTR was used to substantiate the presence of an integrated LHL provirus. Instead of a normal 8.1 kb fragment for PstI-digested DNA, a replacement 7 kb fragment was observed suggesting an introduction of a new PstI site between exons 6 and 8 (Figure 3b, also shown diagrammatically in Figure 3a). The position of the provirus 5′ to exons 6–9 was determined by a wild-type BglII digestion for this region (Figure 3a and c). The provirus and flanking genomic DNA within clone 9 was retrieved using PCR primers designed to amplify DNA outwardly from the provirus LTR and to hprt exon 6. The 1 kb amplification product was cloned into a PCR II cloning vector and sequence analysis from the 5′ end of this product showed 105 bp of provirus DNA including the integration site and the LHL primer and 67 bp of the hprt intron sequence.20 From the 3′ end of the 1 kb fragment, a 215 bp sequence contained 109 bp of the hprt exon 6 sequence and 106 bp of hprt intron sequence including the exon 6 forward primer used in the amplification procedure. PCR between the hygromycin gene of LHL and hprt exon 6 also amplified a product of 2.6 kb in size as expected (data not shown).
Retroviruses are ideal vectors for gene therapy because they can integrate into host DNA for permanent gene transfer. Integrative vectors, however, carry a risk of insertion mutagenesis. Wild-type retroviruses have been previously implicated in mutagenesis following observations that chromosomal rearrangements occur after recombination between integrated viral elements, activation of genes by promoter or enhancer insertion or by direct insertion within gene sequences interrupting gene expression.2,5,19,21,22 Rearrangements have also been demonstrated between AAV virus vector sequences at the sites of vector integration1 although the mechanism leading to such rearrangements may not be the same for different vector types.
Unfortunately, none of the retrovirus vectors presently available can avoid proviral involvement in chromosomal rearrangements or inactivation of essential cellular genes by direct insertion. Mutagenesis caused by integration events would obviously be more likely where multiple integrations have occurred. This study was prompted by our hypothesis that attenuated viruses may also cause insertional mutagenesis and that the existence of several integrated vectors may be an important factor that increases the risk of rearrangements or inactivation of single-copy genes.
V79 cells used in this study are an established male cell line and have been used previously in hprt mutagenesis.23,24,25 The risk of insertional mutagenesis has previously been estimated to be in the region of 10−6 for a haploid target locus of approximately 30 kb in size with about 2 kb of exon sequence, such as the hprt gene. In practice, however, using replication-competent viruses only one in 108 provirus insertions caused mutagenesis in this region.19 Evaluating the risk of insertional mutagenesis using immortal cell lines must be treated with caution, however. For example several immortal cell lines have been shown to carry mutant p53, which has been implicated in many chromosomal abberations in tumours.26,27,28,29,30 Also, several other mutations of genetic loci associated with chromosomal rearrangements may exist. However, although V79 cells are also p53 negative, these cells have a stable karyotype. As an alternative, primary cells may be considered useful to study insertional mutagenesis but they have a restricted proliferative capacity, low infectability and reduced cloning efficiencies needed to select mutant cells.
Provirus-induced mutagenesis which contributed to genetic events leading to cellular immortilization and progression to malignancy has recently been shown in an animal model where a single-copy retrovirus vector insertion has been implicated in myeloid leukaemia as a result of a combination of activation of a transcription factor, Evi-1, and the expression of the transgene product, a truncated nerve growth factor receptor, from the defective retrovirus vector.13 In addition, most recently a leukaemia-like condition has been found in two SCID patients treated with a retrovirus vector where insertion of the vector has been identified in the LMO-2 gene known to be involved in childhood cancers.31 This has resulted in the suspension of SCID trials in France, the UK and the United States.32
Our hypothesis to investigate the effects of an attenuated MLV-based retrovirus vector on mutagenesis of the mammalian genome was tested on the hprt locus because the hprt model provides positive and negative selection using HAT and 6TG, respectively. The possible cause of the relatively small increase in hprt mutagenesis to 7.4 × 10−6 (2.3-fold above the spontaneous mutation frequency) observed following the cocultivation infection procedure became more apparent after Southern analysis revealed between 1 and 19 provirus insertion events and rearrangements in the hprt gene in 63% of the mutants examined. Chromosomal rearrangements have also been observed using replication competent retroviruses by King et al.19 Between 5 and 50 provirus insertions were identified per mutant and the frequency of mutagenesis was 10−8 at the hprt locus in ES cells. Interestingly, rearrangements of the hprt gene in V79 cells used in this study were only observed in mutants harbouring three or more provirus insertions, and not all mutants with high-copy provirus insertions had altered hprt banding patterns. This suggests that loss of functional hprt expression in these mutants may have been due to point mutations, which is presumably also the cause of mutagenesis in low copy number infected mutants or spontaneous hprt-negative mutants which also showed normal hprt-gene profiles. Mock-infected coculture procedures did not yield a significant increase in the level of hprt-negative V79 mutants over the spontaneous level and no rearrangements of the hprt fragment profile were observed.
No increase in the hprt mutation frequency was observed also for mutants where the virus copy number was between 1 and 2 following supernatant infection. Interestingly, Grosovsky et al33 showed a frequency of mutagenesis in a lymphoblastoid cell line to be 2 × 10−5 at the heterozygous thymidine kinase locus (five-fold above the spontaneous mutation frequency) with only 1–2 provirus insertions per cell. These findings, in addition to those described in this study, show obvious differences in the frequencies of mutagenesis at different loci and for different cell lines, which makes accurate quantification of the risk of mutagenesis by provirus insertion difficult.
The presence of the LHL provirus in intron 5 of the hprt gene demonstrates that vector insertion could have been the reason for the hprt-negative phenotype observed in the clone used to retrieve the provirus and flanking genomic DNA. Although hprt insertion was the probable cause of mutagenesis, a more detailed analysis is required at the molecular level to determine the actual cause of 6TG resistance of this mutant. Since the provirus appeared to have inserted into an intron sequence, the presence of both splice donor and acceptor sites in the vector could have led to alternative splicing and loss of hprt expression as has been previously shown for the dilute mutation.4 Alternatively, loss of gene expression may have been caused by viral transcription in the opposite orientation to transcription of the hprt gene by the LHL virus, which has also been shown previously to interfere with c-myc oncogene expression,22 or for the remaining mutants not examined here, provirus insertion(s) within putative hprt promoter/enhancer regions outside the hprt locus may have occurred. Several novel DNA fragments also identified by Southern analysis in coculture-infected mutants using the hprt cDNA probe may have been the result of recombination events between integrated LHL proviruses. Interestingly, a 4.5 kb fragment observed in clone 9 was also seen in clones 2, 4, 10 and 12, and may represent remnants of virus elements such as solo LTR sequences, known to remain at sites of virus recombination, and are currently under investigation.
The coculture infection procedure has been used previously as a means of achieving optimal gene delivery by retrovirus vectors.34 This is mainly to enable efficient ex vivo infection by virus vectors of cells removed from the host which are then returned. Since no selection procedures are used following this mode of infection to identify mutagenized cells, our data suggest that chromosomal rearrangements may occur and this may lead to cellular abnormalities, immortality or even later progression to the malignant phenotype. Although supernatant infection appears to be the preferred method for gene transfer, multiple infection protocols may also result in high provirus copy numbers in infected cells. Since consistently low provirus copy numbers appeared, albeit in the relatively few mutants examined in this study, following a single supernatant infection procedure without hprt rearrangements or induction of mutagenesis, our findings suggest that this mode of infection by retrovirus vectors can be used with some degree of confidence in gene therapy protocols. We therefore suggest that keeping provirus insertion at about 1–2 copies per host genome and increasing transgene expression to physiological levels from single random integration sites or locus targeted integration will be needed to circumvent the potential risk of insertional mutagenesis in gene therapy by integrating vectors.
Materials and methods
Infection of V79 cells by LHL ecotropic viruses
To test the effects of the LHL retrovirus vector on the frequency of insertional mutagenesis, V79 cells were first made permissive to infection by ecotropic MLV pseudotyped LHL. The LHL vector carries the hygromycin B resistance marker driven by the Moloney murine leukaemia virus LTR and contains normal splice donor and acceptors.35 The ecotropic receptor gene on the vector pJET (kindly supplied by JM Cunningham) was used to construct a vector for high-level expression of the ecotropic receptor with a selectable marker gene. This vector, pMETbsr, was generated using the plasmids pMSG and pSV2bsr, which contain the MMTVLTR promoter and a blasticidin selectable marker respectively. Restriction sites SalI and SmaI on pMSG and on pJET allowed the 2.3 kb ecotropic receptor to be inserted into pMSG producing pMETgpt.
The plasmid pSV2bsr (kindly supplied by F Hanaoka of the Riken Institute, Japan) was digested with PvuII and EcoRI releasing the bsr gene, the SV40 early promoter, the SV40 small t intron and the SV40 polyadenylation signal sequence (bsr cassette) on a 2.5 kb fragment. This was purified from agarose and ligated into EcoRI/SmaI-restricted pUC19 DNA to produce the vector pUC 19bsr. This vector was restricted with BamHI, and the 1.8 kb bsr cassette was ligated into pMETgpt after removal of the gpt gene from this vector by BamHI digestion. The resulting expression vector was named pMETbsr.
V79 cells were transfected with the pMETbsr plasmid expressing the ecotropic receptor gene (Materials and methods) to produce V79E cells. Infection by LHL virus produced 20% transduction of these cells using an MOI of 10. To increase this level of infection, cells were treated with tunicamycin which has been previously shown to overcome the refractory nature of Chinese hamster ovary cells by inhibition of N-linked glycosylation.36 This resulted in transduction levels to 70 and 90% for untransfected and V79E cells, respectively. To reach levels of infection to almost 100%, V79E cells were infected with virus complexed with the polycation DEAE dextran instead of polybrene at an MOI of 10. This level of infection was also achieved by the cocultivation procedure with LHL producer cells in the absence of DEAE dextran. Following infection cells were selected by growth in hygromycin B.
Cell culture and generation of retrovirus producer cell lines
All cell lines were grown in Dulbecco's modified Eagle's medium (DMEM, Gibco BRL) supplemented with penicillin and streptomycin antibiotics, 10% foetal calf serum (FCS, Gibco BRL) and 2 mM glutamine at 37°C in an atmosphere of 10% CO2.
GPE+86 retrovirus producer cells were used to generate high-titre ecotropic virus stocks, respectively, as previously described37. Briefly, producer cell lines were transfected using LipofectamineTM reagent (Gibco, BRL) according to the manufacturer's instructions with 3 μg of pLHL retrovirus plasmid DNA. Culture medium was filtered using 0.45 mm filters (Whatmann) and used to infect nontransfected GPE+86 and PA317 cells in the presence of polybrene at 4 μg/ml to generate stable pLHL-producing cell lines.
Clones of retrovirus-producing cells were isolated in the presence of hygromycin B (Calbiochem) at 455 U/ml. Retrovirus stocks were titred and examined for the presence of helper virus against NIH3T3 cells as previously described37. No helper virus contamination was observed from any of the LHL-producing clones.
V79 cells were used for insertional mutagenesis experiments. These cells are HAT resistant and 6TG sensitive. Cells were positively selected (to remove hprt-negative cells) prior to insertional mutagenesis by seeding 5 × 104 cells per 90 mm dish into HAT medium for 72 h. V79 cells transfected with the vector pMETbsr were also treated in the same manner.
Retrovirus insertional mutagnesis of V79 cells
Cells were infected using virus supernatants harvested from GPE+86 LHL producer cells or by coculturing V79 cells with GPE+86 LHL producers. HAT-treated V79 cells were infected for 24 h with filtered virus supernatants removed from confluent producer cell monolayers. Enhanced infection was achieved by two means: by pretreating V79 cells in a medium containing tunicamycin for 18 h at 37°C and 0.01 M dexamethazone for 24 h to induce ecotropic receptor gene expression by pMETbsr-transfected cells. Cells were infected in the presence of 4 μg/ml polybrene or 5 μg/ml DEAE dextran. After removal of virus supernatant, cells were allowed a 24 h recovery period before selection in hygromycin B at 400 U/ml and 6TG at 10 μg/ml at clonal density.
Producer cells to be used in coculture infection were treated with 10 μg/ml mitomycin C for 2 h, then washed twice with PBS before trypsinization and centrifugation. Following a further PBS wash and centrifugation, 2 × 106 producer cells were seeded with 1 × 106 V79 cells/100 mm dish and left for 48 h for coculture infection. Mitomycin C-treated producer cells were also seeded alone on 100 mm dishes to ensure total cell death which occurred at around 7 days. After infection, cells were also allowed a 24 h recovery period prior to selection for virus infection and loss of the hprt phenotype. Selection in hygromycin B took between 10 and 14 days for control non-6TG-selected cells and cells dual-selected in hygromycin B and 6TG. To determine the frequency of retrovirus insertional mutagenesis, 1 × 107 V79 cells were infected.
Analysis of provirus copy number and insertion mutants
Southern and PCR analysis was used to determine proviral copy number, and to identify integration at the hprt locus, genomic DNA preparation and Southern analysis of hprt-negative V79 mutants was carried out. Briefly, 10 μg of genomic DNA was digested with the appropriate restriction enzyme fragments separated on agarose gels before transfer to Nylon membranes (Hybond-N™). Probes were prepared using a random primer labelling kit (Mega-prime™ system Amersham) according to the manufacture's instructions, plus 50 mCi of α-32P-CTP (3000 Ci/mM) to generate probes with a specific activity of 109 cpm/mg. Bio-Spin™ 6 column (Bio-rad) was used to remove unincorporated nucleotides. A 750 bp hygromycin gene probe prepared from SacII/EcoRI-digested pLHL was used to identify integrated LHL provirus DNA. An hprt probe was prepared from the pHpt21 vector (a kind gift from M Fox of Cristie) after PstI digestion which contains the hamster hprt cDNA.
A PCR-amplified exon 1 probe sequence was also used to examine the hprt profile of each mutant clone. PCR was also used to amplify hprt exon sequences to determine provirus integration within exon DNA using hprt primers. In addition, PCR primers were designed from LHL retrovirus sequences and used in combination with hprt primers to amplify provirus elements flanking hprt genomic DNA.
PCR primers used in the identification of retrovirus insertion within the hprt gene and for retrieval of genomic DNA are given below. Sequences are listed with the expected product sizes (where known), final primer concentrations and the optimum MgCl2 concentration required are also given. PCR analysis was carried out using a HYBAID OmniGene™ thermocycler with 0.25 U Taq polymerase (Promega) and 0.25 mM dNTPs (Promega).
hprt exon primers:
Exon 1 (size: 321 bp, 1.5 mM MgCl2)
(66.7 nM) Forward GTA CCT GGC CCC AGG AGC CAC C
(66.7 nM) Reverse TCC GCT CTG CTG AAG AGT CCC G
Cycling parameters: 94°C/5 min, 70°C/19 min; one cycle
94°C/min, 60°C/min, 70°C/45 s; 30 cycles
Exon 2 (size: 166 bp, 1.5 mM MgCl2)
(1.50 mM) Forward AGC TTA TGC TCT GAT TTG AAA TCA GCT G
(1.50 mM) Reverse ATT AAG ATC TTA CTT ACC TGT CCA TAA TC
Exon 3 (size: 220 bp, 1.5 mM MgCl2)
(1.33 mM) Forward CCG TGA TTT TAT TTT TGT AGG ACT GAA AG
(1.33 mM) Reverse AAT GAA TTA TAC TTA CAC AGT AGC TCT TC
Exon 4 (size: 191 bp, 1.5 mM MgCl2)
(1.33 mM) Forward GTG TAT TCA AGA ATA TGC ATG TAA ATG ATG
(1.33 mM) Reverse CAA GTG AGT GAT TGA AAG CAC AGT TAC
Exon 5 (size: 247 bp, 1.5 mM MgCl2)
(1.33 mM) Forward AAC ATA TGG GTC AAA TAT TCT TTC TAA TAG
(1.33 mM) Reverse GGC TTA CCT ATA GTA TAC ACT AAG CTG
Exon 6 (size: 145 bp, 1.5 mM MgCl2)
(1.33 mM) Forward TTA CCA CTT ACC ATT AAA TAC CTC TTT TC
(1.33 mM) Reverse CTA CTT TAA AAT GGC ATA CAT ACC TTG C
Exons 7 and 8 (size: 423 bp, 1.5 mM MgCl2)
(1.33 mM) Forward GTA ATA TTT TGT AAT TAA CAG CTT GCT GG
(1.33 mM) Reverse TCA GTC TGG TCA AAT GAC GAG GTG C
Exon 9 (size: 734 bp, 1.5 mM MgCl2)
(1.67 mM) Forward CAA TTC TCT AAT GTT GCT CTT ACC TCT C
(1.67 mM) Reverse CAT GCA GAG TTC TAT AAG AGA CAG TCC
Cycling parameters for hprt exons 2–9:
94°C/5 min, 70°C/13 min; one cycle
94°C/min, 58°C/min, 70°C/45 s; 30 cycles
LHL retrovirus primers:
Hygromycin gene (size: 187 bp, 1.5 mM MgCl2)
(20 pM) Forward GCC TGA CCT ATT GCA TCT CCC G
(20 pM) Reverse GCC ATG TAG TGT ATT GAC CGA TTC C
94°C/15 s, 66°C/15 s, 70°C/15 s; 35 cycles
Outwardly facing primers:
AN (5′ LTR out): (1.5 mM MgCl2)
(2 pM) GAC CTT GAT CTG AAC TTC TC
CN (3′ LTR out): (1.5 mM MgCl2)
(2 pM) AAT ATC ACC AGC TGA AGC C
where primers AN and CN were used in conjunction with hprt primers, the parameters given above for each of the respective hprt exon primers were used.
Generation of exon 1 radioactive probes by PCR
A radiolabelled exon 1 PCR probe was prepared using exon primers as described by Rossiter et al.20 The reaction mix consisted of 5 mM of d(AGT)TPs plus 50 mCi of α-32P-CTP (3000 Ci/mM). A total of 30 cycles of PCR produced radioactive exon 1 probe which was purified using Bio-Spin™ 6 columns prior to hybridization on nylon membranes.
Retrieval and analysis of provirus DNA within hprt intron 5
PCR products were cloned using an Invitrogen kit™ (Stratagene). Positive tranformants identified on ampicillin plates with X-gal solution at 25 mg/ml revealed white colonies for cloning into the PCR II vector. Positive transformants were analysed by endonuclease digestion and hybridization with the hygromycin gene probe. Sequencing of the retrovirus/hprt DNA retrieved DNA was carried out using the Sequenase™ kit (United States Biochemical) using 32P dATP (3000 Ci/mM) label (Amersham International). Acrylamide gels were prepared and analysed as described by Sambrook et al (1989). Autoradiography was carried out using Kodak XAR-2 X-ray film.
Comparative analysis was based on the Studentized range of t at probability (P) values quoted in the results; confidence intervals assumed 1.96 × s.e. (95%) for a normal distribution.
We are grateful to Dr JM Cunningham, Dr F Hanaoka and Dr AD Miller for kindly providing the pJET, pSV2bsr and pLHL plasmids, respectively. We also thank our colleagues Dr A Cuthbert and Dr J Arrand for helpful suggestions and support during this work and Dr B Bigger for statistical analysis. This work was supported by Grants from the Cancer Research UK SP2133/0301) and the European Commission (FIGH-CT1999-00002; QLG1-1999-01341).