Control of Parvovirus DNA Replication by a Tetracycline-Regulated Repressor

Home , Viral replication

Gene Therapy (1999) 6, 309–313  1999 Stockton Press All rights reserved 0969-7128/99 $12.00 http://www.stockton-press.co.uk/gt Control of parvovirus DNA replication by a tetracycline-regulated repressor

IH Maxwell and F Maxwell Department of Dermatology and University of Colorado Cancer Center and Health Sciences Center, Denver, CO, USA

Autonomous parvoviruses are small, single strand DNA construct unexpectedly showed constitutive expression in viruses which preferentially replicate in transformed and transiently transfected cells, as indicated by efficient exci- tumor cells, causing cell death by expression of the cyto- sion and amplification of viral replicative form (RF) DNA. toxic nonstructural protein, NS1. Several parvoviruses of This was apparently due to self-stimulatory transcriptional the rodent group, including LuIII, efficiently infect human transactivation by NS1. This problem was overcome by transformed cell lines. The potential for systemic use of cotransfection with a plasmid expressing a chimera of the these viruses in targeting metastases might be enhanced repressor of the tetracycline operon with a KRAB trans- if NS1 expression and viral replication could be controlled repression domain. These conditions allowed efficient con- by an innocuous drug such as tetracycline. We therefore trol of transcription and RF amplification by the tetracycline substituted prokaryotic tetracycline operator sequences for derivative, doxycycline. These observations form a basis part of P4 of LuIII, the promoter responsible for transcrip- for developing a therapeutic agent based on a drug- tion of the mRNAs for nonstructural proteins. The resulting controlled parvovirus.

Keywords: parvovirus; transcriptional regulation; tetracycline; KRAB repressor

Introduction by an appropriate drug. We therefore modified the P4 promoter of LuIII to try to impose regulation by a drug- Parvoviruses are small, icosahedral viruses containing a controlled mechanism. We chose to test the tetracycline single strand DNA genome of approximately 5 kb which regulatory system, based on the use of an artificial trans- replicates via double strand linear replicative forms (RF) activator consisting of the prokaryotic repressor of the 1 in the nuclei of infected cells. These viruses can be classi- tetracycline operon, fused with the strong transactivation fied as dependoviruses (adeno-associated viruses), domain from the herpes simplex VP16 protein.12 requiring a helper virus for replication, or as auton- Expression of this chimeric transactivator in mammalian 1,2 omous. The latter occur widely, require proliferating cells enables tight control by tetracycline derivatives over cells and replicate efficiently in, and preferentially kill, the transcription of foreign genes from a minimal pro- 3 malignantly transformed cells. Cell killing appears to be moter linked with operator sequences from the tetracyc- due largely to the cytotoxicity of the viral nonstructural line operon.12 In the original version of this regulatory protein, NS1, which shows both enhanced levels of system, the transactivator, tTA, was released from oper- expression in transformed cells and preferential tox- ator DNA by tetracycline so that the regulated gene was 4,5 icity. NS1 has multiple functions, including site-specific transcribed only in the absence of the drug.12 More endonuclease activity, essential for replication of viral recently, a ‘reverse transactivator’ (rtTA) has been con- DNA, and transcriptional activation of the viral structed from a mutagenized repressor that requires 1,6 promoters. tetracycline for operator binding, thus providing positive Several members of the rodent parvovirus group can response to the drug in mammalian cells.13 Here, we efficiently infect human transformed cell lines. These report the construction of a recombinant LuIII genome viruses include parvovirus H1 of rats, the minute virus with partial replacement of the P4 promoter by tetracy- of mice (MVM), and the closely related parvovirus LuIII cline operator and minimal promoter sequences. Unex- (natural host unknown) which we have previously used pectedly, these sequences, when placed in the context of 7–9 to construct recombinant vectors. There have been the parvovirus DNA, showed constitutive activity isolated reports of attempts to use parvovirus H1 for regardless of the presence of tTA or rtTA. We present 10,11 treatment of cancer patients. results showing that this problem could be circumvented The potential for therapeutic use of parvoviruses, by the coexpression of a tetracycline-regulated based on their oncolytic properties, might be enhanced if repressor,14 as opposed to a transactivator, thus enabling NS1 expression and hence replication could be controlled tetracycline control over parvovirus RF amplification.

Results Correspondence: IH Maxwell, Dermatology B153, University of Colorado Health Sciences Center, 4200 E 9th Avenue, Denver, CO 80262, USA We previously described a recombinant of the rodent Received 24 April 1998; accepted 23 September 1998 parvovirus, LuIII, which expressed luciferase, in a TET control of parvovirus DNA replication IH Maxwell et al 310 manner regulated by tetracycline, in tTA expressing cells of the TO promoter when placed adjacent to the LuIII left transduced by the virus.8 This recombinant genome, in terminus (see Figure 1). As shown previously,8 the plasmid pTO-Lu-LUC, is shown in Figure 1. To gen- expression of this reporter was strongly activated by erate a conditionally replication-competent virus which cotransfection with a tTA expression plasmid and this should also be controllable by tetracycline, we introduced activation was abolished by tetracycline (Figure 3). How- the same substitution of the P4 promoter region into a derivative of pGLu883, an infectious clone of the LuIII genome,15 generating pTO-Lu (Figure 1). Transfection of pGLu883 into the permissive human cell line, NB324K, results in excision of the viral genome and abundant replication of linear, double-stranded replicative form DNA (RF).7,15 These processes require expression of the nonstructural protein, NS1, whose mRNA is transcribed from the viral left end promoter, P4.1 We therefore expected that RF excision and amplification in cells transfected with pTO-Lu would be dependent on the presence of tTA, to transactivate NS1 expression from the substituted TO promoter. This transactivation should be abolished by addition of tetracycline. However, as shown in Figure 2, similarly efficient RF amplification was seen in pTO- Lu transfected cells regardless of the presence of tTA or tetracycline. This unexpected, constitutive RF production from pTO-Lu can be explained by the results shown in Figure 3, where pTO-Lu-LUC was used as a reporter of activity Figure 2 Southern analysis of Hirt extracts from NB324K cells, prepared 39 h following electroporation with pTO-Lu (2 ␮g in 0.2 ml) plus the indicated amounts of the tTA expression plasmid, pUHD 15–1,12 with or without exposure of the cells to tetracycline (1 ␮g/ml), as indicated (− or +TET), from immediately following electroporation until harvested. The extracted low molecular mass nucleic acid samples were treated with RNa- seA and with DpnI, and were then electrophoresed in a 1% agarose gel, transferred to nylon, and probed with a digoxigenin-labeled probe for the NS1 region of the viral genome.7 The arrow indicates the position of the monomer RF. Lane M, digoxigenin-labeled marker DNA (Boehringer Mannheim, Indianapolis, IN, USA) with sizes indicated in kilobase pairs on the left.

Figure 1 Structure of recombinant LuIII genomes described in the text. The recombinants were constructed in pUC18 (plasmid sequences represented by flanking thin lines). The parental, infectious plasmid clone of the LuIII genome, pGLu883,15 is shown in the top row, pTO-Lu-LUC, previously described,8 contains the luciferase reporter (replacing viral coding sequences), driven by minimal sequences from the cytomegalovirus Figure 3 Transactivation of luciferase expression from pTO-Lu-LUC by immediate–early promoter (−53 to +75), linked with oligomeric tetra- coexpression of either tTA (abolished by tetracycline) or of NS1 cycline operator sequences.12 pTO-Lu contains the same promoter/operator (tetracycline independent). NB324K cells (1 × 106 cells in 0.1 ml) were sequences and also retains the complete viral coding sequences. Black bars: electroporated with pTO-Lu-LUC (0.5 ␮g) together with either the tTA left and right, non-identical,1 terminal inverted repeat sequences, neces- expression plasmid12 pUHD15–1 (0.5 ␮g), or an expression plasmid for sary for DNA replication and packaging. NS, VPs, coding sequences for LuIII NS1,9 pRSVhel (2.5 ␮g). The electroporated cell suspensions were LuIII non-structural and capsid proteins, respectively; An, LuIII poly- divided in half and the cells were cultured with or without tetracycline adenylation signal; P4, P38, LuIII promoters located at four and 38 map (1 ␮g/ml) for 16 h. Cell extracts were then prepared and appropriately units, responsible for generating mRNAs for nonstructural proteins and diluted samples were assayed for luciferase activity. Values on the ordinate capsid proteins, respectively. The plasmids are represented for convenience are expressed as light units per 5 × 105 cells electroporated, calculated in linear form; all transfections used the covalently closed circular forms. from measured values ranging from 200 to 1300 light units. TET control of parvovirus DNA replication IH Maxwell et al 311 ever, similarly strong transactivation was observed when but had no effect in cells that were exposed to a low con- pTO-Lu-LUC was cotransfected with an expression plas- centration of doxycycline (Figure 4). Furthermore, as also mid for NS1 and this activation was unaffected by the shown in Figure 4, the repressor could be used in combi- presence of tetracycline (Figure 3). These results suggest nation with a tetracycline-dependent transactivator (the that an initially low basal level of expression of NS1 from ‘reverse’ transactivator, rtTA13) to achieve an even higher pTO-Lu may have sufficed to initiate a self-stimulatory level of doxycycline control (150-fold) over trans- loop that resulted in abundant NS1 production and activation of luciferase expression from pTO-Lu-LUC consequent efficient RF amplification (Figure 2). This (Figure 4). interpretation was further supported by the finding We next examined whether the TetR-KRAB repressor that luciferase expression from pTO-Lu-LUC was also would allow similar drug control over RF excision and activated by cotransfected pTO-Lu (Figure 4). amplification from transfected pTO-Lu. The results Since pTO-Lu was thus apparently capable of self- shown in Figure 5 demonstrate that this was the case. activation, we considered the possibility that tetracycline Hirt extracts were prepared at 28 or 50 h after transfec- control might be imposable by using a repression, as tion of NB324K cells with pTO-Lu, with or without opposed to activation, mechanism. We tested this possi- cotransfection with the TetR-KRAB expression plasmid. bility using a tetracycline-controlled transcriptional In the absence of doxycycline, the latter repressed RF pro- repressor (TetR-KRAB) consisting of a fusion of the bac- duction to a low level (Figure 5, tracks 3, 5, 9, 11). terial tetracycline repressor with the KRAB repression Addition of doxycycline resulted in RF amplification domain from Kox1, a human Kru¨ ppel-type zinc finger (tracks 4, 10) to a level similar to that seen in cells trans- protein.14 It was previously shown that this fusion con- fected with pTO-Lu alone (with or without doxycycline) struct could repress transcription from a strong CMV (tracks 1, 2, 7, 8). The additional cotransfection of the promoter, placed adjacent to tetracycline operator rtTA expression plasmid further increased the level of RF sequences and that the repression was relieved by tetra- amplification in the presence of doxycycline (Figure 5, cycline.14 Figure 4 shows results of cotransfecting an tracks 6, 12), an observation consistent with the increased expression plasmid for TetR-KRAB together with pTO- transactivation of reporter expression from pTO-Lu-LUC Lu-LUC and pTO-Lu. TetR-KRAB expression abolished by cotransfection with this plasmid (Figure 4). transactivation by pTO-Lu in the absence of tetracycline Discussion In several of the major human cancers, including breast, prostate and melanoma, successful resection of the pri- mary tumor is possible but patients frequently succumb to intractable metastatic disease. The possibility of thera-

Figure 5 Southern analysis of Hirt extracts from NB324K cells, prepared Figure 4 TetR-KRAB enables tetracycline control of NS1 expression from 28 h (a), or 50 h (b) following electroporation with pTO-Lu (2 ␮g) with pTO-Lu, as measured by transactivation of luciferase expression from or without pCMV-tetR-KRAB14 (4 ␮g) and the rtTA expression plasmid13 pTO-Lu-LUC. NB324K cells (1 × 106 cells in 0.1 ml) were electroporated pUHD172-lneo (2 ␮g). The cells were exposed to doxycycline (1 ␮g/ml), with pTO-Lu-LUC (0.5 ␮g), with or without additional plasmids, as indi- as indicated (− or + Doxy), immediately following electroporation until cated (pTO-Lu, 1 ␮g; the rtTA expression plasmid13 pUHD172-lneo, harvested. Extracts were prepared and analyzed as in Figure 1. The arrow 0.2 ␮g; pCMV-tetR-KRAB, 2 ␮g). The electroporated cell suspensions indicates the position of the monomer RF (the apparently faster migration were divided in half and the cells were cultured with or without doxycy- on the right hand side of B is an artifact of bowing of migration in the cline (1 ␮g/ml) for 16 h. Cell extracts were then prepared and appropri- gel). Dots in the center of the Figure indicate positions of digoxigenin- ately diluted samples were assayed for luciferase activity. Values on the labeled marker DNA (21.2, 5.15/4.97 and 4.27 kilobase pairs). An internal ordinate are expressed as light units per 5 × 105 cells electroporated, calcu- control for equal loading of the gel tracks was provided by DpnI fragments lated from measured values ranging from 500 to 3000 light units. 24× of the input pTO-Lu, migrating in the lower part of the gel (not shown), and 150× refer to fold activation by doxycycline in the presence of TetR- which showed a similar intensity of hybridization with the probe in KRAB, respectively, without and with rtTA in addition. each track. TET control of parvovirus DNA replication IH Maxwell et al 312 peutically exploiting virus infection of metastases, with regions. However, the latter functions might be par- subsequent viral replication and cytopathicity, is an titioned between two mutually dependent viruses, one of attractive idea of long standing.16 Parvoviruses are good which could encode a transactivator or repressor. Such a candidates for this approach because of their small size, parvovirus-based, drug-controlled binary system may preference for replicating, malignantly transformed cells hold promise for therapeutic targeting of metastases. and well-documented oncolytic effects in experimental animals.1,3 These viruses are amenable to genetic manipu- lation as a possible means of targeting and enhancing Materials and methods their potential oncolytic properties in patients. A desir- able goal, in terms of both efficacy and safety, would be Plasmid constructs 20 to achieve controlled, conditional replication of the Standard cloning methods were used to generate the therapeutic virus using an appropriate drug. Our results constructs shown in Figure 1. pTO-Lu-LUC has been 8 demonstrating tetracycline control over amplification of described previously. pTO-Lu was made as follows. The ⌬ 7 parvovirus LuIII RF DNA represent a first step in this left end of the LuIII genome from pGLu883 Xba, direction. The tetracycline system is attractive for this extending to the MluI site at nt 277, was subcloned into approach because it offers tight regulation of gene an intermediate plasmid and a BglII adapter oligonucleo- expression by low concentrations of clinically used tetra- tide was then ligated between the StyI site at nt 144 and cycline derivatives with well-defined pharmacology. the NcoI site at nt 265 (which overlaps the initiation Our initial attempts to impose drug control over repli- codon for the nonstructural proteins). A DNA fragment cation of LuIII were compromised by the transactivation (0.48 kb BglII to BamHI) containing seven copies of the properties of the essential viral nonstructural protein, tetracycline operator, linked with a minimal CMV pro- 12 NS1. This multifunctional protein strongly transactivates moter, was isolated from a derivative of pUHC13–3 and transcription of mRNAs encoding the capsid proteins was ligated into the BglII site of the adapter. The from the viral promoter, P38,1 but has been reported to resulting, modified LuIII left end was then substituted as affect the left-end promoter, P4, only relatively weakly.17 a PstI–MluI fragment for the left end in pGLu883 to gen- We expected that our replacement of most of the P4 pro- erate pTO-Lu. Further details are available on request. 14 moter by tetracycline operator sequences, linked with a pCMV-tetR-KRAB was a gift from Dr U Deuschle, Hoff- minimal promoter, would render transcription depen- man-La Roche Ltd, Basel, Switzerland. To minimize the dent on the tTA transactivator. However, the resulting possible occurrence of deletions, Escherichia coli strain construct, pTO-Lu, did not show this dependence; this SURE (Stratagene, La Jolla, CA, USA) was used for was apparently due to an initially low basal level of NS1 propagation and isolation of plasmids containing the expression which then acted in a self-stimulatory man- LuIII inverted terminal repeats. Other plasmids were ner. Since we observed strong NS1 transactivation of grown in E. coli strain HB101. Plasmids were isolated by luciferase expression from pTO-Lu-LUC which lacked alkaline lysis and were purified by banding in CsCl in 20 the entire viral coding sequences, this apparent self- the presence of ethidium. stimulation by NS1 was probably mediated by the NS1 binding site in the viral left terminus.18 Although this site Cell culture and transfection by electroporation is essential for viral DNA replication, and so could not The SV40-transformed human kidney cell line, NB324K21 be deleted or mutated, it is possible that further manipu- (a gift from Dr P Tattersall, Yale University School of lation of the P4 region (eg of sequence content or spacing Medicine, New Haven, CT, USA) was maintained in between the NS1 binding site and tetracycline operator either OptiMEM (Life Technologies, Gaithersburg, MD, sequences) might enable tTA control of transcription. USA) with 3.8% fetal bovine serum, or in RPMI1640 with We found that tetracycline control over pTO-Lu could either 10% NuSerum IV (Collaborative Biomedical Pro- be imposed by an alternative means, using a chimeric ducts; Becton Dickinson, Bedford, MA, USA) or 7.5% fetal repressor protein, TetR-KRAB, instead of a trans- bovine serum (‘Tet System Approved’; Clontech, Palo activator.14 In the presence of this protein, we observed Alto, CA, USA). efficient RF amplification from transfected pTO-Lu only Electroporation used a capacitance discharge unit (IBI in the presence of doxycycline. In the absence of this Gene Zapper, Eastman-Kodak, New Haven, CT, USA). drug, when TetR-KRAB was functional as a repressor by NB324K cells were grown to 60–70% confluency, before binding to the operator sequences, RF production was being trypsinized and suspended with cold medium. only minimal. Thus, under these conditions the binding Samples of cell suspension (0.8–1.0 × 107 cells/ml) were of TetR-KRAB was evidently able to repress trans- electroporated in complete medium22 with the indicated activation by NS1 bound to the adjacent parvoviral ter- amounts of plasmids, made up with a ‘filler’ plasmid minal DNA sequence. These results provide the first (pUC18 or equivalent, lacking eukaryotic sequences) to demonstration of ability to impose drug regulation on total DNA concentration 50 ␮g/ml. Voltage was set at replication of a modified parvovirus genome. The TetR- 230 volts and the capacitance was set to give a time con- KRAB has recently been used in a similar way to restrict stant of approximately 30 ms (250 ␮F or 500 ␮F, respect- DNA replication of an adenovirus vector.19 ively, for pulsing of 0.1 ml or 0.2 ml cell suspension). Successful application of drug control over virus replication using this type of system requires that the relevant Analysis of replicative form (RF) DNA amplification trans-acting molecule, whether a drug-controlled acti- Excision and amplification of RF were analyzed by Hirt vator or repressor, be present in the infected cell. Packag- extraction of low molecular mass DNA, followed by gel ing size constraints would preclude encoding such a pro- electrophoresis and Southern blotting as described.7 At tein within a replication-competent parvovirus genome the indicated times, the transfected cells were treated retaining both the nonstructural and capsid-coding with DNaseI (to degrade residual plasmid) before Hirt TET control of parvovirus DNA replication IH Maxwell et al 313 extraction. The resulting low molecular mass nucleic acid a transient transducing vector for human cells. Hum Gene Ther samples were treated with RNaseA and with DpnI and 1993; 4: 441–450. were then electrophoresed in a 1% agarose gel, trans- 8 Maxwell IH, Spitzer AL, Long CJ, Maxwell F. Autonomous par- ferred to nylon and hybridized with a digoxigenin- vovirus transduction of a gene under control of tissue-specific or inducible promoters. Gene Therapy 1996; 3: 1175–1179. labeled probe.7 9 Spitzer AL, Maxwell F, Corsini J, Maxwell IH. Species specificity for transduction of cultured cells by a recombinant LuIII rodent Assay of luciferase activity parvovirus genome encapsidated by canine parvovirus or feline Cell extracts were prepared and assayed for luciferase panleukopenia virus. J Gen Virol 1996; 77: 1787–1792. activity using a commercial kit (Luciferase Assay System; 10 Toolan HW et al. H1 virus viremia in the human. Proc Soc Exp Promega, Madison, WI, USA) and a model TD-20/20 Biol Med 1965; 119: 711–715. luminometer (Turner Designs, Sunnyvale, CA, USA). 11 Le Cesne A et al. Intra-lesional administration of a live virus, parvovirus H1 (PVH-1) in cancer patients: a feasibility study. Proc Annu Meet Am Soc Clin Oncol 1993; 12 (Abstr. 970). Acknowledgements 12 Gossen M, Bujard H. Tight control of gene expression in mammalian cells by tetracycline-responsive promoters. Proc Natl This work was supported by an NIH Institutional grant Acad Sci USA 1992; 89: 5547–5551. to the Breast Cancer Research Program of the University 13 Gossen M et al. Transcriptional activation by tetracyclines in of Colorado Cancer Center and by a generous grant from mammalian cells. Science 1995; 268: 1766–1769. the Denver Metropolitan Chapter of the Susan G Komen 14 Deuschle U, Meyer WK, Thiesen HJ. Tetracycline-reversible Breast Cancer Foundation. We thank Drs Ulrich Deuschle silencing of eukaryotic promoters. Mol Cell Biol 1995; 15: and Hermann Bujard (ZMBH, Heidelberg, Germany) for 1907–1914. supplying plasmids. 15 Diffoot N, Chen KC, Bates RC, Lederman M. The complete nucleotide sequence of parvovirus LuIII and localization of a unique sequence possibly responsible for its encapsidation pat- References tern. Virology 1993; 192: 339–345. 16 Sinkovics JG, Horvath J. Can virus therapy of human cancer be 1 Cotmore SF, Tattersall P. The autonomously replicating parvo- improved by apoptosis induction? Med Hypoth 1995; 44: 359–368. viruses of vertebrates. Adv Virus Res 1987; 33: 91–173. 17 Doerig C, Hirt B, Antonietti JP, Beard P. Nonstructural protein 2 Malik P et al. Recombinant adeno-associated virus mediates a of parvoviruses B19 and minute virus of mice controls transcrip- high level of gene transfer but less efficient integration in the tion. J Virol 1990; 64: 387–396. K562 human hematopoietic cell line. J Virol 1997; 71: 1776–1783. 18 Cotmore SF, Christensen J, Nuesch JP, Tattersall P. The NS1 3 Rommelaere J, Tattersall P. Oncosuppression by parvoviruses. polypeptide of the murine parvovirus minute virus of mice In: Tijssen P (ed). Handbook of Parvoviruses. CRC Press: Boca binds to DNA sequences containing the motif [ACCA]2–3. J Ratan, 1990, volume 2, chapter 3. Virol 1995; 69: 1652–1660. 4 Van Hille B et al. Limitations to the expression of parvoviral 19 Rittner K, Schultz H, Pavirani A, Mehtali M. Conditional nonstructural proteins may determine the extent of sensitization repression of the E2 transcription unit in E1-E3-deleted adeno- of EJ-ras-transformed rat cells to minute virus of mice. Virology virus vectors is correlated with a strong reduction in viral DNA 1989; 171: 89–97. replication and late gene expression in vitro. J Virol 1997; 71: 5 Salome N, van Hille B, Geuskens M, Rommelaere J. Partial 3307–3311. reversion of conditional transformation correlates with a 20 Maniatis T, Fritsch EF, Sambrook J (eds). Molecular Cloning: A decrease in the sensitivity of rat cells to killing by the parvovirus Laboratory Manual. Cold Spring Harbor Laboratory Press: Cold minute virus of mice but not in their capacity for virus pro- Spring Harbor, NY, 1982. duction: effect of a temperature-sensitive v-src oncogene. J Virol 21 Shein HM, Enders JF. Multiplication and cytopathogenicity of 1989; 63: 4797–4807. simian vacuolating virus 40 in cultures of human tissues. Proc 6 Christensen J, Cotmore SF, Tattersall P. Minute virus of mice Soc Exp Biol Med 1962; 109: 495–500. transcriptional activator protein NS1 binds directly to the trans- 22 Maxwell IH, Maxwell F. Electroporation of mammalian cells activation region of the viral P38 promoter in a strictly ATP- with a firely luciferase expression plasmid: kinetics of transient dependent manner. J Virol 1995; 69: 5422–5430. expression differ markedly among cell types. DNA 1988; 7: 7 Maxwell IH et al. Recombinant LuIII autonomous parvovirus as 557–562.