Antiretroviral Effect of a Gag-Rnase HI Fusion Gene

Home , Nuclease

G Schumann1, K Cannon1, W-P Ma2, RJ Crouch2 and JD Boeke1 1Department of Molecular Biology and Genetics, 725 North Wolfe Street, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA; and 2Laboratory of Molecular Genetics, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD, USA

We have previously shown that a molecule consisting of a useful for therapeutic purposes. A second nuclease, fusion of a Ca2+-dependent nuclease (from Staphylococcus Escherichia coli RNase HI was found to be nontoxic and aureus) to a retroviral coat protein speciﬁes a potent anti- highly effective against a murine leukemia virus when it viral speciﬁc for that retrovirus. Genes specifying such was fused to the leukemia virus coat protein. The fusion fusion proteins can be delivered to virus-susceptible cells, protein was enzymatically active and stably expressed, providing an antiviral gene therapy aimed at limiting virus without apparent toxicity to host cells. Reduction in infec- spread. We report here the results of experiments to vary tious virus output was as high as 97–99%. These studies the nuclease moiety of such fusion proteins. We found that provide a model system for the development of gene thera- one nuclease, Serratia marcescens nuclease, was peutic agents aimed at combating retroviral infections in extremely toxic to host cells and hence not likely to be vivo.

Keywords: antiviral gene therapy; murine leukemia virus; RCAS vectors; RNase HI; Serratia nuclease

Introduction susceptible avian retrovirus vector (RCAS-BP), whose gag gene is very different from that of murine Mo(4070A) is Retroviruses assemble their virion cores from coat pro- used as the delivery vector to introduce the antiviral gag- teins specified by their gag genes. We have exploited the SN fusion gene into cells. In the gene therapy model, we rules of this assembly process to devise a novel antiviral first infected cells chronically with Mo(4070A), and then strategy in which the gag genes (or other viral structural treated the culture with the RCAS-gag-SN vector, and genes) are fused to nuclease genes; the fusion proteins observed a 20- to 60-fold reduction in infectious viral tit- encoded by these antiviral genes are destructive to the ers as a result of the treatment. The treated cells viral genomic RNA. This strategy, referred to as capsid- expressed large amounts of the Gag-SN protein, which targeted viral inactivation (CTVI) has been reduced to are encapsidated into nascent virus particles, in which 1 practice against the yeast retroelement Ty1 and the Gag-SN is enzymatically active and hence degrades the 2,3 murine leukemia virus Mo(4070A), and we have begun Mo(4070A) viral RNA.3 4,5 efforts to target immunodeficiency viruses. Ultimately, The SN antiviral moiety was chosen based on some of we envision delivering CTVI constructs therapeutically, its basic biochemical properties. SN is small, lacks disul- targeting either the appropriate virus-susceptible cell fide bonds, and is active against single- and double- type(s) or the hematopoietic stem cells that sustain such stranded nucleic acids, without extensive sequence speci- 6 populations in infected individuals. If appropriate ficity.9 An especially interesting property from the per- human immunodeficiency virus (HIV)-specific genetic spective of gene therapy applications is its absolute antivirals can be delivered efficiently to the appropriate dependence on relatively high concentrations of Ca2+. cells, it should be possible to have a significant impact Because intracellular concentrations of calcium ion are on viral load in infected individuals, and recent data tightly regulated at submicromolar levels, this enzyme regarding viral load suggest that it is a most important shows no apparent activity or toxicity when expressed in prognosticator of disease progression in HIV-positive yeast, chicken or mammalian cells.1,2 Extracellular fluids 7,8 individuals. typically contain calcium concentrations of several milli- Thus far, the system in which we have the best data molar,10 and SN works best at concentrations of about 1 on antiviral effects is with the amphotropic Mo(4070A) millimolar. A down side to the use of bacterial nucleases system, in which we have modeled both antiviral is that they might give rise to an immune response, and 2 3 prophylaxis and therapy. All of our work to date has might pose significant regulatory hurdles for been with a fusion gene encoding the Mo(4070A) Gag implementing a gene therapy trial. Regrettably, a mam- fused to Staphylococcal nuclease (the fusion protein pro- malian analog of SN (that is, a similarly Ca2+-dependent duced is referred to as Gag-SN). In this system, a non- nuclease) has not been described. Therefore, we are exploring other classes of nucleases that are potentially nontoxic and might also have mammalian homologs, in Correspondence: JD Boeke order to explore alternative versions of this antiviral KC and W-PM contributed equally to this work strategy. Received 18 October 1996; accepted 27 January 1997 Encouraged by our success with SN, we have begun Antiviral Gag-RNase HI fusion protein G Schumann et al 594 to investigate other classes of nucleases for eventual prac- fusion proteins can be used both prophylactically (by tical application of CTVI in a gene therapy setting. In this infecting first with the RCAS derivative and then chal- study, we have examined two additional nucleases for lenging with Mo(4070A)) and therapeutically by persist- their possible utility in eventual gene therapy appli- ently infecting the culture with Mo(4070A) and then tre- cations. These are: (1) the Serratia marcescens nuclease ating the culture with the RCAS derivative. (SMN), which is representative of a class of nucleases that We constructed a set of RCAS proviruses that express is reported to be absolutely dependent on disulfide bond fusion proteins consisting of Mo(4070A) Gag (nuc- formation;11 and (2) RNase HI (RHI) from Escherichia coli, leocapsid) protein fused to SMN and RHI in a manner a ribonuclease predicted to attack retroviral replication similar to that previously used to generate the Gag-SN intermediates.12 Both of these enzymes have mammalian fusion2 (Figure 1). First, the nuclease coding regions were homologs or analogs. amplified by PCR and sequenced. Then, they were joined The S. marcescens SMN enzyme, like SN, attacks both to the BamHI site near the 3′-end of the Mo(4070A) gag RNA and DNA relatively nonspecifically, but unlike SN, gene to create fusion cassettes. Finally, the fusion cas- is active with physiological concentrations of Mg2+ and settes were inserted into the RCAS-BP vector. These vec- does not require calcium ions.13 Unlike many exotoxic tors can then be transfected into chick embryo fibroblasts proteins produced by bacteria, there is apparently no (CEF), in which they efficiently replicate until every cell ‘immunity protein’ produced to inactivate SMN. Genetic in the population is infected (at about day 15 after studies indicate that SMN can be tolerated by the cell that transfection) and expresses the encoded fusion protein. produces it because the enzyme is exported via a signal sequence and furthermore, the two disulfide bonds in the Serratia marcescens nuclease construct is highly enzyme are required for activity.11 Disulfide bonds are cytotoxic rare in cytoplasmic proteins, but common in secreted pro- We found that when the RCAS gag-SMN virus was transfected into cells, the CEF culture reproducibly and dra- teins. The reducing environment of the cytoplasm, which Ͼ normally contains a 500:1 ratio of reduced to oxidized matically died out, with 90% cell killing by day 14 after glutathione, is not conducive to the formation of disul- transfection, a time at which complete infection of the fide bonds. CEF culture is achieved. This lethality was observed The E. coli RNase HI enzyme is a representative of an whether or not the cells were previously infected with ubiquitous cellular enzyme whose cellular role is not Mo(4070A). We examined this further by investigating unambiguously established, but probably plays roles in the time course of loss of viability (Figure 2). The lethality DNA replication and repair.14 RNase H specifically recog- required expression of the Gag-SMN protein, because nizes RNA–DNA hybrid (hence H) duplexes and specifi- cells transfected with a nonexpressing construct grew cally degrades the RNA strand. Retroviruses encode an normally. The observed lethality presumably results from RNase H which is essential for their replication; the residual intracellular nuclease activity of SMN. SMN is activity is required for strand transfer during minus not a good candidate for the CTVI approach and this strand DNA synthesis, for generation of the plus strand emphasizes the need for individual testing of nuclease primer, and for degradation of the RNA strand to allow candidates. 15 synthesis of the plus DNA strand. The specific activity RNase HI is expressed and incorporated into virions of the cellular enzyme (RHI) is approximately 2500-fold In contrast, the gag-RHI construct, like the gag-SN con- higher than that of the viral enzyme, suggesting that if it struct, showed no signs of overt toxicity to the CEF cul- could be targeted to retroviral replication complexes, it ture at any time (Figure 2). We collected culture super- might abrogate retroviral replication processes. We natants from the infected cells and pelleted them to look recognized that RNase HI might be a useful nuclease for evidence of expression and encapsidation of the Gag- moiety for CTVI and provided evidence that it inhibited RHI protein and its proteolytically processed derivatives. murine leukemia virus RT in vitro and had a potent anti- 12 The Gag-RHI protein could be readily detected in the Ty1 effect in vivo. Thus, RHI represented an excellent viral pellet fractions by immunoblotting with antibodies candidate antiviral moiety to test for therapeutic effects raised against E. coli RNase HI (Figure 3a) as well as anti- in the Mo(4070A) murine leukemia virus system, as we bodies against Mo(4070A) NC protein (Figure 4). Both of have done here. these antibodies react with an approximately 85 kDa low abundance species that corresponds to full-length Gag- Results RHI protein that has not been processed by the Mo(4070A) protease. In addition, there is a very abun- dant species recognized by both antibodies at about 27 New Gag-nuclease fusion constructs targeted against kDa that corresponds to the NC-RHI processed fusion Mo(4070A) virus protein. These results are similar to those previously To obtain uniform delivery of the antiviral transgenes, obtained with Gag-SN, and provide strong evidence that and high-level expression of the putative antiviral pro- the Gag-RHI protein is not only being expressed, but is teins, we used the RCAS vector system developed by 2,16–21 being efficiently encapsidated into Mo(4070A) virions, Hughes and his collaborators. In this system, inserts where it is acted upon by the Mo(4070A) protease and of up to 2 kb can be delivered in a straightforward way released as an NC-RHI fusion protein. to cells sensitive to the avian-specific Rous sarcoma virus or its amphotropic derivatives, which can infect mam- Gag-RHI fusion protein is enzymatically active malian cells.22 We have previously shown that RCAS-BP We investigated whether the Gag-RHI fusion protein was can provide a potent antiviral effect against the unrelated active by an in-gel activity assay. This assay is especially murine leukemia virus Mo(4070A) when it expresses a useful because it not only provides evidence for activity Gag-SN fusion protein in chicken cells.2,3 These antiviral but also indicates which of the species observed by Antiviral Gag-RNase HI fusion protein G Schumann et al 595

Figure 1 Diagram of the constructs. Replication-competent retrovirus vector to introduce and express Mo(4070A)-gag-SMN and Mo(4070A)-gag-RHI fusions in CEF cells; the structure of the RSV-based RCAS-vector (top cartoon) has been described.21 The structures of the inserts in (a) Mo(4070A)- gag-SN (pGN1600),2 (b) Mo(4070A)-gag-SMN (pGS137) and (c) Mo(4070A)-gag-RHI (pCAN255) are indicated. In these constructs, the Mo(4070A) gag is symbolized by a hatched box and the nuclease by a black box; fusion junction DNA sequences are given as insets. Expression of these fusion proteins is under control of the RSV-LTR. The four different domains of Mo(4070A) Gag, matrix (MA), p12, capsid (CA) and nucleocapsid (NC) are indicated. Boxed triangles, long terminal repeats; open box, RSV coding regions; hatched box, murine leukemia virus sequences; solid box, nuclease coding sequences; arrow with V, spliced mRNA produced in RCAS; SA, splice acceptor; relevant restriction sites are indicated, BamHI (B), ClaI (C), NcoI (N), SalI (S). immunoblotting are active. The activity gel (Figure 3b) ing out quantitative immunoblots on viral pellet fractions contains viral pellet samples from the gag-RHI transfec- from parallel cultures using anti-NC (Figure 4, Table 1) tion as well as from mock-transfected cells. The latter and anti-CA (not shown) antibodies. The anti-NC immu- control shows that we are not detecting the viral RNase noblots confirmed the results of the anti-RHI antibodies; H activity which has different biochemical optima and the species assumed to be NC-RHI indeed reacted hence is not detectably active under the conditions used strongly with the anti-NC antibody (and not with the here. In addition, we include a standard curve of purified anti-CA antibody), and as expected, the 85 kDa species E. coli RNase HI which allows an assessment of the spe- reacted with both anti-NC and anti-CA. The levels of NC- cific activity of the fusion proteins. We conclude from RHI and NC-SN can be directly compared and were less comparative analysis of the immunoblot and the activity than two-fold different, thus the NC-RHI protein is gel that the specific activities of the fusion proteins are expressed and encapsidated most efficiently. The total very similar (within a factor of 5) to native RNase HI levels of bulk virions produced in both the gag-SN and from E. coli. Furthermore, we note that three additional gag-RHI transfected cultures were three- to six-fold active species are visible. These presumably represent higher than those in the mock-transfected culture (Table additional processed forms of the fusion protein. 1). Thus the specific infectivity of these virions is con- siderably lower than is apparent from inspection of the Expression levels of Gag-SN, Gag-RHI and viral raw infectivity data presented in Figure 5, and the proteins reduction in specific activities of Gag-SN and Gag-RHI We examined the relative expression levels of Gag-SN are similar to each other. and Gag-RHI. In addition, the expression of Gag-SN was previously shown not to interfere with production of Antiviral effect of Gag-RHI fusion protein viral proteins or viral assembly. We investigated whether The cultures that had been treated by transfection of gag- the Gag-RHI fusion protein behaved similarly by carry- SN and gag-RHI proviruses (or mock-treated) were Antiviral Gag-RNase HI fusion protein G Schumann et al 596 The Gag-RHI protein is efficiently expressed without apparent toxicity to host cells. The protein is efficiently encapsidated into virions, where it undergoes processing by the Mo(4070A) protease. We do not know whether this proteolytic cleavage is required for the antiviral effect, but it seems likely that untethering the enzyme from the inner capsid surface would allow it better access to the replicating RNA–DNA hybrids that it targets. The NC- RHI fusion protein as well as its precursor (and a very small amount of native RHI-sized material visible in the activity gel) all have enzymatic activity, as assayed both in the gel assay (Figure 3b) and in a quantitative in vitro RNase HI assay carried out on permeabilized virion preparations (data not shown). The gag-RHI fusion gene is as effective in this tissue culture gene therapy model system as is the previously developed gag-SN fusion gene. We presume that the RNase HI results in an antiviral effect via premature degradation of RNA–DNA hybrids, as expected from in vitro studies12 or interference with specific steps in the complex reverse transcription pathway. Also, this previous study on Ty1 showed that the active site of RHI was necessary to obtain an ‘antiviral’ effect.12 Previous studies of our Mo(4070A) system Figure 2 Gag-SMN expression is toxic to CEF cells. CEF were trans- have shown that foreign proteins without nuclease fected with RCAS vector (square), pCAN255 (expressing Gag-RHI, activity fused to the C-terminus of Gag result only in triangle), pGS137 (expressing Gag-SMN, diamond) or pGS145 (circle), modest antiviral effects, much less dramatic than those which is identical to pGS137 except it is cloned in the RCAN vector in observed here.3 The action of the RHI on Mo(4070A) which the insert is not expressed, because the splice acceptor (SA in Figure 1) is absent. Transfected cells were passaged and split 1:3 every 2–4 days . could occur either in the infected cell or prior to that, as Cells were trypsinized, resuspended in a final volume of 10 ml and the there is now considerable evidence that reverse transcrip- number of viable (Trypan blue-negative) cells per ml medium was determ- tion can occur, at least in part, before infection.24–26 How- ined (Cells/ml). The mean cell numbers were determined from two inde- ever, further studies will be needed to investigate the pendent transfection series and monitored every 2–4 days for 12 days after detailed mechanism by which these fusion proteins inter- transfection. Error bars indicate standard deviations. fere with viral multiplication in vivo. Finally, unlike SN, RNase H is widely conserved assayed for infectious virus production every 4 days, among prokaryotes and eukaryotes and thus the more using the sensitive S+L− focus formation assay.23 As can appropriate human RNase H cDNA useful for human be seen in Figure 5, both the Gag-SN and Gag-RHI pro- gene therapy can be obtained and built into antiviral con- teins have similar antiviral effects on infectivity, that is, structs. Such efforts are currently underway. about a 90% decrease in raw titer, and about a 97–99% reduction when the effect of increased number of bulk particles in the transfected samples (see previous section Materials and methods and Table 1) is accounted for. As is the case with Gag- SN, Gag-RHI maintained its antiviral effect throughout Plasmid constructs the lifetime of the culture (about 15 days after the drop The SMN gene fragment was kindly provided by Dr in titer was noted). Thus RNase H based fusion proteins Michael Benedik (University of Houston, TX, USA). We show significant promise as antivirals. amplified the SMN coding region as a BamHI–SalI restriction fragment using PCR and subcloned this fragment (exactly as previously described for SN2) as an in-frame Discussion gag-SMN fusion gene into the RCAS-BP vector,21 Our studies have shown that not all nucleases are appro- resulting in the RCAS-gag-SMN plasmid pGS137. The priate to mediate capsid-targeted viral inactivation with- construction and detailed maps of the RCAS replication- competent avian retrovirus vector and its derivative pGN out affecting the viability of host cells. The extreme tox- 2,21 icity of the gag-SMN construct indicates that SMN is 1600 have been described previously. For the RHI con- probably killing cells by significant residual nonspecific struct, we first constructed a gag-RHI gene with appropri- nuclease activity. This is somewhat unexpected because ate restriction sites at the ends. pRNH1.4 contains the disulfide bond formation is required for maximal enzyme rnhA gene of E. coli and additional nucleic acid sequences AAG CTT CAT and ACT AGT at each end of the gene activity and (presumably) does not occur in the cyto- ′ plasm or nucleus. This does not mean that all enzymes for cloning purposes, namely a HindIII site just 5 of the of the general class of disulfide-bond containing nucle- ATG codon and has a downstream flanking XhoI site. The ases are necessarily inappropriate, but that each will have relevant sequences are: to be tested empirically. RNase H on the other hand is a very good candidate ...... AAG CTT CAT ATG. . .rnhA. . .TAA ACT for CTVI, as it has now been shown to be active against AGT...... two extremely dissimilar retroelements, the Ty1 retro- transposon and the murine leukemia virus Mo(4070A). pCAN 60, containing the gag-RHI fusion gene, was made Antiviral Gag-RNase HI fusion protein G Schumann et al 597

Figure 3 Gag-RHI fusion proteins are expressed and secreted in particulate form. (a) Immunoblot analysis. Incorporation of Gag-RNH fusion proteins into Mo(4070A) virions. Five microliters of Gag-RNH viral pellets from gag-RNH-transfected CEF cells harvested 7 days after transfection were subjected to SDS-polyacrylamide gel electrophoresis and examined by immunoblot analysis with anti-E. coli RNase HI antibodies. Virions (5 ␮l) from the Mo(4070A)-infected CEF cells (MOCK) and purified E. coli RNase HI (0.5–10 ng) were loaded as negative controls and for estimation of E. coli RNase HI expression. Molecular mass markers and estimated masses of the detectable processing products of Gag-RNH fusion proteins are shown. (b) Activity gel analysis. RNase H renaturation gel assay on Mo(4070A) virions containing Gag-RNH fusion proteins. Mo(4070A) virions and purified E. coli RNase HI were electrophoresed in a 15% polyacrylamide gel which contains 32P-poly(rA)-poly(dT). The picture of the gel for RNase H activity is given as a negative of the autoradiogram; thus RNase H activity appears as a dark band. Molecular mass markers and estimated masses of the detectable processing products of Gag-RNH fusion proteins are shown. by Klenow filling the HindIII site of pRNH-1.4, prior to purified antibody raised in rabbit against pure E. coli digestion with XhoI. The HindIII–XhoI fragment was then RNase HI as previously described.12 The reaction was ligated to pGN1542 (a pCla 12-Nco derivative17 contain- developed using enhanced chemiluminescence ing Mo(4070A) gag) that had been Klenow filled at the (Amersham Life Sciences, Buckinghamshire, UK). The BamHI site prior to SalI digestion, generating the appro- molecular mass of each protein was determined using priate fusion gene (Figure 1). The fusion gene was then prestained protein markers (Rainbow markers RPN 756, subcloned into the pGN1600 plasmid,2 substituting gag- 14.3 to 200 kDa; Amersham Life Sciences). RHI for gag-SN. The clones were sequenced to confirm Immunoblotting with anti-p30CA, detection with 125I- the nuclease and junction sequences. Protein G and quantification was exactly as described previously,3 and immunoblotting with anti-p10NC fol- Immunoblotting lowed the same protocol except that the anti-p30CA anti- Resuspended viral pellets and purified E. coli RNase HI body was diluted 1500-fold and the anti-p10NC antibody 27 were mixed with Laemmli buffer, boiled, separated on was diluted 5000-fold. The anti-p10NC and anti-p30CA SDS-polyacrylamide gel (10–18%) and electro-blotted on antibodies were the kind gift of Dr Alan Rein (NCI- to Immobilon P membrane (Millipore Corporation, Frederick Cancer Research and Development Center, Bedford, MA, USA) in Tris-glycine-SDS buffer for 2 h at Frederick, MD, USA). 5 V. Gag-RNH fusion proteins were detected by using Antiviral Gag-RNase HI fusion protein G Schumann et al 598

Figure 4 Immunoblot analysis with anti-NC. Quantiﬁcation of bulk Mo(4070A) virions. Various volumes of the same viral pellet fractions from gag- SN, gag-RHI and mock-transfected cells (Gag-SN, Gag-RHI, MOCK), used for the immunoblot analysis of Figure 3 were examined by immunoblot analysis using anti-p10NC antibody as probe. Dilutions of a titrated Mo(4070A) stock were loaded as positive controls. The quantity of infectious particles that corresponds to the amount of loaded bulk particles is indicated (FIU(×102)). Processing intermediates and ﬁnal products resulting from processing of the wild-type Gag precursor and the Gag-SN and Gag-RHI fusion proteins are indicated as boxes, together with their predicted molecular weights. The identities of the processing intermediates were obtained from additional blots done with anti-p30CA anti-SN and anti-RHI antibodies (not shown).

2-mercaptoethanol) at room temperature. Renaturation of RNase H was achieved by incubating the gel overnight in renaturation buffer. After renaturation the gel was exposed to a radiograph ﬁlm. The band demonstrating RNase H activity was detected in a dark background as a clear area in which the substrate had been degraded and the soluble radioactivity had leached from the gel. In Figure 3b RNase H activity appears as a dark band because it represents a negative of the autoradiogram.

Cell growth, infection with Mo(4070A), transfection with RCAS derivatives, and Mo(4070A) virus harvesting and titration Growth of CEF cells and infection with Mo(4070A) virus was as described previously.3 In order to perform trans- fections CEF cells were grown to a density of 7 × 106 to 1 × 107 cells per dish, passaged at a 1:3 dilution with tryp- sin and transfected the next day with 10 ␮g of the recombinant replication-competent avian retrovirus vectors using the calcium-phosphate precipitation method.29 Spread of the recombinant virus by infection was achieved by passaging the transfected cells every 3–4 days. Collection and resuspension of viral pellets was also as previously described.2 Titration of infectious Mo(4070A) Figure 5 Antiviral effect of Gag-RHI and Gag-SN fusion genes. Infec- 23 tious Mo(4070A) titers of the supernatants of gag-RHI, (pCAN255, was essentially as described by Bassin et al with diamonds), gag-SN (pGN1600, circles) and mock-transfected cells (Mock, minor modifications.3 squares), were assayed by the S+L− focus formation assay.23 The mean titers were determined from two independent transfection series and moni- Note added in proof tored every 3–6 days for 34 days after transfection. Error bars indicate Independent work showing the antiviral efficacy of standard deviations. FIU, focus inducing units. RNAse H fusion proteins is described by Van Brocklin et al. ‘Expression of a murine leukemia virus gag-Escherichia RNase H renaturation gel assay coli RNAse hi fusion polyprotein significantly inhibits Resuspended viral pellets and purified E. coli RNase HI virus spread’. J Virol 1997 (in press). were electrophoresed on a 15% SDS-polyacrylamide gel using the buffer system of Laemmli.27 The separating gel Acknowledgements contained 2–10 nmol (5 × 107 c.p.m.) of 32P-poly(rA)- poly(dT). The renaturation gel assay of RNase H was per- Mark Federspiel and his laboratory (Mayo Foundation, formed as described previously.28 Electrophoresis was Rochester, MN, USA) have independently shown that conducted at 50 V for 16 h. After electrophoresis, the gel Gag-RHI fusions can interfere with retrovirus multipli- was soaked in several changes of renaturation buffer (50 cation. We thank Mark Federspiel for sharing the results m m m m m Tris-HCl (pH 7.9), 50 m NaCl, 10 m MgCl2,10m of his experiments prior to publication and for helpful Antiviral Gag-RNase HI fusion protein G Schumann et al 599 Table 1 Quantification of data on bulk virion production in mock-transfected and transfected cells

Total pixel signal above Signal relative to mock Reduction in speciﬁc background activity (%)

Mock 9.6 × 104 ± 2.5 × 104 10 Gag-SN 25.9 × 104 ± 7.3 × 104 2.7 96.9 Gag-RHI 62.2 × 104 ± 18.9 × 104 6.5 98.6

Independent phosphorimager measurements of total NC protein found in viral pellet preparations were made based on the four dilutions done on each sample. These data agree with our earlier work that shows that NC (and other viral Gag proteins) are overproduced when transfected with gag-based constructs.3 In these experiments, they have a 2.7- to 6.5-fold higher signal than mock-transfected cells. We assume that this represents an equivalent increase in the number of virions in these cultures. The specific infectivity per virion takes this discrepancy into account and is based on the data from day 7 after transfection (Figure 5). discussions. The Johns Hopkins University and JD Boeke 14 Hostomsky Z, Hostomska Z, Mathews DA. Ribonucleases H. In: have a financial interest in the CTVI method described Linn SM, Lloyd RS, Roberts RJ (eds). Nucleases, 2nd edn. Cold in this publication. The terms of this arrangement have Spring Harbor Laboratory Press: Cold Spring Harbor, NY, 1993, been reviewed and approved by the School of Medicine pp 341–376. 15 Champoux JJ. Roles of ribonuclease H in reverse transcription. Committee on Conflict of Interest. Supported in part by In: Skalka A, Goff S (eds). Reverse Transcriptase. Cold Spring a DFG postdoctoral fellowship Schu-1014 to GS, the Intra- Harbor Laboratory Press: Cold Spring Harbor, NY, 1993, 103– mural AIDS Targeted Antiviral Program to RC and 118. NCDDG-AIDS grant AI35282 to JDB. 16 Hughes SH, Kosic E. Mutagenesis of the region between env and src of the SR-A strain of Rous sarcoma virus for the purpose References of constructing helper-independent vectors. Virology 1984; 136: 89–99. 1 Natsoulis G, Boeke JD. New antiviral strategy using capsid- 17 Hughes SH, Greenhouse JJ, Petropoulos CJ, Sutrave P. Adaptor nuclease fusion proteins. Nature 1991; 352: 632–635. plasmids simplify the insertion of foreign DNA into helper- 2 Natsoulis G et al. Targeting of a nuclease to murine leukemia independent retroviral vectors. J Virol 1987; 61: 3004–3012. virus capsids inhibits viral multiplication. Proc Natl Acad Sci 18 Salter DW et al. Transgenic chickens: insertion of retroviral USA 1995; 92: 364–368. genes into the chicken germ line. Virology 1987; 157: 236–240. 3 Schumann G et al. Therapeutic effect of Gag-nuclease fusion pro- 19 Greenhouse JJ, Petropoulos CJ, Crittenden LB, Hughes SH. tein on retrovirus-infected cell cultures. J Virol 1996; 70: 4329– Helper-independent retrovirus vectors with Rous-associated 4337. virus type O long terminal repeats. J Virol 1988; 62: 4809–4812. 4WuXet al. Targeting foreign proteins to human immunodefi- 20 Federspiel MJ, Crittenden LB, Hughes SH. Expression of reticul- ciency virus particles via fusion with Vpr and Vpx. J Virol 1995; oendotheliosis virus envelope confers host resistance. Virology 69: 3389–3398. 1989; 173: 167–177. 5 Fletcher TM III et al. Virion associated accessory proteins: new 21 Petropoulos CJ, Hughes SH. Replication-competent retrovirus targets for antiviral therapy. In: Girard M, Dodet B (eds). Retro- vectors for the transfer and expression of gene cassettes in avian viruses of Human AIDS and Related Animal Diseases. Dixieme Col- cells. J Virol 1991; 65: 3728–3737. loque des Cent Gardes-1995, 1995, pp 99–106. 22 Barsov X, Hughes S. Amphotropic derivatives of RCAS vector 6 Boeke JD, Hahn B. Destroying retroviruses from within. Trends system. J Virol 1996; 70: 3922–3929. Microbiol 1996; 4: 421–426. 23 Bassin RH, Tuttle N, Fischinger JR. Rapid cell culture assay tech- 7 Perelson AS et al. HIV-1 dynamics in vivo: virion clearance rate, nique for murine leukaemia viruses. Nature 1971; 229: 564–566. infected cell life-span, and viral generation time. Science 1996; 24 Trono D. Partial reverse transcripts in virions from human 271: 1582–1586. immunodeficiency and murine leukemia viruses. J Virol 1992; 8 Wei X et al. Viral dynamics in human immunodeficiency virus 66: 4893–4900. type 1 infection. Nature 1995; 373: 117–122. 25 Lori F et al. Viral DNA carried by human immunodeficiency 9 Tucker PW, Hazen EE, Cotton FA. Staphylococcal nuclease virus type 1 virions. J Virol 1992; 66: 5067–5074. reviewed: a prototypic study in modern enzymology. Isolation; 26 Zhang H, Dornadula G, Pomerantz RJ. Endogenous reverse physical and enzymatic properties. Mol Cell Biochem 1978; 22: transcription of human immunodeficiency virus type 1 in 67–77. physiological microenviroments: an important stage for viral 10 Altman PL. Blood and Other Body Fluids. Federation of American infection of nondividing cells. J Virol 1996; 70: 2809–2824. Societies for Experimental Biology: Washington, DC, 1961. 27 Laemmli UK. Cleavage of structural proteins during the 11 Ball TK, Suh Y, Benedik MJ. Disulfide bonds are required for assembly of the head of bacterophage T4. Nature 1970; 227: Serratia marcescens nuclease activity. Nucleic Acids Res 1992; 20: 680–685. 4971–4974. 28 Carl PL, Bloom L, Crouch RJ. Isolation and mapping of a 12 Ma W-P, Crouch RJ. Escherichia coli RNAse HI inhibits murine mutation in Escherichia coli with altered levels of ribonuclease leukaemia virus reverse transcription in vitro and yeast retro- H. J Bacteriol 1980; 144: 28–35. transposon Ty1 transposition in vivo. Genes to Cells 1996; 1: 29 Kingston RE, Chen CA, Okayama H, Rose JK. Introduction of 581–593. DNA into eukaryotic cells. In: Ausubel FM et al (eds). Current 13 Eaves GN, Jeffries CD. Isolation and properties of an extracellu- Protocols in Molecular Biology. John Wiley & Sons: New York, lar nuclease of Serratia marcescens. J Bacteriol 1963; 85: 273–278. 1996, pp 9.1.4–9.1.6.