Proc. Natl. Acad. Sci. USAVol. 81, pp. 7212-7215, November 1984Microbiology

Slow virus visna: Reproduction in vitro of virus fromextrachromosomal DNA

(lentiviruses/integration/restriction enzyme analysis)

JEFFREY D. HARRIS*, HUBERT BLUM, JANE SCOTTt, BETTY TRAYNOR*, PETER VENTURA,AND ASHLEY HAASE§Infectious Disease Section, V.A. Medical Center, San Francisco, CA 94121

Communicated by Howard M. Temin, July 27, 1984

ABSTRACT Under permissive conditions of growth in tis-sue culture, the retrovirus visna multiplies over the course of afew days to high titer and kills the host cell. We show that inthis lytic life cycle, viral DNA is tightly associated with, but notcovalently linked to, chromosomal DNA. This finding providesexplanations for a number of the unusual properties of the len-tivirus subfamily of retroviruses, and suggests potential mech-anisms for the block in virus gene expression in vivo responsi-ble for the slow infection in nature.

Visna virus is the prototype of the subfamily of retrovirusesthat cause slow infections in sheep characterized by patho-logical changes in the lungs and central nervous system (1,2). Both the persistence of virus in the face of the immuneresponse mounted by the host and the slow evolution of dis-ease can be explained by restriction in virus gene expressionimposed at the transcriptional level (3, 4). Most of the cellsthat harbor virus genomes have insufficient antigen to be de-tected and destroyed by immune surveillance, and limitationin synthesis of virus gene products allows the host cell tosurvive for the extended periods characteristic of slow infec-tions.The mechanism of the block in transcription is unknown

but has generally been assumed to be related to a lysogenicstate, since visna virus is a retrovirus, and viral DNA is asso-ciated with high molecular weight host DNA (5-7). Howev-er, we recently found that transcription in vitro is governedby the extent of early DNA synthesis and suggested that ex-trachromosomal DNA might be the template for viral RNA(8). Moreover, Panganiban and Temin (9) have shown thatproduction of spleen necrosis virus, an avian C-type retrovi-rus, can occur from unintegrated viral DNA. For these rea-sons, we reexamined the role of integration in the visna lifecycle in vitro and found that in the vast majority of cells un-integrated DNA must serve as the template for transcriptionand for virus production.

MATERIALS AND METHODSInfection of Cells and Isolation of DNA. Confluent cultures

of sheep choroid plexus (SCP) cells were infected at a multi-plicity of 3 plaque-forming units per cell as described (8, 10).At 60-70 hr after infection, cells were removed by trypsini-zation and either replated under agarose for infectious centerassay or separated into nuclear and cytoplasmic fractions(11). High molecular weight DNA was isolated by the Hirtfractionation procedure (12). In some experiments, high mo-lecular weight DNA was purified further by electrophoresisin 0.5% low-gelling temperature agarose; the portion of thegel containing DNA 20 kilobase pairs (kbp) or greater in sizewas excised and melted, and DNA was isolated by phen-ol/chloroform extraction.

Cloning of Visna DNA. Viral DNA for probes and for re-construction experiments was obtained by cloning. The rele-vant restriction enzyme sites in visna DNA (9) are shown inFig. 1. DNA from the Hirt supernatant fraction from SCPcells infected for 60 hr was digested with Sst I, and the frag-ments were inserted into a X Charon 10 vector (13). Fifteenclones of recombinant bacteriophage were identified by hy-bridization with a 32P-labeled probe transcribed from viralRNA with random primers (11). Four clones with the small(Vs) Sst I fragment and two clones with the large (VL) Sst Ifragment were identified by restriction enzyme mapping.These clones were amplified and purified by banding inCsCl. DNA from purified bacteriophage was digested withSst I, and the small and large fragments were isolated afterseparation in agarose gels.

Hybridization Procedures. Restriction enzyme digestions,electrophoretic separation of DNA in agarose, and transferto diazophenylthioether-paper followed published protocols(11). Visna-specific probes were radiolabeled to 10 cpm/,gby nick-translation of cloned DNA. The number of copies ofviral RNA in individual cells was evaluated by in situ hybrid-ization (8).

RESULTSVisna DNA Is Tightly Associated with High Molecular

Weight DNA. Most visna DNA extracted from cells is a lin-ear duplex molecule of 9.5 kbp with a nick or gap near thecenter of the molecule (11). This DNA is found in the nucle-us of infected cells within the first few hours of infection andthereafter (8), and a signficant proportion (about 25%) parti-tions into the Hirt pellet. The association with high molecu-lar weight cellular DNA is apparently quite stable, since thesame fraction of DNA is associated with high molecularweight DNA prepared by other procedures (data not shown),such as network formation (5) in alkali (14) or sedimentationof DNA through gradients after lysing the cells in detergentand 2 M NaCl (15).

Restriction Enzyme Analysis. However, by restriction en-zyme tests for integration (16, 17), visna DNA is not cova-lently linked to host DNA. We first suspected that this wasthe case when we found that we could no longer detect viralDNA associated with host DNA 20 kpb or larger after elec-trophoresis through low-gelling temperature agarose (Fig.2C). Full-length viral DNA in the Hirt pellet (Fig. 2B, lane U)freed by this procedure comigrated with uncut viral DNA inthe Hirt supernatant (Fig. 2A), and identical species weregenerated by digestion with BamHI, EcoRI, and Hindill

Abbreviations: kbp, kilobase pair(s); SCP, sheep choroid plexus.*Present address: Microgenics, Concord, CA 94520.tPresent address: Rockefeller University, New York, NY 10021.tPresent address: Academic Research Information System, Inc.,San Francisco, CA 94115.§To whom correspondence should be addressed at: Department ofMicrobiology, University of Minnesota, Minneapolis, MN 55455.

7212

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Proc. Natl. Acad. Sci USA 81 (1984) 7213

1-Vs +

Sst I Sst I Bam Hil lIl- VL

Bam HI

5'kbp LTR

Eco RIHind III

Sst'

13'2 3 4 5 6 7 8 91 LTRI

Hind IIIEco RI

Hind IIIHind III

FIG. 1. Map of the visna genome. Visna DNA is a linear duplex of 9.5 kbp comprised of an intact full-length minus strand, a plus strandinterrupted by a nick or gap (I-4) in the center of the molecule, and two long repeat sequences (LTR; long terminal repeats) at the termini (11).Restriction enzyme sites in the DNA relevant to cloning and other experiments described in this article are indicated; the clone of visna DNA inX Charon 10 corresponding to the small Sst I fragment from 0.8-9.4 kbp is designated Vs; the clone corresponding to the large Sst I fragmentfrom 0.8-9.4 kbp is designated VL-

(Fig. 2 A and B, lanes B, RI, and HI11). Most important, wecould not detect in DNA 20 kbp or larger (Fig. 2C) fragmentsof viral DNA internal to putative viral-host junctions thatshould have been released by these three restriction en-zymes.Reconstruction Experiments, in Situ Hybridization, and In-

fectious Center Assay. The sensitivity of the hybridizationmethods we used was more than sufficient to detect a singlecopy of visna DNA integrated in high molecular weightDNA. We established this by reconstruction experiments inwhich we mixed the equivalent of 0.25 copy of viral DNAper cell in 106 cells (3 pg of viral DNA) with high molecularweight DNA equivalent to that number of cells (5 Ag) fromuninfected cultures. After electrophoresis of uncut DNA,

U B RI HIII U B RI HIII U B RI HllOrigin _kbp

Z.J-~~~~~~~~

2.0-

A

HirtSupernatant

'II.kX

B

HirtPellet

IC

Gel PurifiedHigh MolecularWeight DNAfrom Hirt Pel let

DNA cut with enzymes that do not cleave visna DNA, orDNA cut with enzymes that do cleave visna DNA, we coulddetect 0.25 copy of full-length virus DNA per cell, or evenfragments of viral DNA (Fig. 3A). To further substantiate thesensitivity of the method, we also showed that we could de-tect a single copy of the sheep growth hormone gene on aBamHI fragment of ovine DNA (Fig. 3B) and smaller frag-ments of that gene, using 5 pug of DNA and 0.7-kbp clonedprobe (18).At a time when DNA from infected cells lacks a single

copy of integrated viral DNA, the cells are producing 50-100plaque-forming units of virus per cell and several thousandcopies of viral RNA (8). We assured ourselves that this wasthe case for the cultures from which DNA has been ana-lyzed, by showing by in situ hybridization that most cells hadviral RNA (Fig. 4) and by infectious center assay that 90% ofthe cells were producing virus (data not shown).

DISCUSSIONThe production in vitro of visna virus from extrachromosom-al DNA clearly distinguishes this representative of the sub-family of lentiviruses from retroviruses that transform cellsand cause tumors and, by implication, highlights aspects ofDNA structure important to integration. The vast majority ofvisna DNA molecules are linear duplexes comprised of afull-length minus strand and a plus strand interrupted by anick or gap located at the center of the molecule (11). Only0.1% of the molecules are covalently closed circles, inmarked contrast to oncogenic retroviruses (11). If circularDNA is the topologically preferred form for integration, as anumber of lines of evidence suggest (19-23), visna DNA pre-dictably would not be expected to integrate.The insignificant role evidently played by integration in

the visna life cycle in vitro holds the explanation we think fora number of peculiar features of the lentivirus subfamily,

FIG. 2. Restriction Enzyme Analysis of visna DNA Associatedwith high molecular weight Cell DNA. SCP cells infected for 66 hrwere separated into nuclear and cytoplasmic fractions. DNA in thenucleus was partitioned by the Hirt procedure into supernatant (A)and pellet (B) fractions. (C) DNA in the Hirt pellet was electropho-resed in a gel of 0.5% low-gelling temperature agarose, and DNA 20kbp or larger was identified from the position of markers in a parallellane (bars at the left indicate the position and the size in kbp of XDNA cut with HindIll), and excised and purified. Five microgramsof this DNA, equivalent to 1.2 x 106 SCP cells, was rerun in a gel of1.2% agarose, either uncut (U) or after digestion with BamHI (B),EcoRI (RI), or HindIll (Hill). DNAs in the Hirt pellet and superna-tant were treated similarly. After electrophoresis, the DNA was blottransferred to diazophenylthioether-paper and hybridized to a 32p_labeled visna-specific probe (Vs + VL, 109 dpm/pg). The blots wereexposed at -70'C to x-ray film in a cassette with intensifying screenfor periods of hours to 9 days (C). The asterisks (*) in B indicate thepieces of viral DNA internal to the ends of the genome that shouldhave been evident in C after release from putatively integrated viralDNA.

Microbiology: Harris et aL

Proc. NatL. Acad. Sci. USA 81 (1984)

12 3 4 5 6 7 8 1 2 3 4

;Ls i

23.5-

9.5-

6.6-

4.4

2.3 -

2.0- A

FIG. 3. Reconstruction experiment to estimate the sensitivity ofhybridization. (A) Five micrograms of DNA from uninfected SCPcells, equivalent to 1.2 x 106 cells, was mixed with 3 pg of clonedvisna DNA (VL) equivalent to 0.25 copy of viral DNA per cell in 106cells. DNA was electrophoresed, blot transferred, and hybridized asdescribed in the legend to Fig. 2; exposure time was 5 days. Lane 1,uncut DNA; lane 2, DNA digested with BamHI; lane 3, DNA digest-ed with Bgl II; lane 4, DNA digested with EcoRI; lane 5, DNA di-gested with HindflI; lanes 6-8, DNA digested with enzymes that donot cleave visna DNA-Sal I, Sst II, and Xho I. Bars at the left indi-cate the position of X HindIll markers in kbp. (B) Five microgramsof DNA from uninfected cells was digested with BamHI (lane 1), BglII (lane 2), EcoRI (lane 3), or HindIII (lane 4), electrophoresed, blottransferred, and hybridized with a 700-nucleotide bovine growthhormone insert in pBR322 labeled with 32p to 109 dpm/tzg. Exposuretime was 5 days.

such as the remarkable resistance of lentiviruses to interfer-on (24) and the lack of endogenous viral sequences (5). Ifamong its other actions, interferon coordinately inhibits for-mation of circular DNA, integration, and growth of oncogen-ic retroviruses (25), its effects on visna predictably would beminimal, and the exogenous life cycle of visna in vitro pro-

vides a ready explanation for our failure to detect homologybetween lentivirus DNA and cellular genes (5).

Visna virus in tissue culture fuses cells from without and,in the course of its growth cycle, causes giant cell formationwith subsequent degeneration of the cells (26). We have at-tributed the cytopathic effects of visna virus to this demon-strably damaging effect of the virus on the cell membrane(1), although, by analogy with other cytopathic retroviruses(27, 28), the large amount of unintegrated viral DNA in thecell also could contribute to cell damage. What we can ex-clude now in visna as the mechanism of cytotoxicity is multi-ple insertions of viral DNA into host DNA, a mechanism

B

.w,Z.

.4~~~~~,_ f,

As. .s. f* :i

itiL,,,

74

At.p

FIG. 4. Detection of Visna RNA by in situ Hybridization. (A) Analiquot of the SCP cells infected for 66 hr described in the legend toFig. 2 was deposited by cytocentrifugation on coated glass slides.The cells were fixed, treated to enhance diffusion of the probe, andhybridized in situ (8) with a representative visna-specific probe re-verse transcribed from visna RNA with [3H]TTP to a specific activi-ty of 1.5 x 108 dpm/,g. After coating with NTB-2 emulsion andexposure at 40C for 3 hr, the slides were developed and stained.(Photomicrograph taken with transillumination and incident polariz-ing light; original magnification, x500.) (B) Uninfected SCP cells;hybridization and exposure identical to that of A.

initially proposed by Battula and Temin (29) to account forthe cytopathic effects of spleen nicrosis virus.The persistence of visna virus and the slow pace of repli-

cation in nature can be considered derivatives of restrictedgene expression in vivo (1-4, 8). We have proposed twomechanisms to account for this restriction, one equivalent tolysogeny on a grand scale in an animal, and the other a genedosage model, in which transcription is regulated by the ex-tent of early viral DNA synthesis (8).The findings reported here are consistent with both mech-

anisms of gene regulation. Under natural conditions of infec-tion with small inocula, too few particles may enter cells tosupport the extensive early synthesis of viral DNA requiredto achieve the high levels of transcription characteristic ofthe in vitro life cycles (8). Alternatively, virus DNA might beintegrated in vivo at loci in host DNA where transcription isrepressed.

We thank Yvonne Guptill and Rosaria Cardella for preparation ofthis manuscript and Walter Miller for bovine growth hormone DNA.This work was supported by grants from the American Cancer Soci-ety and the National Institute of Neurological and CommunicativeDisorders and Stroke, by basic institutional support from the V.A.,and by a Heisenberg fellowship from the German government(H.B.). In the course of these studies, A.H. was a medical investiga-tor of the Veterans Administration.

kbpA

7214 Microbiology: Harris et aL

BQ:

Proc. Natl. Acad Sci. USA 81 (1984) 7215

1. Haase, A. T. (1975) Curr. Top. Microbiol. Immunol. 72, 101-156.

2. Brahic, M. & Haase, A. T. (1981) in Comparative Diagnosis ofViral Disease, eds. Kurstak, B. E. & Kurstak, C. (Academic,New York), Vol. 4, Part B, p. 619.

3. Haase, A. T., Stowring, L., Narayan, O., Griffin, D. & Price,D. (1977) Science 195, 175-177.

4. Brahic, M., Stowring, L., Ventura, P. & Haase, A. T. (1981)Nature (London) 292, 240.

5. Haase, A. T. & Varmus, H. E. (1973) Nature (London) NewBiol. 245, 237-239.

6. Haase, A. T., Traynor, B. L. & Ventura, P. E. (1976) Virology70, 65-79.

7. Clements, J. E., Narayan, O., Griffin, D. E. & Johnson, R. T.(1979) Virology 93, 377-386.

8. Haase, A. T., Stowring, L., Harris, J. D., Traynor, B., Ventu-ra, P., Peluso, R. & Brahic, M. (1982) Virology 119, 399-410.

9. Panganiban, A. & Temin, H. (1983) Nature (London) 306, 155-160.

10. Haase, A. T. & Baringer, R. (1974) Virology 57, 238-250.11. Harris, J. D., Scott, J. V., Traynor, B., Brahic, M., Stowring,

L., Ventura, P., Haase, A. T. & Peluso, R. (1981) Virology113, 573-583.

12. Hirt, B. (1967) J. Mol. Biol. 26, 365-369.13. Williams, B. G. & Blattner, F. R. (1978) J. Virol. 29, 555-575.14. Tyndall, C., Younghusband, H. B. & Bellet, A. J. D. (1978) J.

Virol. 25, 1-10.

15. Razin, S. V., Mantieva, V. L. & Georgiev, G. P. (1979) Nucle-ic Acids Res. 7, 1713-1735.

16. Botchan, M., Topp, W. & Sambrook, J. (1976) Cell 9, 269-287.17. Ketner, G. & Kelly, T. J. (1976) Proc. Natl. Acad. Sci. USA

73, 1192-1206.18. Miller, W., Martial, J. & Baxter, J. (1980) J. Biol. Chem. 255,

7521-7524.19. Ramareddy, V., Guntaka, B., Mahy, B. W. J., Bishop, J. M.

& Varmus, H. (1975) Nature (London) 253, 507-511.20. Jolicoeur, P. & Rassart, E. (1980) J. Virol. 33, 183-195.21. Jolicoeur, P., Flamand, J. & Hamelin, D. (1980) Virology 103,

255-258.22. Yang, W. K., Kiggans, J. O., Yang, D., Ou, C., Tennant,

R. W., Brown, A. & Bassin, R. H. (1980) Proc. Natl. Acad.Sci. USA 77, 2994-2998.

23. Panganiban, A. & Temin, H. (1984) Cell 36, 673-679.24. Carroll, D., Ventura, P., Haase, A., Rinaldo, C. R., Overall,

J. C. & Glasgow, L. A. (1978) J. Infect. Dis. 138, 614-617.25. Huleihel, M. & Aboud, M. (1983) J. Virol. 48, 120-126.26. Harter, D. & Choppin, P. (1967) Virology 31, 279-288.27. Temin, H. M., Keshet, E. & Weller, S. K. (1980) Cold Spring

Harbor Symp. Quant. Biol. 44, 773-778.28. Weller, S. K. & Temin, H. M. (1981) J. Virol. 39, 713-721.29. Battula, N. & Temin, H. M. (1978) J. Natl. Cancer Inst. 13,

387-398.

Microbiology: Harris et aL