Defective viral particles in caprine arthritis encephalitis virus infection

VIROLOGY

189, 344-349 (1992)

Defective Viral Particles in Caprine Arthritis Encephalitis Virus Infection ARNONA

GAZIT,”

RONIT SARID,* PNINA MASHIAH,* DENIS ARCHAMBAULT,t” STEVEN R. TRoNIcK,tv3AND ABRAHAM YANIV*

JOHN

E. DAHLBERG,t’*

*Department of Human Microbiology, Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel; and tLaboratory of Cellular and Molecular Biology, National Cancer Institute, Bethesda, Maryland 20892 Received January 2, 1992; accepted

March 6, 1992

Attempts to isolate full-length unintegrated circular forms of the caprine arthritis encephalitis virus (CAEV) genome yielded only a large number of molecules with deletions. The 3’ borders of most of these deletions were near the U3 region of the long terminal repeat whereas the 5’edges were found at various upstream sites withinpolor env. With one exception, gag sequences were always present. Analysis of molecular clones derived from integrated proviral CAEV genomes from the same infected cells showed a similar spectrum of deletions. The presence of transcriptionally active elements within the U3 domain of the defective genomes, as well as cis-acting elements within the leader sequences known to be required for efficient encapsidation of viral RNA, suggested that the defective viral DNA genomes could be transcribed into defective RNA molecules which could then be packaged into virions. lsopycnic density gradient centrifugation of supernatants of infected cell cultures indicated the presence of particles with densities less than that expected for intact virions (1.16 g/cc). Northern analysis revealed the presence of smaller viral-specific RNAs that lacked env sequences. These data, along with the structures of the molecular clones, suggest that CAEV stocks contain particles with defective genomes. The role of these particles in influencing the course of virus infection remains to be determined.

The lentiviruses induce chronic, degenerative diseases in their hosts. The major characteristics of lentiviral infections are a long latent period, virus persistence and spread in the face of a strong host immune response, and a slow progression of pathological events (1, 2). Although the basis for viral persistence is unknown, one possible mechanism might involve down-regulation of viral gene expression which could enable virus-infected cells to evade host defense mechanisms. There is also evidence that antigenic variation could account for virus persistence. Caprine arthritis encephalitis virus (CAEV) is a lentivirus that causes a slowly progressive multiorgan disease characterized by leukoencephalomyelitis, pneumonia, and arthritis in goats (1, 2). In attempts to isolate infectious molecular clones of CAEV derived from unintegrated viral DNA, a large number of clones were obtained but all contained deletions. RNAvirus genomes commonly undergo rapid mutation during repeated undiluted passage, generating subgenomic deletion mutants (3, 4). Deletion mutants have also been observed to arise

spontaneously at relatively high frequency in retrovirus infections (5-12). There is evidence that retroviralinduced immunodeficiency diseases of mice (13, 14) and domestic cats (15- 17) are induced by defective viral genomes. The circular unintegrated form of the CAEV genome was molecularly cloned from Hirt extracts of CAEVinfected cells after co-cultivation with uninfected tahr cells (see legend to Fig. 1). Sixteen CAEV-containing clones were isolated. Insert sizes (1.5 to 8 kb) were all less than that expected for the full-length CAEV provirus (10 kb) (18). Detailed restriction enzyme and Southern blotting analyses using probes representing various regions of the CAEV genome (18) (Fig. 1) demonstrated that all clones contained one copy or less of LTR sequences. The 3’ boundary of most of the deletions mapped within U3 and the 5’ boundary of the deletions extended upstream for variable distances toward gag. Several clones (51, 59, 79, 127, 45, and 56) lacked detectable LTR sequences. We sought next to determine whether deleted CAEV genomes were integrated into host genomic DNA. Since the majority of CAEV DNA is unintegrated under the conditions described above, it was necessary to enrich the cultures for cells containing integrated forms. This was accomplished by establishing a persistent infection. Thus, CAEV was added to tahr cells at a multiplicity of infection (m.o.i.) of 1 PFU/cell, and after 7 days the surviving cells were diluted 1:2 and

’ Present address: Universitk du Quebec 6 Mont&al, Dkpartment des Sciences Biologiques, P.O. Box 8888, Station A, Mont&al, QuBbec, Canada, H3C 3P8. ’ Present address: Advanced Biotechnologies, Inc., 12150 Tech Road, Silver Spring, MD 20904. 3 To whom correspondence and reprint requests should be addressed.

0042-6822192 $5.00

344

SHORT COMMUNICATIONS

I

I I

2 1

3 1

4 1

5 ’

8 ’

7 ’

8 ’ P

?,Kb K

P

Cl.51 Xb -----------d

Cl.59 Cl.79 Cl.127 Cl.91 Cl.55 Cl.140

----d

“ev----------XbSBaR

Ba

XbSBaPs

8a

XbSBaPs

Ba BX

XbSBaPs

Ba

Cl.116 Cl.45 Cl.93 Cl.56 Cl.142 Cl.118

---d

--_--_-_--------

Xb

-----------------------____---_---_--------KPBXb

8 --P 8

Cl.124b Cl.lZ&

______

Xb

E

PKPXb

PE EPsBX ---xi-

XbSBaPs 8a ------------------

8

XbSBaR

8

8a

E

PE EPspX

KP

Xb

-----IF XbSBaPs

Ba

B

XbSBaR

8a

B

XbSBaR

Ba

8

E

E

E ERBXKEBa r WI

Xb 1

PE ERBXKEBa -----------I

XbSBaR

Ba

8

E

PE EPs8X r/

XbSBaR

Ba

B

E

PE E Pa8XKEBa

Xb

KE

FIG. 1 Physical maps of CAEV clones derived from the unintegrated form. Cells acutely infected with CAEV are a poor source of unintegrated viral DNA(Yanivetal., unpublished observations). However, suitable yields can be obtained when uninfected cells are added to 7-day-old cultures of acutely infected cells (18). Molecular cloning of unintegrated CAEV DNA was performed as follows: Hirt supernatants were prepared and subjected to sucrose gradient centrifugation in order to enrich for circular unintegrated genomes. Gradient fractions were analyzed for the presence of viral sequences by Southern blotting using a CAEV DNA (clone 1244) probe (18). Viral DNA was digested with Xbal which cuts once within the viral genome (18); the linearized molecules were inserted into the unique Xbal site of bacteriophage XWesB, packaged, and plated onto fscherichia co/i strain BNN45; and recombinant plaques were detected using a CAEV probe. Restriction sites were identified by Southern analysis and hybridization to subgenomic probes representing various regions of the viral genome (a) nucleotide (nt) 601 to 1306, (b) nt 5401-6348, (c) nt 7724-8018, (d) nt 8018-8940, (e) nt 89409 167, (f) nt 1855525, where + 1 represents the transcription inltlatlon site (29). For these analyses cloned double-stranded DNA probes were labeled by using the nick-translation kit of Amersham and [am3’P]dCTP (3000 Ci/Mmol). Anti-sense nboprobes were syntheslzed from pGEM subclones (Promega) by using SP6 polymerase and [L~-~‘P]UTP (3000 Ci/Mmol) after llnearizatlon with an appropriate restriction enzyme. Anti-sense riboprobes representing U5 or

345

trypsinized. This routine was continued for 2 months. DNA from these cultures was digested with HindIll, an enzyme that does not cleave CAEV, and cloned by using X 2001. Of 33 clones detected with a CAEV probe, 16 contained genome-length inserts, 8 contained inserts of 2.5 to 8 kb, and 9 contained inserts largerthan 10 kb. Since this analysis was performed by using fagl, which cleaves once in each LTR, the larger-thangenomic size of the latter group could be due to loss of one LTR and, thus, one Eagl site. None of the clones yielded infectious virus upon transfection of tahr cells. Five representatives of each group were analyzed in more detail by using the phage X Quik-kit (Collaborative Research, Inc.) as well as with Hindlll and Bglll, which produce internal fragments. Southern blots were then hybridized to probes representing gag, pal, env, and LTR sequences. These analyses (Fig. 2) indicated that the deletions were not random but followed a pattern similar to that of the unintegrated CAEV clones. Thus, clone 26 contained the 5’ LTR, gag, PO/, and env regions, but did not contain a second copy of the LTR; clone 55 contained the 5’ LTR, gag, and pol regions but was missing env sequences; and clones 31, 36, and 46 contained the 5’ LTR and gag sequences, but lacked PO/, env, and the 3’ LTR. To confirm that these clones were derived from deleted integrated proviruses, they were examined for the presence of host flanking sequences. Hybridization of the same Southern blots to (32P)-labeled high molecular weight DNA of noninfected tahr cells indicated that clones from each group (31, 36, 46, and 55) were linked to host flanking sequences (Fig. 2). These data suggested that these deleted clones were derived from integrated CAEV proviruses and not from autointegrated molecules as suggested to exist in HIV-infected cells (19). Southern blotting of DNA isolated from infected cultures indicated size heterogeneity of integrated proviruses. Thus, no discrete virus-specific bands were detected, but a faint smear was present (but not in uninfected cells) which extended below the size expected for an intact, integrated genome (10 kb, assuming minimal amounts of host flanking sequences) (data not shown). In addition to size heterogeneity, we attribute the difficulty in detecting individual bands to the small

U3 sequences (30) were prepared by using the polymerase chain reaction (PCR) to ampltfy these regions from CAEV clone 1244 (19) and the resulting fragments were subcloned Into bluescnpt KS vettors (Stratagene). The maps of the various cloned genomes are allgned with that of the complete CAEV genome permuted at the Xbal site (18). Dashed lines indicate the extent of the deletions. B, Bglll; Ba, Ball; E, EcoRI; K, Kpnl; P. Pvull; Ps, Pstl; S, Sphl; X, Xhol; Xb, Xbal.

346

SHORT COMMUNICATIONS

CAEV abcde

LTR abcde

GAG abcde

POL abcde

ENV abcde

HOST abcde

23.1 9.6 6.4

2.3 2.0

1.3 1.0 0.9 FIG. 2. Southern analysis of CAEV clones derived from high molecular weight DNA. Integrated CAEV molecules were cloned from high molecular weight DNA (37) of tahr cells persistently infected with CAEV. The DNA was digested with HindIll [which does not cut within the viral genome (Is)], fractionated on a sucrose gradient, and ligated to HindIll-digested bacteriophage X 2001 DNA, packaged, and plated on E. co/i strain LE392. Recombinant phage plaques were identified as described above. DNAs derived from CAEV clones 26 (a), 55 (b), 31 (c), 36 (d), and 46 (e) were digested to completion with HindIll and /3g/ll. Southern blots were hybridized with a probe representing the entire CAEV (clone 1244) (18) genome (probe CAEV), riboprobes representing nucleotides (29) -176 to 46 (probe LTR), nucleotides 1 136 to 1674 (probe GAG), nucleotides 3897 to 4025 (probe POL), nucleotides 7142 to 7959 (probe ENV), and a probe (probe HOST), representing the host flanking sequences. The insert in the lower left corner is a photograph of the ethidium bromide-stained gel. Numbers indicate positions of sizes standards in Kb.

fraction of integrated genomes present in chronic CAEV infections (18). Furthermore, the cell cultures are not clonally derived and represent a mixed population of infected and uninfected cells, the latter being added at intervals to sustain the cytopathic infection. Defective integrated genomes would not necessarily influence the course of virus infection if they were transcriptionally silent. This could be the case for some of the clones described above, since restriction analysis showed deletions of LTR U3 sequences in several clones. In order to establish whether such an LTR retained its transcriptional potential, the U3 region of the LTR of a representative clone (91) was inserted upstream of U5 sequences in a construct containing U5R sequences linked to the chloramphenicol acetyltransferase gene (CAT) (20). This plasmid, designated pCAEV-A U3R UWAT was used to transfect tahr cells and the levels of CAT activity were compared to those of cells transfected with a construct containing the complete CAEV LTR [pCAEV LTR-CAT (20)]. The data shown in Fig. 3 support the conclusion that the sequences retained in the deleted U3 of clone 91 were fully transcriptionally active.

Packaging of viral RNA into virions depends on the presence in cis of encapsidation signals of 150 to 450 bp that have been localized to a region downstream of the 5’ LTR in the vast majority of retroviruses [reviewed in (21)]. Moreover, in several cases it was shown that sequences extending into the gag region were responsible for the increased efficiency of the packaging of viral RNA. The presence of the 5’ leader as well as the gag region in the defective CAEV genomes would allow packaging of these defective RNA molecules. Experiments were therefore performed to establish the existence of defective virions in CAEV stocks. Defective particles have been described in most virus isolates (3). Since successive passages at high m.o.i. favor the generation of defectives, CAEV was serially passaged 16 times at high m.o.i. (50 PFU/cell). Defective particles of RNA viruses are characterized by the presence of nucleocapsids of subgenomic size which have a lower buoyant density (22). Supernatants were collected from infected cultures at the 1st and the 16th passages. Virions were pelleted and subjected to isopycnic density centrifugation and the fractions tested for RT activity (Fig. 4). A broad profile of several peaks,

347

SHORT COMMUNICATIONS

0.5

1

2

4

6

0.5

1

2

could be detected, most of the short RNA species that hybridized to the U5 probe did not show up by using the U3 probe. This suggested that the short RNA species presumably contained RU5 at the 5’ end, but did not contain the 3’ regions of the genomic RNA. There was one species (Fig 5, lane b) that hybridized to the env but not the U5 probe. Since other subgenomic probes failed to detect this species it may have been generated by nuclease cleavage or from transcription

FIG. 3. Transcriptional activity of the deleted CAEV LTRs. LTR sequences derived from CAEV clones were inserted into a plasmrd containrng the chloramphenicol acetyltransferase gene (CAT) (32), transfected (33) into tahr cells, and assayed for CAT activity (34). Specifrcally, the Ball-Kpnl fragment of the deleted clone 91 (see Fig. 1) was inserted in the Kpnl site (located within the R of CAEV LTR) of pCAEV LTR-CAT (30) thus generating a plasmid, pCAEV AU3RU5CAT (A), in which the upstream U3R region was replaced by the homologous, although deleted, U3R region of clone 91. Each plasmrd, pCAEV AUBRU5CAT, and pCAEV LTR-CAT (5) was introduced In increasing amounts into tahr cells, and 48 hr later cells were harvested and cell extracts (50 pg protein) were analyzed for CAT activity. Numbers Indicate fig DNA used for transfectron.

with the major ones at the densities of 1.15 1 and 1.142 g/cc, was observed in virion preparations from either the 1st or 16th passage. In order to establish whether this banding pattern was due to particles containing subgenomic RNA species, virions from gradient fractions were pelleted and the viral RNA was extracted and subjected to Northern analysis (Fig. 5). A major band was detected (35s) along with several smaller species. The same pattern was observed with virion RNA extracted at the 24th passage (data not shown). Since the same pattern was seen in virions of various passages, different RNA preparations and in various conditions of gel analyses (data not shown), it is unlikely that the smaller species were degradation products. Assuming that the small RNA species are the transcripts derived from defective viral genomes, differential patterns of hybridization with subgenomic CAEV probes should be detectable. Small RNA species derived from transcriptionally active packagable genomes should contain 5’leadersequences. The restriction maps of the majority of clones derived from integrated or unintegrated DNA predict that most shorter RNA species will lack env sequence. Therefore, gradient RNA preparations were first hybridized to an env probe and then to a U5 probe (following exposure and stripping of blots). As seen in Fig. 5a, all CAEV virion RNA species hybridized to the U5 probe. In contrast, the short subgenomic species failed to hybridize to the env probe (lane b). Subsequent hybridization of the stripped blots to the U3 probe (data not shown) showed that although the 35s genomic RNA species

0' 0

I

I

/

10

20

30

FRACTION

1.20

g

1.18

5

1.16

?

1.14

Q

1.12

2

NO.

FIG. 4. lsopycnrc sedimentation profile of CAEV vinons in sucrose gradients, Tahr lung cells obtained from the Amencan Type Culture collection (ATCC 6277) were grown in Dulbecco’s modified Eagle’s medium (DMEM) and supplemented with 10% fetal calf serum (FCS). Infections were performed by incubating subconfluent tahr cells for 3 hr at 37” in medium containing 10 pg/ml polybrene and then with CAEV (1 PFU/cell) In DMEM containing 2% (FCS) and 8 pg/ml polybrene for 2 hr at 37”. Unadsorbed virus was then removed and fresh medium containrng 10% FCS was added. Sequential passaging of undiluted CAEV-containing supernatants was Initiated by Infecting cells at 50 PFUkell, according to the procedure described above. Three days later, viral particles were harvested from the culture medium by centnfugatron, suspended in DMEM, and added to uninfected cells at 50 PFUkell. This procedure was repeated at 72.hr Intervals until the vrrus had been passaged 16 to 24 times. Cell-free culture supernatants were collected at 2-hr intervals from Infected cells and the pelleted virions were layered onto 20-55% (wtiwt) SW crose gradients and centrifuged at 95,000 g for 17 hr. Fractions (1 ml) were collected dropwise from the bottom. Reverse transcnptase (RT) assays were performed on 25.~1 samples of each fraction, using oligo(dT),,_,,:poly(rA) and (3H)TTP. After Incubation for 30 min at 37”, acid precrprtable radroactivrty was determined.

348

SHORT COMMUNICATIONS

a

b

- 28s

- 18s

FIG. 5. Northern blot analysis of virion RNAs. Virions isolated from the sucrose gradient described in the legend to Fig. 4 (fractions 22 to 30) were lysed with a solution containing 1% sodium dodecyl sulfate (SDS) and 100 pg/ml of carrier tRNA and then immediately extracted with phenol and chloroform. RNA was precipitated with ethanol and analyzed by Northern blotting (35) except that a formaldehydeagarose vertical gel (36) run at 15OV was used. The 28s and 18s ribosomal RNAs were used as size standards. The viral RNA applied to each lane corresponds to the total amount isolated from 100 ml of tissue culture supernatant medium. The blot was hybridized to an env probe representing nucleotides 7142 to 7959 (b). Following hybridization, the probe was stripped and rehybridized to the U5 probe representing nucleotides 86 to 163 (a). Positions of 28s and 18s ribosomal RNAs, as revealed by ethidium bromide staining, are marked by bars.

initiation within env. Attempts to isolate the cDNA corresponding to this transcript have been unsuccessful. It should be noted that although less virion RNA could be isolated following passages at high m.o.i., the subgenomic RNA species did not seem to be overrepresented (data not shown). In summary, our attempts to isolate molecular clones of the unintegrated circular form of the CAEV genome yielded only deleted molecules. Restriction enzyme analysis of 16 clones revealed that none of the clones contained two LTRs. With one exception, gag sequences were retained but the 5’-end points of the deletions varied greatly. The 3’-end points, in nine clones, were located either at the 5’ border of the remaining LTR or within. In six clones, the entire LTR was deleted. The existence of deleted unintegrated forms has been also reported for other retroviruses. For example, in the avian system deletions extended from the 5’ LTR toward the 5’ border of gag (23, 24). In Mo-

MULV (25), a significant number of the deletions started at the LTR and extended further downstream toward gag and PO/. In the case of human adult T-cell leukemia virus (HTLV-I), most deletions also started at the 5’ LTR and extended downstream as far as env( 12). A different pattern was observed in studies of visna virus (26). Thus, the 5’ borders were localized to gag, PO/, or env and the 3’ borders to the 3’ LTR. These deletions are more similar to those we observed for CAEV. One other difference is that while only 50% of the cloned unintegrated circular forms of Rous sarcoma virus (27) and 20% of the cloned MO-MULV DNAs (25) were deleted, all CAEV and most visna virus clones (26) were extensively deleted. Analysis of molecular clones of the integrated form of the CAEV genome revealed a similar pattern of deletions. Braun et al. (1988) (27) also reported the existence of a high proportion of defective proviruses in cells infected with bovine immunodeficiency-like virus. The mechanism by which deletions are introduced both into unintegrated and integrated forms of the CAEV genome is unknown. The precursor for the integrated form of MoMuLV was shown to be a linear molecule containing two LTRs (28). If CAEV DNA integrates by the same mechanism, then the deletions must occur upon joining LTR sequences to host target sites since none of the clones of integrated CAEV genomes contained two LTRs. Similar nucleolytic events could take place upon circularization of unintegrated linear DNA. The presence of deleted genomes could likely affect the course of viral infection only if they were transcribed, packaged, and transmitted to uninfected cells. The majority of clones contained LTR and gag sequences so it is probable that they could be transcribed and encapsulated. LTR sequences of a clone derived from unintegrated DNA were found to be capable of driving the transcription of a reporter gene. Consistent with this is our finding that virion preparations contained short, viral-specific RNA species that lacked env sequences as did most of the molecular clones. It was surprising to find, however, that the proportion of subgenomic species did not increase following serial passage of virus at high m.o.i. In order to determine with certainty how these defective genomes are generated and to what extent they influence the course of CAEV-induced disease, it will be necessary to isolate a full-length, infectious CAEV molecular clone. ACKNOWLEDGMENTS We thank search was U.S.-Israel Agency for

S. Aaronson for continued advice and support. This resupported by Grant DPE-5544.G-SS-6007-00, from the Cooperative Development Research Program, U.S. International Development.

SHORT COMMUNICATIONS

REFERENCES 1. CHEEVERS,W. P., and MCGUIRE, T. C. (1988). The Lentiviruses: MaediiVlsna, Caprine Arthritis-Encephalitis and Equine Infectious Anemia. Adv. virus Res. 34, 189-215. 2. NAW\YAN, 0. (1990). lmmunopathology of lentlviral infections in Ungulate animals. Curr. Opinion Immunol. 2, 399-402. 3. BARRETT,A. D. T., and DIMMOCK. N. J. (1986). Defective interfering viruses and infections of animals. Current Top. Microb,b/. lmmunol. 128, 55-84. 4. WHITAKER-DOWLING,P., and YOUNGNER,J. S. (1987). Viral interference-dominance of mutant viruses over wild type virus In mixed infections. Microbial. Rev. 51, 179-l 9 1, 5. SHIELDS, A., WITTE, 0. N., ROTHENBERG,E., and BALTIMORE, D. (1978). High frequency of aberrant expression of Moloney murune leukemia virus in clonal infections. Cell 14, 601-609. 6. DIERKS. P. M., HIGHFIELD, P. E., and PARSONS,I. T. (1979). Deletion mutant of the Bratislava-77 strain of Rous sarcoma virus containing a fusion of the group-specific antigen and envelope genes. J. Wol. 32, 567-582. 7. SHOEMAKER,C. S., GOFF, S., GILBOA. E., PASKIND, M., MITRA, S., and BALTIMORE, D. (1980). Structure of a cloned circular Moloney murine leukemia virus DNA molecule containing an inverted segment: implications for retrovirus integration. Proc. Nat/. Acad. Sci. USA 77, 3932-3936. 8. VERMA, I. M., LAI, M. H. T., BOSSELMAN,R. A., MCKENNET, M. A., FAN, N., and BERNS, A. (1980). Molecular cloning of unintegrated Moloney sarcoma virus DNA in bacteriophage X. Proc. Nat/. Acad. Sci. USA 77, 1773-l 777. 9. YOSHIMURA, F. K., and YAMAMURA, J. M. (1981). Four Moloney murine leukemia virus-infected rat cell clones producing replication-defective particles: protein and nucleic acid analysis. 1. Viral. 38, 895-905. 10. RASSART,E., DESGRSEILLERS,L., and JOLICOEUR,P. (198 1). Molecular cloning of B- and N-tropic endogenous BALB/c murine leukemta virus circular DNA intermediates: isolation and charactenzatlon of Infectious recombinant clones. /. Virol. 39, 162-171. 11. HOLLAND, J., SPINDLER,K., HOWDYSKI,F., GRABAU, E., NICHOL, S., and VANDEPOL, S. (1982). Rapid evolution of RNA genomes. Science 215, 1577-l 585. 12. HIRAMATSU, K., and YOSHIJURA,H. (1986). Frequent partial deletion of human adult T-cell leukemia virus type I proviruses in experimental transmission: pattern and possible Implication. 1. Viral. 58, 508-512. 13. HARTLEY,1. W., FREDERICKSON,T. N., YE~ER, R. A., MALINO, M., and MORSE, H. C. (1989). Retrovirusinduced murine acquired lmmunodeficiency syndrome: natural history of Infection and differing susceptibility of Inbred mouse strains. /. tirol. 63, 1223-1231. 14. AZIZ, D. C., HANNA, Z., and JOLICOEUR,P. (1989). Severe immunodeficiency disease induced by a defective murine leukemia virus. Nature 338, 505-508. 15. MULLINS, 1. I., CHEN, C. S., and HOOVER, E. A. (1986). Diseasespecific and tissue-specific products of unintegrated feline leukemia virus variant DNA In feline AIDS. Nature 319, 333336. 16. OVERBAUGH,J., DONAHUE, P. R., QUACKENBUSH, S. L., HOOVER, E. A., and MULLINS, 1. I. (1988). Molecular cloning of a feline leukemia virus that induces fatal immunodeficiency disease in cats. Science 239, 906-9 10. 17. Poss, M. L., MULLINS, J. I., and HOOVER, E. A. (1989). Posttranslatlonal modifications distinguish the envelope glycoprotein of the immunodeficiency disease-inducing feline leukemia virus retrovirus. /. Lhol. 63, 189-195.

349

18. YANIV, A., DAHLBERG,1. E., TRONICK, S. R., CHIU, I. M., and AARONSON, S. A. (1985). Molecular cloning of Integrated caprine arthritis encephalitis virus. virology 145, 340-345. 19. PAUZA, C. D., and GALINO, J. (1989). Persistent human immunodeficiency virus Type I infectjon of monoblastoid cells leads to accumulation of self-integrated viral DNA and to production of defective virions. /. Viral. 63, 3700-3707. 20. SHERMAN, L., YANIV, A., LICHTMAN-PLEBAN, H., TRONICK, S. R., and GAZIT, A. (1989). Analysis of regulatory elements of the equine infectious anemia virus and capnne arthritis encelphalitis virus long terminal repeats. /. Lhrol. 63, 4925-4931. 27. LINIAL, M. L., and MILL, A. D. (1990). Retroviral RNA packaging: sequence requirements and Implications. Current Top. ML crab/o/. Immunol. 157, 125-l 52 22. REICHMANN, M. E., and SCHNITZLEIN. W. M. (1979). Defective Interfering particles of rhabdovlruses. Cur. Topcs Microbial. Immunol. 86, 123-l 62. 23. OLSEN, J. C., and SWANSTROM, R. (1985). A new pathway in the generation of defective retrovirus DNA. /. V/ro/. 56, 779-789. 24. OLSEN, J. C., BOVA-HILL, C., GRANDGENE~, D. P., QUINN, T. P , MANFREDI, J. P., and SWANSTROM, R. (1990). Rearrangements In unintegrated DNA are complex and are the result of multiple genetic determinants. /. Viral. 64, 5474.-5484. 25. SHOEMAKER, C., HOFFMANN, J., GOFF, S. P., and BALTIMORE, D. (1981). Intramolecular Integration within moloney murine leukemia virus DNA. /. Vkol. 40, 164-l 72. 26. MOLINEAUX, S., and CLEMENTS, J. E. (1983). Molecular cloning of unintegrated visna viral DNA and characterization of frequent deletions in the 3’ terminus. Gene 23, 137--l 48. 27. BRAUN. M. J., IAHN, S.. BOYD, A. L., KOST, T. A., NAGASHIMA, K., and GONDA, M. A. (1988). Molecular cloning of biologically active proviruses of bovine immunodehciency-like virus. Wrology 167, 515-523. 28. BROWN, P. 0. (1990). lntegratlon of retrovlral DNA. Current Top. Microbial. lmmunol. 157, 19-48. 29. SALTARELLI, M., QUERAT, G., KONINGS, D. A. M., VIGNE, R., and CLEMENT% J. E. (1990). Nucleotide sequence and transcriptlonal analysis of molecular clones of CAEV which generate infectious virus. Virology 179, 347-364. 30. SHERMAN, L., GAZIT, A., YANIV, A.. DAHLBERG,J E., and TRONICK, S. R. (1986). Nucleotlde sequence analysis of the long term!nal repeat of integrated caprlne arthntls encephalitis virus. Virus Res. 5, 145-l 55. 37. BLIN, N., and STAFFORD, D. W. (1976). A general method for isolation of high molecular weight DNA from eukaryotes. Nucl. Acids Res. 3, 2302-2380. 32. GORMAN, C. M., MERLINO, G. T., WILLINGHAM, M. C., PASTAN, I., and HOWARD, B. H. (1982). The Rous sarcomavirus long termlnal repeat IS a strong promoter when introduced into a variety of eukaryotic cells by DNA mediated transfectlon. hoc. Nat/. Acad. SC;. USA 79, 6777-6781. 33. GRAHAM, F. L., and VAN DER EB, A. J. (1973). A new technique for the assay of Infectivity of human adenovlrus 5 DNA. Virology 52,456&467. 34. SHERMAN, L., GAZIT, A., YANIV, A., KAWAKAMI,T., DAHLBERG,J. E., and TRONICK, S. R. (1988). Localizatton of sequence responslble for trans.activation of the equine infectious anemia virus long terminal repeat. /. Viral. 62, 120.. 126. 35. NOIMAN, S., YANIV, A., SHERMAN, L., TRONICK, S. R., and GAZIT, A. (1990). Pattern of transcrtption of the genome of equine Infectlous anemia virus. J. Viroi. 64, 1839-l 843. 36. AMBULOS, JR., N. P., DUVAL, E J., and LOVETT. P. S. (1987). Methods for blot-hybridization analysis of mRNA molecules from Bacillus subtik. Gene 51, 281~ 286.