Complementation between SV40 and RFV defectives and acquisition of SV40 origins by late RFV genomes

VIROLOGY

154,344-356 (1986)

Complementation between SV40 and RFV Defectives and Acquisition of SV40 Origins by Late RFV Genomes FRANK J. O’NEILL,’ THOMAS H. MILLER,

AND

ROBERT STEVENS

Research Service, VA Medical Center, and University of Utah, Department of Cellular, Viral and Molecular Biology, and Department of Pathology, Salt Lake City, Utah 841@ Received April

L&1986; accepted July 8, 1986

EL SV40 and RFV are variants of SV40 and BKV which contain bipartite or dual genomes. One molecule contains all the early viral sequences (E-SV40, E-RFV) and the other all the late viral sequences (L-SV40, L-RFV). Early and late genomes complement one another during productive infection. Experiments were designed to determine if E-genomes of one virus could complement L-genomes of another virus. If complementation did occur, intermolecular recombination events which lead to a more efficient infection or an altered host range might occur, and the sequences involved could then be identified. Two combinations were generated by direct transfection of BSC-1 green monkey cells. E-RFV and L-SV40 DNA complementation resulted in hybrid virus growth and cell killing. The hybrid demonstrated a narrow host range. Following serial passage, some E-RFV genomes contained SV40 origin region sequences but these recombinants did not overgrow prototype E-RFV genomes, even after many virus passages. In addition, no significant alterations in host range could be detected. Complementation between E-SV40 and L-RFV yielded a virus with a relatively wider host range. Virus growth and cell killing appeared very slowly at first. However, with each passage of E-SV40/L-RFV, cell killing occurred progressively more rapidly, until passage 7 when it became extensive in ‘7 days rather than 6-8 weeks. Infected cells contained lo-20 times more E-SV40 than L-RFV DNA during the first passage. However, by passage 7, both genomes were equally represented. During serial passage, LRFV DNA acquired SV40 sequences from around the origin and the terminus of replication, such that recombinant (r) L-RFV genomes contained 33 SV40 origins (including the 72bp repeat) and 2 termini, and prototype L-RFV DNA was lost. E-SV40/rL-RFV demonstrated an altered host range propagating in some cell lines which did not support ESV40/L-RFV growth. Both the host range change and the increased growth of rL-RFV genomes were shown to be at least partly caused by the acquisition of the SV40 sequences. 0 19% Academic

Press, Inc.

and MGV which are variants of BKV, and JCV(HEK), and EL SV40 (Pater et d, 1980, 1981; Yoshiike et al, 1982; O’Neill et al, 1982). In these variants, one molecule contains the entire early region and the other the entire late region but each molecule resembles the genomes of defective interfering particles. They contain reiterations of the viral origin and reiterations of either the viral terminus or of cellular sequences (Brockman et a& 1973; Mertz et cd, 1975; Carroll and O’Neill, 1978). Early and late genomes complement each other during the infectious cycle. It has been suggested that the containment of the viral DNA in two molecules resembles that seen for the seg-

INTRODUCTION

The DNA of polyomaviruses ordinarily exists as a single supercoiled circular molecule (Tooze, 1980). These genomes contain both early and late sequences, the expression of which is controlled by the interaction of viral coded T antigen with the origin of replication and the influence of adjacent enhancer sequences. Recently, several variants have been identified which contain bipartite or dual genomes, each of which is a supercoiled circle. These include RFV ‘To whom dressed.

requests

for

reprints

0042-6822/86 $3.00 Copyright All rights

0 1986 by Academic Press, Inc. of reproduction in any form reserved.

should

be ad-

344

COMPLEMENTATION

BETWEEN

SV40 AND

RFV

DEFECTIVES

345

(HEK, MG178). Some are permissive for both SV40 and RFV (HFB and BSC-1; see O’Neill and Miller, 1985). E- and L-SV40 genomes were previously cloned into the Pet1 site of pBR322 (O’Neill et aL, 1982). Eand L-RFV genomes were secured from Dr. Mary M. Pater and had been cloned into the EcoRI site of pML, a derivative of pBR322 (Pater et al, 1983). Viral DNAs were prepared from plasmid clones by digestion with PstI and EcoRI and then ligated as described (Maniatis et al, 1982). Transfectim Confluent monolayers of BSC-1 cells were cotransfected with ERFV/L-SV40 and E-SV40/L-RFV combinations by modifications (Kawai and Niskizawa 1984; O’Neill and Miller, 1985) of the DEAE-dextran method (Sompayrac and Danna, 1982). Anal@s of viral DNAs from Hirt supernatants. Viral DNAs were extracted from virus infected cells and displayed by electrophoresis in agarose gels (Carroll and O’Neill, 1978). Because of the known sizes of the various viral genomes, it was sometimes possible to identify each viral DNA band by simple inspection of agarose gels. However, it was often necessary to digest the DNA with one or more restriction endonucleases in order to positively identify the viral genomes and their relative proportions. EL-SV40 and RFV genomes can be distinguished because they are uniquely susceptible to digestion with different enzymes. For instance, &$I cleaves EL-SV40 at the origin but does not cleave EL-RFV genomes. In contrast &$I1 and SstI cleave both E- and L-RFV genomes and do not MATERIALS AND METHODS cleave E- or L-SV40 DNAs. In addition, SstI Virus, cells, and DNA. EL-SV40, and cleaves the L-RFV genome near the origin RFV have been described (Pater et al, 1980; of replication. O’Neill et al., 1982). Both viruses were T Antigen assay. T antigen was assayed propagated in green monkey BSC-1 cells. by indirect immunofluorescence, using RFV was also propagated in human em- serum from Syrian hamsters bearing tubryonic kidney (HEK) cells or human fetal mors induced by SV40 or BKV transformed brain (HFB) cells. Other cell lines used hamster cells (O’Neill and Carroll, 1978, were A172 and MG178 which are human 1983). glioblastoma cell lines; MBA9812, a human Molecular cloning of defective genornes. bronchogenic carcinoma cell line; and TC- Recombinant L-RFV genomes (rL-RFV) ‘7, a green monkey kidney cell line. These which appeared after serial undiluted pascell lines and cell strains (O’Neill and Car- sage of E-SV40/L-RFV were cloned into roll, 1981) are permissive for SV40 (TC-‘7, the EcoRI site of pBR322 using convenBSC-1 and MBA9812) or RFV and BKV tional methodology (Maniatis et al, 1982; mented genomes of influenza and reoviruses and could allow for a more rapid evolution and the generation of diversity (O’Neil1 and Miller, 1985). From this it may be predicted that new viruses could be generated if an early defective of one virus could complement the late defective of another virus. We may further predict that, following complementation, recombination between the different genomes may occur to generate additional diversity. Correlation of functional diversification with the sequences involved in intermolecular recombination could help to assign functions to specific viral sequences. The recombinational events can be permitted to occur naturally in coinfected cells rather than by constructing SV40/RFV recombinants in vitro which would be technically more difficult. In a previous report, we demonstrated that defective early and late SV40 genomes could interact with wtBKV. Following the interaction of L-SV40 and wtBKV, the wtBKV genomes were lost and replaced by a defective, E-BKV, which complemented L-SV40 to produce a new hybrid papovavirus (O’Neill and Miller, 1985). In the present report, we demonstrate complementation between E-RFV and L-SV40, and E-SV40 and L-RFV. In addition, during serial passage of the hybrid E-SV40/L-RFV, the L-RFV genomes acquire SV40 origin regions and this apparently results in an increased growth capacity of RFV genomes and in an alteration of host range.

346

O’NEILL,

MILLER,

O’Neill et al, 1982;O’Neill and Miller, 1985). E-SV40 genomes from passage 7 E-SV40/ rL-RFV infected BSC-1 cells were also cloned into pBR322 at the PstI site. Cloned genomes were assayed for biological activity by digestion with EcoRI or PstI, recircularization with T4 DNA ligase and cotransfection into BSC-1 cells with complementing viral genomes. Several rL-RFV clones were produced. Two clones were analyzed by Southern blot hybridization and restriction endonuclease digestion. Both clones were infectious in BSC-1 cells when cotransfected with E-SV40 genomes. Southern blot hybridization. These experiments were performed as described previously (O’Neill and Miller, 1985) with the following exceptions or important modifications. Molecularly cloned or Hirt supernatant viral DNAs were nicked, using DNAseI digestion, to approximately 50%. The DNAs were then labeled with [32P]dCTP by Escherichia coli DNA polymerase (holoenzyme) treatment to a specific activity of 10’ dpm. In some experiments 5 X lo7 dpm were used for labeling and this was sufficient as there were several thousand copies of viral DNA loaded onto agarose gels for subsequent hybridization to the probe. The DNA was transferred to zetabind blotting paper and then hybridized with the probe in 5X SSC and 50% formamide at 42”. Removal of SV40 origins from recombinant rLRFV DNA. pBR322 cloned rL-RFV DNA was digested with EcoRI (cloned into EcoRI site) to completion and then ligated with T4 DNA ligase. rL-RFV genomes were then digested with BglI to completion and circularized by T4 DNA ligase treatment. Religated rL-RFV genomes were shown to contain only one (as opposed to three) BgZI cleavage site as subsequent BglI digestion produced a linear rL-RFV genome and no BglI cleavage fragments appeared near the bottom of agarose gels. BglI, T4 DNA ligase treated rL-RFV genomes were then cotransfected with E-SV40 into BSC-1 cells. RESULTS

E-S UO/GRF V Transfection Following cotransfection of BSC-1 cells efthere was little evidence of cytopathic

AND

STEVENS

fects (CPE) until approximately 10 weeks. Virus from these cultures could be serially passed to fresh cells and with each passage CPE appeared more rapidly, until passage 7 when it occurred in seven days. Hirt supernatant viral DNA was analyzed on agarose gels (Fig. 1). At passage 1 the E-SV40 DNA was in excess of L-RFV genomes by at least lo-fold. However, at passage ‘7both genomes appeared in near equal amounts. Restriction endonuclease digestion dem-

E94O/L~RFV

FIG. 1. Identification of complementing E-SV40 and L-RFV (or rL-RFV) genomes in infected BSC-1 cells. For passage 1, BSC-1 cells were transfected with both DNAs isolated from molecular clones. After 2 months, when cell killing was extensive, newly formed virus was used to infect 5 X lo7 fresh BSC-1 cells. The viral DNA was then extracted and run on agarose gels, before and after cleavage with BglI or BgZII. In the uncut lanes the complementing genomes are difficult to distinguish but in the endonuclease digested lanes the genomes are clearly separated. BgZII digestion does not cut E-SV40 but some of the molecules are nicked (form II). BgZII linearizes the L-RFV genomes. Note the increase in size of the linear L-RFV genomes (rLRFV) in passage 7 and the increased intensity of the band (compare lanes 3 and 6). The increase in the size and relative amounts of rL-RFV DNA are also observed in the BgZI lanes. Arrowhead outside right margin identifies large BglI fragment of rL-RFV genomes in lane 5. The highest band in lane 2 (faint) appearing just below form II E-SV40 (lanes 1 and 3) is a partial BglI cleavage of E-SV40.

COMPLEMENTATION

BETWEEN

onstrated very significant alterations in LRFV genomes during serial passage of ESV40/L-RFV. At passage 3 the L-RFV genomes were linearized by Bglr digestion (not shown) and by passages 5-7, BglI digestion cleaved L-RFV several times (Fig. 1). Apparently all L-RFV genomes were cleaved by BglI digestion. This suggested that rL-RFV genomes had acquired one and then several SV40 origin-containing sequences from E-SV40 genomes. We also analyzed two pBR322 cloned rL-RFV genomes (generated from BSC-1 cells persistently infected with passage ‘7 virus) with a variety of other enzymes. SphI cleaves SV40 DNA within the 72-bp repeats on the late side of the origin and this region is considered to contain an enhancer function for late transcription (Benoist and Chambon, 1981; Gruss et aL, 1981). This region also contains host range determinants (Laimins et al, 1982, Kriegler and Botchan, 1983;Byrne et al, 1983). The EL-SV40 variant contains only a single 72-bp sequence in each of the three reiterations in each of the complementing genomes. EL-SV40 therefore contains three SphI sites in each genome. MspI (HpoII) cleaves SV40 DNA at position 0.725 map units, distal to the SphI site (0.686). This site is also contained within the reiterated origin regions of ELSV40 genomes. Prototype L-RFV DNA does not contain a MspI site. Analysis of rL-RFV genomes prior to and after pBR322 cloning demonstrated that they contained at least three copies of an origin-containing SV40 sequence spanning from a locus between the MspI and SphI sites (of E-SV40 DNA) to a locus on the early side of the SV40 origin (BgZI site), prior to the Hind111 site. The sequence containing the MspI site is not included but the sequence containing the SphI site is included in the SV40 DNA acquired by rL-RFV. In addition, the SV40 terminus (BumHI site) was reiterated in tandem with two of the three SV40 origin regions such that there were three SV40 origins and two SV40 termini. Two pBR322 clones of rL-RFV were analyzed by BgZII and SstI cleavage and both of these contain deletions of the already truncated E-RFV sequence. One of the recombinants is missing one of the two L-RFV origins. We

SV40 AND

RFV

DEFECTIVES

347

therefore conclude that the minimum size of each the SV40 origin-containing sequences present in rL-RFV genomes is 450 bp. Each of the two SV40 termini contribute 100 bp. In the two rL-RFV clones, the minimum amount of SV40 DNA acquired is approximately 1500 bp. Molecular

Cloning

From passage 7 E-SVRO/rL-RFV virus two rL-RFV clones and a single E-SV40 clone were analyzed extensively. Both rLRFV clones were infectious in BSC-1 cells when cotransfected with E-SV40 DNA. E-SV40 DNA originally cloned from ELSV40 infected cells (O’Neill et ak, 1982) or cloned from passage 10 E-SV40/rL-RFV infected cells could be used for complementation. This recloned E-SV40 DNA was compared to the original E-SV40 clone by restriction analysis and no alterations could be detected. Also, it did not hybridize to RFV probes in Southern analysis (not shown). Southern Blot Hybridization

In order to demonstrate that the apparent origin-containing sequences present in rL-RFV genomes were indeed derived from SV40, we hybridized SV40 and RFV probes to rL-RFV DNA immobilized on zetabind paper (Fig. 2). We used two rL-RFV genomes which we had cloned in pBR322. The SV40 probes hybridized to uncleaved rLRFV DNA clones but not to the prototype L-RFV DNA clone. When rL-RFV was cleaved with BglI, the SV40 probe hybridized to the BgZI fragments migrating near the bottom of agarose gels (-800-bp fragments). SphI digestion also generated origin containing fragments and these hybridized with the SV40 probe (not shown). In contrast, PvuII digestion cleaves outside the SV40 sequences of rL-RFV but cleaves near the L-RFV origin. The large PvuII fragment (-1700 bp) does not hybridize with the SV40 probe while the small fragment does. Also, when codigested with SphI (PvuII + SphI), the smallest fragment (300 bp) hybridizes with the SV40 probe (not shown). When prototype L-RFV is digested with PvuII or BgZI there is still no hybrid-

348

O’NEILL, EcoRl+4u

EC0 RI

EC0 RI + P”” II

l 0.ll

MILLER,

II EcoRI+B.I

I

L.RVV Probe

El940

Probe

FIG. 2. Southern blot hybridization of cloned L-RFV and rL-RFV genomes with EL-SV40 and L-RFV probes. L-RFV, rL-RFV (clones A and D) and pBR322 DNA were electrophoresed in 1% agarose gels. The cloned viral DNAs were linearized (but removed from pBR322 plasmids) or digested with enzymes noted above lanes. The DNAs were then transferred to zetabind filter paper, hybridized with an EL-SV40 DNA

AND

STEVENS

ization with the SV40 probe. (BglI does not cleave prototype L-RFV). These observations suggest that each of the three SV40 origin regions in rL-RFV contain one or a portion of one 72-bp repeat and that the reiterated SV40 sequences are located next to the remaining L-RFV origin. The SV40 origins do not interrupt the L-RFV late coding region, rather they replace one of the two L-RFV origins (clone A) or a portion of one L-RFV origin (clone D). In addition, there is an alteration in the L-RFV late region, near the L-RFV origin. This appears as an enlarged EcoRI/PvuII fragment (Fig. 2). This region of rL-RFV does not hybridize to SV40 probes (not shown). Similar results were obtained when rLRFV genomes from E-SV40 coinfected cells were analyzed prior to cloning, but were more difficult to interpret because of the presence of E-SV40 (not shown).

probe, and exposed on Kodak XAR-5 film (bottom panel). The probe was removed by NaOH treatment and the blot was then hybridized with a L-RFV plasmid probe (middle panel). The SV40 and L-RFV probes were nick translated to a specific activity of 1 X 10’ cpm (Rigby et al, 1977; Maniatis et al, 1982). The SV40 probe was generated from EL-SV40 DNA isolated from infected BSC-1 cells and the L-RFV probe was generated from L-RFV/pBR322 plasmids. White arrowheads in top panel and black arrowheads in lower panels indicate position of SV40 origin fragments. Note that the L-RFV probe does not hybridize to this band. The small arrows indicate the BglI/PvuII fragment in rL-RFV which contains the SV40/L-RFV junction. This fragment hybridizes to both probes. The larger bands indicated by brackets contain SV40 sequences not released by BgZI digestion. A portion of these sequences appear in the smallest fragments (small arrows) after BgZI/PvuII cleavage. As expected the bracketed bands hybridize to both probes. The last lane contains an E-SV401pBR322 plasmid digested with EcoRI + BglI. The bottom band, which hybridizes only with the SV40 probe represents origin containing fragments released by BgZI cleavage. Note that these fragments are slightly larger than those contained in rL-RFV genomes. Lanes marked STDS contain SV40 size standards which are (in bp) 5243, 2994,2249,2924,1790,1446,993,673,552,444 and 375. These size standards do not hybridize uniformly with the EL-SV40 probe as EL-SV40 contains six origin and terminus regions rather than one.

COMPLEMENTATION

BETWEEN

The Southern blot of Fig. 2 was then rehybridized with a L-RFV probe. Both prototype L-RFV and rL-RFV genomes along with their restriction fragments hybridized with this probe (Fig. 2, middle panel). The SV40 origin fragments of approximately 800 bp, generated by &$I cleavage of rLRFV DNA did not hybridize with the LRFV probe while, as expected, the 350-bp fragment (&$I + PvuII) did. Restriction Map of rL-RFV From the restriction endonuclease digestion and hybridization data a map of the rL-RFV DNA clones was deduced. As shown in Fig. 3 the late region and at least

SV40 AND

RFV

DEFECTIVES

349

one of the L-RFV origins remain, but there is an insertion of non-SV40 sequences or a duplication of L-RFV sequences near the 5’ end of the late coding region. The SV40 origin-containing sequences have been inserted into the small early region of LRFV. In clone A, one of the L-RFV origins has been deleted while in clone D only a part of one of the L-RFV origin regions has been deleted. The SV40 sequences are separated from the rL-RFV late coding region by the L-RFV origin(s) and the further truncated rL-RFV early region. The sizes of rL-RFV clone A and clone D are approximately 4875 and 5075 bp, respectively, as compared to the 4500-bp size of the prototype L-RFV genome.

FIG. 3. Restriction map of rL-RFV genomes. rL-RFV-D is depicted. However, recombinant A is very similar except for the retention of a part of a second L-RFV origin in D but not in A. The recombinant contains a 500-bp deletion of sequences between the two L-RFV origins. The arrows just inside and parallel to the circles indicate the extent and direction of viral transcription. Enzyme designations: Gl, BglI; G2, BgZII, B, BarnHI; H, HhI; K, KpnI; M, MspI; P, PvuII; R, Em-RI; S, ,%I; Sp, SphI; T, ToqI.

350

O’NEILL,

MILLER,

Host Range Anal@ BSC-1, TC-7, MBA9812, MG178, HEK, and HFB cells were infected with E-SVIO/ L-RFV and E-SV40/rL-RFV viruses. It should be recalled that TC-7 and MBA9812 cells are fully permissive for only SV40 and not BKV (or RFV). HEK cells are permissive for only BKV or RFV while BSC-1 and HFB cells are permissive for both viruses (O’Neill and Carroll, 1981). Virus infected cells were monitored for T antigen and CPE up to 5 weeks after infection. The evolution of E-SV40/L-RFV to E-SV40/rL-RFV resulted in an altered host range (Table 1). In addition to propagating in HFB, and of course in BSC-1 cells, E-SV40/rL-RFV propagates in MBA9812 but not in HEK cells. This represents a host range shift toward cells exclusively permissive for SV40. However, E-SV40/rL-RFV could not propagate in TC-7 cells. Therefore, since the ac-

TABLE

AND

STEVENS

quisition of SV40 origins appeared to allow E-SV40/rL-RFV to propagate in MBA9812 cells but not in TC-7, cells the shift towards cells exclusively permissive for SV40 was incomplete. Role of SV40 @igins in the Phenotype of H&-id E-SV4O/rLRFV Virus E-SV40/rL-RFV demonstrated altered phenotypes (host range and relative amounts of E- and L-DNAs) following the acquisition of SV40 origins by L-RFV genomes. In order to determine if this was the etiologic event, or some other change in either the two complementing genomes was responsible for the altered phenotypes, we did the following: (1) We focused on ESV40 of E-SV40/rL-RFV and from Hirt supernatant DNA we cloned E-SV40 in pBR322. Cloned E-SV40 DNA was then introduced into BSC-1 cells with prototype

1

HOST RANGE ANALYSIS OF NEWLY GENERATED AND RECOMBINANT SV40/RFV

HYBRID VIRUSES

Cell lines BSC-1

ELSVIO RFV E-RFV/L-SV40 E-RFV/L-SV40” E-SV40/L-RFV E-SV40/rL-RFV E-SV40/L-RFV” E-SV40/drLRFVd

TC-7

HFB

MBA9812

HEK

CPE

T-ag’

CPE

T-ag

CPE

T-ag

CPE

T-ag

CPE

T-ag

+++ +++ ++ ++ ++ ++ ++

+++ ++ ++ ++ ++ ++ ++

+++ +++ ++ ++ ++ ++ ND

+++ ++ ++ ++ ++ ++ ND

+++ -

++ -

+++ -

+++ + + + + +++ +

-

f +

+++ + + ++ + + ND

++++ + + ++ +

+ +++ ++ + + + +

++

++

ND

ND

ND

ND

+

+

ND

ND

+ -

-

Note. -: no cell killing, no T-antigen. +: lo-20% cell killing, lo-20% Tag positive cells. ++: 25-50% cell killing, 25-50% T-ag positive cells. +++: 60-75% cell killing and T-ag containing cells. ++++: 95-99% cell killing, 80-95% T-ag containing cells. CPE and T-ag assays were made up to 3 weeks after infection. ND: not done. r: Recombinant genomes containing three SV40 origin sequences, demonstrated by susceptibility to &$I and SphI cleavage. ’ T-ag: T-antigen (identified by immunofluorescence). * E-RFV/L-SV40: Virus passage No. 15. ‘E-SVIO/L-RFV: E-SV40 was “recloned” from E-SV40/rL-RFV Hirt DNA and used to generate a new hybrid virus. d drL-RFV: Contains only a single SV40 origin, others were deleted (d) by BQ~I cleavage. Growth in MBA9812 cells was minimal. However, within 4 weeks variants of drL-RFV, containing 2 or 3 SV40 origins, arose and by 5 weeks cell killing was more extensive. E-SV40/rL-RFV produced extensive cell-killing in 2 weeks.

COMPLEMENTATION

BETWEEN

~WV40 farm II - rl-RFV form III oann 4.6LI)

ori fragments -

12345678

FIG. 4. Relative abundance of E-SV40 and complementing L-RFV, rL-RFV, and deleted rL-RFV genames in infected BSC-1 cells. The DNA was digested with enzymes (BgZI or &I) which clearly delineated each of the complementing genomes. The first lane contains prototype EL-SV40 DNA digested with &,%I. From top to bottom are the large E-SV40 fragment, the large L-SV40 fragment and origin fragments. Lanes 23, and 4 contain E-SVIO/L-RFV combinations but the E-SV40 genome of lanes 3 and 4 is derived from E-SV40/rL-RFV by cloning in pBR322. This newly cloned E-SV40 is indistinguishable from prototype E-SVIO, not only by Bgll cleavage but by digestion with BaaI and MboI. It also does not hybridize to RFV probes (not shown). Note also that in lanes 2-4 the L-RFV genome is almost undetectable. SstI digestion nicks some E-SV40 molecules (form II) and cuts L-RFV DNA once. Linear L-RFV DNA is barely detectable and appears just below nicked ESV40 DNA. L-RFV is also just barely detectable in the Bglr digest of lane 3, appearing as a heterogenous group of molecules just above the large BgZI E-SV40 fragment. Actually, these L-RFV genomes have a single BgZI site. True prototype L-RFV genomes are too infrequent to be seen here. In the last two lanes, ESV40/rL-RFV DNA has been digested with BglI (lane 7) or S&I (lane 8). The large BglI fragment of rL-RFV DNA appears just above the E-SV40 large fragment. rL-RFV DNA is cleaved once by Sat1 digestion and the linear molecule appears just below the nicked circles of E-SV40. Note that the intensity of the rL-RFV bands are equal to that of the E-SV40. In lane 8 the two bands appear slightly lower because lanes ‘7 and 8 were spliced from another photograph where the electrophoresis conditions were slightly different. In lanes 5 and 6, two of the three SV40 origins or rLRFV had been removed by BgZI digestion of the pBR322 plasmic insert and was then allowed to com-

SV40 AND

RFV

DEFECTIVES

351

L-RFV. If E-SV40 DNA had sustained an important etiologic mutation then it should be unable to overgrow complementing LRFV genomes. As shown in Fig. 4, newly cloned E-SV40 still overgrows L-RFV genomes. Also, the host range is identical to that of the original E-SV4O/L-RFV and the hybrid did not propagate in MBA9812 cells (Table 1). (2) We analyzed the rL-RFV genome. From a pBR322 clone of rL-RFV we removed two of the three SV40 origins and generated a new hybrid virus. As shown in Fig. 4, deleted rL-RFV (one SV40 origin) lost its ability to grow to near equal levels with E-SVIO. Restriction digests show ESV40 genomes in about a lo-fold excess of SV40 origin deleted rL-RFV genomes (Fig. 4). However, during growth of the virus one can observe the reappearance of a second and a third SV40 origin. (Fig. 4). Eventually, after serial virus passage, L-RFV genomes containing three SV40 origins again overgrow those with a single SV40 origin. In addition, deletion of two SV40 origins from rL-RFV caused ESV40/rLRFV to lose its ability to grow efficiently in MBA9812 cells. These results suggest that the acquisition of SV40 origins by rLRFV results in the increased ability of rLRFV to grow in BSC-1 cells and the altered host range of E-SV40/rL-RFV. E-RFV/LSVIO

Transfection

Within 6-8 days following cotransfection of BSC-1 cells with cloned E-RFV and LSV40 genomes CPE became evident and were extensive within lo-14 days. Lysates from these cultures could be serially passed

plement E-SV40 DNA in BSC-1 cells (see Methods). The viral DNA from infected cells was then cleaved with BglI (lane 5) or Sat1 (lane 6). The deleted rL-RFV genomes again become barely detectable. However, L-RFV genomes with reiterated BglI sites reappear and these can be seen as faint bands just above the large Bglr E-SV40 fragment and just below the S&I nicked circles of E-SV40 DNA. The reappearance of rL-RFV genomes with reiterated BgZI sites may be a result of incomplete BgA digestion of cloned rL-RFV molecules prior to transfection or it may result from reiteration of the remaining BgZI site occurring after transfection.

352

O’NEILL,

MILLER,

E RFV L 5140 '3 = ; 14

E.RFVformY UrurZk7~

STEVENS EL.RFV

E RFV kSV40 '7 =

2 = =-; r2Zf

-

AND

P L=--=-v) = Z ; z IuIIeav)ca(Iu)III

;

3

;

e

-

1

2

3

4

5

6

7

9

9

IO 11

12 13 14

L,RFV Probe

kSV40 9611 t or0 IraKmcti t Iher 2.3.766l

SW0

Plobr

FIG. 5. Southern blot hybridization of E-RFV/L-SV40 hybrid virus genomes contained in infected BSC-1 cells. Viral DNA was digested with BgZI, BgZII, or S&I. The DNA was hybridized with a wt SV40/pBR322 plasmid probe, and then with a L-RFV/pBR322 plasmid probe. L-RFV DNA contains some E-RFV sequences (Pater et al, 1980). Both probes were nick translated with PPJdCTP to a specific activity of l-2 X 10s cpm. Lanes 11-13 contain controls of EL-RFV DNA digested with Bgll, BgZII, or with %I. BglI does not cut EL-RFV and the lower bands represent supercoiled circles while the upper bands are nicked circles. BgnI does cleave EL-RFV DNA and produces several different size fragments. Lane 14 contains SV40 size standards. In the blot probed with L-RFV there

COMPLEMENTATION

BETWEEN

into fresh BSC-1 cells. CPE produced from virus became extensive in 9-12 days and did not appear more rapidly following 15 passages in BSC-1 cells. Virus from infected cells was neutralized by anti-SV40 serum but not by anti-BKV serum (data not shown). Viral DNA from Hirt supernatants of infected cells was periodically examined on agarose gels (Fig. 5, top). Restriction analysis demonstrated that both E-RFV and L-SV40 genomes were present in near equal amounts. Most molecules contained few apparent alterations after two passages in BSC-1 cells (Fig. 5, top). There was some heterogeneity in the size of the molecules but this heterogeneity was also present in both EL-SV40 and RFV (EL-RFV) infected cells (Pater et al, 1983; O’Neill et al, 1982). L-SV40 DNA contained alternating reiterations of both the viral origin and the viral terminus and contained the entire late SV40 coding region. E-RFV genomes contained a complete copy of the early region and digestion with BgZII, SstI, and PwuII demonstrated no significant changes. E-RFV/LSV40 was subjected to 15 serial passages in BSC-1 cells and there was no significant alteration in the relative proportions of both genomes (not shown). Smthemz Blot Hybridization

E-RFV/L-SV40 DNAs were prepared and digested with restriction endonucleases such that E-RFV and L-SV40 sequences

SV40 AND

RFV

DEFECTIVES

353

became clearly separated on agarose gels. For instance, BglI digestion cleaved L-SV40 DNA into three fragments, two of these being origin containing fragments and the other the large late coding region fragment. E-RFV DNA from early passage hybrid virus was not cleaved but was nicked. Nicking caused E-RFV DNA to migrate higher. Therefore BgZI digests contained (beginning from the top of the gel) form II E-RFV, the L-SV40 large fragment (middle of the gel) and small origin fragments (bottom of gel). In contrast B&I + EcoRI digestion linearized L-SV40 DNA but fragmented E-RFV DNA. The DNAs from these agarose gels were transferred to zetabind paper and subjected to hybridization with RFV DNA. RFV probes hybridized only to E-RFV fragments and not to L-SV40. SV40 probes hybridized only to LSV40 DNA. This suggested that there was no detectable recombination between the RFV and SV40 genomes during early passage of E-RFV/LSV40 virus. However, by passage 3 there was evidence of E-RFV/LSV40 intermolecular recombinants representing a fraction of the total viral DNA. Some genomes could now be cleaved by BgZI and B&II digestion. BgZII cleavage produced an anomalously large molecule (Fig. 5). Digestion with BglI plus BglII demonstrated that the anomalously large BgZII fragment contained a BgZI site. After 7-15 passages, recombinant E-RFV/LSV40 genomes were still present, were heterogeneous in size but did not overgrow proto-

is an indication that some L-SV40 genomes contain one or perhaps two BglII sites, as early as the third virus passage. The largest BgZII fragment disappears when the DNA is codigested with BgZI. However, since BgZII cuts E-RFV three times and the largest fragment is less than 2000 bp in size, the large BgZII fragment in lane 4, middle panel cannot be prototype E-RFV genomes. Since this fragment of approximately 4850 bp migrates near nicked L-SVIO, it is difficult to determine how well it hybridizes with the SV40 probe. In addition, in lane 2 (BgZI) there is a band which migrates just below form II molecules and it hybridizes with both probes. When digested with BgZI + BgZII, this molecule is further cut into two additional fragments of approximately 2600 and 1500 bp. The 1500-bp band is just barely detectable in this photograph and not in the RFV probed blot, as it is obscured by the E-RFV BgZII cleavage fragments. Both of these fragments (2600 and 1500 bp) again hybridize with both probes. Therefore, some of the viral DNA molecules consists of E-RFV/L-SV40 recombinants which contain both BgZI and BgZII recognition sites. Similar observations were made in passage 7 virus. BglII digestion produced several anomalous fragments and many of these can be further cleaved by BglI digestion. Digestion with BgZI alone produces several fragments (smear) migrating below nicked circles. These linear molecules hybridize with both probes and are cleaved further by Bg.!II digestion. Passage 7 virus contains more RFV/SV40 recombinants but these still are outnumbered by prototype E-RFV and L-SV40 genomes.

354

O’NEILL,

MILLER,

type E-RFV or L-SV40 genomes. The possibility that these recombinant molecules are infectious and possess altered biological properties awaits isolation and purification by molecular cloning. Host Range Anal&s of E-RFV/L.-SVLO Virus from infected BSC-1 cells was added to monolayers of TC-‘7, MBA9812, MG178, HEK, and HFB cells. Cultures were monitored for CPE up to 4 weeks after infection. They were also analyzed for intranuclear T antigen during this period. In isolates from early passages in BSC-1 cells, virus propagation could only be detected in HFG, HEK, and of course BSC-1 cells (Table 1). There was some growth in HEK but no growth was detected in TC-7 cells. These results are consistent with previous observations on the E-BKV/L-SV40 hybrid virus which demonstrated a narrow host range, propagating in only BSC-1 and HFB cells (O’Neill and Miller, 1985). Passage 15 ERFV/LSV40 propagated in BSC-1, HFB, and somewhat in HEK and TC-7 cells. However, this increased host range was not great as growth in HEK was less efficient than with E-RFV/LSV40. In TC-7 growth was limited as only approximately 10% of infected cells were lysed and virus could not be serially passed in these cells. DISCUSSION

It has been suggested that the existence of papovavirus genomes in the bipartite state, where one molecule contains a complete early region and the other a complete late region, might lead to rapid recombination and the generation of diversity (O’Neil1 and Miller, 1985). Two types of recombination may occur. In one type, early and late genomes of different polyomaviruses may be reassorted to produce a new hybrid virus. In a second type, true intermolecular recombination among complementing genomes may occur. This would be expected to generate additional diversity among hybrid papovaviruses. New recombinant hybrid papovaviruses might be expected to exhibit an altered host range if sequences specifying host range determinants are involved in recombination. In a previous report we demonstrated that wtBKV would complement E-SV40 or L-

AND

STEVENS

SV40 genomes and that in wtBKV/LSV40 combinations, the wtBKV genome was gradually replaced by a defective, E-BKV (O’Neill and Miller, 1985). This indicated the formation of a BKV/SV40 hybrid virus with a bipartite genome. Here we have cotransfected BSC-1 cells with early and late RFV and SV40 and have produced two new hybrid papovaviruses, E-RFV/L-SV40 and E-SV40/L-RFV. As with E-BKV/L-SV40, the E-RFV/L-SV40 hybrid also has a narrow host range. In contrast, the E-SV40/ L-RFV hybrid possesses a wider host range, propagating in cells permissive for SV40 or RFV. In both cases there are cell lines which do not support the growth of the original hybrid. We believe this is a result of the permissiveness of the cell line for only one rather than both of the parental genomes. In addition to reassortment we have found that in SV40/RFV hybrids, true intermolecular recombination has occurred. In the case of E-SV40/L-RFV, sequences from the E-SV40 genome have been introduced into L-RFV. These sequences contain the SV40 origin of replication, the 72-bp enhancer element or a portion of it and the SV40 terminus region. As E-SV40/rL-RFV evolves, several copies of the SV40 origin region and two copies of the terminus appear in rL-RFV genomes. The SV40 sequences displace part of the already truncated L-RFV early region and various portions of the second L-RFV origin region. Similarly, after three passages of E-RFV/ LSV40, there is evolution of some E-RFV/ L-SV40 recombinant genomes. The recombinants demonstrate one or two SV40 origins, but these do not become the dominant species after 15 passages of E-RFV/ L-SV40 in BSC-1 cells. True intermolecular recombination between different papovaviruses has not been previously described. Green monkey cells have been coinfected with a tsA mutant of SV40 and wtBKV, and BKV T-antigen apparently complemented the growth of tsA SV40 at the nonpermissive temperature (Mason and Takemoto, 1976). However, no recombination between the two viruses was detected. As suggested earlier this may in part be a result of the inability of BKV to grow in green monkey CV-1 cells (O’Neill

COMPLEMENTATION

BETWEEN

and Miller, 1985). Also, only wt viruses have been analyzed for recombination, and recombination may be difficult to detect because small changes in wt viruses can result in loss of infectivity. In contrast, papovaviruses with bipartite genomes contain large sequences which are not essential for infectivity (Pater et al, 1983,O’Neill and Miller, 1985). The recombination demonstrated here has occurred in these socalled nonessential regions. The acquisition of SV40 origins by LRFV genomes appears to have been followed by at least three important biological alterations. First, at passage 1 of E-SVIO/ L-RFV there is a relative paucity of L-RFV genomes and an abundance of E-SV40 genomes, in the Hirt supernatant. As the LRFV genome acquires SV40 origins its relative abundance increases, such that by passage 7 the E-SV40 and rL-RFV genomes are in nearly equal amounts. This is consistent with our observations that SV40 genomes grow better in BSC-1 cells than RFV genomes do and that RFV T-antigen, RFV DNA and RFV CPE appear more slowly in BSC-1 cells. Also, even under maximum conditions, there is 75% less RFV DNA in infected BSC-1 cells, as compared to EL-SV40 infections. Therefore the growth of rL-RFV DNA is apparently enhanced by the acquisition of SV40 origins. Second, the ability of the hybrid virus to propagate rapidly and produce cell killing of BSC-1 cells is also enhanced. E-SV40/ L-RFV produces extensive cell killing in 2 months and E-SV40/rL-RFV does so in 7 days. Third, the host range of the hybrid viruses has been altered. E-SV40/rL-RFV propagates in a cell line that E-SV40/LRFV does not. As the 72-bp sequence located near the SV40 origin is present in rL-RFV genomes, these observations are consistent with studies demonstrating that the 72-bp repeats contain host range determinants (Laimins et aL, 1982; Kriegler and Botchan, 1983; Byrne et al, 1983). However, E-SV40/rL-RFV also loses its ability to efficiently grow in HEK cells. This may have resulted from the loss of at least part of one of the RFV origins and/ or the alteration in the late coding region of rL-RFV. Although E-SV40/rL-RFV grows in one

SV40 AND

RFV

DEFECTIVES

355

cell line exclusively permissive for SV40 it does not grow in TC-7 cells, which are also permissive for SV40. This suggests that other host range determinants may be located elsewhere in the viral DNA. Recently, it has been shown that a small alteration in the VpI coding region of the lymphotropic papovavirus (LPV) produces a host range shift which allows LPV to grow in T leukocytes (Kanda and Takemoto, 1985). There are alternate explanations for the changes in phenotype exhibited by ESV40/rL-RFV. It is unlikely that E-SV40 has undergone some recombinational event since the E-SV40 genomes cloned from ESV40/rL-RFV infected cells demonstrate no differences from prototype E-SV40, in either the restriction pattern or in host range (Fig. 4, and data not shown). Also, “recloned” E-SV40 genomes are still able to overgrow complementing L-RFV genomes. Another possibility is that rL-RFV has undergone a sequence alteration in the late coding region, and this is responsible for the phenotypic changes. This explanation may be relevant as some changes in the late coding region have been identified. However, removal of two of the three origins from rL-RFV DNA restores the original phenotype. Therefore, the SV40 origins are at least partly responsible for the phenotypic changes exhibited by E-SVIO/ rL-RFV. We believe that the finding that E-SV40 DNA can be present in lo- to 20-fold greater amounts than L-RFV genomes in E-SV40/L-RFV infected BSC-1 cells is significant. It is of course a phenomenon unique to polyomaviruses with bipartite genomes. If the early genome can appear in such great amounts, packaging and cell killing may be inefficient and transformation of permissive cells may follow. We have found recently that E-SV40/L-RFV infected BSC-1 cells undergo morphological transformation at a high rate. In contrast, survivors of E-SV40/rL-RFV infection show not a single transformant. ACKNOWLEDGMENTS We thank Laurence Renzetti and Wendy Hopson for expert technical assistance. Dana Carroll generously provided advice on mapping of rL-RFV genomes. We are grateful to Mary M. Pater for the molecular

356

O’NEILL, MILLER, AND STEVENS

clones of E- and L-RFV DNAs. We thank David Hansen and Grant Bagley of the Utah Womens Clinic for

supplying us with tissue from therapeutically aborted fetuses. Supported by Veterans Administration Research Funds, Special VA grants on Innovative Aging and Innovature Carcinogenesis, and by R. K. Reynolds Industries.

of individual clones of simian virus 40 mutants containing deletions, duplications and insertions in their DNA. Cold Sp&g Harbor Symp. f&ant. BioL 39,69-U

NORKIN,L. C. (1982).Papovaviral persistent infections. MicrobioL

Rev. 46,384-425.

O’NEILL, F. J., and CARROLL,D. (1978). Appearance of defective simian virus 40 following infection of culREFERENCES tured human glioblastoma cells. virdogy 87, 109BENOIST,C., and CHAMBON,P. (1981).In tivo sequence 119. requirements of the SV40 early promotor region. O’NEILL, F. J., and CARROLL,D. (1981). Amplification Nature (London) 290,304-310. of papovavirus defectives during serial low multiBOTCHAN,M., TOPP,W. C., and SAMBROOK,J. (1976). plicity infections. l%rology 112, 800-803. The arrangement of SV40 sequences in the DNA of O’NEILL,F. J., and CARROLL,D. (1983).Role of defective transformed cells. CeU 9,269. simian virus 40 genomes in establishment and BROCKMAN,W. W., LEE, T. N. H., and NATHANS, D. maintenance of persistently infected primate cell (1973). The evolution of new species of viral DNA lines. InterviroZogg 19,181-194. during serial passage of simian virus 40 at high O’NEILL, F. J., MARYON,E. B., and CARROLL,D. (1982). multiplicity. virology 54,384-397. Isolation and characterization of defective SV40 BYRNE,B. J., DAVIS, M. S., YAMAGUCHI,J., BERGMSA, genomes which complement for infectivity. J. vird D., and SUBRAMANIAN,K. N. (1983). Definition of 43,18-25. O’NEILL, F. J., and MILLER, T. H. (1985). Isolation of the SV40 early promotor region and demonstration a papovavirus with a bipartite genome containing of a host range bias in the enhancement effect of the simian virus 40 72 base pair repeat. Proc. NatL unlinked SV40 and BKV sequences. firology 143, Acad. Sci USA 80.721-725.

CARROLL,D., and O’NEILL, F. J. (1978). Genome maps and simian virus 40 defectives propagated in human glioblastoma cells. virology 87,120-129. GRUSS,P., DHAR, R., and KHOURY,G. (1981). Simian virus 40 tandem repeated sequences as an element of the early promotor. Proc. NatL Acud Sci USA 78.943-947.

HIRT, B. (1967). Selective extraction of polyoma DNA from infected mouse cell cultures. J. Mol BioL 26, 365-369.

KANDA, T., and TAKEMOTO,K. K. (1985). Monkey Blymphotropic papovavirus mutant capable of replicating in T-lymphoblastoid cells. J. ViroL 55,96100. KAWAI, S., and NISKIZAWA,M. (1984). New procedure for DNA transfection with polycation and dimethyl sulfoxide. MoL Cell Bid 4.1172-1174. KRIEGLER,M., and BOTCHAN,M. M. (1983). Enhanced transformation by a simian virus 40 recombinant virus containing a Harvey murine sarcoma virus long terminal repeat. MoL CeLL BioL 3,325-329. LAIMINS,L. A., KHOURY,G., GORMAN,C., HOWARD,B., and GRUSS,P. (1982). Host specific activation of transcription by tandem repeats from SV40 and Moloney murine sarcoma virus. Proc NatL Ad Sci USA 79,6453-6457.

MANIATIS,T., FRITSCH,E. F., and SAMBROOK, J. (1982). “Molecular Cloning: A Laboratory Manual.” Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. MASON,D. H., and TAKEMOTO,K. K. (1976). Complementation between BK human papovavirus and a simian virus 40 ts A mutant. J. vird 17,1060-1062. MERTZ,J. E., CARBON,J., HERZBERG,M., DAVIS, R. W., and BERG,P. (1975). Isolation and characterization

75-87.

PATER, A., PATER, M. M., CHANG,L. S., SLOWIN, K., and DI MAYORCA,G. (1983). Multiple origins of the complementary defective genomes of RF and origin proximal sequences of GS, two human papovavirus isolates. virology 131,426-436. PATER,A., PATER,M. M., and DI MAYORCA,G. (1980). Arrangement of the genome of the human papovavirus RF virus. .J. ViroL 36,480-487. PATER,A., PATER,M. M., and DI MAYORCA,G., (1981). Genome analysis of Mg virus, a human papovavirus. J. Viral

39.968-972.

RIGBY,P. W. J., DIEKMANN,M., RHODES,C., and BERG, P. (1977). Labeling deoxyribonucleic acid to high specific activity in vitro by nick translation with DNA polymerase I. J. MoL BioL 113,237-251. ROSENTHAL,N., KRESS,M., GRUSS,P., and KHOURY, G. (1983). BK viral enhancer element and a human cellular homolog. Stience 222, 749-755. RYDER,K., DELUCIA, A. L., and TEGTMEYER,P. (1983). Binding of SV40 A protein to the BK virus origin of DNA replication. virology 129,239-245 (1983). SOMPAYRAC, L. M., and DANNA, K. J. (1982). Efficient infection of mammalian cells with SV40 DNA. Proc NatL Acad Sk USA 78,7575-7578.

SOUTHERN,E. (1975). Detection of virus specific sequences among DNA fragments separated by gel electrophoresis. J. MoL BioL g&503-509. TOOZE,J. (1980). Molecular biology of tumor viruses. In “DNA Tumor Viruses,” Part 2, 2nd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. YOSHIIKE,K., MIYAMURA,T., CHAN, H. W., and TAKEMOTO,K. K. (1982). Two defective DNAs of human polyomavirus JC adapted to growth in human embryonic kidney cells. J. ViroL 42,395-401.