- Email: [email protected]
Tumorigenic poxviruses: genomic organization of malignant rabbit virus, a recombinant between shope fibroma virus and myxoma virus
VIROLOGY
140, 113-124 (1985)
Tumorigenic Poxviruses: Genomic Organization of Malignant Rabbit Virus, a Recombinant between Shope Fibroma Virus and Myxoma Virus W. BLOCK, C. UPTON, AND G. MCFADDEN’ Department
of Biochemistry,
University
of Alberta,
Edmonton,
Alberta
T6G 2H7, Canada
Received July 20, 1984; accepted September 12, 1984 The genome of malignant rabbit virus (MRV), a newly discovered tumorigenic poxvirus of rabbits, has been analyzed using cloned DNA probes from Shope fibroma virus (SFV) and myxoma virus. Under high stringency conditions for Southern blotting such that SFV probes do not cross-hybridize with myxoma virus DNA, it is demonstrated that greater than 90% of the MRV genome has been derived from myxoma virus, and that approximately 10 kb of SFV-derived sequences have substituted for a similar amount of myxoma sequences. Mapping of the MRV genome indicates that the SFV sequences are present in two regions of the genome, one in each copy of the MRV terminal inverted repeat sequence. Furthermore, fine mapping studies of the integration sites for SFV into the myxoma background show that these SFV sequences are not symmetrical with respect to the left and right genomic termini. At the left end, 4 kb of SFV-derived DNA maps between 6 and 10 kb from the terminus, while at the right end about 5.5 kb of SFV sequences are found to extend at least 1 kb further toward the unique internal sequences. Based on this asymmetrical bipartite distribution of SFV sequences in MRV, a two-stage model to rationalize the origin of MRV is proposed. This model postulates an initial recombination event similar to gene conversion between myxoma and SFV at the right terminus of myxoma, followed by an incomplete transposition of only part of these SFV sequences to the left terminus. o 1985 Academic PRESS, IN. INTRODUCTION
Certain members of the genus leporipoxvirus cause benign tumors in their natural animal host (for reviews, see Febvre, 1962; Fenner and Ratcliffe, 1965). For example, Shope fibroma virus (SFV) causes a local, self-limited, histologically benign proliferation of fibroblasts in adult feral rabbits or experimentally infected domestic rabbits (Shope, 1932; Smith et al., 1973). However, if the rabbit host has an impaired cellular immune response or has been immunosuppressed, the tumors progess onto invasive fibrosarcomas which kill the host (Allison and Friedman, 1966; Scott et aZ., 1981; Sell and Scott, 1981). Recently, a novel leporipoxvirus designated malignant rabbit virus (MRV) was isolated from a SFV-infected rabbit (Strayer et al, 1983b). Unlike the parental r Author addressed.
to whom requests for reprints
should be
SFV, which routinely induces tumors that spontaneously regress in adult rabbits, MRV causes extensive immunosuppression which results in a disseminated malignancy in immunocompetent adults and is invariably fatal (Strayer et al, 1983a, b, c; Strayer and Sell, 1983; Skaletsky et al., 1984). MRV is antigenically very similar to both SFV and the closely related myxoma virus, the agent of rabbit myxomatosis. Analysis of MRV has indicated that it possesses biological properties which appear to have been derived from both SFV and myxoma: (1) at early times during MRV infection, the histological profile of the induced fibromas are indistinguishable from those caused by SFV, but the subsequent invasion of internal organs by the disease is a characteristic feature of myxomatosis; (2) MRV, like myxoma virus, has a broader host range than SFV and infects certain types of mucosal cells and lymphocytes, as well as fibroblasts, 113
0042-6822/85 $3.00 Copyright All rights
0 1985 by Academic Press. Inc. of reproduction in any form reserved.
BLOCK,
114
UPTON,
whereas SFV infects the latter only; (3) restriction enzyme digests of MRV DNA indicate a close but nonidentical relationship with the myxoma virus DNA genome (Strayer et al., 1983a, b, c; Strayer and Sell, 1983). MRV is of particular interest not only because it induces immunosuppression in the adult rabbit, but also because the capacity to induce characteristic proliferative fibromas has apparently been transferred out of SFV and into a myxoma genetic background. Since the BamHI restriction fragments of the SFV genome have recently been cloned in bacterial plasmid vectors (Wills et ah, 1983) and the physical map of the genome deduced (Delange et al, 1984), we have investigated the nature of the MRV genome by restriction enzyme mapping and Southern blotting and show that MRV is indeed a bona fide recombinant between SFV and myxoma virus. Furthermore, the organization of SFV sequences within MRV suggest a model to rationalize the acquisition of SFV sequences by myxoma virus to produce the resultant MRV genome. MATERIALS
AND
METHODS
Cells and viruses. SFV (strain Kasza) and Myxoma virus (strain Lausanne) were obtained from the American Type Culture Collection (ATCC). MRV was generously provided by D. S. Strayer and J. Leibowitz (Strayer et al., 1983a, b, c; Strayer and Sell, 1983). Titrations of viral infectivity were done in rabbit SIRC cells (obtained from the ATCC) grown in Dulbecco’s minimal Eagle’s (DME) medium supplemented with 10% fetal calf serum. SFV formed characteristic foci 3-5 days after infection; myxoma and MRV produced plaques after 2-4 days. Routine high titer growth of the virus stocks was performed in BGMK cells (from S. Dales) grown in DME medium plus 5% calf serum. DNA extraction, electrophoresis, blotting, and hybridization, Viral infections, DNA extraction procedures, conditions of restriction enzyme digestions, agarose gel electrophoresis, nick translation, and
AND
MC FADDEN
“standard” Southern blotting conditions have been described (Wills et aL, 1983; Delange et al, 1984). “Low” stringency Southern blotting conditions used for mapping terminal restriction sites in MRV and myxoma with cloned SFV probes refers to the following washing conditions: 1 hr in 2~ SSC (SSC = 150 mM NaCl, 15 mM sodium citrate, pH 7) at room temperature, followed by 1 hr in 2X SSC at 50 ‘. “High” stringency Southern blotting used to specifically monitor SFV sequences in MRV refers to 1 hr in 2X SSC at room temperature, followed by 1 hr in 0.1X SSC plus 0.1% sodium dodecyl sulfate at 60”. Acrylamide gel electrophoresis used for separating DNA fragments less than 1 kb was performed in 5-8% polyacrylamide, using a Tris-Borate-EDTA buffer system (Maniatis et al., 1982). DNA cloning. Bacterial plasmid vectors used were pBR322, propagated in Escherichia coli HBlOl, and pUC8 (gift of R. Hodgetts), grown in E. coli JM83. General conditions for ligation, transformation, and colony hybridization have been described (Wills et al., 1983). The subclones of SFV BamHI fragments E and C, which map at the junction of the terminal inverted repeats with the unique internal sequences, have been described (Wills et ab, 1983). The MRV BamHI fragments A and H were cloned into the BamHI site of pBR322; their corresponding BamHIBglII and BglII-BglII subfragments were cloned directly in the BamHI site of pUC8 (see Fig. 5 for positions of all subclones used). The MRV subclone AA grew very poorly in all clones tested and was distinctly unstable, although sufficient plasmid could be obtained for these studies. The myxoma BamHI fragments K, V, and E were identified with SFV-BamHI-E probe and with MRV unique left (MRVBamHI Hc) plus MRV unique right (MRVBamHI A*) probes. All attempts to clone the myxoma BamHI fragment E, which corresponds in location to the unstable MRV subclone A*, were unsuccessful and mapping of these myxoma sequences was accomplished using native viral DNA fragments isolated by preparative agarose gel electrophoresis.
GENOME
OF MALIGNANT
RESULTS
Restriction enzyme digestions of MRV DNA. It has been previously demonstrated that the XhoI digestion pattern of MRV DNA is very similar to that of myxoma virus (Strayer et UC, 1983a). In order to facilitate mapping of the MRV genome, the MRV restriction profiles were determined for BarnHI, PstI, and Sat1 because accurate mapping data for these enzymes has been previously determined for SFV (Delange et ah, 1984). The data in Fig. 1 and Table 1 illustrate the extent of similarity between the MRV and myxoma genomes. Both are about 160 kb in length, which is the same as that of the SFV genome (Wills et aa, 1983; Delange et al., 1984), and the differences can be localized to a relatively small number of restriction fragments. In the BumHI digests, MRV has acquired three new fragments (A, C, and H), while myxoma possesses five fragments not found in MRV (E, G, K, Tl, and V). Similarly the MRV unique frag-
VIRUS
:! 5
115
ments generated by PstI (C, D, and G) and SstI (D, E, and K) and the myxoma unique fragments (PstI Cl/C2 and SstI D, E, and J) are all confined to a relatively small fraction of their respective genomes. As will be detailed in the last section in Results, all but one of these changes can be explained by the recombination events to be described. The one exception relates to the MRV BamHI fragment C (10.8 kb) and the myxoma BumHI fragments G (7.7 kb) and Tl (3.2 kb). To confirm the suspicion that this represents a case of strainspecific polymorphism at a single BamHI site present in the myxoma standard (strain Lausanne) but not in MRV, 32P probe to purified MRV BamHI fragment C was shown by Southern blotting to hybridize only with MRV BamHI fragments G and Tl (not shown).
Detection of SFV sequences in the MRVgenome. Hybridization studies using cloned SFV DNA probes have shown that under the standard stringencies used during Southern blotting, considerable cross-
Pst I ‘, Y
BamH I h I
RABBIT
-Tl
FIG:. 1. Restriction enzyme digestions of MRV and myxoma DNA. MRV and myxoma DNA were digested with BumHI, PstI, and SstI, electrophoresed in 0.7% agarose, and stained with ethidium bromide. The lettered fragments represent differences between MRV and myxoma. Fragments smaller than 1.7 kb are not shown here but are indicated in Table 1.
BLOCK,
116
UPTON,
AND
TABLE
MC FADDEN 1
SUMMARYOFRESTRICTIONENZYMEDIGESTIONSOFMRV AND MYXOMADNA” Myxoma
MRV (1) BarnHI* A B C D E F G H I J K
(13.5)* (12.0) (lo.@* (10.0) (9.4) (9.0) (8.4) (8.0)* (7.3) (6.5) (5.7) (5.1) (4.7) (4.5) (4.4) (4.0) x 2 (3.8) (3.5) (3.4) x 2 (3.2) X 2
U
(2.2) x 2
V w X Y z AA
(1.4) (1.3) (1.0) (0.9)
Sum
160.5
(1.9)
(1.8)
A
(12.0)
B C D E F
(10.0) (9.4 (9.0) (8.6)* (8.4)
G H I J K L M N 0 P
Q R s T Tl u V W X Y z AA BB
(7.7)* (7.3) (6.5) (5.7) (5.2) X (5.1) (4.7) (4.5) (4.4) (4.0) x (3.8) (3.5) (3.4) x (3.2) X (3.2)* (2.2) x (2.1)* (1.9)
Myxoma
MRV (2) PstIb A B C D E F G H
2*
(3) sst1b A B C D
(>40) (>30) (19.5)* (18.5)* (11.0) (9.7) (5.6) X 2* (1.7) x 2
G-30) (24.0) X 2 (22.0) (18.0)*
A B Cl, c2
D E
(1.7) x 2
A B C
(>30) (24.0) X 2 (22.0) (17.0)*
(10.8) X 2* (9.6) X 2*
2
2 2 2
F G H I
Q3.8)
c3.8)
(7.2) (7.0) (3.1)
J K
(1.1) (0.9):
(7.2) (7.0) (3.1) (2.1) x 2* (1.1)
(1.8) (1.4) (1.3) (1.0) (0.9) 160.2
a Sizes in kilobases. The asterisk refers to unique fragments. *The following fragments in Fig. 1 are mitochondrial DNA contamination: BarnHI-17 SstI-6.9,
(11.0) (9.7)
F
D E
(>40) e-30) (24.0) X 2*
kb, P&I-17
kb,
5.6, 4.9 kb.
hybridizations are detectable with the majority of the BamHI fragments of myxoma virus (Wills et ah, 1983, Fig. 1). However, if the temperature of blot washing is increased from 50”, the variable amounts of hybridization of SFV probe with myxoma fragments is progressively reduced until, as shown in Fig. 2, it is completely abolished at 60” (in 0.1X SSC). Under these conditions of high stringency, two MRV fragments, BamHI-A and H,
strongly hybridized to the SFV DNA probe. We have also observed that under these conditions all MRV BamHI fragments still cross-hybridize with myxoma DNA (not shown). In order to map these MRV BamHI fragments A and H in the MRV genome, the reciprocal blotting experiment using MRV probe against SFV DNA was performed to exploit the fact that the SFV genomic map has been recently deduced (Delange et cd, 1984). In
GENOME
Ethidium
Bromide
SFV x 5
OF MALIGNANT
Probe z H
2 v)
RABBIT
VIRUS
117
sequences in MRV might also be located close to each copy of the MRV terminal inverted repeat.
SFV sequences in MRV are located at the terminal inverted repeats. Using SFV terminal IT probe under standard low stringency conditions (to permit crosshybridization between SFV and both MRV and myxoma), blotting experiments (data not shown) have mapped the MRV BamHI fragments A and H to positions overlapping the MRV terminal inverted repeat sequences. As shown in Fig. 4B, MRV and myxoma share identical restriction enzyme sites up to the Hind111 site at 5.25 kb from both termini, and these sites are clearly distinguishable from those found in SFV. The cloned MRV BamHI fragments A and H were used as probes for Southern blot analysis of digested myxoma DNA (not shown) to identify the appropriate myxoma BamHI fragments
FIG. 2. Hybridization of SFV probe with BornHIdigested MRV DNA under conditions of high stringency. DNA from myxoma virus (MYX), malignant rabbit virus (MRV), and Shope fibroma virus (SFV) was digested with BamHI, electrophoresed in 0.7% agarose, stained with ethidium bromide, and analyzed by Southern blotting under high stringency (see Materials and Methods), using SFV DNA probe. The two MRV fragments which hybridize to the probe are designated MRV BumHI A and H.
Fig. 3, when the cloned SFV DNA fragments were probed with MRV DNA under high stringency, the homology between MRV and SFV was localized to within SFV BamHI fragments C and E. These two SFV fragments are located at opposite ends of the linear SFV genome and span the junctions of the 12-kb terminal inverted repeat sequences with the unique single-copy internal sequences (Fig. 4A). This observation suggested that the SFV
1
2
3 4
5
6 7
E 9
1011
12 13 14 15 16 17 18 19 20
FIG. 3. Hybridization of MRV probe with the cloned BamHI restriction fragments of SFV under conditions of high stringency. The nomenclature of the cloned SFV fragments has been described (Wills et ah, 1983). Hybridization was performed as in Fig. 1, except genomic MRV DNA was used as probe. The following lanes are identified by the SFV BamHI fragment: lane 1, B; lane 2, C, lane 3, D, lane 4, E; lane 5, F2; lane 6, Fl; lane 7, G, lane 8, H; lane 9, I; lane 10, IT; lane 11, total SFV DNA digested with BarnHI; lane 12, 52; lane 13, Jl; lane 14, Kl; lane 15, K2; lane 16, Ll; lane 17, L2; lane 18, M; lane 19, N; lane 20, 0. The smallest SFV BamHI fragments-P, Q, R, S, and T (Delange et al, 1984) were not tested.
118
BLOCK,
UPTON,
AND
MC FADDEN
,
FIG. 4. Genomic locations of SFV, MRV, and myxoma BamHI fragments used to map the SFV sequences in MRV. (A) The deduced BumHI map of SFV (Delange et ab, 1984). (B) The BarnHI, BgZII, and Hind111 sites of MRV and myxoma were determined by partial digestions and by hybridizations with the SFV-IT terminal probe under low stringency conditions to permit crosshybridization between these genomes (see Materials and Methods). Only the BumHI sites are indicated at the right termini. The sizes of the terminal inverted repeats of SFV and myxoma were determined to be 12.2 and 11.5 kh, respectively. The size of the MRV inverted repeat (10.0 kb) is determined in Figs. 7 and 8, but is also indicated here for clarity. B = BarnHI; Bg = Bg111; S = SmaI; Bc = BclI; H = HindHI.
at the left end (K and V) and the right end (K and E) necessary for fine mapping of the myxoma terminus. Also indicated in Fig. 4B is the observation that the size of the terminal inverted repeat of myxoma (11.5 kb) is very close to that of SFV (12.2 kb). The significance of this will be shown to have relevance to the structure of the MRV inverted repeat (last section).
The SFV/myxoma junctions closest to the termini are identical at either end of the MRVgenome. Figure 5 shows the map locations of the subcloned fragments of SFV, MRV, and myxoma DNA used to determine the integration site of SFV sequences in MRV. Note that both the left and right ends of MRV are missing characteristic myxoma restriction sites, indicating that MRV was not generated by simple integration of SFV sequences into the myxoma genome. The fusion regions between SFV and myxoma DNA in MRV were first deduced
in the outermost junctions nearest to the termini (between 5 and 7 kb): note that some SFV subclones (EG/CG; ED/CD; EC/ CB; EE/CE; EF/CF) and MRV subclones (HD/AD; HB/AC; HA/AB) are identical because they map entirely within the respective inverted repeats of these viruses (Figs. 5 and 6). Inasmuch as the SFV sequences in MRV are 200-300 bp more distant from the genomic termini than in SFV itself, the maps in Figs. 6, 7, and 8 have been drawn so as to align these comparable SFV restriction sites. In Fig. 6, analysis of restriction sites within SFV ED and MRV HB fragments maps the position where SFV DNA replaces myxoma sequences to between the ClaI site at 6.1 kb in MRV and the unique SFV HpaI site 150 bp closer to the SFV left terminus. Identical results were obtained using subclones CB and AC from the right termini of SFV and MRV, respectively, thus establishing that the outermost SFV/
GENOME
OF MALIGNANT
RABBIT
VIRUS
LEFT
RIGHT
BSS
MVX
119
yyi+--y
SS
s E
B
S$b
s
B
I(
I ’ S,
’ St,
H
FIG. 5. Fine structure of SFV, MRV,and myxoma DNA subclones at the left and right termini. The letters with superscripts indicate subcloned fragments of BarnHI fragments SFV-E and C and MRV fragments A and H. The arrows indicate the terminal inverted repeats of SFV and myxoma. B = BumHI; Bg = BglII; S = SmaI; Pv = PvuII; Si = SstI; Sn = SstII; B1 = BglI; H = HindIII.
myxoma junctions are probably at each terminus of MRV.
identical
compared to the relevant SFV and myxoma sequences. The junction between SFV and myxoma sequences distal from the terminus was mapped to within the 300-bp segment between 10.05 and 10.35
Internal SF V/myxomajunctions are different at either terminus of MRL! In Fig. 7 the left
end of the MRV
genome
is
C Sa
SI
P
Sa
X D SaD
ED/CB
SFV Bg
@iI D
Hi
Hp
j
1
69 MRV
Bg :
5 5 15
. .
1 6
I C Sa
HE /AC
I SII
I P
, Sa
6
I I I X D SaD 7
I
1 7
1 kb
FIG. 6. Integration sites of SFV sequences in MRV: outermost junction closest to the MRV left and right termini. SFV BumHI subclone ED (left terminus) and CB (right terminus) are identical, as are MRV clones HB (left) and AC (right). The SFV restriction sites extending from 6 kb in MRV have been aligned with their counterparts in SFV such that the two maps are approximately 250 bp out of phase with respect to their termini. The shaded area between HpaII at 5.7 kb in SFV and ClaI at 6.1 kb in MRV denotes the beginning of SFV sequences in MRV. Bg = BgZII; D = DdeI; Hi = HincII; Hp = HpuII; C = ClaI; Sa = SalI; S11 = Sst11; P = PstI; x = XhoI; H = HindIII.
BLOCK,
120
DS,
UPTON, c
CD
AND
MC FADDEN
c
EE
SFV
S
Bg
DSI
:
CD
MRV
..
SAD
‘D
.\.:.
SI
A
HC
1 B
Bg 11
10
945 9 75
10
, kb 1165
11
!
CD
D
SAD
A
SI
V
MYX
I B
B 10
I FIG. ‘7. Integration terminus of MRV. As with their counterparts with respect to their transition to myxoma B = RamHI.
11
sites of SFV sequences in MRV: innermost in Fig. 6, the SFV and myxoma restriction in SFV and myxoma such that the three termini. The shaded area in MRV denotes sequences. Bg = BglII; D = DdeI; S1 = S&I;
kb from the left MRV end and establish that the total length of SFV sequences at the left end of MRV is between 3.95 and
1145 ’ kb
junction distal from the left sites in MRV have been aligned maps are slightly out of phase the end of SFV sequences and C = ClaI; S = SmaI; A = AmI;
4.45 kb. In addition, alignment of the flanking myxoma BumHI sites at the left end of the MRV genome with those in
SFV
S
PV
s
MRV
s
A”
s
SIC s
St
/!
B kb
B 13
, 176
I 17
i
I
0 15:
12 I
10 B
; PV
S
MYX
s
11
B 17
s, MYY Y-S 1
4
E
kb 17% I
1 10
12
15
B 12
10
FIG. 8. Integration sites of SFV sequences in MRV: innermost junction distal from the right terminus of MRV. As in Fig. 6, the SFV restriction sites in MRV have been aligned with their counterparts in SFV such that the two maps are out of phase with respect to their termini. The shaded area in MRV denotes the transition zone between SFV and myxoma sequences in MRV. The myxoma VB probe (inset) was derived from the junction between the myxoma inverted repeat and internal unique sequences at the left terminus and used to show that the inverted repeat junction of myxoma is still maintained at the MRV right terminus as well.
#
GENOME
OF MALIGNANT
myxoma DNA itself indicates that there is a net 200-300 bp insertion of SFV sequences in excess of the myxoma sequences deleted in MRV. Thus the BumHI site at 11.45 kb in the myxoma inverted repeat is now at 11.65 kb in MRV. On the other hand, the SFV/myxoma junction distal from the right end of the MRV genome was found to be further from the terminus than that observed at the left end. The fine mapping of this junction in MRV-AA was hindered by the fact that no useful restriction sites were found in the 4-kb region between 11 and 15 kb from the right MRV terminus. Furthermore, attempts to clone the myxoma BamHI-E fragment in this region were unsuccessful: all positive clones identified by colony hybridization to contain these myxoma sequences grew very slowly and quickly lost their inserted DNA fragments. The exact position of the sequences within this region that appear to be unstable in E. coli was not determined. Instead, mapping of this region (Fig. 8) was accomplished by high stringency Southern blotting analysis of myxoma and MRV DNA, using CE and CF SFV subclones and myxoma BamHI-SstI fragment (VB) as probes. Two important conclusions were made from such experiments: first, that CE and CF SFV sequences (the latter containing the characteristic S&II site) are present within MRV-AA and, second, that the original myxoma IR/unique sequence junction (as determined with myx-VB probe) is maintained in MRV-A*. Therefore, the SFV-derived sequences at the right end of MRV extend for 5.1-5.4 kb, ending between 11.1 and 11.4 kb from the right MRV terminus. Thus the SFV homology at the right end is at least 1 kb longer than that found at the left end of MRV. On the basis of this mapping information, it is now possible to rationalize the differences in BamHI, PstI, and SstI restriction profiles between MRV and myxoma (Fig. 1, Table 1) as follows: (1) The MRV BamHI fragments A and H were generated by the loss of BamHI sites between myxoma fragments K and V at the left end and E and K at the right end
RABBIT
VIRUS
121
(Fig. 4), while MRV BamHI fragment C is a polymorphic variant of myxoma BamHI fragments Tl and G. (2) Myxoma PstI fragments Cl, C2 have had PstI sites inserted into them due to the single Pst I site mapping within each copy of the acquired SFV sequences in MRV. (3) Myxoma virus contains SstI sites at 9 and 11 kb from each terminus (Fig. 5), of which the 9-kb site is replaced by SFV sequences at the left end, while at the right end both the 9- and 11-kb sites are replaced by the SFV sequences. Note that both copies of the acquired SFV sequences contain a new SstI site as well. Therefore at the left end, myxoma SstI fragments E and J have been replaced by MRV SstI fragments E and K, while at the right end myxoma SstI fragments E, J, and D have been replaced with MRV SstI fragments E and D. DISCUSSION
MRV was first isolated from rabbits that had been infected with SFV and, after plaque purification, was shown to be biologically distinct from SFV (Strayer et al, 198313). Unlike SFV, which causes benign fibromas in adult rabbits, the new MRV isolate induced an invasive malignant disease and produced such a profound immunosuppression in the adult rabbit that death would routinely occur because of extensive secondary gram-negative infections (Strayer et aL, 1983a, b, c). Further analysis of the biology of MRV has suggested that this virus possesses properties reminiscent of both SFV and the related leporipoxivirus myxoma, and it has been proposed that MRV might be a natural recombinant between these two viruses (Strayer et aL, 1983a, b; Strayer and Sell, 1983). We have utilized a bank of cloned SFV DNA restriction fragments (Wills et aL, 1983) and the known physical maps of SFV (Delange et ah, 1984) to verify this suggestion and to map the integration sites of SFV sequences to within both copies of the MRV terminal inverted repeat sequences. Interestingly, the distribution of SFV sequences in each copy of the MRV inverted repeats was
BLOCK,
122
UPTON,
AND
markedly asymmetric. At the left end of MRV approximately 4 kb of SFV sequences mapped between 6 and 10 kb from the terminus, while at the right terminus at least 5.5 kb of SFV sequences were detected, beginning at an identical position 6 kb from the terminus and extending more than 1 kilobase further inward than at the left terminus. Because of this nonidentity between the left and right end, the inverted repeat of MRV is now defined to stop at the junction of SFV and myxoma sequences at the left end. To account for the physical distribution of SFV sequences at the MRV termini, a two-stage model for the generation of MRV is presented (Fig. 9). The first event in this model is a recombination event between SFV and myxoma at the right terminus of myxoma such that approximately 5.5 kb of myxoma sequences are
MC FADDEN
replaced by homologous SFV sequences. The length of the SFV DNA is about 250300 bp longer than the replaced myxoma sequences, such that there is a small net increase in the distance of myxoma sequences internal to the SFV integration site from the molecular terminus of MRV. Although the precise molecular mechanism of this initial integration/substitution event is not known it is consistent with either a double recombination event, or a process similar to gene conversion. The second stage of the model postulates that 10 kb of the recombinant right terminus was then copied to the left terminus by a well-documented but poorly understood phenomenon of sequence duplication between the termini. Transposition of sequences from one terminus to the other has been observed in members of the orthopoxvirus genus (Moyer et al., 1980; Dumbell and Archard, 1980; Esposito et
MYX-IR
c
L
c
MYX
0
R
(
MYX
0
t L
C
MYX
t L
c
Recombination
MYX
1
SFY
wth
SFV
Q,
Partial
I
Transposntmn
0
MYX
H
lkb
FIG. 9. Model to rationalize the origin of MRV. The structure of SFV integration sites in MRV suggest that the initial event was recombination, or gene conversion, at the myxoma right terminus such that about 5.5 kb of myxoma sequences within the inverted repeat was replaced by almost the same amount of partially homologous SFV sequences. This was subsequently followed by an incomplete duplication of the 4 kb of the SFV sequences closest to the terminus to the left end of MRV such that a new inverted repeat (IR) junction was created. Despite this, notice that the original myxoma IR junction (denoted $J in the map) is still present at both MRV termini.
GENOME
OF MALIGNANT
aL, 1981; Archard et al., 1984), and the reproduction of identical deletions in a mirror-image fashion has been observed in the terminal inverted repeats of vaccinia virus (McFadden and Dales, 1979). Note that although the original junctions at 11.5 kb between the myxoma terminal inverted repeat and unique internal sequences are maintained in MRV, the MRV inverted repeat/unique junction is 1.5 kb shorter due to this incomplete transposition event. Thus MRV sequences between 10 to 11 kb from the left terminus are from myxoma but at the right terminus are derived from SFV. Since the MRV genome appears different from its myxoma parent only by the SFV sequences mapped in this paper the biological differences between MRV and either SFV or myxoma (Strayer et al., 1983a; Strayer and Sell, 1983) could be rationalized in a number of ways: (1) the addition of SFV gene products which replace the presumably analogous myxoma counterparts could alter the viral tropism for certain cellular targets or modulate cell-specific cytopathologies; (2) the recombination or transposition events might have truncated or inactivated either SFV or myxoma genes at the junction regions in MRV; (3) hybrid SFV/myxoma fusion proteins may have been generated at any one of the three different junction sites in MRV. Transcriptional mapping and DNA sequencing studies may resolve this issue. The absence of extensive genetics of leporipoxviruses has hampered the dissection of the molecular basis of poxvirus tumorigenicity. The fact that the capacity to stimulate fibroblast proliferation seems to have been transferred from SFV to MRV, and that the origin of MRV can be explained by recombination events which transfer only 5-6 kb of SFV sequence information into a myxoma genetic background, makes this region of the SFV genome of particular interest for analysis. It is also intriguing that a subset of this same 5- to 6-kb region of the SFV inverted repeat shares homology with an endogenous plasmid-like DNA species in uninfected rabbit cells, and thus may itself
RABBIT
VIRUS
123
have been originally acquired from the host cell (Upton and McFadden, submitted). ACKNOWLEDGMENTS We thank R. Maranchuk for technical assistance, A. M. Delange for helpful discussions, and C. Macaulay for assistance with the subclone mapping. G.M. is an Alberta Heritage Foundation for Medical Research (AHFMR) Scholar, C.U. and W.B. are AHFMR postdoctoral fellows. This work was supported by the NC1 of Canada. REFERENCES ALLISON, A. C., and FRIEDMAN, R. M. (1966). Effects of immunosuppressants on Shope fibroma virus. J. Natl Cancer Inst. 36, 8594368. ARCHARD, L. C., MACKETT, M., BARNES, D. E., and DUMBELL, K. R. (1984). The genome structure of cowpox white pock variants. J. Gen Viral 65,875886. DELANGE, A. M., MACAULAY, C., BLOCK, W., MUELLER, T., and MCFADDEN, G. (1984). Tumorigenic poxviruses: Construction of the composite physical map of the Shope fibroma virus genome. J. Viral. 50, 408-416. DUMBELL, K. R., and ARCHARD, L. C. (1980). Comparison of white pock (h) mutants of monkeypox virus with parental monkeypox and with variolalike viruses isolated from animals. Nature (London) 286, 29-32. ESPOSITO, J. J., CABRADILLA, C. D., NAKANO, J. H., and OBIJESKI, J. F. (1981). Intragenomic sequence transposition in monkeypox virus. Virology 109, 231-243. FEBVRE, H. (1962). In “Tumors induced by viruses: Ultrastructural studies” (A. J. Dalton and F. Haguenau, eds.), pp. 79-111. Academic Press, New York. FENNER, F., and RATCLIFFE, F. N. (1965). “Myxomatosis.” Cambridge Univ. Press, Cambridge, Mass. MANIATIS, T., FRITSCH, E. F., and SAMBROOK, J. (1982). “Molecular Cloning. A Laboratory Manual.” Cold Spring Harbor Laboratory, Cold Spring Harbor, N. Y. MCFADDEN, G., and DALES, S. (1979). Biogenesis of poxviruses: Mirror image deletions in vaccinia virus DNA. Cell 18, 101-108. MOYER, R. W., GRAVES, R. L., and ROTHE, C. T. (1980). The white pock(u) mutants of rabbit poxvirus. III. Terminal DNA sequence duplication and transposition in rabbit poxvirus. Cell 22, 545-553. SCOTT, C. B., HOLDBROOK, R., and SELL, S. (1981). Cell-mediated immune response to Shope fibroma virus-induced tumors in adult rabbits. J. Natl. Cancer Inst. 66, 681-689.
124
BLOCK,
UPTON,
SELL, S., and Sco?~, C. B. (1981). Immunohistology of malignant rabbit fibroma virus-a comparative study with rabbit myxoma virus. J. Natl. Cancw Inst. 66, 363-373. SHOPE, R. E. (1932). A transmissible tumor-like condition in rabbits. J. Exp. Med 56, 793-802. SKALETSKY, E., SHARP, P. A., SELL, S., and STRAYER, D. S. (1984). Immunologic dysfunction during viral oncogenesis. II. Inhibition of cellular immunity to viral antigens by malignant rabbit fibroma virus. Cell Immunol. 86, 64-74. J. W., TEVETHIA, S. S., LEVY, B. M., and RAWLS, W. E. (1973). Comparative studies on host responses to Shope fibroma virus in adult and newborn rabbits. J. NatL Cancer Inst. 50, 15291539.
SMITH,
STRAYER, D. S., CABIRAC, G., SELL, S., and LEIBOWITZ, J. L. (1983a). Malignant rabbit fibroma virus: Observations on the culture and histopathologic
AND
MC FADDEN
characteristics of a new virus-induced rabbit tumor. J. Natl. Cancer Inst. 71, 91-104. STRAYER, D. S., and SELL, S. (1983). Immunohistology of malignant rabbit fibroma virus-a comparative study with rabbit myxoma virus. J. Nat1 Cancer znst. 71,105-111. STRAYER, D. S., SKALETSKY, E., CABIRAC, G. F., SHARP, P. A., CORBEIL, L. B., SELL, S., and LEIBOWITZ, J. L. (1983b). Malignant rabbit fibroma virus causes secondary immunosuppression in rabbits. .I Immunol 130, 399-404. STRAYER, D. S., SELL, S., SKALETSKY, E., and LEIBOWITZ, J. L. (1983~). Immunologic dysfunction during viral oncogenesis. 1. Nonspecific immunosuppression caused by malignant rabbit fibroma virus. J. Immunol. 131, 2595-2600. WILLS, A., DELANGE, A. M., GREGSON, C., MACAULAY, C., and MCFADDEN, G. (1983). Physical characterization and molecular cloning of the Shope fibroma virus DNA genome. Virology 130, 403-414.
140, 113-124 (1985)
Tumorigenic Poxviruses: Genomic Organization of Malignant Rabbit Virus, a Recombinant between Shope Fibroma Virus and Myxoma Virus W. BLOCK, C. UPTON, AND G. MCFADDEN’ Department
of Biochemistry,
University
of Alberta,
Edmonton,
Alberta
T6G 2H7, Canada
Received July 20, 1984; accepted September 12, 1984 The genome of malignant rabbit virus (MRV), a newly discovered tumorigenic poxvirus of rabbits, has been analyzed using cloned DNA probes from Shope fibroma virus (SFV) and myxoma virus. Under high stringency conditions for Southern blotting such that SFV probes do not cross-hybridize with myxoma virus DNA, it is demonstrated that greater than 90% of the MRV genome has been derived from myxoma virus, and that approximately 10 kb of SFV-derived sequences have substituted for a similar amount of myxoma sequences. Mapping of the MRV genome indicates that the SFV sequences are present in two regions of the genome, one in each copy of the MRV terminal inverted repeat sequence. Furthermore, fine mapping studies of the integration sites for SFV into the myxoma background show that these SFV sequences are not symmetrical with respect to the left and right genomic termini. At the left end, 4 kb of SFV-derived DNA maps between 6 and 10 kb from the terminus, while at the right end about 5.5 kb of SFV sequences are found to extend at least 1 kb further toward the unique internal sequences. Based on this asymmetrical bipartite distribution of SFV sequences in MRV, a two-stage model to rationalize the origin of MRV is proposed. This model postulates an initial recombination event similar to gene conversion between myxoma and SFV at the right terminus of myxoma, followed by an incomplete transposition of only part of these SFV sequences to the left terminus. o 1985 Academic PRESS, IN. INTRODUCTION
Certain members of the genus leporipoxvirus cause benign tumors in their natural animal host (for reviews, see Febvre, 1962; Fenner and Ratcliffe, 1965). For example, Shope fibroma virus (SFV) causes a local, self-limited, histologically benign proliferation of fibroblasts in adult feral rabbits or experimentally infected domestic rabbits (Shope, 1932; Smith et al., 1973). However, if the rabbit host has an impaired cellular immune response or has been immunosuppressed, the tumors progess onto invasive fibrosarcomas which kill the host (Allison and Friedman, 1966; Scott et aZ., 1981; Sell and Scott, 1981). Recently, a novel leporipoxvirus designated malignant rabbit virus (MRV) was isolated from a SFV-infected rabbit (Strayer et al, 1983b). Unlike the parental r Author addressed.
to whom requests for reprints
should be
SFV, which routinely induces tumors that spontaneously regress in adult rabbits, MRV causes extensive immunosuppression which results in a disseminated malignancy in immunocompetent adults and is invariably fatal (Strayer et al, 1983a, b, c; Strayer and Sell, 1983; Skaletsky et al., 1984). MRV is antigenically very similar to both SFV and the closely related myxoma virus, the agent of rabbit myxomatosis. Analysis of MRV has indicated that it possesses biological properties which appear to have been derived from both SFV and myxoma: (1) at early times during MRV infection, the histological profile of the induced fibromas are indistinguishable from those caused by SFV, but the subsequent invasion of internal organs by the disease is a characteristic feature of myxomatosis; (2) MRV, like myxoma virus, has a broader host range than SFV and infects certain types of mucosal cells and lymphocytes, as well as fibroblasts, 113
0042-6822/85 $3.00 Copyright All rights
0 1985 by Academic Press. Inc. of reproduction in any form reserved.
BLOCK,
114
UPTON,
whereas SFV infects the latter only; (3) restriction enzyme digests of MRV DNA indicate a close but nonidentical relationship with the myxoma virus DNA genome (Strayer et al., 1983a, b, c; Strayer and Sell, 1983). MRV is of particular interest not only because it induces immunosuppression in the adult rabbit, but also because the capacity to induce characteristic proliferative fibromas has apparently been transferred out of SFV and into a myxoma genetic background. Since the BamHI restriction fragments of the SFV genome have recently been cloned in bacterial plasmid vectors (Wills et ah, 1983) and the physical map of the genome deduced (Delange et al, 1984), we have investigated the nature of the MRV genome by restriction enzyme mapping and Southern blotting and show that MRV is indeed a bona fide recombinant between SFV and myxoma virus. Furthermore, the organization of SFV sequences within MRV suggest a model to rationalize the acquisition of SFV sequences by myxoma virus to produce the resultant MRV genome. MATERIALS
AND
METHODS
Cells and viruses. SFV (strain Kasza) and Myxoma virus (strain Lausanne) were obtained from the American Type Culture Collection (ATCC). MRV was generously provided by D. S. Strayer and J. Leibowitz (Strayer et al., 1983a, b, c; Strayer and Sell, 1983). Titrations of viral infectivity were done in rabbit SIRC cells (obtained from the ATCC) grown in Dulbecco’s minimal Eagle’s (DME) medium supplemented with 10% fetal calf serum. SFV formed characteristic foci 3-5 days after infection; myxoma and MRV produced plaques after 2-4 days. Routine high titer growth of the virus stocks was performed in BGMK cells (from S. Dales) grown in DME medium plus 5% calf serum. DNA extraction, electrophoresis, blotting, and hybridization, Viral infections, DNA extraction procedures, conditions of restriction enzyme digestions, agarose gel electrophoresis, nick translation, and
AND
MC FADDEN
“standard” Southern blotting conditions have been described (Wills et aL, 1983; Delange et al, 1984). “Low” stringency Southern blotting conditions used for mapping terminal restriction sites in MRV and myxoma with cloned SFV probes refers to the following washing conditions: 1 hr in 2~ SSC (SSC = 150 mM NaCl, 15 mM sodium citrate, pH 7) at room temperature, followed by 1 hr in 2X SSC at 50 ‘. “High” stringency Southern blotting used to specifically monitor SFV sequences in MRV refers to 1 hr in 2X SSC at room temperature, followed by 1 hr in 0.1X SSC plus 0.1% sodium dodecyl sulfate at 60”. Acrylamide gel electrophoresis used for separating DNA fragments less than 1 kb was performed in 5-8% polyacrylamide, using a Tris-Borate-EDTA buffer system (Maniatis et al., 1982). DNA cloning. Bacterial plasmid vectors used were pBR322, propagated in Escherichia coli HBlOl, and pUC8 (gift of R. Hodgetts), grown in E. coli JM83. General conditions for ligation, transformation, and colony hybridization have been described (Wills et al., 1983). The subclones of SFV BamHI fragments E and C, which map at the junction of the terminal inverted repeats with the unique internal sequences, have been described (Wills et ab, 1983). The MRV BamHI fragments A and H were cloned into the BamHI site of pBR322; their corresponding BamHIBglII and BglII-BglII subfragments were cloned directly in the BamHI site of pUC8 (see Fig. 5 for positions of all subclones used). The MRV subclone AA grew very poorly in all clones tested and was distinctly unstable, although sufficient plasmid could be obtained for these studies. The myxoma BamHI fragments K, V, and E were identified with SFV-BamHI-E probe and with MRV unique left (MRVBamHI Hc) plus MRV unique right (MRVBamHI A*) probes. All attempts to clone the myxoma BamHI fragment E, which corresponds in location to the unstable MRV subclone A*, were unsuccessful and mapping of these myxoma sequences was accomplished using native viral DNA fragments isolated by preparative agarose gel electrophoresis.
GENOME
OF MALIGNANT
RESULTS
Restriction enzyme digestions of MRV DNA. It has been previously demonstrated that the XhoI digestion pattern of MRV DNA is very similar to that of myxoma virus (Strayer et UC, 1983a). In order to facilitate mapping of the MRV genome, the MRV restriction profiles were determined for BarnHI, PstI, and Sat1 because accurate mapping data for these enzymes has been previously determined for SFV (Delange et ah, 1984). The data in Fig. 1 and Table 1 illustrate the extent of similarity between the MRV and myxoma genomes. Both are about 160 kb in length, which is the same as that of the SFV genome (Wills et aa, 1983; Delange et al., 1984), and the differences can be localized to a relatively small number of restriction fragments. In the BumHI digests, MRV has acquired three new fragments (A, C, and H), while myxoma possesses five fragments not found in MRV (E, G, K, Tl, and V). Similarly the MRV unique frag-
VIRUS
:! 5
115
ments generated by PstI (C, D, and G) and SstI (D, E, and K) and the myxoma unique fragments (PstI Cl/C2 and SstI D, E, and J) are all confined to a relatively small fraction of their respective genomes. As will be detailed in the last section in Results, all but one of these changes can be explained by the recombination events to be described. The one exception relates to the MRV BamHI fragment C (10.8 kb) and the myxoma BumHI fragments G (7.7 kb) and Tl (3.2 kb). To confirm the suspicion that this represents a case of strainspecific polymorphism at a single BamHI site present in the myxoma standard (strain Lausanne) but not in MRV, 32P probe to purified MRV BamHI fragment C was shown by Southern blotting to hybridize only with MRV BamHI fragments G and Tl (not shown).
Detection of SFV sequences in the MRVgenome. Hybridization studies using cloned SFV DNA probes have shown that under the standard stringencies used during Southern blotting, considerable cross-
Pst I ‘, Y
BamH I h I
RABBIT
-Tl
FIG:. 1. Restriction enzyme digestions of MRV and myxoma DNA. MRV and myxoma DNA were digested with BumHI, PstI, and SstI, electrophoresed in 0.7% agarose, and stained with ethidium bromide. The lettered fragments represent differences between MRV and myxoma. Fragments smaller than 1.7 kb are not shown here but are indicated in Table 1.
BLOCK,
116
UPTON,
AND
TABLE
MC FADDEN 1
SUMMARYOFRESTRICTIONENZYMEDIGESTIONSOFMRV AND MYXOMADNA” Myxoma
MRV (1) BarnHI* A B C D E F G H I J K
(13.5)* (12.0) (lo.@* (10.0) (9.4) (9.0) (8.4) (8.0)* (7.3) (6.5) (5.7) (5.1) (4.7) (4.5) (4.4) (4.0) x 2 (3.8) (3.5) (3.4) x 2 (3.2) X 2
U
(2.2) x 2
V w X Y z AA
(1.4) (1.3) (1.0) (0.9)
Sum
160.5
(1.9)
(1.8)
A
(12.0)
B C D E F
(10.0) (9.4 (9.0) (8.6)* (8.4)
G H I J K L M N 0 P
Q R s T Tl u V W X Y z AA BB
(7.7)* (7.3) (6.5) (5.7) (5.2) X (5.1) (4.7) (4.5) (4.4) (4.0) x (3.8) (3.5) (3.4) x (3.2) X (3.2)* (2.2) x (2.1)* (1.9)
Myxoma
MRV (2) PstIb A B C D E F G H
2*
(3) sst1b A B C D
(>40) (>30) (19.5)* (18.5)* (11.0) (9.7) (5.6) X 2* (1.7) x 2
G-30) (24.0) X 2 (22.0) (18.0)*
A B Cl, c2
D E
(1.7) x 2
A B C
(>30) (24.0) X 2 (22.0) (17.0)*
(10.8) X 2* (9.6) X 2*
2
2 2 2
F G H I
Q3.8)
c3.8)
(7.2) (7.0) (3.1)
J K
(1.1) (0.9):
(7.2) (7.0) (3.1) (2.1) x 2* (1.1)
(1.8) (1.4) (1.3) (1.0) (0.9) 160.2
a Sizes in kilobases. The asterisk refers to unique fragments. *The following fragments in Fig. 1 are mitochondrial DNA contamination: BarnHI-17 SstI-6.9,
(11.0) (9.7)
F
D E
(>40) e-30) (24.0) X 2*
kb, P&I-17
kb,
5.6, 4.9 kb.
hybridizations are detectable with the majority of the BamHI fragments of myxoma virus (Wills et ah, 1983, Fig. 1). However, if the temperature of blot washing is increased from 50”, the variable amounts of hybridization of SFV probe with myxoma fragments is progressively reduced until, as shown in Fig. 2, it is completely abolished at 60” (in 0.1X SSC). Under these conditions of high stringency, two MRV fragments, BamHI-A and H,
strongly hybridized to the SFV DNA probe. We have also observed that under these conditions all MRV BamHI fragments still cross-hybridize with myxoma DNA (not shown). In order to map these MRV BamHI fragments A and H in the MRV genome, the reciprocal blotting experiment using MRV probe against SFV DNA was performed to exploit the fact that the SFV genomic map has been recently deduced (Delange et cd, 1984). In
GENOME
Ethidium
Bromide
SFV x 5
OF MALIGNANT
Probe z H
2 v)
RABBIT
VIRUS
117
sequences in MRV might also be located close to each copy of the MRV terminal inverted repeat.
SFV sequences in MRV are located at the terminal inverted repeats. Using SFV terminal IT probe under standard low stringency conditions (to permit crosshybridization between SFV and both MRV and myxoma), blotting experiments (data not shown) have mapped the MRV BamHI fragments A and H to positions overlapping the MRV terminal inverted repeat sequences. As shown in Fig. 4B, MRV and myxoma share identical restriction enzyme sites up to the Hind111 site at 5.25 kb from both termini, and these sites are clearly distinguishable from those found in SFV. The cloned MRV BamHI fragments A and H were used as probes for Southern blot analysis of digested myxoma DNA (not shown) to identify the appropriate myxoma BamHI fragments
FIG. 2. Hybridization of SFV probe with BornHIdigested MRV DNA under conditions of high stringency. DNA from myxoma virus (MYX), malignant rabbit virus (MRV), and Shope fibroma virus (SFV) was digested with BamHI, electrophoresed in 0.7% agarose, stained with ethidium bromide, and analyzed by Southern blotting under high stringency (see Materials and Methods), using SFV DNA probe. The two MRV fragments which hybridize to the probe are designated MRV BumHI A and H.
Fig. 3, when the cloned SFV DNA fragments were probed with MRV DNA under high stringency, the homology between MRV and SFV was localized to within SFV BamHI fragments C and E. These two SFV fragments are located at opposite ends of the linear SFV genome and span the junctions of the 12-kb terminal inverted repeat sequences with the unique single-copy internal sequences (Fig. 4A). This observation suggested that the SFV
1
2
3 4
5
6 7
E 9
1011
12 13 14 15 16 17 18 19 20
FIG. 3. Hybridization of MRV probe with the cloned BamHI restriction fragments of SFV under conditions of high stringency. The nomenclature of the cloned SFV fragments has been described (Wills et ah, 1983). Hybridization was performed as in Fig. 1, except genomic MRV DNA was used as probe. The following lanes are identified by the SFV BamHI fragment: lane 1, B; lane 2, C, lane 3, D, lane 4, E; lane 5, F2; lane 6, Fl; lane 7, G, lane 8, H; lane 9, I; lane 10, IT; lane 11, total SFV DNA digested with BarnHI; lane 12, 52; lane 13, Jl; lane 14, Kl; lane 15, K2; lane 16, Ll; lane 17, L2; lane 18, M; lane 19, N; lane 20, 0. The smallest SFV BamHI fragments-P, Q, R, S, and T (Delange et al, 1984) were not tested.
118
BLOCK,
UPTON,
AND
MC FADDEN
,
FIG. 4. Genomic locations of SFV, MRV, and myxoma BamHI fragments used to map the SFV sequences in MRV. (A) The deduced BumHI map of SFV (Delange et ab, 1984). (B) The BarnHI, BgZII, and Hind111 sites of MRV and myxoma were determined by partial digestions and by hybridizations with the SFV-IT terminal probe under low stringency conditions to permit crosshybridization between these genomes (see Materials and Methods). Only the BumHI sites are indicated at the right termini. The sizes of the terminal inverted repeats of SFV and myxoma were determined to be 12.2 and 11.5 kh, respectively. The size of the MRV inverted repeat (10.0 kb) is determined in Figs. 7 and 8, but is also indicated here for clarity. B = BarnHI; Bg = Bg111; S = SmaI; Bc = BclI; H = HindHI.
at the left end (K and V) and the right end (K and E) necessary for fine mapping of the myxoma terminus. Also indicated in Fig. 4B is the observation that the size of the terminal inverted repeat of myxoma (11.5 kb) is very close to that of SFV (12.2 kb). The significance of this will be shown to have relevance to the structure of the MRV inverted repeat (last section).
The SFV/myxoma junctions closest to the termini are identical at either end of the MRVgenome. Figure 5 shows the map locations of the subcloned fragments of SFV, MRV, and myxoma DNA used to determine the integration site of SFV sequences in MRV. Note that both the left and right ends of MRV are missing characteristic myxoma restriction sites, indicating that MRV was not generated by simple integration of SFV sequences into the myxoma genome. The fusion regions between SFV and myxoma DNA in MRV were first deduced
in the outermost junctions nearest to the termini (between 5 and 7 kb): note that some SFV subclones (EG/CG; ED/CD; EC/ CB; EE/CE; EF/CF) and MRV subclones (HD/AD; HB/AC; HA/AB) are identical because they map entirely within the respective inverted repeats of these viruses (Figs. 5 and 6). Inasmuch as the SFV sequences in MRV are 200-300 bp more distant from the genomic termini than in SFV itself, the maps in Figs. 6, 7, and 8 have been drawn so as to align these comparable SFV restriction sites. In Fig. 6, analysis of restriction sites within SFV ED and MRV HB fragments maps the position where SFV DNA replaces myxoma sequences to between the ClaI site at 6.1 kb in MRV and the unique SFV HpaI site 150 bp closer to the SFV left terminus. Identical results were obtained using subclones CB and AC from the right termini of SFV and MRV, respectively, thus establishing that the outermost SFV/
GENOME
OF MALIGNANT
RABBIT
VIRUS
LEFT
RIGHT
BSS
MVX
119
yyi+--y
SS
s E
B
S$b
s
B
I(
I ’ S,
’ St,
H
FIG. 5. Fine structure of SFV, MRV,and myxoma DNA subclones at the left and right termini. The letters with superscripts indicate subcloned fragments of BarnHI fragments SFV-E and C and MRV fragments A and H. The arrows indicate the terminal inverted repeats of SFV and myxoma. B = BumHI; Bg = BglII; S = SmaI; Pv = PvuII; Si = SstI; Sn = SstII; B1 = BglI; H = HindIII.
myxoma junctions are probably at each terminus of MRV.
identical
compared to the relevant SFV and myxoma sequences. The junction between SFV and myxoma sequences distal from the terminus was mapped to within the 300-bp segment between 10.05 and 10.35
Internal SF V/myxomajunctions are different at either terminus of MRL! In Fig. 7 the left
end of the MRV
genome
is
C Sa
SI
P
Sa
X D SaD
ED/CB
SFV Bg
@iI D
Hi
Hp
j
1
69 MRV
Bg :
5 5 15
. .
1 6
I C Sa
HE /AC
I SII
I P
, Sa
6
I I I X D SaD 7
I
1 7
1 kb
FIG. 6. Integration sites of SFV sequences in MRV: outermost junction closest to the MRV left and right termini. SFV BumHI subclone ED (left terminus) and CB (right terminus) are identical, as are MRV clones HB (left) and AC (right). The SFV restriction sites extending from 6 kb in MRV have been aligned with their counterparts in SFV such that the two maps are approximately 250 bp out of phase with respect to their termini. The shaded area between HpaII at 5.7 kb in SFV and ClaI at 6.1 kb in MRV denotes the beginning of SFV sequences in MRV. Bg = BgZII; D = DdeI; Hi = HincII; Hp = HpuII; C = ClaI; Sa = SalI; S11 = Sst11; P = PstI; x = XhoI; H = HindIII.
BLOCK,
120
DS,
UPTON, c
CD
AND
MC FADDEN
c
EE
SFV
S
Bg
DSI
:
CD
MRV
..
SAD
‘D
.\.:.
SI
A
HC
1 B
Bg 11
10
945 9 75
10
, kb 1165
11
!
CD
D
SAD
A
SI
V
MYX
I B
B 10
I FIG. ‘7. Integration terminus of MRV. As with their counterparts with respect to their transition to myxoma B = RamHI.
11
sites of SFV sequences in MRV: innermost in Fig. 6, the SFV and myxoma restriction in SFV and myxoma such that the three termini. The shaded area in MRV denotes sequences. Bg = BglII; D = DdeI; S1 = S&I;
kb from the left MRV end and establish that the total length of SFV sequences at the left end of MRV is between 3.95 and
1145 ’ kb
junction distal from the left sites in MRV have been aligned maps are slightly out of phase the end of SFV sequences and C = ClaI; S = SmaI; A = AmI;
4.45 kb. In addition, alignment of the flanking myxoma BumHI sites at the left end of the MRV genome with those in
SFV
S
PV
s
MRV
s
A”
s
SIC s
St
/!
B kb
B 13
, 176
I 17
i
I
0 15:
12 I
10 B
; PV
S
MYX
s
11
B 17
s, MYY Y-S 1
4
E
kb 17% I
1 10
12
15
B 12
10
FIG. 8. Integration sites of SFV sequences in MRV: innermost junction distal from the right terminus of MRV. As in Fig. 6, the SFV restriction sites in MRV have been aligned with their counterparts in SFV such that the two maps are out of phase with respect to their termini. The shaded area in MRV denotes the transition zone between SFV and myxoma sequences in MRV. The myxoma VB probe (inset) was derived from the junction between the myxoma inverted repeat and internal unique sequences at the left terminus and used to show that the inverted repeat junction of myxoma is still maintained at the MRV right terminus as well.
#
GENOME
OF MALIGNANT
myxoma DNA itself indicates that there is a net 200-300 bp insertion of SFV sequences in excess of the myxoma sequences deleted in MRV. Thus the BumHI site at 11.45 kb in the myxoma inverted repeat is now at 11.65 kb in MRV. On the other hand, the SFV/myxoma junction distal from the right end of the MRV genome was found to be further from the terminus than that observed at the left end. The fine mapping of this junction in MRV-AA was hindered by the fact that no useful restriction sites were found in the 4-kb region between 11 and 15 kb from the right MRV terminus. Furthermore, attempts to clone the myxoma BamHI-E fragment in this region were unsuccessful: all positive clones identified by colony hybridization to contain these myxoma sequences grew very slowly and quickly lost their inserted DNA fragments. The exact position of the sequences within this region that appear to be unstable in E. coli was not determined. Instead, mapping of this region (Fig. 8) was accomplished by high stringency Southern blotting analysis of myxoma and MRV DNA, using CE and CF SFV subclones and myxoma BamHI-SstI fragment (VB) as probes. Two important conclusions were made from such experiments: first, that CE and CF SFV sequences (the latter containing the characteristic S&II site) are present within MRV-AA and, second, that the original myxoma IR/unique sequence junction (as determined with myx-VB probe) is maintained in MRV-A*. Therefore, the SFV-derived sequences at the right end of MRV extend for 5.1-5.4 kb, ending between 11.1 and 11.4 kb from the right MRV terminus. Thus the SFV homology at the right end is at least 1 kb longer than that found at the left end of MRV. On the basis of this mapping information, it is now possible to rationalize the differences in BamHI, PstI, and SstI restriction profiles between MRV and myxoma (Fig. 1, Table 1) as follows: (1) The MRV BamHI fragments A and H were generated by the loss of BamHI sites between myxoma fragments K and V at the left end and E and K at the right end
RABBIT
VIRUS
121
(Fig. 4), while MRV BamHI fragment C is a polymorphic variant of myxoma BamHI fragments Tl and G. (2) Myxoma PstI fragments Cl, C2 have had PstI sites inserted into them due to the single Pst I site mapping within each copy of the acquired SFV sequences in MRV. (3) Myxoma virus contains SstI sites at 9 and 11 kb from each terminus (Fig. 5), of which the 9-kb site is replaced by SFV sequences at the left end, while at the right end both the 9- and 11-kb sites are replaced by the SFV sequences. Note that both copies of the acquired SFV sequences contain a new SstI site as well. Therefore at the left end, myxoma SstI fragments E and J have been replaced by MRV SstI fragments E and K, while at the right end myxoma SstI fragments E, J, and D have been replaced with MRV SstI fragments E and D. DISCUSSION
MRV was first isolated from rabbits that had been infected with SFV and, after plaque purification, was shown to be biologically distinct from SFV (Strayer et al, 198313). Unlike SFV, which causes benign fibromas in adult rabbits, the new MRV isolate induced an invasive malignant disease and produced such a profound immunosuppression in the adult rabbit that death would routinely occur because of extensive secondary gram-negative infections (Strayer et aL, 1983a, b, c). Further analysis of the biology of MRV has suggested that this virus possesses properties reminiscent of both SFV and the related leporipoxivirus myxoma, and it has been proposed that MRV might be a natural recombinant between these two viruses (Strayer et aL, 1983a, b; Strayer and Sell, 1983). We have utilized a bank of cloned SFV DNA restriction fragments (Wills et aL, 1983) and the known physical maps of SFV (Delange et ah, 1984) to verify this suggestion and to map the integration sites of SFV sequences to within both copies of the MRV terminal inverted repeat sequences. Interestingly, the distribution of SFV sequences in each copy of the MRV inverted repeats was
BLOCK,
122
UPTON,
AND
markedly asymmetric. At the left end of MRV approximately 4 kb of SFV sequences mapped between 6 and 10 kb from the terminus, while at the right terminus at least 5.5 kb of SFV sequences were detected, beginning at an identical position 6 kb from the terminus and extending more than 1 kilobase further inward than at the left terminus. Because of this nonidentity between the left and right end, the inverted repeat of MRV is now defined to stop at the junction of SFV and myxoma sequences at the left end. To account for the physical distribution of SFV sequences at the MRV termini, a two-stage model for the generation of MRV is presented (Fig. 9). The first event in this model is a recombination event between SFV and myxoma at the right terminus of myxoma such that approximately 5.5 kb of myxoma sequences are
MC FADDEN
replaced by homologous SFV sequences. The length of the SFV DNA is about 250300 bp longer than the replaced myxoma sequences, such that there is a small net increase in the distance of myxoma sequences internal to the SFV integration site from the molecular terminus of MRV. Although the precise molecular mechanism of this initial integration/substitution event is not known it is consistent with either a double recombination event, or a process similar to gene conversion. The second stage of the model postulates that 10 kb of the recombinant right terminus was then copied to the left terminus by a well-documented but poorly understood phenomenon of sequence duplication between the termini. Transposition of sequences from one terminus to the other has been observed in members of the orthopoxvirus genus (Moyer et al., 1980; Dumbell and Archard, 1980; Esposito et
MYX-IR
c
L
c
MYX
0
R
(
MYX
0
t L
C
MYX
t L
c
Recombination
MYX
1
SFY
wth
SFV
Q,
Partial
I
Transposntmn
0
MYX
H
lkb
FIG. 9. Model to rationalize the origin of MRV. The structure of SFV integration sites in MRV suggest that the initial event was recombination, or gene conversion, at the myxoma right terminus such that about 5.5 kb of myxoma sequences within the inverted repeat was replaced by almost the same amount of partially homologous SFV sequences. This was subsequently followed by an incomplete duplication of the 4 kb of the SFV sequences closest to the terminus to the left end of MRV such that a new inverted repeat (IR) junction was created. Despite this, notice that the original myxoma IR junction (denoted $J in the map) is still present at both MRV termini.
GENOME
OF MALIGNANT
aL, 1981; Archard et al., 1984), and the reproduction of identical deletions in a mirror-image fashion has been observed in the terminal inverted repeats of vaccinia virus (McFadden and Dales, 1979). Note that although the original junctions at 11.5 kb between the myxoma terminal inverted repeat and unique internal sequences are maintained in MRV, the MRV inverted repeat/unique junction is 1.5 kb shorter due to this incomplete transposition event. Thus MRV sequences between 10 to 11 kb from the left terminus are from myxoma but at the right terminus are derived from SFV. Since the MRV genome appears different from its myxoma parent only by the SFV sequences mapped in this paper the biological differences between MRV and either SFV or myxoma (Strayer et al., 1983a; Strayer and Sell, 1983) could be rationalized in a number of ways: (1) the addition of SFV gene products which replace the presumably analogous myxoma counterparts could alter the viral tropism for certain cellular targets or modulate cell-specific cytopathologies; (2) the recombination or transposition events might have truncated or inactivated either SFV or myxoma genes at the junction regions in MRV; (3) hybrid SFV/myxoma fusion proteins may have been generated at any one of the three different junction sites in MRV. Transcriptional mapping and DNA sequencing studies may resolve this issue. The absence of extensive genetics of leporipoxviruses has hampered the dissection of the molecular basis of poxvirus tumorigenicity. The fact that the capacity to stimulate fibroblast proliferation seems to have been transferred from SFV to MRV, and that the origin of MRV can be explained by recombination events which transfer only 5-6 kb of SFV sequence information into a myxoma genetic background, makes this region of the SFV genome of particular interest for analysis. It is also intriguing that a subset of this same 5- to 6-kb region of the SFV inverted repeat shares homology with an endogenous plasmid-like DNA species in uninfected rabbit cells, and thus may itself
RABBIT
VIRUS
123
have been originally acquired from the host cell (Upton and McFadden, submitted). ACKNOWLEDGMENTS We thank R. Maranchuk for technical assistance, A. M. Delange for helpful discussions, and C. Macaulay for assistance with the subclone mapping. G.M. is an Alberta Heritage Foundation for Medical Research (AHFMR) Scholar, C.U. and W.B. are AHFMR postdoctoral fellows. This work was supported by the NC1 of Canada. REFERENCES ALLISON, A. C., and FRIEDMAN, R. M. (1966). Effects of immunosuppressants on Shope fibroma virus. J. Natl Cancer Inst. 36, 8594368. ARCHARD, L. C., MACKETT, M., BARNES, D. E., and DUMBELL, K. R. (1984). The genome structure of cowpox white pock variants. J. Gen Viral 65,875886. DELANGE, A. M., MACAULAY, C., BLOCK, W., MUELLER, T., and MCFADDEN, G. (1984). Tumorigenic poxviruses: Construction of the composite physical map of the Shope fibroma virus genome. J. Viral. 50, 408-416. DUMBELL, K. R., and ARCHARD, L. C. (1980). Comparison of white pock (h) mutants of monkeypox virus with parental monkeypox and with variolalike viruses isolated from animals. Nature (London) 286, 29-32. ESPOSITO, J. J., CABRADILLA, C. D., NAKANO, J. H., and OBIJESKI, J. F. (1981). Intragenomic sequence transposition in monkeypox virus. Virology 109, 231-243. FEBVRE, H. (1962). In “Tumors induced by viruses: Ultrastructural studies” (A. J. Dalton and F. Haguenau, eds.), pp. 79-111. Academic Press, New York. FENNER, F., and RATCLIFFE, F. N. (1965). “Myxomatosis.” Cambridge Univ. Press, Cambridge, Mass. MANIATIS, T., FRITSCH, E. F., and SAMBROOK, J. (1982). “Molecular Cloning. A Laboratory Manual.” Cold Spring Harbor Laboratory, Cold Spring Harbor, N. Y. MCFADDEN, G., and DALES, S. (1979). Biogenesis of poxviruses: Mirror image deletions in vaccinia virus DNA. Cell 18, 101-108. MOYER, R. W., GRAVES, R. L., and ROTHE, C. T. (1980). The white pock(u) mutants of rabbit poxvirus. III. Terminal DNA sequence duplication and transposition in rabbit poxvirus. Cell 22, 545-553. SCOTT, C. B., HOLDBROOK, R., and SELL, S. (1981). Cell-mediated immune response to Shope fibroma virus-induced tumors in adult rabbits. J. Natl. Cancer Inst. 66, 681-689.
124
BLOCK,
UPTON,
SELL, S., and Sco?~, C. B. (1981). Immunohistology of malignant rabbit fibroma virus-a comparative study with rabbit myxoma virus. J. Natl. Cancw Inst. 66, 363-373. SHOPE, R. E. (1932). A transmissible tumor-like condition in rabbits. J. Exp. Med 56, 793-802. SKALETSKY, E., SHARP, P. A., SELL, S., and STRAYER, D. S. (1984). Immunologic dysfunction during viral oncogenesis. II. Inhibition of cellular immunity to viral antigens by malignant rabbit fibroma virus. Cell Immunol. 86, 64-74. J. W., TEVETHIA, S. S., LEVY, B. M., and RAWLS, W. E. (1973). Comparative studies on host responses to Shope fibroma virus in adult and newborn rabbits. J. NatL Cancer Inst. 50, 15291539.
SMITH,
STRAYER, D. S., CABIRAC, G., SELL, S., and LEIBOWITZ, J. L. (1983a). Malignant rabbit fibroma virus: Observations on the culture and histopathologic
AND
MC FADDEN
characteristics of a new virus-induced rabbit tumor. J. Natl. Cancer Inst. 71, 91-104. STRAYER, D. S., and SELL, S. (1983). Immunohistology of malignant rabbit fibroma virus-a comparative study with rabbit myxoma virus. J. Nat1 Cancer znst. 71,105-111. STRAYER, D. S., SKALETSKY, E., CABIRAC, G. F., SHARP, P. A., CORBEIL, L. B., SELL, S., and LEIBOWITZ, J. L. (1983b). Malignant rabbit fibroma virus causes secondary immunosuppression in rabbits. .I Immunol 130, 399-404. STRAYER, D. S., SELL, S., SKALETSKY, E., and LEIBOWITZ, J. L. (1983~). Immunologic dysfunction during viral oncogenesis. 1. Nonspecific immunosuppression caused by malignant rabbit fibroma virus. J. Immunol. 131, 2595-2600. WILLS, A., DELANGE, A. M., GREGSON, C., MACAULAY, C., and MCFADDEN, G. (1983). Physical characterization and molecular cloning of the Shope fibroma virus DNA genome. Virology 130, 403-414.