- Email: [email protected]
Analysis of the binding sites for the varicella-zoster virus gene 51 product within the viral origin of DNA replication
VIROLOGY
177, 570-577
(1990)
Analysis of the Binding Sites for the Varicella-Zoster Virus Gene 51 Product within the Viral Origin of DNA Replication NIGEL D. STOW,’ Medical
Research
Council
Virology
HAZEL M. WEIR,
Unit, Institute
Received
December
of Virology,
AND Church
7 1, 1989; accepted
ELIZABETH Street, March
Glasgow
C. STOW G 11 5JR, United
Kingdom
27, 1990
The C-terminal 322 amino acids of the varicella-zoster virus (VZV) gene 51 product were expressed in Escherichia co/i and shown to bind to specific DNA sequences within the VZV origin of DNA replication. The gene 51 product and its herpes simplex virus type 1 homolog (the ULQ protein) are capable of recognizing identical DNA sequences but the arrangement of binding sites within the origin regions of the two viruses differs. Three binding sites for the VZV gene 51 protein were identified within the VZV origin region and these lie in the same orientation and on the same side of the origin palindromic DNA sequence. DNA replication assays in transfected tissue culture cells demonstrated that the site closest to the palindrome is essential for origin activity whereas the most distal site is dispensable. The middle binding site may play an auxiliary role in DNA synthesis. o 1990 Academic press. IW.
INTRODUCTION
indromic DNA sequences in which the central regions contain only A and T residues and include repeats of the dinucleotide AT. Following the observation that HSV-1 encoded functions can activate VZV oris, albeit at low efficiency, it was suggested that an 1 1-bp sequence (CGTKGCACTT) common to the above origins might play a role in origin recognition (Stow and Davison, 1986). In independent experiments, using DNase footprinting, Elias and colleagues demonstrated the presence within HSV-1 oris of two binding sites for an (as then) unidentified viral protein (Elias et al., 1986; Elias and Lehman, 1988). The two sites, one of which includes the conserved 1 1-bp element and the other a closely related sequence, are present in opposite orientations and overlap the ends of the palindromic DNA sequence. More recent work has identified a consensus binding site YGYTCGCACT (Y equals pyrimidine nucleotide) for the HSV-1 protein, the middle 8 residues probably constituting the recognition sequence (Koff and Tegtmeyer, 1988). Transfection assays and studies with virus mutants have shown that replication of HSV-1 DNA directly involves the participation of seven viral gene products and specific functions have now been ascribed to most of these (see review by Challberg and Kelly, 1989; Wu et at., 1988; McGeoch et a/., 1988b; Olivo et al,, 1988; Weir et al., 1989; Crute et a/., 1989). Of special relevance to this communication are the observations that the HSV-1 UL9 gene encodes the origin binding protein described above (Olivo et al., 1988), and that a fusion protein expressed in E. co/i which contains only the Cterminal 3 17 amino acids (37%) of the UL9 polypeptide exhibits the sequence-specific DNA binding activity (Weir et al., 1989).
Complete DNA sequences for two human herpesviruses belonging to the subfamily Alphaherpesvirinae have been determined (Davison and Scott, 1986; McGeoch et a/., 1988a). The genomes of herpes simplex virus type 1 (HSV-1) and varicella-zoster virus (VZV) differ by 22% in their G + C content but nevertheless exhibit extensively collinear gene arrangements and encode products which in most instances exhibit high levels of amino acid identity (McGeoch et a/., 1988a). Thus, although molecular studies with HSV-1 are generally at a more advanced stage, it is possible to predict the functions of many VZV proteins and signals from a knowledge of their HSV-1 counterparts. Replication of the HSV-1 and VZV genomes depends upon the presence of c&-acting signals (presumed to function as origins of DNA synthesis) and frans-acting proteins. The origins of replication of HSV-1 and VZV have been functionally defined by using assays based upon their requirement, in cis, for the amplification of plasmid DNA molecules in transfected tissue culture cells expressing viral replication functions. The HSV-1 origins are specified by two distinct but related sequences, one (ori,) lies close to the center of the long unique (U,) region (Weller eta/., 1985), while two identical copies of the second (or&) are present within the inverted repeats, TRs and IRS (Stow and McMonagte, 1983). An equivalent of oriL is lacking from the VZV genome, but the TRs and IRS repetitions (Fig. la) each contain a copy of oris in a position equivalent to that occupied by HSV-1 oris (Stow and Davison, 1986). These origins are characterized by the presence of pal’ To whom 0042.6822/90
requests
for reprints
$3.00
Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
should
be addressed. 570
VZV GENE 51 PRODUCT
BINDING
SITES
571
(a)
IRS
UL
(b)
Kpnl GGTACCCCGC CCATGGGGCG
CGATGTTTAT GCTACAAATA
AACCATAATT TTGGTATTAA
CTCTAAACCG GAGATTTGGC
AAACAAGTCG TTTGTTCAGC
AAGAACTTCA TTCTTGAAGT
TATCTGAGGC ATAGACTCCG
ATGTAAACC TACATTTGG
GGTGGGGGGG CCACCCCCCC
TGAAAAAGGG GGGGGGTTAA ACTTTTTCCC CCCCCCAATT Bsml C
AACCA TTGGT
GTT CGCACT CAA GCGTG
TTTCTTGTTC
(cl
AAAGAACAAG
GCGCGTGTTC CGCGCACAAG
CCGCGATGTC GGCGCTACAG
TATATTCCAA ATATAAGGTT
ATGGAGCGGC TACCTCGCCG
AGGCTTTTTA TCCGAAAAAT
Kpnl
Us
TRs
TTGTAGAAAA AACATCTTTT
TCACAAAAAA AGTGTTTTTT
ATTTATTCAA TAAATAAGTT
ATTGGGCGTC TAACCCGCAG
CGCATGTCTG GCGTACAGAC
TGGTGTACGC ACCAECG
CAATCGGATA GTTAGCCTAT
TTTTTTCACT AAAMAGTGA
GTATGGGTTT CATACCCAAA
TCATGTTTTG AGTACAAAAC
GCATGTGTCC CGTACACAGG
280
ATATATATAT ATATATATAT TATATATATe_TATATATATA
ATATAGAGAA TATATCTCTT
AGAGAGAGAG TC~TCTCTC
350
GGGGTGTGGG
CGGGCTTTTC
ACAGAATATA
420
CCCCACACCC
GCCCGAAAAG
TGTCTTATAT
GCGGTTTTAT CGCCAAAATA CM AAATCGAT TTTAGCTA
458
Clal
Rsal &ml n 0
C
A
P pV6
A
p’J6 AA pv20
A
pV20AA
A
pV20 AC A
A
pV20 AAC PV3 pvo2
LH
RH
pvo37
FIG. 1. (a) Structure of the VZV genome showing the locations of the two copies of or&. (b) Sequence of the Kpnl plus C/al fragment containing oris (nucleotides 109, 893-l 10, 350 in the sequence of Davison and Scott, 1986). Motifs A, B, and C are boxed and the position of relevant restriction enzyme recognition sequences shown. The palindromic sequence is marked by arrows. (c) Structure of plasmid inserts. The Kpnl plus C/al fragment is represented diagrammatically in the top line and the positions of motifs A, B, and C and the palindromic sequence (P) are indicated. Shown below are the DNA sequences present in the various plasmids (solid lines) with internal deletions represented by carets. The cloned fragments are all flanked by EcoRl and HindIll sites at their left and right ends, respectively. The left-most nucleotide of the pV20 and pV20AA inserts is at position 226 in (b) (Le., the first residue of motif C). Plasmids pVO2 and pVO37 have been previously described (Stow and Davison, 1986). The latter contains an Xbal site at the position of the internal deletion which was utilized for the preparation of the LH and RH fragments.
In the present series of experiments we have examined the sequence-specific DNA binding properties of the VZV homolog of the HSV-1 UL9 protein. The binding sites for this protein, which is encoded by VZV gene 51 (McGeoch et a/., 1988a), have been located and their importance in origin activity assessed. The results reveal that although both the origin binding protein and its recognition sequence are highly conserved, there nevertheless exist significant differences between the
arrangement oris regions.
of binding sites within the HSV-1 and VZV
MATERIALS
AND
METHODS
Cells and virus Human fetal lung (HFL) fibroblasts (Flow Laboratories, 2002) were grown in Eagle’s medium supplemented with 10% fetal calf serum, nonessential amino
572
STOW,
WEIR.
acids, penicillin (100 units/ml), and streptomycin (100 units/ml). VZV of the strain described by Dumas et a/. (1981) was propagated in HFL cells as previously described (Stow and Davison, 1986). Plasmid
replication
assays
Subconfluent monolayers of HFL cells in 35-mm plastic petri dishes were used for plasmid replication assays which were performed as described previously (Stow and Davison, 1986). Briefly, monolayers were transfected with 0.25 pg circular plasmid DNA in the presence of calf thymus carrier DNA using the calcium phosphate technique followed by a DMSO boost. The cells were subsequently either mock-infected or superinfected by adding approximately 1 O6 HFL cells from a VZV-infected flask to each monolayer. Cellular DNA was prepared when the infected plates exhibited approximately 80% cpe (usually 3 days postinfection). Samples of DNA were cleaved with EcoRl plus Dpnl (Dpnl digests the methylated unreplicated input molecules but the products of plasmid amplification are resistant to its action) and analyzed by Southern blot hybridization using plasmid vector pTZl9U DNA, 32P-labeled in vitro, as probe. Plasmids The plasmids containing DNA fragments from the VZV oris region are illustrated in Fig. 1 c. The Kpnl plus C/al, Kpnl plus Bsml, and Bsml plus C/al fragments were inserted initially into pUC8 (Viera and Messing, 1982) and subsequently transferred to pTZ19U (Mead eta/., 1986) to facilitate site-directed mutagenesis. The orientation of each fragment is such that the EcoRl and HindIll sites of the polylinker lie at the left and right ends (as depicted in the figure), respectively. Plasmids pVGAA, pV20AA, pV20AC, and pV20AAC, which contain specific 11 -bp deletions of conserved sequences ‘A’ or ‘C’, were generated by site-specific mutagenesis using synthetic oligonucleotides. Single-stranded templates for mutagenesis were prepared by superinfection with Ml 3 K07 helper phage (Mead et al., 1986).
I (motif A) II (motif B)
5’ 5’
AND
STOW
Mutagenesis was performed by the method of Nakamaye and Eckstein (1986) using a commercially available kit (Amersham International Ltd., Bucks, UK) and oligonucleotides consisting of 15-l 7 nucleotides from each side of the sequence to be deleted. The identity of mutant plasmids was confirmed by DNA sequencing. Plasmids pVO2 and pVO37 have been previously described (Stow and Davison, 1986). As a result of Jhe deletion event a novel Xbal site is present in plasmid pVO37 (Fig. 1c). Expression
in E. co/i
The C-terminal region of VZV gene 51 was expressed as a fusion protein in E. co/i using the system previously employed for the corresponding region of the HSV-1 UL9 protein (Weir et a/., 1989). The Pvull plus Pstl fragment encoding the C-terminal 322 amino acids was inserted, in frame, between the Smal and Pstl sites of the vector pRIT2T+4 (Weir et a/., 1989) generating a hybrid gene which also encodes the Nterminal 260 amino acids. of the Staphylococcus aureus A protein. Propagation of bacteria (E. co/i strain K12AHlAtrp) containing the recombinant plasmid, heat induction of fusion protein expression, and the preparation of extract from sonicated cells (KV51 extract) were performed as previously described (Weir et a/., 1989). Extracts were similarly prepared from previously described transformants containing the parental vector pRIT2T (KR2 extract) or a plasmid encoding the corresponding HSV-1 UL9 gene fusion protein (KBl extract; Weir et al., 1989). DNA fragments gel retardation
and oligonucleotides analysis
DNA fragments for use as probes in gel retardation experiments were 3’ labeled with [32P]deoxyribonucleoside triphosphates and their overhanging ends filled in prior to purification from nondenaturing acrylamide gels. Duplex oligonucleotides containing the conserved motifs A and B (Fig. 1b) were formed as previously described (Weir et al., 1989) by annealing the following single-stranded synthetic oligonucleotides:
GATCAACCACCGTTCGCACTTTCTTTCT TTGGTGGCAAGCGTGAAAGAAAGACTAG GATCGTAAACCCGTTCGCACTTCCTGGGG CATTTGGGCAAGCGTGAAGGACCCCCTAG
Binding reactions were set up using bacterial extracts freshly diluted to 200 pg/ml in buffer C (Weir et al., 1989) supplemented with 600 mM KCI, 0.5 mM DlT, and 0.5 mM PMSF. Diluted extract (6 ~1) was added to 18 ~1 of 15 mlVl Hepes, pH 7.5, 1 mlVl EDTA,
for
5’ 5’
0.5 mM DlT containing 5 ng of labeled DNA fragment and 2 pg sonicated calf thymus DNA. Reactions were incubated at 25’ for 20 min and analyzed on 5% polyacrylamide gels as previously described (Weir et al., 1989).
VZV
DNase
GENE
51 PRODUCT
SITES
573
% amino acid Identity 60 -
I protection
DNase I protection experiments were performed essentially as described by Preston et al. (1988). KV5 1 or KBI extract was incubated with approximately 1O5 cpm of the plasmid pVO2 EcoRl plus HindIll fragment, uniquely 3’ labeled at either end, under the conditions employed for gel retardation analysis (Weir et al., 1989). After incubation at 25” for 20 min MgCI, and DNase I were added to concentrations of 5 mlL1 and 6 pg/ml, respectively, and incubation continued for a further 3 min. EDTA was added to a final concentration of 10 mM and the samples loaded on an 8% nondenaturing polyacrylamide gel. After electrophoresis the gel was exposed to autoradiographic film for 1 hr at 4’. Bands containing complexed and free DNA were excised and the DNA was recovered by electroelution, extracted sequentially with phenol and chloroform, and precipitated with ethanol. DNA fragments were analysed on a 6% denaturing polyacrylamide gel together with markers obtained by performing G or G plus A sequencing reactions (Maxam and Gilbert, 1977) on the end-labeled DNAs. RESULTS A candidate its possible
BINDING
VZV origin binding protein binding sites
and
In HSV-1 the UL9 gene encodes a polypeptide which binds to specific sites within the viral origins of replication (Olivo et a/., 1988). The VZV homolog of this protein is encoded by gene 51. The two proteins exhibit an overall 44% identity of amino acid residues, the distribution of which is shown in Fig. 2. A fragment of the UL9 polypeptide, comprising the C-terminal 317 amino acids and produced in E. co/i as a fusion protein, was previously shown to exhibit the same sequence-specific DNA binding ability as the authentic UL9 gene product (Weir et a/., 1989). The corresponding region of the VZV gene 51 (Fig. 2) was therefore expressed in the same system and its binding to the VZV oris region examined. We previously noted that the sequence CGTTCGCACTT, which corresponds closely to one of the binding sites within HSV-1 oris for the UL9 polypeptide, also occurs within a 259-bp Rsal plus C/al VZV DNA fragment (Fig. 1 b) which exhibits origin activity (Stow and Davison, 1986). Furthermore, Koff and Tegtmeyer (1988) proposed a consensus binding sequence YGYTCGCACT (Y equals pyrimidine) for the UL9 protein. This latter sequence occurs only four times within the VZV genome, in each instance as part of a CGTTCGCACTT motif. Two of these occurrences (A and B in Fig. 1 b) are close to the IRs copy of oris and the remaining two occupy corresponding positions near the TRs
60 -
40 J
20 -
OL
I N
1
<
1
I
I
I
I
I
200
400
600
600
c
ittTdn0 acid
residue
FIG. 2. Comparison of the amino acid sequences of the HSV-1 UL9 and VZV 51 gene products. The predicted amino acid sequences for these proteins (McGeoch et a/., 1988a; Davison and Scott, 1986) were aligned using the HOMOL program (Taylor, 1984) and the percentage of identical amino acids in adjacent 40 amino acid segments determined. These values are plotted across the length of the aligned proteins (which overall show 44% identical amino acids) with the N-terminus at the left. The bar in the lower portion of the figure shows the positions of identical amino acids (central vertical lines) and gaps inserted into the aligned UL9 and gene 51 protein sequences (marked above and below the central region respectively). The arrow indicates the regions of these proteins which were expressed In f. co/i and exhibit sequence-specific binding activity.
copy. Interestingly, the sequence CfilXGCACTT, which differs only at the indicated residue, is also present within each repeat region (C in Fig. 1b). A variety of VZV DNA fragments containing the motifs A, B, and C were therefore cloned and specifically mutated in order to assess the role of these sites in protein binding and origin activity. Detection of binding activity of the VZV gene 51 product Extracts were prepared from heat-induced cultures of E. co/i harboring either the plasmid encoding the VZV gene 5 1 fusion protein (KV51 extract) or the parental vector pRIT2T (KR2 extract) and examined in gel retardation assays using fragments from the VZV oris region as probes. Figure 3 shows that retarded complexes were observed only when KV51 extract was incubated with 32P-labeled pV3 and pVO37 LH fragments (lanes 3 and 8). The presence of an excess of unlabeled doublestranded oligonucleotides I or II, which contain the sequence motifs A and B, respectively, competed efficiently for binding to the labeled probe (lanes 4, 5, 9, and 10). No interaction was detected between the KV51 extract and labeled pVO37 RH fragment (lane 13). These results indicate the presence of functional
574
STOW,
pV037 6
7
LH 8
j
pVO37 12
WEIR,
RH 13
14
15
FIG. 3. Gel retardation analysis of the binding of the VZV gene 51 fusion protein to VZV oris. The pV3 EcoRl plus HindIll. pVO37 EcoRl plusMa (pVO37 LH), and pVO37 Xbal plus HindIll (pVO37 RH)fragments were used as labeled probes as indicated. Fragments were incubated without extract (lanes 1,6. and 1 1). with KR2 extract (lanes 2, 7. and 12) or with KV51 extract (lanes 3-5, 8-l 0, and 13-l 5). A 1 OO-fold molar excess unlabeled oligonucleotide I (lanes 4, 9, and 14) or oligonucleotide II (lanes 5, 10, and 15) was added to compete for sequence-specific binding. The positions of free probe (F) and the major retarded complexes (C, CU, and C,) are indicated. The fainter bands visible in tracks 3 and 8 probably result from interactions between fusion protein molecules and from the presence of a small amount of degraded fusion protein in the extract.
binding sites for the gene 51 fusion protein within oligonucleotide II, which contains sequence B from within the pV3 fragment, and oligonucleotide I, which contains sequence A from within the pVO37 LH fragment. Furthermore, in contrast to HSV-1, no binding site was detected at the right hand side of the palindromic DNA sequence (pVO37 RH fragment).
Fine-mapping of binding sites within the pV20 fragment The formation of two major retarded complexes following incubation of labeled pVO37 LH fragment, but not pV3 fragment, with KV5l extract (C, and CL in Fig. 3) suggests that the fragment may contain two binding sites, probably involving the related sequences A and C. To investigate further this possibility, a series of derivatives of plasmid pV20 containing specific 1 1-bp deletions eliminating sequence A, sequence C, or both were constructed and the appropriate fragments used in gel retardation assays. The results are presented in Fig. 4. As with the pVO37 fragment, the pV20 fragment yielded two retarded complexes specific for the KV5 1 extract (lane 3). Both the pV2OAA and pV2OAC fragments generated only a single complex, and no retarded band was observed with the pV2OAAC probe. These results demonstrate that the VZV gene 51 product binds to two
AND
STOW
sites within the pV20 fragment, one involving sequence A and the other involving sequence C., A striking feature is the difference in mobility between the complexes obtained with the pV2OAA and pV2OAC fragments. These fragments are of identical length and, as expected, the bands representing the free probe comigrate. The reduced mobility of the complex with protein bound near its center (pV2OAC fragment) compared to that with protein bound near one end (pV2OAA fragment) is consistent with the binding of protein resulting in bending of the DNA (Wu and Crothers, 1984; Zahn and Blattner, 1985; Mukherjee er al., 1985). The correspondence of sequences A and C with gene 51 product binding sites was confirmed by DNase footprinting using fragment pVO2 uniquely 3’ end-labeled on either strand. When DNase l-treated binding reactions were applied to a preparative nondenaturing gel both KBI (which contains the HSV-1 UL9 gene fusion protein) and KV5 1 extracts yielded two major complexes with each probe, essentially similar to the pattern observed with the pV20 fragment (Fig. 4). DNA from bands corresponding to the upper and lower (i.e., lower and higher mobility) KV51 complexes, the upper KBI complexes, and the free probes was analyzed on a denaturing gel (Fig. 5). Protection of sequences in the regions of both motifs A and C was apparent in the upper complex formed between KV5 1 extract and the probe labeled at its Hindill terminus. The corresponding complex formed with the probe labeled at the EcoRl site also showed protection in the regions of motifs A and C. The patterns of protection obtained with the upper complexes formed with KBI extract were almost indistinguishable demonstrating that the two viral polypeptides are capable of binding to the same 1
2
3
4
5
6
FIG. 4. Gel retardation analysis of binding to fragments lacking motifs A or C. The probes consisted of HindIll plus EcoRl fragments from plasmids pV20 (lanes l-3), pV2OAA (lane 4), pV2OAC (lane 5) or pV2OAAC (lane 6) which had been end-labeled to comparable specific activities. The pV20 probe was incubated without extract (lane l), with KR2 extract (lane 2) or with KV5 1 extract (lane 3). The remaining probes were each incubated with KV5 1 extract. Complex formation was analyzed as described in the legend to Fig. 3, and the major complexes are indicated by arrowheads.
VZV
Hmdlll
I
GENE
51 PRODUCT
ECORI
FIG. 5. DNase I protection by HSV-1 UL9 and VZV gene 51 fusion proteins. The pVO2 insert was uniquely 3’ end-labeled at either the fcoRl or Hindlll terminus, as indicated, and allowed to form complexes with KBl and KV51 extracts. Partial digestion with DNase I was performed prior to the resolution of complexes on a nondenaturing gel. DNA was extracted from bands corresponding to the pooled free DNA (F), upper KBI complex (Bl). lower KV51 complex (51L). and upper KV51 complex (51,) and analyzed on a 6% denaturing gel together with markers generated by partial cleavage at G and A residues (GA) or at G residues (G). The positions of sequence motifs A and C on the two strands are shown and the G residues within these motifs indicated (strands labeled at the HindIll and EcoRl sites correspond to the top and bottom strands in Fig. 1 b, respectively). The GA lanes contain some unexpected cleavages which were easily identified by comparison with other experiments.
BINDING
SITES
575
the cells was examined by Southern blotting for the presence of Dpnl-resistant (replicated) plasmid molecules. Figure 6 shows the result of an experiment using plasmids pV6, pV20, pV3, and the vector pTZ19U. VZV-encoded functions facilitated the replication of plasmids pV6 and pV20 but not pV3 or pTZlSU, indicating that the former two plasmids contain functional viral origins of replication. The insert in plasmid pV20 is similar to that in plasmid pVO2 (Fig. 1) confirming our earlier mapping of an origin within the &al plus CM fragment. Quantification by autodensitometry of this and other experiments revealed no significant difference between the replicative abilities of plasmids pV6 and pV20. The presence of binding site B therefore appears not to influence origin activity. The failure of plasmid pV3 to be replicated demonstrates that the presence of a gene 51 protein binding site in the absence of the palindromic DNA sequence is insufficient for activity. Figure 7 shows experiments assessing the roles of sequences A and C in origin function. Deletion of binding site A from either pV6 or pV20 abolished detectable replication. Binding at this locus is therefore absolutely essential for DNA synthesis, and absence of site A cannot be compensated for by the presence of sites B and C. Following deletion of site C from plasmid pV20 there was a significant but reproducible reduction in origin function. In repeat experiments plasmid pV2OAC replicated with between 20 and 50% of the efficiency of its parent pV20. Binding site C may therefore play some auxiliary role in DNA amplification in the presence of site A. DISCUSSION The results presented in this paper have identified the VZV gene 51 product as a sequence-specific origin pTZlQU pV6 abababab
pV3
pV20
DNA sequences. The lower complexes formed with KV51 extract exhibited partial protection over both the A and C motifs, suggesting that they contain DNA molecules with the fusion protein bound to one or the other site. This conclusion is supported by the relative mobility and diffuse character of the pV20 lower complex seen in Fig. 4. Role of sequences
A, 6, and C in origin activity
To test for origin activity plasmids containing DNA fragments were transfected into HFL cells subsequently either mock-infected or infected VZV to provide helper functions. DNA extracted
VZV and with from
FIG. 6. Replication of plasmids containing fragments from the VZV oris region. HFL cells were transfected with the indicated plasmids and either mock-infected (a) or infected with VZV (b). Samples containing 10% of the DNA recovered from each monolayer were digested with fcoRl plus Dpnl and the fragments were separated by electrophoresis, transferred to nitrocellulose, and hybridized to W labeled plZl9U. An autoradiograph of the washed filter is shown.
576 pTZlQU ababababab
STOW, PV6
6AA
pV20
20AA
pv20 a b
WEIR,
20AC a
b
FIG. 7. Replication of plasmids containing specific deletions. HFL cells were transfected with the indicated plasmids (the pV prefix has been omitted from plasmids pVGAA, pV20AA. and pV2OAC) and analyzed as described in the legend to Fig. 6. Lanes a and b contain DNA from mock-infected and VZV-infected cells, respectively.
binding protein, located its binding sites within the VZV oris region, and investigated their roles in origin activity. Because we were unable to reproducibly detect origin binding activity in extracts of VZV-infected HFL cells (data not shown), we chose to express the region of the VZV gene 51 protein homologous to the DNA binding domain of its HSV-1 counterpart, the UL9 polypeptide, in f. co/i. Using this product we have shown that, although the origin binding proteins of HSV-1 and VZV exhibit high levels of amino acid identity and contain sequence-specific DNA binding domains within their C-terminal regions which are capable of interacting with identical sequences, there nevertheless exist interesting differences between the interactions of these proteins with their cognate origins. The homology plot for the two proteins (Fig. 2) reveals that the C-terminal binding domains exhibit lower similarity than the remaining portions. It is tempting to speculate that the region of greatest similarity within the binding domains (amino acids 750-800) may be involved in sequence recognition, and further experiments are addressing this question. The regions of maximum amino acid identity within the N-terminal portions contain several motifs found in many purine nucleotide-utilizing helicases (Gorbalenya et al., 1989) suggesting that the HSV-1 UL9 and VZV gene 51 proteins are probably able to unwind DNA strands (either at the origin or elsewhere) in an energy-dependent reaction. Indeed, recent experiments using model substrates have confirmed that the UL9 protein exhibits helicase activity (M. D. Challberg and J. M. Calder, personal communications). The ability of the truncated fusion proteins to bind to the origins of replication suggests that the helicase function is probably dispensable for the sequence-specific interaction with DNA. We have demonstrated that the DNA sequence mo-
AND
STOW
tifs A, B, and C which conform closely to the consensus binding sites for the HSV-1 UL9 protein are involved in the binding of the VZV gene 51 product. These must represent the only major sites of interaction within the Kpnl plus C/al fragment (Fig. 1) since no retarded band was observed with the pV20AAC probe and only a single prominent complex with the pV3 probe (Figs. 3 and 4). The absence of complex formation with the pVO37 RH probe indicates that the sequence TGTTCGCGCGT (nucleotides 356-366 in Fig. 1 b) which matches motifs A and B in 8 of 1 1 positions does not function as a binding site for the gene 51 product. Previous work demonstrated the presence in HSV1 of two UL9 binding sites in opposite orientation and overlapping the ends of the palindromic DNA sequence (Elias and Lehman, 1988; Koff and Tegtmeyer, 1988; Olivo et al., 1988; Weir et a/., 1989). In contrast, in VZV oris the three binding sites are all located on the same side of the palindromic sequence and are arranged in the same relative orientation (Fig. 1). Sites B and C lie 168 and 62 bp from the palindromic DNA sequence, respectively. The most distal site, B, appears dispensable for DNA synthesis while site C possibly plays an auxiliary role. Nevertheless the ability of plasmid pV20AC to be replicated demonstrates that a single gene 5 1 protein binding site alongside the palindromic DNA sequence suffices for efficient origin activity. Whether sites B and C are of greater importance for DNA synthesis in the context of the viral genome or play some alternative role is not known. The roles of the two binding sites in HSV-1 oris function have also been investigated. The higher affinity site (site I) is absolutely essential for origin activity (Deb and Deb, 1989; Weir and Stow, 1990) while the absence of a functional site II apparently leads to inefficient replication (Lockshon and Galloway, 1988; Weir and Stow, 1990). Conflicting results on the importance of site II have, however, been presented by Deb and Doelberg (1988) who found no impairment in activity with an origin fragment completely lacking this binding site. Both DNA binding studies (results of M. D. Challberg cited in Challberg and Kelly, 1989) and experiments studying the effects upon origin activity of inserting various numbers of copies of the dinucleotide AT into the center of the HSV oris palindrome (Lockshon and Galloway, 1988) suggest that cooperative interactions between UL9 protein molecules bound to sites I and II can occur. These interactions are possibly important for allowing the initial opening of the HSV-1 origin region. In VZV, gene 51 protein bound to a single site adjacent to the palindromic sequence may perform a similar function, the longer AT-containing sequence within
VZV
GENE
51 PRODUCT
VZV ori, possibly facilitating easier unwinding of the DNA strands. Alternatively another viral or cellular protein could bind to the right-hand side of the VZV oris palindrome and substitute for the role played by UL9 protein interacting with site II of HSV-1 or&, or significantly different events might follow the sequence-specific binding of the UL9 and gene 51 proteins. These could include DNA bending, DNA unwinding, or interactions with other components of the replicative machinery. Such possibilities should be amenable to experimental investigation.
ACKNOWLEDGMENTS We thank J. H. Subak-Sharpe, A. J. Davison, R. D. Everett, R. M. Elliott, and C. M. Preston for helpful suggestions and criticisms, and 1. McLauchlan for the synthesis of oligonucleotides. H.M.W. was the recipient of a Medical Research Council Studentship.
REFERENCES CHALLBERG, M. D., and KELLY, T. J. (1989). Animal virus DNA replication. Annu. Rev. Biochem. 58, 67 l-7 17. CRUTE. 1. J., TSURUMI, T., ZHU, L., WELLER, S. K., Owe, P. D., CHALLBERG, M. D., MOCARSKI, E. S., and LEHMAN, I. R. (1989). Herpes simplex virus 1 helicase-primase: A complex of three herpes-encoded gene products. Proc. Nat/. Acad, Sci. USA 86, 2186-2189. DAVISON, A. J., and Scorr. J. E. (1986). The complete DNA sequence of varicella-zoster virus. J. Gen. Viral. 67, 1759-l 816. DEB, S., and DEB. S. P. (1989). Analysis of oris of HSV-1: Identification of one functional DNA binding domain. Nucleic Acids Res. 17, 2733-2752. DEB, S., and DOELBERG, M. (1988). A 67-base pair segment from the oris region of herpes simplex virus type 1 encodes origin function. 1. Viral. 62, 2516-2519. DUMAS, A. M., GEELEN, J. L. M. C., WESTRATE, M. W.. WERTHEIM, P., and VAN DER NOORDAA, J. (1981). Xbal. Pstl and Bglll restriction enzyme maps of the two orientations of the varicella-zoster virus genome. /. viral. 39, 390-400. ELIAS, P., and LEHMAN, I. R. (1988). Interaction of origin binding protein with an origin of replication of herpes simplex virus 1. Proc. Natl. Acad. Sci. USA 85,2959-2963. ELIAS, P.. O’DONNELL, M. E., MOCARSKI, E. S., and LEHMAN, I. R. (1986). A DNA binding protein specific for an origin of replication of herpes simplex virus type 1. Proc. Nat/. Acad. SC;. USA 83, 63226326. GORBALENYA. A. E., KOONIN, E. V., DONCHENKO, A. P., and BLINOV, V. M. (1989). Two related superfamilies of putative helicases involved in replication, recombination, repair and expression of DNA and RNA genomes. Nucleic Acids Res. 17,47 13-4730. KOFF. A., and TEGTMEYER, P. (1988). Characterisation of major recognition sequences for a herpes simplex virus type 1 origin-binding protein. J. Viral. 62, 4096-4103. LOCKSHON, D., and GALLOWAY, D. A. (1988). Sequence and structural requirements of a herpes simplexviral DNA replication origin. Mol. Cell. Biol. 8, 4018-4027.
BINDING
SITES
577
MAXAM, A. M., and GILBERT, W. (1977). A new method for sequencing DNA. Proc. Natl. Acad. Sci. USA 74, 560-564. MCGEOCH, D. J., DALRYMPLE, M. A., DAVISON, A. J., DOLAN, A., FRAME, M. C., MCNAB, D., PERRY, L. J., Scan, 1. E., and TAYLOR, P. (1988a). The complete DNA sequence of the long unique region in the genome of herpes simplex virus type 1.1. Gen. Viral. 69, 1531-l 574. MCGEOCH, D. J., DALRYMPLE, M. A., DOLAN, A., MCNAB, D., PERRY, L. J., TAYLOR, P., and CHALLBERG, M. D. (1988b). Structures of herpes simplex virus type 1 genes required for replication of virus DNA. 1. viral. 62,444-453. MEAD. D. A., SZCZESNA-SKORUPA, E., and KEMPER, B. (1986). Single stranded DNA ‘blue’ T7 promoter plasmids: A versatile tandem promoter system for cloning and protein engineering. Protein Eng. 1 I 67-74. MUKHERJEE, S., PATEL, I., and BASTIA, D. (1985). Conformational changes in a replication origin induced by initiator protein. Cell43, 189-l 97. NAKAMAYE. K. L., and ECKSTEIN, F. (1986). Inhibition of restriction endonuclease Neil cleavage by phosphorothioate groups and its application to oligonucleotide-directed mutagenesis. Nucleic Acids Res. 14,9679-9698. Ouvo, P. D., NELSON, N. J., and CHALLBERG, M. D. (1988). Herpes simplex virus DNA replication: The UL9 gene encodes an origin blnding protein. Proc. Nat/. Acad. Sci. USA 85, 5414-5418. PRESTON, C. M., FRAME, M. C., and CAMPBELL, M. E. M. (1988). A complex formed between cell components and an HSV structural polypeptide binds to a viral Immediate early gene regulatory DNA sequence. Ce//52,425-434. STOW, N. D., and DAVISON, A. J. (1986). Identification of a varicellazoster virus ongin of DNA replication and its activation by herpes simplex virus type 1 gene products. /. Gen. Viral. 67, 16 13-l 623. STOW, N. D., and MCMONAGLE. E. C. (1983). Characterisation of the TRs/lRs origin of DNA replication of herpes simplex virus type 1. Virology 130, 427-438. TAYLOR, P. (1984). A fast homology program for aligning biological sequences. Nucleic Acids Res. 12,447-455. VIERA, J., and MESSING, J. (1982). The pUC plasmids, an Ml3 mp7derived system for insertion mutagenesis and sequencing with synthetic universal primers. Gene 19, 259-268. WEIR, H. M.. and STOW, N. D. (1990). Two binding sites forthe herpes simplex virus UL9 protein are required for efficient activity of the oris replication origin. /. Gen. Viral. 71, in press. WEIR, H. M., CALDER, J. M., and STOW, N. D. (1989). Binding of the herpes simplex virus type 1 UL9 gene product to an ongin of viral DNA replication. Nucleic Acids. Res. 17, 1409-l 425. WELLER, S. K., SPADARO, A., SCHAFFER. J. E., MURRAY, A. W., MAXAM, A. M., and SCHAFFER, P. A. (1985). Cloning, sequencing and functional analysis of or&, a herpes simplex virus type 1 origin of DNA synthesis. Mol. Cell. Biol. 5, 930-942. Wu, C. A., NELSON, N. J., MCGEOCH, D. J., and CHALLEERG, M. D. (1988). Identification of herpes simplex virus type 1 genes required for origin-dependent DNA synthesis. J. K-o/. 62,435-443. Wu, H.-M., and CROTHERS, D. M. (1984). The locus of sequence-directed and protein-induced DNA bending. Nature (London) 308, 509-513. ZAHN, K., and BLAT~NER, F. R. (1985). Binding and bending of the lambda replication origin by the phage 0 protein. EMBO /. 4, 3605-36 16.
177, 570-577
(1990)
Analysis of the Binding Sites for the Varicella-Zoster Virus Gene 51 Product within the Viral Origin of DNA Replication NIGEL D. STOW,’ Medical
Research
Council
Virology
HAZEL M. WEIR,
Unit, Institute
Received
December
of Virology,
AND Church
7 1, 1989; accepted
ELIZABETH Street, March
Glasgow
C. STOW G 11 5JR, United
Kingdom
27, 1990
The C-terminal 322 amino acids of the varicella-zoster virus (VZV) gene 51 product were expressed in Escherichia co/i and shown to bind to specific DNA sequences within the VZV origin of DNA replication. The gene 51 product and its herpes simplex virus type 1 homolog (the ULQ protein) are capable of recognizing identical DNA sequences but the arrangement of binding sites within the origin regions of the two viruses differs. Three binding sites for the VZV gene 51 protein were identified within the VZV origin region and these lie in the same orientation and on the same side of the origin palindromic DNA sequence. DNA replication assays in transfected tissue culture cells demonstrated that the site closest to the palindrome is essential for origin activity whereas the most distal site is dispensable. The middle binding site may play an auxiliary role in DNA synthesis. o 1990 Academic press. IW.
INTRODUCTION
indromic DNA sequences in which the central regions contain only A and T residues and include repeats of the dinucleotide AT. Following the observation that HSV-1 encoded functions can activate VZV oris, albeit at low efficiency, it was suggested that an 1 1-bp sequence (CGTKGCACTT) common to the above origins might play a role in origin recognition (Stow and Davison, 1986). In independent experiments, using DNase footprinting, Elias and colleagues demonstrated the presence within HSV-1 oris of two binding sites for an (as then) unidentified viral protein (Elias et al., 1986; Elias and Lehman, 1988). The two sites, one of which includes the conserved 1 1-bp element and the other a closely related sequence, are present in opposite orientations and overlap the ends of the palindromic DNA sequence. More recent work has identified a consensus binding site YGYTCGCACT (Y equals pyrimidine nucleotide) for the HSV-1 protein, the middle 8 residues probably constituting the recognition sequence (Koff and Tegtmeyer, 1988). Transfection assays and studies with virus mutants have shown that replication of HSV-1 DNA directly involves the participation of seven viral gene products and specific functions have now been ascribed to most of these (see review by Challberg and Kelly, 1989; Wu et at., 1988; McGeoch et a/., 1988b; Olivo et al,, 1988; Weir et al., 1989; Crute et a/., 1989). Of special relevance to this communication are the observations that the HSV-1 UL9 gene encodes the origin binding protein described above (Olivo et al., 1988), and that a fusion protein expressed in E. co/i which contains only the Cterminal 3 17 amino acids (37%) of the UL9 polypeptide exhibits the sequence-specific DNA binding activity (Weir et al., 1989).
Complete DNA sequences for two human herpesviruses belonging to the subfamily Alphaherpesvirinae have been determined (Davison and Scott, 1986; McGeoch et a/., 1988a). The genomes of herpes simplex virus type 1 (HSV-1) and varicella-zoster virus (VZV) differ by 22% in their G + C content but nevertheless exhibit extensively collinear gene arrangements and encode products which in most instances exhibit high levels of amino acid identity (McGeoch et a/., 1988a). Thus, although molecular studies with HSV-1 are generally at a more advanced stage, it is possible to predict the functions of many VZV proteins and signals from a knowledge of their HSV-1 counterparts. Replication of the HSV-1 and VZV genomes depends upon the presence of c&-acting signals (presumed to function as origins of DNA synthesis) and frans-acting proteins. The origins of replication of HSV-1 and VZV have been functionally defined by using assays based upon their requirement, in cis, for the amplification of plasmid DNA molecules in transfected tissue culture cells expressing viral replication functions. The HSV-1 origins are specified by two distinct but related sequences, one (ori,) lies close to the center of the long unique (U,) region (Weller eta/., 1985), while two identical copies of the second (or&) are present within the inverted repeats, TRs and IRS (Stow and McMonagte, 1983). An equivalent of oriL is lacking from the VZV genome, but the TRs and IRS repetitions (Fig. la) each contain a copy of oris in a position equivalent to that occupied by HSV-1 oris (Stow and Davison, 1986). These origins are characterized by the presence of pal’ To whom 0042.6822/90
requests
for reprints
$3.00
Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
should
be addressed. 570
VZV GENE 51 PRODUCT
BINDING
SITES
571
(a)
IRS
UL
(b)
Kpnl GGTACCCCGC CCATGGGGCG
CGATGTTTAT GCTACAAATA
AACCATAATT TTGGTATTAA
CTCTAAACCG GAGATTTGGC
AAACAAGTCG TTTGTTCAGC
AAGAACTTCA TTCTTGAAGT
TATCTGAGGC ATAGACTCCG
ATGTAAACC TACATTTGG
GGTGGGGGGG CCACCCCCCC
TGAAAAAGGG GGGGGGTTAA ACTTTTTCCC CCCCCCAATT Bsml C
AACCA TTGGT
GTT CGCACT CAA GCGTG
TTTCTTGTTC
(cl
AAAGAACAAG
GCGCGTGTTC CGCGCACAAG
CCGCGATGTC GGCGCTACAG
TATATTCCAA ATATAAGGTT
ATGGAGCGGC TACCTCGCCG
AGGCTTTTTA TCCGAAAAAT
Kpnl
Us
TRs
TTGTAGAAAA AACATCTTTT
TCACAAAAAA AGTGTTTTTT
ATTTATTCAA TAAATAAGTT
ATTGGGCGTC TAACCCGCAG
CGCATGTCTG GCGTACAGAC
TGGTGTACGC ACCAECG
CAATCGGATA GTTAGCCTAT
TTTTTTCACT AAAMAGTGA
GTATGGGTTT CATACCCAAA
TCATGTTTTG AGTACAAAAC
GCATGTGTCC CGTACACAGG
280
ATATATATAT ATATATATAT TATATATATe_TATATATATA
ATATAGAGAA TATATCTCTT
AGAGAGAGAG TC~TCTCTC
350
GGGGTGTGGG
CGGGCTTTTC
ACAGAATATA
420
CCCCACACCC
GCCCGAAAAG
TGTCTTATAT
GCGGTTTTAT CGCCAAAATA CM AAATCGAT TTTAGCTA
458
Clal
Rsal &ml n 0
C
A
P pV6
A
p’J6 AA pv20
A
pV20AA
A
pV20 AC A
A
pV20 AAC PV3 pvo2
LH
RH
pvo37
FIG. 1. (a) Structure of the VZV genome showing the locations of the two copies of or&. (b) Sequence of the Kpnl plus C/al fragment containing oris (nucleotides 109, 893-l 10, 350 in the sequence of Davison and Scott, 1986). Motifs A, B, and C are boxed and the position of relevant restriction enzyme recognition sequences shown. The palindromic sequence is marked by arrows. (c) Structure of plasmid inserts. The Kpnl plus C/al fragment is represented diagrammatically in the top line and the positions of motifs A, B, and C and the palindromic sequence (P) are indicated. Shown below are the DNA sequences present in the various plasmids (solid lines) with internal deletions represented by carets. The cloned fragments are all flanked by EcoRl and HindIll sites at their left and right ends, respectively. The left-most nucleotide of the pV20 and pV20AA inserts is at position 226 in (b) (Le., the first residue of motif C). Plasmids pVO2 and pVO37 have been previously described (Stow and Davison, 1986). The latter contains an Xbal site at the position of the internal deletion which was utilized for the preparation of the LH and RH fragments.
In the present series of experiments we have examined the sequence-specific DNA binding properties of the VZV homolog of the HSV-1 UL9 protein. The binding sites for this protein, which is encoded by VZV gene 51 (McGeoch et a/., 1988a), have been located and their importance in origin activity assessed. The results reveal that although both the origin binding protein and its recognition sequence are highly conserved, there nevertheless exist significant differences between the
arrangement oris regions.
of binding sites within the HSV-1 and VZV
MATERIALS
AND
METHODS
Cells and virus Human fetal lung (HFL) fibroblasts (Flow Laboratories, 2002) were grown in Eagle’s medium supplemented with 10% fetal calf serum, nonessential amino
572
STOW,
WEIR.
acids, penicillin (100 units/ml), and streptomycin (100 units/ml). VZV of the strain described by Dumas et a/. (1981) was propagated in HFL cells as previously described (Stow and Davison, 1986). Plasmid
replication
assays
Subconfluent monolayers of HFL cells in 35-mm plastic petri dishes were used for plasmid replication assays which were performed as described previously (Stow and Davison, 1986). Briefly, monolayers were transfected with 0.25 pg circular plasmid DNA in the presence of calf thymus carrier DNA using the calcium phosphate technique followed by a DMSO boost. The cells were subsequently either mock-infected or superinfected by adding approximately 1 O6 HFL cells from a VZV-infected flask to each monolayer. Cellular DNA was prepared when the infected plates exhibited approximately 80% cpe (usually 3 days postinfection). Samples of DNA were cleaved with EcoRl plus Dpnl (Dpnl digests the methylated unreplicated input molecules but the products of plasmid amplification are resistant to its action) and analyzed by Southern blot hybridization using plasmid vector pTZl9U DNA, 32P-labeled in vitro, as probe. Plasmids The plasmids containing DNA fragments from the VZV oris region are illustrated in Fig. 1 c. The Kpnl plus C/al, Kpnl plus Bsml, and Bsml plus C/al fragments were inserted initially into pUC8 (Viera and Messing, 1982) and subsequently transferred to pTZ19U (Mead eta/., 1986) to facilitate site-directed mutagenesis. The orientation of each fragment is such that the EcoRl and HindIll sites of the polylinker lie at the left and right ends (as depicted in the figure), respectively. Plasmids pVGAA, pV20AA, pV20AC, and pV20AAC, which contain specific 11 -bp deletions of conserved sequences ‘A’ or ‘C’, were generated by site-specific mutagenesis using synthetic oligonucleotides. Single-stranded templates for mutagenesis were prepared by superinfection with Ml 3 K07 helper phage (Mead et al., 1986).
I (motif A) II (motif B)
5’ 5’
AND
STOW
Mutagenesis was performed by the method of Nakamaye and Eckstein (1986) using a commercially available kit (Amersham International Ltd., Bucks, UK) and oligonucleotides consisting of 15-l 7 nucleotides from each side of the sequence to be deleted. The identity of mutant plasmids was confirmed by DNA sequencing. Plasmids pVO2 and pVO37 have been previously described (Stow and Davison, 1986). As a result of Jhe deletion event a novel Xbal site is present in plasmid pVO37 (Fig. 1c). Expression
in E. co/i
The C-terminal region of VZV gene 51 was expressed as a fusion protein in E. co/i using the system previously employed for the corresponding region of the HSV-1 UL9 protein (Weir et a/., 1989). The Pvull plus Pstl fragment encoding the C-terminal 322 amino acids was inserted, in frame, between the Smal and Pstl sites of the vector pRIT2T+4 (Weir et a/., 1989) generating a hybrid gene which also encodes the Nterminal 260 amino acids. of the Staphylococcus aureus A protein. Propagation of bacteria (E. co/i strain K12AHlAtrp) containing the recombinant plasmid, heat induction of fusion protein expression, and the preparation of extract from sonicated cells (KV51 extract) were performed as previously described (Weir et a/., 1989). Extracts were similarly prepared from previously described transformants containing the parental vector pRIT2T (KR2 extract) or a plasmid encoding the corresponding HSV-1 UL9 gene fusion protein (KBl extract; Weir et al., 1989). DNA fragments gel retardation
and oligonucleotides analysis
DNA fragments for use as probes in gel retardation experiments were 3’ labeled with [32P]deoxyribonucleoside triphosphates and their overhanging ends filled in prior to purification from nondenaturing acrylamide gels. Duplex oligonucleotides containing the conserved motifs A and B (Fig. 1b) were formed as previously described (Weir et al., 1989) by annealing the following single-stranded synthetic oligonucleotides:
GATCAACCACCGTTCGCACTTTCTTTCT TTGGTGGCAAGCGTGAAAGAAAGACTAG GATCGTAAACCCGTTCGCACTTCCTGGGG CATTTGGGCAAGCGTGAAGGACCCCCTAG
Binding reactions were set up using bacterial extracts freshly diluted to 200 pg/ml in buffer C (Weir et al., 1989) supplemented with 600 mM KCI, 0.5 mM DlT, and 0.5 mM PMSF. Diluted extract (6 ~1) was added to 18 ~1 of 15 mlVl Hepes, pH 7.5, 1 mlVl EDTA,
for
5’ 5’
0.5 mM DlT containing 5 ng of labeled DNA fragment and 2 pg sonicated calf thymus DNA. Reactions were incubated at 25’ for 20 min and analyzed on 5% polyacrylamide gels as previously described (Weir et al., 1989).
VZV
DNase
GENE
51 PRODUCT
SITES
573
% amino acid Identity 60 -
I protection
DNase I protection experiments were performed essentially as described by Preston et al. (1988). KV5 1 or KBI extract was incubated with approximately 1O5 cpm of the plasmid pVO2 EcoRl plus HindIll fragment, uniquely 3’ labeled at either end, under the conditions employed for gel retardation analysis (Weir et al., 1989). After incubation at 25” for 20 min MgCI, and DNase I were added to concentrations of 5 mlL1 and 6 pg/ml, respectively, and incubation continued for a further 3 min. EDTA was added to a final concentration of 10 mM and the samples loaded on an 8% nondenaturing polyacrylamide gel. After electrophoresis the gel was exposed to autoradiographic film for 1 hr at 4’. Bands containing complexed and free DNA were excised and the DNA was recovered by electroelution, extracted sequentially with phenol and chloroform, and precipitated with ethanol. DNA fragments were analysed on a 6% denaturing polyacrylamide gel together with markers obtained by performing G or G plus A sequencing reactions (Maxam and Gilbert, 1977) on the end-labeled DNAs. RESULTS A candidate its possible
BINDING
VZV origin binding protein binding sites
and
In HSV-1 the UL9 gene encodes a polypeptide which binds to specific sites within the viral origins of replication (Olivo et a/., 1988). The VZV homolog of this protein is encoded by gene 51. The two proteins exhibit an overall 44% identity of amino acid residues, the distribution of which is shown in Fig. 2. A fragment of the UL9 polypeptide, comprising the C-terminal 317 amino acids and produced in E. co/i as a fusion protein, was previously shown to exhibit the same sequence-specific DNA binding ability as the authentic UL9 gene product (Weir et a/., 1989). The corresponding region of the VZV gene 51 (Fig. 2) was therefore expressed in the same system and its binding to the VZV oris region examined. We previously noted that the sequence CGTTCGCACTT, which corresponds closely to one of the binding sites within HSV-1 oris for the UL9 polypeptide, also occurs within a 259-bp Rsal plus C/al VZV DNA fragment (Fig. 1 b) which exhibits origin activity (Stow and Davison, 1986). Furthermore, Koff and Tegtmeyer (1988) proposed a consensus binding sequence YGYTCGCACT (Y equals pyrimidine) for the UL9 protein. This latter sequence occurs only four times within the VZV genome, in each instance as part of a CGTTCGCACTT motif. Two of these occurrences (A and B in Fig. 1 b) are close to the IRs copy of oris and the remaining two occupy corresponding positions near the TRs
60 -
40 J
20 -
OL
I N
1
<
1
I
I
I
I
I
200
400
600
600
c
ittTdn0 acid
residue
FIG. 2. Comparison of the amino acid sequences of the HSV-1 UL9 and VZV 51 gene products. The predicted amino acid sequences for these proteins (McGeoch et a/., 1988a; Davison and Scott, 1986) were aligned using the HOMOL program (Taylor, 1984) and the percentage of identical amino acids in adjacent 40 amino acid segments determined. These values are plotted across the length of the aligned proteins (which overall show 44% identical amino acids) with the N-terminus at the left. The bar in the lower portion of the figure shows the positions of identical amino acids (central vertical lines) and gaps inserted into the aligned UL9 and gene 51 protein sequences (marked above and below the central region respectively). The arrow indicates the regions of these proteins which were expressed In f. co/i and exhibit sequence-specific binding activity.
copy. Interestingly, the sequence CfilXGCACTT, which differs only at the indicated residue, is also present within each repeat region (C in Fig. 1b). A variety of VZV DNA fragments containing the motifs A, B, and C were therefore cloned and specifically mutated in order to assess the role of these sites in protein binding and origin activity. Detection of binding activity of the VZV gene 51 product Extracts were prepared from heat-induced cultures of E. co/i harboring either the plasmid encoding the VZV gene 5 1 fusion protein (KV51 extract) or the parental vector pRIT2T (KR2 extract) and examined in gel retardation assays using fragments from the VZV oris region as probes. Figure 3 shows that retarded complexes were observed only when KV51 extract was incubated with 32P-labeled pV3 and pVO37 LH fragments (lanes 3 and 8). The presence of an excess of unlabeled doublestranded oligonucleotides I or II, which contain the sequence motifs A and B, respectively, competed efficiently for binding to the labeled probe (lanes 4, 5, 9, and 10). No interaction was detected between the KV51 extract and labeled pVO37 RH fragment (lane 13). These results indicate the presence of functional
574
STOW,
pV037 6
7
LH 8
j
pVO37 12
WEIR,
RH 13
14
15
FIG. 3. Gel retardation analysis of the binding of the VZV gene 51 fusion protein to VZV oris. The pV3 EcoRl plus HindIll. pVO37 EcoRl plusMa (pVO37 LH), and pVO37 Xbal plus HindIll (pVO37 RH)fragments were used as labeled probes as indicated. Fragments were incubated without extract (lanes 1,6. and 1 1). with KR2 extract (lanes 2, 7. and 12) or with KV51 extract (lanes 3-5, 8-l 0, and 13-l 5). A 1 OO-fold molar excess unlabeled oligonucleotide I (lanes 4, 9, and 14) or oligonucleotide II (lanes 5, 10, and 15) was added to compete for sequence-specific binding. The positions of free probe (F) and the major retarded complexes (C, CU, and C,) are indicated. The fainter bands visible in tracks 3 and 8 probably result from interactions between fusion protein molecules and from the presence of a small amount of degraded fusion protein in the extract.
binding sites for the gene 51 fusion protein within oligonucleotide II, which contains sequence B from within the pV3 fragment, and oligonucleotide I, which contains sequence A from within the pVO37 LH fragment. Furthermore, in contrast to HSV-1, no binding site was detected at the right hand side of the palindromic DNA sequence (pVO37 RH fragment).
Fine-mapping of binding sites within the pV20 fragment The formation of two major retarded complexes following incubation of labeled pVO37 LH fragment, but not pV3 fragment, with KV5l extract (C, and CL in Fig. 3) suggests that the fragment may contain two binding sites, probably involving the related sequences A and C. To investigate further this possibility, a series of derivatives of plasmid pV20 containing specific 1 1-bp deletions eliminating sequence A, sequence C, or both were constructed and the appropriate fragments used in gel retardation assays. The results are presented in Fig. 4. As with the pVO37 fragment, the pV20 fragment yielded two retarded complexes specific for the KV5 1 extract (lane 3). Both the pV2OAA and pV2OAC fragments generated only a single complex, and no retarded band was observed with the pV2OAAC probe. These results demonstrate that the VZV gene 51 product binds to two
AND
STOW
sites within the pV20 fragment, one involving sequence A and the other involving sequence C., A striking feature is the difference in mobility between the complexes obtained with the pV2OAA and pV2OAC fragments. These fragments are of identical length and, as expected, the bands representing the free probe comigrate. The reduced mobility of the complex with protein bound near its center (pV2OAC fragment) compared to that with protein bound near one end (pV2OAA fragment) is consistent with the binding of protein resulting in bending of the DNA (Wu and Crothers, 1984; Zahn and Blattner, 1985; Mukherjee er al., 1985). The correspondence of sequences A and C with gene 51 product binding sites was confirmed by DNase footprinting using fragment pVO2 uniquely 3’ end-labeled on either strand. When DNase l-treated binding reactions were applied to a preparative nondenaturing gel both KBI (which contains the HSV-1 UL9 gene fusion protein) and KV5 1 extracts yielded two major complexes with each probe, essentially similar to the pattern observed with the pV20 fragment (Fig. 4). DNA from bands corresponding to the upper and lower (i.e., lower and higher mobility) KV51 complexes, the upper KBI complexes, and the free probes was analyzed on a denaturing gel (Fig. 5). Protection of sequences in the regions of both motifs A and C was apparent in the upper complex formed between KV5 1 extract and the probe labeled at its Hindill terminus. The corresponding complex formed with the probe labeled at the EcoRl site also showed protection in the regions of motifs A and C. The patterns of protection obtained with the upper complexes formed with KBI extract were almost indistinguishable demonstrating that the two viral polypeptides are capable of binding to the same 1
2
3
4
5
6
FIG. 4. Gel retardation analysis of binding to fragments lacking motifs A or C. The probes consisted of HindIll plus EcoRl fragments from plasmids pV20 (lanes l-3), pV2OAA (lane 4), pV2OAC (lane 5) or pV2OAAC (lane 6) which had been end-labeled to comparable specific activities. The pV20 probe was incubated without extract (lane l), with KR2 extract (lane 2) or with KV5 1 extract (lane 3). The remaining probes were each incubated with KV5 1 extract. Complex formation was analyzed as described in the legend to Fig. 3, and the major complexes are indicated by arrowheads.
VZV
Hmdlll
I
GENE
51 PRODUCT
ECORI
FIG. 5. DNase I protection by HSV-1 UL9 and VZV gene 51 fusion proteins. The pVO2 insert was uniquely 3’ end-labeled at either the fcoRl or Hindlll terminus, as indicated, and allowed to form complexes with KBl and KV51 extracts. Partial digestion with DNase I was performed prior to the resolution of complexes on a nondenaturing gel. DNA was extracted from bands corresponding to the pooled free DNA (F), upper KBI complex (Bl). lower KV51 complex (51L). and upper KV51 complex (51,) and analyzed on a 6% denaturing gel together with markers generated by partial cleavage at G and A residues (GA) or at G residues (G). The positions of sequence motifs A and C on the two strands are shown and the G residues within these motifs indicated (strands labeled at the HindIll and EcoRl sites correspond to the top and bottom strands in Fig. 1 b, respectively). The GA lanes contain some unexpected cleavages which were easily identified by comparison with other experiments.
BINDING
SITES
575
the cells was examined by Southern blotting for the presence of Dpnl-resistant (replicated) plasmid molecules. Figure 6 shows the result of an experiment using plasmids pV6, pV20, pV3, and the vector pTZ19U. VZV-encoded functions facilitated the replication of plasmids pV6 and pV20 but not pV3 or pTZlSU, indicating that the former two plasmids contain functional viral origins of replication. The insert in plasmid pV20 is similar to that in plasmid pVO2 (Fig. 1) confirming our earlier mapping of an origin within the &al plus CM fragment. Quantification by autodensitometry of this and other experiments revealed no significant difference between the replicative abilities of plasmids pV6 and pV20. The presence of binding site B therefore appears not to influence origin activity. The failure of plasmid pV3 to be replicated demonstrates that the presence of a gene 51 protein binding site in the absence of the palindromic DNA sequence is insufficient for activity. Figure 7 shows experiments assessing the roles of sequences A and C in origin function. Deletion of binding site A from either pV6 or pV20 abolished detectable replication. Binding at this locus is therefore absolutely essential for DNA synthesis, and absence of site A cannot be compensated for by the presence of sites B and C. Following deletion of site C from plasmid pV20 there was a significant but reproducible reduction in origin function. In repeat experiments plasmid pV2OAC replicated with between 20 and 50% of the efficiency of its parent pV20. Binding site C may therefore play some auxiliary role in DNA amplification in the presence of site A. DISCUSSION The results presented in this paper have identified the VZV gene 51 product as a sequence-specific origin pTZlQU pV6 abababab
pV3
pV20
DNA sequences. The lower complexes formed with KV51 extract exhibited partial protection over both the A and C motifs, suggesting that they contain DNA molecules with the fusion protein bound to one or the other site. This conclusion is supported by the relative mobility and diffuse character of the pV20 lower complex seen in Fig. 4. Role of sequences
A, 6, and C in origin activity
To test for origin activity plasmids containing DNA fragments were transfected into HFL cells subsequently either mock-infected or infected VZV to provide helper functions. DNA extracted
VZV and with from
FIG. 6. Replication of plasmids containing fragments from the VZV oris region. HFL cells were transfected with the indicated plasmids and either mock-infected (a) or infected with VZV (b). Samples containing 10% of the DNA recovered from each monolayer were digested with fcoRl plus Dpnl and the fragments were separated by electrophoresis, transferred to nitrocellulose, and hybridized to W labeled plZl9U. An autoradiograph of the washed filter is shown.
576 pTZlQU ababababab
STOW, PV6
6AA
pV20
20AA
pv20 a b
WEIR,
20AC a
b
FIG. 7. Replication of plasmids containing specific deletions. HFL cells were transfected with the indicated plasmids (the pV prefix has been omitted from plasmids pVGAA, pV20AA. and pV2OAC) and analyzed as described in the legend to Fig. 6. Lanes a and b contain DNA from mock-infected and VZV-infected cells, respectively.
binding protein, located its binding sites within the VZV oris region, and investigated their roles in origin activity. Because we were unable to reproducibly detect origin binding activity in extracts of VZV-infected HFL cells (data not shown), we chose to express the region of the VZV gene 51 protein homologous to the DNA binding domain of its HSV-1 counterpart, the UL9 polypeptide, in f. co/i. Using this product we have shown that, although the origin binding proteins of HSV-1 and VZV exhibit high levels of amino acid identity and contain sequence-specific DNA binding domains within their C-terminal regions which are capable of interacting with identical sequences, there nevertheless exist interesting differences between the interactions of these proteins with their cognate origins. The homology plot for the two proteins (Fig. 2) reveals that the C-terminal binding domains exhibit lower similarity than the remaining portions. It is tempting to speculate that the region of greatest similarity within the binding domains (amino acids 750-800) may be involved in sequence recognition, and further experiments are addressing this question. The regions of maximum amino acid identity within the N-terminal portions contain several motifs found in many purine nucleotide-utilizing helicases (Gorbalenya et al., 1989) suggesting that the HSV-1 UL9 and VZV gene 51 proteins are probably able to unwind DNA strands (either at the origin or elsewhere) in an energy-dependent reaction. Indeed, recent experiments using model substrates have confirmed that the UL9 protein exhibits helicase activity (M. D. Challberg and J. M. Calder, personal communications). The ability of the truncated fusion proteins to bind to the origins of replication suggests that the helicase function is probably dispensable for the sequence-specific interaction with DNA. We have demonstrated that the DNA sequence mo-
AND
STOW
tifs A, B, and C which conform closely to the consensus binding sites for the HSV-1 UL9 protein are involved in the binding of the VZV gene 51 product. These must represent the only major sites of interaction within the Kpnl plus C/al fragment (Fig. 1) since no retarded band was observed with the pV20AAC probe and only a single prominent complex with the pV3 probe (Figs. 3 and 4). The absence of complex formation with the pVO37 RH probe indicates that the sequence TGTTCGCGCGT (nucleotides 356-366 in Fig. 1 b) which matches motifs A and B in 8 of 1 1 positions does not function as a binding site for the gene 51 product. Previous work demonstrated the presence in HSV1 of two UL9 binding sites in opposite orientation and overlapping the ends of the palindromic DNA sequence (Elias and Lehman, 1988; Koff and Tegtmeyer, 1988; Olivo et al., 1988; Weir et a/., 1989). In contrast, in VZV oris the three binding sites are all located on the same side of the palindromic sequence and are arranged in the same relative orientation (Fig. 1). Sites B and C lie 168 and 62 bp from the palindromic DNA sequence, respectively. The most distal site, B, appears dispensable for DNA synthesis while site C possibly plays an auxiliary role. Nevertheless the ability of plasmid pV20AC to be replicated demonstrates that a single gene 5 1 protein binding site alongside the palindromic DNA sequence suffices for efficient origin activity. Whether sites B and C are of greater importance for DNA synthesis in the context of the viral genome or play some alternative role is not known. The roles of the two binding sites in HSV-1 oris function have also been investigated. The higher affinity site (site I) is absolutely essential for origin activity (Deb and Deb, 1989; Weir and Stow, 1990) while the absence of a functional site II apparently leads to inefficient replication (Lockshon and Galloway, 1988; Weir and Stow, 1990). Conflicting results on the importance of site II have, however, been presented by Deb and Doelberg (1988) who found no impairment in activity with an origin fragment completely lacking this binding site. Both DNA binding studies (results of M. D. Challberg cited in Challberg and Kelly, 1989) and experiments studying the effects upon origin activity of inserting various numbers of copies of the dinucleotide AT into the center of the HSV oris palindrome (Lockshon and Galloway, 1988) suggest that cooperative interactions between UL9 protein molecules bound to sites I and II can occur. These interactions are possibly important for allowing the initial opening of the HSV-1 origin region. In VZV, gene 51 protein bound to a single site adjacent to the palindromic sequence may perform a similar function, the longer AT-containing sequence within
VZV
GENE
51 PRODUCT
VZV ori, possibly facilitating easier unwinding of the DNA strands. Alternatively another viral or cellular protein could bind to the right-hand side of the VZV oris palindrome and substitute for the role played by UL9 protein interacting with site II of HSV-1 or&, or significantly different events might follow the sequence-specific binding of the UL9 and gene 51 proteins. These could include DNA bending, DNA unwinding, or interactions with other components of the replicative machinery. Such possibilities should be amenable to experimental investigation.
ACKNOWLEDGMENTS We thank J. H. Subak-Sharpe, A. J. Davison, R. D. Everett, R. M. Elliott, and C. M. Preston for helpful suggestions and criticisms, and 1. McLauchlan for the synthesis of oligonucleotides. H.M.W. was the recipient of a Medical Research Council Studentship.
REFERENCES CHALLBERG, M. D., and KELLY, T. J. (1989). Animal virus DNA replication. Annu. Rev. Biochem. 58, 67 l-7 17. CRUTE. 1. J., TSURUMI, T., ZHU, L., WELLER, S. K., Owe, P. D., CHALLBERG, M. D., MOCARSKI, E. S., and LEHMAN, I. R. (1989). Herpes simplex virus 1 helicase-primase: A complex of three herpes-encoded gene products. Proc. Nat/. Acad, Sci. USA 86, 2186-2189. DAVISON, A. J., and Scorr. J. E. (1986). The complete DNA sequence of varicella-zoster virus. J. Gen. Viral. 67, 1759-l 816. DEB, S., and DEB. S. P. (1989). Analysis of oris of HSV-1: Identification of one functional DNA binding domain. Nucleic Acids Res. 17, 2733-2752. DEB, S., and DOELBERG, M. (1988). A 67-base pair segment from the oris region of herpes simplex virus type 1 encodes origin function. 1. Viral. 62, 2516-2519. DUMAS, A. M., GEELEN, J. L. M. C., WESTRATE, M. W.. WERTHEIM, P., and VAN DER NOORDAA, J. (1981). Xbal. Pstl and Bglll restriction enzyme maps of the two orientations of the varicella-zoster virus genome. /. viral. 39, 390-400. ELIAS, P., and LEHMAN, I. R. (1988). Interaction of origin binding protein with an origin of replication of herpes simplex virus 1. Proc. Natl. Acad. Sci. USA 85,2959-2963. ELIAS, P.. O’DONNELL, M. E., MOCARSKI, E. S., and LEHMAN, I. R. (1986). A DNA binding protein specific for an origin of replication of herpes simplex virus type 1. Proc. Nat/. Acad. SC;. USA 83, 63226326. GORBALENYA. A. E., KOONIN, E. V., DONCHENKO, A. P., and BLINOV, V. M. (1989). Two related superfamilies of putative helicases involved in replication, recombination, repair and expression of DNA and RNA genomes. Nucleic Acids Res. 17,47 13-4730. KOFF. A., and TEGTMEYER, P. (1988). Characterisation of major recognition sequences for a herpes simplex virus type 1 origin-binding protein. J. Viral. 62, 4096-4103. LOCKSHON, D., and GALLOWAY, D. A. (1988). Sequence and structural requirements of a herpes simplexviral DNA replication origin. Mol. Cell. Biol. 8, 4018-4027.
BINDING
SITES
577
MAXAM, A. M., and GILBERT, W. (1977). A new method for sequencing DNA. Proc. Natl. Acad. Sci. USA 74, 560-564. MCGEOCH, D. J., DALRYMPLE, M. A., DAVISON, A. J., DOLAN, A., FRAME, M. C., MCNAB, D., PERRY, L. J., Scan, 1. E., and TAYLOR, P. (1988a). The complete DNA sequence of the long unique region in the genome of herpes simplex virus type 1.1. Gen. Viral. 69, 1531-l 574. MCGEOCH, D. J., DALRYMPLE, M. A., DOLAN, A., MCNAB, D., PERRY, L. J., TAYLOR, P., and CHALLBERG, M. D. (1988b). Structures of herpes simplex virus type 1 genes required for replication of virus DNA. 1. viral. 62,444-453. MEAD. D. A., SZCZESNA-SKORUPA, E., and KEMPER, B. (1986). Single stranded DNA ‘blue’ T7 promoter plasmids: A versatile tandem promoter system for cloning and protein engineering. Protein Eng. 1 I 67-74. MUKHERJEE, S., PATEL, I., and BASTIA, D. (1985). Conformational changes in a replication origin induced by initiator protein. Cell43, 189-l 97. NAKAMAYE. K. L., and ECKSTEIN, F. (1986). Inhibition of restriction endonuclease Neil cleavage by phosphorothioate groups and its application to oligonucleotide-directed mutagenesis. Nucleic Acids Res. 14,9679-9698. Ouvo, P. D., NELSON, N. J., and CHALLBERG, M. D. (1988). Herpes simplex virus DNA replication: The UL9 gene encodes an origin blnding protein. Proc. Nat/. Acad. Sci. USA 85, 5414-5418. PRESTON, C. M., FRAME, M. C., and CAMPBELL, M. E. M. (1988). A complex formed between cell components and an HSV structural polypeptide binds to a viral Immediate early gene regulatory DNA sequence. Ce//52,425-434. STOW, N. D., and DAVISON, A. J. (1986). Identification of a varicellazoster virus ongin of DNA replication and its activation by herpes simplex virus type 1 gene products. /. Gen. Viral. 67, 16 13-l 623. STOW, N. D., and MCMONAGLE. E. C. (1983). Characterisation of the TRs/lRs origin of DNA replication of herpes simplex virus type 1. Virology 130, 427-438. TAYLOR, P. (1984). A fast homology program for aligning biological sequences. Nucleic Acids Res. 12,447-455. VIERA, J., and MESSING, J. (1982). The pUC plasmids, an Ml3 mp7derived system for insertion mutagenesis and sequencing with synthetic universal primers. Gene 19, 259-268. WEIR, H. M.. and STOW, N. D. (1990). Two binding sites forthe herpes simplex virus UL9 protein are required for efficient activity of the oris replication origin. /. Gen. Viral. 71, in press. WEIR, H. M., CALDER, J. M., and STOW, N. D. (1989). Binding of the herpes simplex virus type 1 UL9 gene product to an ongin of viral DNA replication. Nucleic Acids. Res. 17, 1409-l 425. WELLER, S. K., SPADARO, A., SCHAFFER. J. E., MURRAY, A. W., MAXAM, A. M., and SCHAFFER, P. A. (1985). Cloning, sequencing and functional analysis of or&, a herpes simplex virus type 1 origin of DNA synthesis. Mol. Cell. Biol. 5, 930-942. Wu, C. A., NELSON, N. J., MCGEOCH, D. J., and CHALLEERG, M. D. (1988). Identification of herpes simplex virus type 1 genes required for origin-dependent DNA synthesis. J. K-o/. 62,435-443. Wu, H.-M., and CROTHERS, D. M. (1984). The locus of sequence-directed and protein-induced DNA bending. Nature (London) 308, 509-513. ZAHN, K., and BLAT~NER, F. R. (1985). Binding and bending of the lambda replication origin by the phage 0 protein. EMBO /. 4, 3605-36 16.