vncvoay 197, 116-124 [1989 A Cosmid-Based System for Constructing Mutants of Herpes Simplex Virus Type 1 CHARLES CUNNINGHAM ano ANDREW J. DAVISON' MRC Virology Unit, stitute of Vrology. Church Sree, Glasgow G11 BIR, United Kingdom Received March 17, 1998; accepted Juy 7, 1903 Cosmids containing large fragments of harpes simplex vue type 1 DNA wore prepared using a vector that allows Intact insors to be excises using the restrieion endonuclease Pact. Two independent gets VA and B) of five comme ‘Were identified whose Inserts overlap and eprasent the entire viral genome, and set C was obtained by replacing two cosmids in set 8, Each set gave rise to viral plaques when digested with Pac and transfected into cells culture, Two ‘cosmids common to sets B and C ostensibly contain one of the origin of viral DNA replication lori) in @ region of overlap between insets, but both actualy consist of a minority of apparently intact fort) forms and 9 majority of deleted (ri?) forms. These sets yildad exclusively on vital progeny. Whon either of these cosmids was replaced by a derivative comprising only or), forms, ai, nd ori; progeny were obtained, and only ori, progeny were producad when ‘both were replaced. One cosmid in set A contains the ri locus in a nonovertapping ragion and lacks or): forms. This set generated only ov; vis, Viral mutants wit lesions in either or both of genes UL2 and UL¢4, which are not essential for growth in cell culture, were constructed using cosmids containing spacifcaly introduced frameshitt ‘mutations. A mutant with a trameshit mutation in an essential gone (UA.29} was solotedby transfecting a complement ing callie, These result indicate that a cosmid-based aystem wil facilitate isolation of arge numbers of dtined vat INTRODUCTION The considerable amount of information now avail: able on the biology and molecular biology of herpes: simplex virus type | (HSV-1), the prototype of the a-her pesviruses, has accumulated largely rom genetic stud: jes. Early mutants were generated without detailed knowledge of viral genes and were selected on the basis of host range properties (e.g., Aurelian and Roiz man, 1864), plaque morphology (e.g. Ejercito et al 1968), drug resistance (e.g. Kit and Dubbs, 1863), or temperature sensitivity (e.g., Schaller et af, 1970). The lesions in drug-resistant mutants map in a very few genes encoding inhibitor-susceptible enzymes (e.g thymidine kinase and DNA polymerase), and those in temperature-sensitive mutants are in genes required for growth under normal cell culture conditions (so called “essential” genes), Later, mutants that contgin lesions in surface glycoproteins and are resistent to immune cytolysis (e.g.. Machtiger et al, 1980) or neu- tralization by monoclonal antibodies (e.g., Holland er al,, 1983) were added to the armory. Determination of the complete DNA sequence of HSV-1, culminating in the paper by McGeoch et al. (1988), opened the way for specific mutation of each viral gene, including those that had previously been inaccessible to genetic analysis, ‘0042-6822799 $5.00 Cent 8100 by Aweerve Ps he ‘aeons 6 ‘Transfection methods have enabled insertion or de- letion mutants in many HSV-1 genes to be isolated, The plasmid-based method usually involves cotrans- {ecting colls with intact viral DNA and a plasmid bear ing the target which has been modified by insertion of a ker. Selectable markers include the HSV-1 thymi- dine kinase gene (Post et al, 1981) and a gene confer- ting resistance to neomycin (Neidhardt et a/, 1987), and detectable markers are exemplified by E, coli /acZ (Goldstein and Welter, 1988). Mutants arise by recom- binotion and are isolated from wildtype (wi) virus by virtue of expression of the marker. The possibilty of reversion can be eliminated by replacing part of the target gene in the plasmid by the marker instead of merely inserting it. Additional manipulations can be cartied out to obtain revertants, true deletion mutants lacking the marker or more complex mutants. ‘An alternative means for constructing mutants was developed by van Zl eal (1988) for another a-herpes- virus, pseudorabies virus (PRY). Cosmids containing large PRY DNA fragments were constructed, and I braries containing four of five cosmids whose inserts, overiap and represent the entire genome were identi fied, Intact cosmid inserts were obtained by cleaving at flanking EcoRI sites (which are absent from PRV DNA) and putiied by sucrose-density centrifugation, When the inserts were cotransfected into rabbit kidney cells, viral plaques were produced via recombination be tween overlapping DNA fragments. The genome pro- files and growth properties in vitro and in vivo of wt and(COSMID-SASED MUTAGENESIS OF HSv-1 7 reconstructed PRV were indistinguishable. Moreover, the utlty of this system for producing PRV mutants by replacing the wt cosmid with a specifically mutated derivative was demonstrated, and there are several in- stances where it has been used to produce mutants (eg,, de Wind et af, 1990). Nevertheless, the unavail ability of the complete PRV DNA sequence poses a present limitation on genetic studies of PRV by this and other means: ‘The plasmid- and cosmid-based systems should be viewed as complementary, since each has different metits. The former allows mutations to be introduced easily at specifically targeted loci and can be ultiized to generate revertants, thus allowing phenotynic changes to be associated directty with the introduced mutation. This possibilty is not inherent in the cosmid based system, but the association can instead be es- tablished by examining several independent mutants. ‘The cosmid-based system does, however, have sev- eral advantages that make it very attractive. Most im portantly, mutants may be isolated in the absence of wt virus. In contrast, transfection of parental DNA in the plasmid-based system may result in cytopathic effect (CPE) or a high background of parental plaques, which may obscure recombinant plaques, particularly those formed by seriously disabled but nonetheless viable mutants, Also, since a marker is not required, the cos- mid-based system is @ powerful too! for directly gener- ating large numbers of random or quasirandom mute tions containing minimally disruptive changes (i... in- sertions or deletions of ¢ few bese pairs). Therefore, development of a cosmid-based system tor HSV-1 is a desirable objective, particularly since its application could usefully exploit the available so- ‘quence data, We have succeeded in constructing such a system, and demonstrate its utility for generat: ing mutant HSV-1 viruses containing single or mutiple lesions in essential or nonessential genes. We have recognized in our strategy that one of the origins of HSV-1 DNA replication (orj) consists of a large pelin- ‘drome that is readily deleted when cloned in bacteria (Gray and Kaerner, 1984; Quinn and MeGeoch, 1985; Weller et al, 1986). This sequence is not required for viral growth in cell culture or for establishment and reactivation of latency in mice (Polvino-Bodnar et al, 1987), but its conservation in HSV-1 and HSV-2 sug- gests that it has @ role at some point in the natural infectious cycle, We have found that ori, can be main- tained unstably in cosmid populations, and that viruses possessing ari, can be constructed without difficulty. MATERIALS AND METHODS Cells BHK C13 cells (MacPherson and Stoker, 1962) were grown in Glasgow modified Eagle's medium supple- mented with 100 U/ml penicillin, 100 ug/ml streptomy- cin, 0.3%6 (wvvol) tryptose phosphate broth, and 10% (vol/vol) newborn bovine serum, Preparation of cosmids DNA was isolated trom cell-released virions of HSV- 1 strain 17 (Brown et al, 1973) 3s described by Wikio (1973) and cloned into a cosmid vector derived from SuperCos 1 (Evans et a., 1988; Stratagene}. SuperCos 1 contains genes conferring resistance 10 ampicilin and neamycin and two adjacent packaging (cos) sig- nals from bacteriophage 4 separated by an Xbal site The region for inserting foreign DNA contains & single Bamiil site flanked by bacteriophage 17 and T3 tran- scriptional promoters and Not! and EcoRI sites in the configuration EcoRI-Notl-T7 _promoter-BemH!-T3 promoter-Notl-EcoRl. Nott and EcoRI cleave the HSV- 1 genome in several locations, so SuperCos 1 was modified by replacing the ‘Vor! fragment with a syn- thetic oligonucleotide duplex to ge SuperCos 1MW, containing a cloning region with the configuration EcoRI-Not|—Asc|-Pacl—BamHi —Asc\~-Pac|—Notl— EcoR, where the sites are contiguous and nonoverlap- ping. Paci does not cleave HSV-1 DNA and can be Used to excise HSV-1 inserts. HSV-1 DNA fragments were cloned into SuperCos 1MW according to insiructions supplied by Strata- geno. Tho vector was treated with Xbal and calf intos- tinal phosphatase and then digested with BemH!, The resulting mixture of two fragments, each terminated by 98 site, was ligated to HSV-1 DNA that had been partially digested with Mol (which recognizes the se- quence GATC) to produce large quasirandom frag: ments and treated with calf intestinal phosphatase. A Gigapack Il Plus kit (Stratagene) was employed to package the ligated DNA into \ heads, which were then used to infect Escherichia coli NM554, Of the ampicillin resistant clones obtained, 112 were ana- lyzed by digestion with restriction endonucleases. Sets of cosinids apparently containing the entire genome were characterized further by digestion with additional restriction endonucleases. Bulk preparations of cos mids were obiained by alkaline lysis (Birnbom and Doly, 1979) of bacteria from 200-ml cultures, treated with 100 ag/ml RNase A and purified by precipitating with 11% (wt/vol) polyethylene glycol in the presence of 1.36 M NaCl. The concentrations of commercial DNA standards were verified by spectrophotometry, and concentrations of cosmid DNAs (as EcoR digests) were estimated by careful comparison on ethicium bromide-stained agarose gels viewed by uv illumina- tion. Sequences at the ends of all he inserts endin the region of ori in one cosmid (cos71) were obtained after subcloning EcoR!-Smal fragments and a BamHl trag- ment, respectively, into bacteriophage M13mp19 RFI118 ‘CUNNINGHAM AND DAVISON NA, The average size of the cosmid inserts used in this study is 37.4 kb. Mutagenesis of cosmids Individual cosmids were linearized at unique Xbal sites, treated with bacteriophage T4 DNA polymerase in the presence of the four deoxynucleoside triphos. phates to produce blunt ends, ligated, and transfected into E. coli MAX efficiency DHBa (Gibco-BRL). This strategy should insert 4 bp into the Xbal site, convert ing the sequence TCTAGA into TCTAGCTAGA. The resulting clones were screened for loss of the target Xbal site, and the precise neture of the insertion muta- tion was confirmed in one (cos14) by sequencing @ Smal tragment subcloned into bacteriophage Miamp19 RFI DNA, Generation of complementing cell lines Call lines were generated by transfecting 40 4a of a cosmid comprising SuperCos 1 and HSV-1 DNA from 43423 to 78287 bp in the published sequence (ie from within gene UL27 to within gene UL36; MeGeoch et al, 1988) into @ 35-mm dish of BHK C13 calls by calcium phosphate precipitation and 20% (vol/vol) di- methyisulfoxide (OMSO) boost (Stow and Wikio, 1976). Alter 24 hr at 37°, the cells were diluted in 96- well rays and selected with 400 ug/ml Geneticin G418 sulfate (Gibco-BRL). Surviving colonies were sub- jacted to two further rounds of isolation by limiting chiu- tion in the presence of G418. The complementing cell line used in this study was designated 208 Generation of viruses from cosmids Cosmids were digested with Pacl (New England Bio: labs) for 16 hr at 37°. Cost7 was also treated with Xbal. The DNA was extracted with 1:1 (volvo) phenol:chloroform, ethanol precipitated and resus- pended in a small volume of 10 mM Tris-HCl (pH 7.5), 1 mM EDTA, and the integrity of the digests was checked by agarose gel electrophoresis. Equimolar amounts of digested cosmids representing the entire HSV-1 genome were mixed, and the DNA was trans- fected into 90% confluent BHK C13 cells in 24-well ‘rays by calcium phosphate precipitation and DMSO boost. A volume of 100 ul of caleium phosphate precip- itated DNA, including 2 ug of calf thymus DNA, was added to each well, The cells were incubated at room temperature for 30 min, 1 ml of medium was added, and incubation was continued for a further 3.6 hr at 37°. After boosting at room temperature for 3 min 45 826, the DMSO solution was replaced with 1 ml of me- dium and the cells were incubated for 2 hr at 37°. The medium was then replaced with 1 ml of 1.5% (wt/vol) carboxymethylcellulose (sodiurn salt) in_ medium. Plaques, which were visible under the microscope after incubation at 37° for about 3 days, were removed. from the wells in 5 ul of carboxymethyloellulase overiay using plastic pipet tips, mixed with 0.5 ml of medium, sonicated, and stored at ~70°. Viral stocks were pre- pared by infecting cell monolayors in 6-well trays with 0.25 miof virus for 1 hr, overlaying with 2 ml of medium and incubating at 37° until CPE was complete (3-4 days). Infected cells were scraped into the medium, sonicated, and stored at ~70°. For comparison with viruses generated trom cosmids, wt virus was obtained by transfecting HSV-1 ONA. ‘Screening of viral genotypes Cell monolayers in 24-well rays were infected with 260 il of viral stock for 1 hr, overiayed with medium, and incubated for 16 hr at 37°. The medium was de canted and infected cell DNA was prepared (Stow et al, 1983). The DNA was digested with appropriate re sttiction endonucleases, electrophoresed on agarose or polyacrylamide gels, and transterred to nitrocellu- lose membranes (Southern, 1975). Immobilized DNA \was hybridized to [*P]-labeled DNA probes prepared by the random-priming method (Feinberg and Vogel stein, 1983) and autoradiographed. The probes used included HSV-1 DNA and cloned HSV-1 Kon frag: ments b,c, d, and v (Preston et él, 1978; Davison and Wilkie, 1983), RESULTS Cosmid libraries As @ result of the initial screening procedure, two independent sets (A and 8) of five cosmids were identi- fied that contain the entire HSV-1 genome (Fig. 1). Two of the cosmids in set B were subsequently evaluated as posing potential problems for generating virus. Cos18 was found to contain a mutation resulting in an additional Xbal site about 700 bp from the wt site at 97,667 bp. The precise nature of this mutation, which is located in UL44 or UL45, was not established. Also, by virtue of the presence in viral DNA of four genome isomers differing in the orientations of U, and Us, the ingert in c0817 consists of the left end of U, linked to ‘he left end of Ug via the flanking inverted repeats (Fig 1), Thus, although the cost7 and cost8 inserts are partially homologous and could recombine, the lack ot homology between their termini might affect recombi nation frequency. The likelihood of this occuring was reduced by digesting cos17 not only with Pac! but also with Xbal, which cleaves at 10,634 bp, in order to ex- cise the majority of nonhomologous terminal se quence. Nevertheless, with these considerations in mind, cos18 and cos17 were replaced by cosS6 and cos4a, respectively, to give set C (Fig. 1), which thus shares 00848 with Set A, Cos66 contains only the wz(COSMID AASED MUTAGENESIS OF HSv-1 119 6 we & % mo 150 kop {eee ec erreur, pe ue eu ws Set A cos32 cos24 cost cost cosa SetB cos6 c0s28 cost4 cost8 cost7 Sete cos6 0828 costs coss6 cosa, Fa. 1. Locations of cosmic inarte with respect io the HSV-1 genome. Tho gonome is depicted a9 wo unigue regione horizontal nes, the largor U, and the smaller Uy, fanked by inverted repeats ectanges). Tha four Xba stes are nlcsted by veral arfows, and he ganes within wiose pr eimoodig regions they are located are neicated, Locations of cosmic inser are indicated by open Sars, an ther precise ovations inthe sequence are given in base pars on the rch, The position of on is Genoted by vertical lines in aporonrate ori; cosmids. Xbal site, and the termini of the cos48 insert are fully homelagous to the regions of neighboring cosmid in: serts that they overlap, Generation of viral plaques When transfected into BHK ceils, each of the three cosmid sets generated viral plaques. Thus. the prob: lems anticipated for set 8 did not materialize, Plaques generated from each set usually took up to 24 hr longer to appear than those from HISV-1 ONA. The absolute numbers of plaques obtained in individual experiments depended on the amount of DNA transfected, and the efficiency of plaque formation by cosmids relative to HSV-1 DNA seemed to be affected by unidentified vari ables associated with the transfected cells. Table 1 shows the results of three independent experiments in which the DNA input of set C or HSV-1 DNA was var- ied, Depending on the experiment and DNA dose con- sidered, the cosmids were at least 259% as efficient at {generating plaques as an equimolar amount of HSV-1 DNA (e.g... 0.2 ug DNA in Experiment 1), and in some ‘cases the efficiency approached that of HSV-1 DNA, (€.g., 0.2 ug DNA in Experiments 2 and 3). The num. bors of plagues generated from defined inputs of DNA were of the same order as those reported by van Zill et al. {1988) for the PRV cosmid-based system, and re- combination of the cosmid inserts to form viable ge- homes appears to be a very efficient process for PRV and HSV-1 Status of ori, ‘The region containing ori is clearly unstable in the cosmids (Fig. 2). Although cos28 and cas14 contain a minor proportion {approximately 20%) of molecules that produced full length Koni v, a major proportion generated a deleted form of this fragment (év). This is characteristic of deletion af the ori, palindrome (Gray and Keerner, 1984; Quinn and MeGeoch, 1986; Weller et.ai., 1985) and, although these cosmids are mixtures, we term them ani’, Derivatives of cos28 and cos14 that completely lack ori, were obtained by colony-puri- fying twice; these are termed ari; . DNA sequencing af 0871, which is also ori, showed that the orj-asso- ciated palindrome has been lost by deletion of 148 bp, from 62,393 to 62,640 bp in tho HSV-1 DNA se- quence, and that the deletion has occurred between120 ‘CUNNINGHAM AND DAVISON TABLE 1 RELATIONSHIP BETWEEN THE AMOUNT OF TRENSEECTED HSV-1 OR Costa DNA AND THe NUMBEROF VIRAL PLAQUES FORMED Number of plaques Micrograms of Experiment 1 Experiment 2 Experiment 3 ‘genomic DNAIwell Hav Costrids HSV Cosmids Hv Cosmids a ° ° ° ° ° ° 0.02 = ° 7 5 ~ , ocs 6 7 19 4 2 - 0.08 a 7 a a ~3 04 = 8 = 2 ~ v 02 3 “4 ~1 «0 ~s0 43 oa ~82 aa cree ~106 CPE a 08 ope 38 PE ~102 re 33 2 GPE 2 OPE ~86 Cre “4 Not cone, A goneral OPE was observad but the plaques were oo numercus fo be counted, Plaques formed using cosmids in the cortesponaing ‘experiments were stained when they were sight smaller and therefore could be courted at higher densities. 8-bp direct repeats, precisely as described for plas- mids and bacteriophage X clones containing the ori, ragion (Gray and Kaerner, 1984; Quinn and McGeoch, 1985; Weller et af, 1985; Polvino-Bodnar et af, 1987) ‘Agarose gel electrophoresis analysis of the cosmid li- brary indicated that this is the most common deletion of ori, but other smaller or larger deletions were noted. The endpoints of rarer deletions were not deter- seen, Fig. 2 Status of orn cosmids. Kan fragments of cos28, an ony ‘cerivative of cos28, cost4, an or Gerivatve of eos14, and cos7* ‘oro transterod rom 30.7% (eval agarose gol and probed witha plasmic containing Kn. which containg the oi, egi0n. Fragmente Fepresenting the Tukiength (v, 1933 9p) and deleted (ay; approx 1800 Dp) farms af Kony ae inicated, Hybridization to lager frag ments resulted rom the presence of vectr-containng secvences in tne probe. mined, but may carrespond to forms nated previously by Weller et a). (1985) and Polvino-Bodnar er al. (1987). ‘The genotypes of viral progeny from each cosmid set were analyzed by blot hybridization of infected cell DNA. Eight plaques were examined for sets A and B, and seven for set C. Representative results for sets A and C are shown in Figs. 3-5, The KonI-Xbal profiles of all the progeny from set C were identical to that of we virus, except for fragments that are known to be natu- rally variable in size (e.g., Kpn| | and k), whereas those {rom set A exhibited Kpni dv rather than Koni v (Fig. 3). Profiles for the set B progeny were wt, except for the presence of the additional Xbal site contributed by ‘cos18 (data not shawn). Probing with Koni v confirmed that all of the progeny trom set C were orit and all of the prageny from set A were oriy (Fig. 4): all of the progeny from set 8 were also orig (data not shown). The resolution of Fig. 4D implies thet if ori? progeny lack any sequence in the vicinity of ori,,no more than 5 bp are missing. f progeny fram the other experiments described below are included, whether mutated or not at other sites in the genome, all of the 26 progeny from sets B or C were oriy and all ofthe 16 progeny from set Awere ori. Those results show that ori molecules in 0814 oF os28 apnear exclusively in viral progeny, even though ori: virus is viable When cos28 was replaced by an oric derivative, six of eight progeny analyzed were orit one was orir ,and ‘one was a mixiure. When cos'4 was replaced by an ori: derivative, four of eight progeny analyzed were ‘ori? one was ori; and three were mixtures. These results imply that orft forms of a cosmid are not at a disadvantage in the absence of ori? forms, and thus become incorporated into viral progeny. When both cosmids were replaced by ori: derivatives, all of the eight progeny analyzed were ori.(COSMID BASED MUTAGENESIS OF HSV-1 124 vs ge 63 BaBaSS BRGNSS bd ii ky di 1 “BL 1B 2B Fic. 3. Restiction proties of we vius, vnuees constucted Laing cosmid seis C of A, and mutants produces using set C. Al the viruses were generated using normal BHK coll. infected cell DNA. wos digaetog wit Kool and Xbal,tonsleted fom a 0.7% twee) ‘agor0g0 9el, end probed with HSV-1 ON, The postions of Kent- “bal ragmens 61 (8973 bp), B2 (5637 bp), 1 (7285 Op), and 2 {6343 bo) and she fagments frm whieh they originate, Ken! b (12,810 bp) and 6 (12,629 0), are indicated an he ght. Kan kt ‘and vate dicoted onthe let ard are naturaly vanabiein sie, and | comigrated in simiar positions 10 61 and 61, and By com rated wath Kon w and x Construction of mutant viruses AAs indicated in Fig, 1, the HSV-+ genome contains four Xbal sites (Wilkie, 1976), each conveniently lo- cated in a single cosmid in at feast one cosmid set. These sites aro in genes UL2, UL22, UL33, and UL44 (MoGeoch eral, 1988), The site in UL 22 when present in cosmids is cefractory to Xbal because it overlaps @ ‘dam site (GATC) which is methylated in the bacterial strain used to propagate the cosmids. Frameshift mu- tations were introduced into the other three genes by linearizing cosmids with Xbal, generating blunt ends and religating. Loss of Xbal sites was established by restriction endonuclease analysis, and the precise mu- {ation in UL33 was confirmed by DNA sequencing. UL2, which encodes uracil-DNA glycosylase, and UL44, which encodes glycoprotein C. are known to be nonessential for growth of virus in cell culture (Mul faney ef af, 1989; Draper et ai, 1984), and single and double mutants were obtained by transfecting normal BHK cells. UL2~ mutants were produced using sets B ‘and C containing mutated cos6, UL44~ mutants were ‘generated using set A containing mutated cos6 1 or set C containing mutated cos6, and the double mutant ‘was obtained using sot C containing mutated cos6 and c0s56. Six plague isolates were analyzed from each. All of the set A UL44" mutants were avi: and have identical KoniiXbal profiles, lacking spectic we frag- tents a8 appropriate (Fig. 3). Similarly, all set G mu- tants in a particular gene were orit (Fig. 4) and dis- played identical profiles (Fig. 3), except for one UL4a~ mutant which yielded 8 novel large Konl fragment of ‘unknown origin (data not shown). The profiles of set B ULZ mutants wore also identical, each containing or and exhibiting the additional Xbal site present in cos {data not shawn), The UL2™ mutants lack the Xbal site in Kpnt b, resulting in replacement of b1 and b2 by b (Fig. 5a), the UL44~ mutants lack the Xbal site in Kon 4, resulting in replacement of dt and d2 by d (Fig. 5b), ULz-ULAs UL33- 20) unas- uL3- uL33- HSV-1 wt HSV-1 set C HSV-1 set A. un UL33- (204) uL33 és 2g rad 1SV-1 wt HSV-I set 3 2 53 in y ur33- + ork 281- me Fig 4. Sratus of a in vt ius, vieuses constructed using cosmid sets C or, and mutants produced using setC. Alle viruses were (gonerated using normal SHK cele excopt one o! ne ULIT" mutors, Which was obtained using 20 cals, Inlected celNA was digested With fa) Ken or) Bare anc BSE, wanetered from (a) 0.7% (wwivol)agerose ge! oF (b) 8 5% (w/v) polyacyismide gel and probed wth a plasmid containing Karl v.Tho poaitans of Koni van by a¢6 showin fa, and tha locations of en? andan raginents and ‘adjacent fragments ere indicated in (sizes ace in base pais,12 CUNNINGHAM AND. DAVISON and the UL2-UL#4~ mutants lack both sites (Figs. 50 and 5b). UL33is known to be essential for virus growth in cel culture (ALKobsisi et @!, 1984), so a complementing call line (204) was prepared by transforming BHK cells with @ cosmid containing UL39. Seven viral plaques obtained using set C containing mutated cosi4 were analyzed and each lacks the Xbal site in Konic, result ing in replacement of ct and c2 by c (Fig. 5c}. Surpris- ingy. plaques were also abtained using normal BHK calls at an efficiency of up to 10% of that obtained using 20A calls, Analysis of seven plaques showed thal they comprise not UL33 virus alone but wt vieus ‘containing the Xbal site in Konic (four plaques} or mox tures of wt and ULAI- vius tthree plaques. Fig. 5c) ‘One ofthe plaques isolated on 20A cells and two of the plaques isolated on norma! BHK cells were analyzed for the presence of or (Fig. 4). The 20A plaque was 8 mixture of ori’ and ori forms, one of the BHK plaques was also a miture, and the other was ori, This is consistent with the observed absence of orit forms from UL33" cos14 (data not shown). High-requency reversion, presumably occurring by oss of 4 bp (CTAG) to regenerate the Xbal site, is probably facilitated by active recombination processes, since the proportion of infectious virus able to form plaques on normal BHK Calis was variable but low in stocks of 10 independent UL3S" mutants isolated on 20A calls. The upper limit was 10° but the lower limit (nominally less than 10°") ‘was dificult to judge, since large amounts of some viral stocks induced CPE without forming detectable plaques DISCUSSION ‘The aim of this study was to develop a cosmid-based ‘system for constructing HSV-1 mutants. We generated three sets of cosmids (two independent and one de- rived), each of which is capable of yielding HSV-1 plaques atter digestion with Pacl and transfection into BHK cells. We then constructed viral mutants contain- ing frameshift mutations in either or both of genes UL2 and UL44, which are not essential for growth in cell culture, and also produced a mutant with a frameshift ‘mutation in UL33, an essential gene, by transfecting a complementing cell line. These genes were not of spe- cific interest but were chosen because they contain convenient Xbal sites. In demonstrating the utility of the system, we have not ruled out the possibility that the cosmids may contain undetected mutations that ae transferred to progeny virus. We note, however, that a virus stock obtained using set C showed no growth impairment in comparison with wt when human diploid fibroblasts (Flow 2002} cells were infected at HSV-1 set C 2 HSV-1 ot HSV-1 sot A L unas 2 3 a8 b aera & 38 Pa Sha ae egggssy c wocewe "Se | worere w0-2 Fria 6. Status of Xbalstos in wt virus, veusos constructed using ‘cosmid cats C oF A, and mutants produced using set C. Al the ‘ise ware generated using normal BHI calls excep ane ofthe ULS9" mutants, which was obtaines using 20A cells. nected call DNA was cigested wih Kon and Xtal, ranstorred rom 0.79%6{wthva) ‘2gnr030 gol, and probod wih plasmid containing (a) Korb, (o} pal 0, oF (e) Kon. The postions of Kort b, 8, snd ¢ (12,798 bp) wren lacking Xbal sites and the Xbal products of these tragments fare shown on tne right by bt, b2, 41, 62, Ct (10,212 bp) ard 2 (2586 bs). Konlb contain part of an inverted repeat and thus hybrid: ized in al dgesta to 9 fragmont (Kon! g,inckeatod onthe let in (a trom the other repoat.(COSMID-BASED MUTAGENESIS OF HSV-1 ea 5PFUlcell and monitored for production of infectious Virus during 2 24-hr period (data not shown). We intend 10 compare the in vivo growth properties of wt and reconstructed viruses in a separate study. The ability of the UL3S- mutation to revort readily when under selec tion during the recombination process indicates that transfection of normal BHK cells with cosmids contain ing small duplications presents @ simple method for determining whether the target region is essential for growth in tissue culture. This assay would not neces sarily succeed with deletion mutants because they cannot revert genotypically to wt, but transfection of ‘cosmids containing minimal deletions into normal cells might form a useful means of selecting phenotypically wt viruses produced by mutation at secondary sites. In developing a cosmic-based system for HSV-1, we investigated the arj region in cosmids and vial prog- eny in some detail. Isolation of viable ori virus con- firms an earlier report that ari, is not essential for growth of HSV-1 in cell culture (Polino-Bodnar et ai, 1987}. The ori locus is ina region of overlap between the insens in cosi4 and cos28 in sets B end C and, ‘even though the majority of DNA molecules in these cosmids lack ori, only ori? viral progeny were ob- tained. ori and ori; progeny were isolated when either of these cosmids was replaced by a fully orc derivative and only ori: progeny were produced when both cos- mids were replaced. Production of orit virus from cos- mids that are predominantly ori, presumably results from the ability of ori! molecules to replicate and pre. dominate in the viral progeny owing to a dosage effect or an enhancad abilty to recombine. The high propor- tion of mixed progeny obtained when cos14 or cos28 was replaced by oric forms may have resuited from having picked overlapping plaques. Some isolates (eg. the UL33™ revertants) were obtained from wells that contained few plaques, however, and it is likely that these were @ consequence of active recombina- tion occurring during generation of viral genomes in @ single infectious center. Mutants containing ori, can be obtained easily if the mutated cosmid does not contain the ori, locus. ori mutants can also be isolated using mutated cosmids that contain the ori region if the palindrome is main- tained (in sets where ori, is in a single cosmid) or the progeny are screened carefully {in sets where on isin € region of overlap}, Of the throe cosmid sets produced during this study, set B is unsuitable for further develop- ment because of the presence of an additional Xbal site in C0818. Set C generated progeny that appeared to be wt, and we are now employing ito. obtain arange of HSV-1 mutants containing minimal lesions. The availabilty of a second cosmid set able to yield wt virus would facilitate isofation of mutants with lesions in re- gions of overtap between inserts in set C. To this end, we are currently replacing cos71 in set A with an oni? alternative. ACKNOWLEDGMENTS We thank John Subak Sharpe for continuing interest, Christine ‘MaoLean and Nigel Stow for erica eading ofthe manuscript, ard Moira Watson for technical assisiance with preparing cosmids REFERENCES ALKoaais, MF. Roxon, F 1, MeDouae, 1, and Preston, ¥. (1981). The herpes sislex vrus UL3S gene orocuct is required {or the asserbly of fll capsids, Yrology 180, 990-208, ‘AUPeELIAN, Land RODMAN, B. (1964). The host range of herpes sr ‘plex virus lncerteron, ural DNA, and antigen synthesis m apart infection of dag key calls. Virology 22, 462-461 Bison, H.C. and Doty (1873), A rapid alkane extraction pro ‘cecure far scrgening recombinant plasmid DNA, Nucleic Acias es, 7, 1813-1923, Brown, 5. M., RICH, D. A., and SuBAK-SKARPS, J H. (1973), Ga netic studies with herpas simplex virus type 1. The isolation of ‘emperature-sensitve mutants. their arrangement inio comple ‘mentation graups and recombination analyss leading 06 linkage map. Gen, Vio! 18, 329-346, ‘Dawson, A, and Witte, N. M. (1983). Location and orentaion of homologous sequences inthe genomes of fe herpestiruses Gen. Virol 64, 1927-1942 De Wawa, N. Zupewveo, A, Guiensun, K., GELKENS, A. ond eens, A (1990), Laker insertion mutaganesis of herpsewruses: Inactivation of single genes within the Ul region of pseudorabes vitvs.J, Viol. 64, 4691-4566, parce, KG. Costa R.H. LEE, G.T-¥., SPEAR, P.G., ond WAGNER, EK. (1880) Motecularbesisa' the glycoprotan C-nagatve phono ‘ype of herpes simplex virus type 1 macroplaque strain. Vira! Bt, 378-585 sera, P. M., Kier, E.D., and Rowuay, B. (1968). Characterize tion of herpes simplex wius string fering n ther elleets on | social Bohaviour of inlociod cals. J Gen. Viol. 2, 357-368. Eves, GA, Lews, X., and RomHensero, B. . (1988). High a= lency vectors for cosmid micrecioning and genomic anayss Gene 79, 8-20, Fenecrs, AP. and Vogmsten, B.(1883}. A technique for race ‘baling ONA rasticton endonuclease fragments to high specifc activity Anal. Boohem, 132, 6-13 Gowwsren, DJ, and WeuLeR,. K. (1988) Herpes simplex veus type Trnduced ribenucleatde reductase ectvty « clepensatle for virus growth and DNA syntheses: galtion and charactenzation ot an IP6 lacZ insercan mutant. J. Yvel 62, 198-205, Gray, C.P., and KaeRNEs, H.C. (1984), Sequence o! the curative ‘origin of rephoatan inthe U, ragion of herpes simplex wrus type ANG DNA. J. Gen. Virol 65, 2108-2178 Houano, TC, MARUN, SO, LEVBE, M,, nd GLOmO8, J (1983). ‘Antigenic variants of eps simplex vitis selected with glyeopre tein pectic monoclonal antibodies. J.Virol. 45, 872-882. ‘er, S., and Duses, D. R. {1983}, Nonfunctional thymidine funase Cision in beomadeosjurdine reeletant strain of herpes siespox ‘rus. iachem: Signtys. Res. Commun. 13, 500-608, McGzocm, 0.1, DALRYWRLE.M.A., DAWSON, A.J, DOLAN, A, FRAME, IM. Molina, D., Penny, LJ, Scorr,).E., and Tavton, . (1888) The complete DNA sequence of the ong unique regian in ho (Jenome of narpos simplex vius ype 1.1. Gan, Vira! 68, 1831 1874. MacPHERGOW, |. and Sroxen, M. (1962) Polyome transformation of124 CUNNINGHAM AND DAVISON hamster call clones—An investigation of genatic actors afecting cell competence. Virology 16, 147-15? MacHTIGeR, N.A. PANCAKE, B.A., FdeRe, R., COURTNEY, A J, TE VET#A, 5.5, and SHAFEER, PA, (1960), Herpes Smplox vius ahcoprotens: isolation of mutans resstontw immune ev osis. J Vir 34, 336-286, Muuuaner, J, Moss, H. W. MeL. anc McGeoor, DJ. (1968), Gene LULZ of herpos simplex vius type 1 ancodos a Urac-DNA glycosy- lase. J, Gen, vrot 70, 449-254 NepHsnor, H., Scw0De, C. H, and Kagmnen, H.C. (1887), Herpas ‘simplex veus tye 1 ghcoprctein Eis ret indispersable lox val infoctaty. J Vr. 61, 600-803. Po.wno-Boowaa, M,, OrseRa, P. K, and SoHAcrER, P. A, (1987). Herpes cimplex vius typo 1 ov, is ot required for vius repication ‘ror theestabishment and reactivation olaten infecton in mice. J Virol 81, 3528-3536, Post. LE, Macken, §, and Rowman, B. (1981), Regulation of @ ‘ones of herpes strplex virus: Expression of chimeric gonoe aro- ‘duvea by fus10n oF thymidine kenaso with» geno promoters. Cot! 24, $55-565. PRESTON, V.G., DAVISON, A.) MARSDEN. H.S.. Taauny, M.C.. Sv fsacSharee, | H., and Wiuae, NM. (1978). Recombinant bo- ‘ween herpes simplex vius types 1 and 2: Analyses of genome secures and exorossion of immeciate-eary polypeptiges. Wok 28, 498-817 (Gun, JP and Meroe, 0. (1986), DNA sequence ofthe region ‘inthe genome ef herpes simoiex vive ype | containing the gones {or ONA polymerase and the major DNA binding protein. Nucl Acids Aes. 13, 8143-8163, ‘ScqAEFER,P..VOMEA, V, LEME, R, and Banvests Mec, M.(1970). Temperaturesensiive mutans of herpes simplex wus. Viology 42, 1144-1146, ‘SouTweRy, E, M. (1975). Detoction of specie sequences among DNA fragments separated ty gel elecvophoresis. Mol Biol. 98, 508-517 ‘STOW, N.D., MeMoNAGie, E.G. and Dawson, A. J. (1983). Frog rents trom Roth termini of the herpes simplex virus ‘ype | ge rome contin gna required forthe encansdation of wal ONA, {Nuoleie Acids Res. 11, 8205-8220. ‘S10W, N. D., ond Wx, NM, (1978). An improved technique for ‘obtaining enhanced infectivity with herpes simpiex vius ype 1 BNA. J Gen. Virol. 33, 447-258. ‘yaw Zi, M:, OUNT, W, BRINRE, J, DE ROWER, T., GIELEENS, A, and ‘Berns, A, (1986), Reganeration OT haypecvituses Kom molecularly cloned subgenomic Hagments. J, Viol. 62, 2191-2195, WELLER, S.K, SPADARO, A, SCHAFFER, JE, MURRAY, BW, MAKIN, 'A.1h, and SoxAFFER, A. (1988). Cloning, sequencing, and func” tional analysis of or. a herpes simplex vius type 1 orgin of DNA syntnesis. Mol Call Biol 8, 990-942, {Wine N. Ml (1973). The symhesis and substructure of hopeevis DONA: The distribution of akaable single strand Interruptions in HBV-1 DNA, J. Gen, Virol 21, 453-467 ‘We, N, M. (1976), Physical maps for herpes simplex virus toe 1 DONA for restction endonucleases Hind Il, Hoa) and X. Qa Virol 20, 222-283,