JOURNAL OF VIROLOGY, May 1980, p. 336-346 Vol. 34, No. 20022-538X/80/05-0336/11$02.00/0

Herpes Simplex Virus Glycoproteins: Isolation of MutantsResistant to Immune Cytolysis

N. A. MACHTIGER,' B. A. PANCAKE,' R. EBERLE,2 R. J. COURTNEY,2 S. S. TEVETHIA,3 ANDP. A. SCHAFFER'*

Sidney Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts 02115'; Department ofMicrobiology, University of Tennessee, Knoxville, Tennessee 379162; and Department ofMicrobiology,

Milton S. Hershey Medical Center, Pennsylvania State University, Hershey, Pennsylvania 17033'

Immune cytolysis mediated by antibody and complement is directed againstcomponents of the major herpes simplex virus (HSV) glycoprotein complex(molecular weight, 115,000 to 130,000), comprised of gA, gB, and gC, and againstglycoprotein gD-all present on the surfaces of infected cells. Tests with atemperature-sensitive (ts) mutant of HSV-1 (tsAl) defective in glycoproteinsynthesis at the nonpermissive temperature (39°C) demonstrated that over 90%of mutant-infected cells maintained at 39°C and treated with antibody andcomplement were not lysed, presumably due to the absence of viral glycoproteinson the surface of infected cells at this temperature. Furthermore, a small numberof tsAl-infected cells could be detected among a large excess of wild-type virus-infected cells by virtue of their failure to be lysed at 39°C by antibody andcomplement. Making use of the involvement of viral glycoproteins in immunecytolysis and the ability of cells infected with glycoprotein-defective mutants toescape cytolysis, we sought mutants defective in the expression of individual viralglycoproteins. For this purpose, antisera directed against the VP123 complex andagainst the gC and combined gA and gB glycoprotein subcomponents of thiscomplex. were first tested for their ability to lyse wild-type virus-infected cells inthe presence of complement. Wild-type virus-infected cells were lysed aftertreatment with each of the three antisera, demonstrating that the gC glycoproteinand the combined gA and gB glycoproteins can act as targets in the immunecytolysis reaction. Next, these antisera were used to select for mutants whichwere resistant to immune cytolysis. Cells infected with wild-type virus which hadbeen mutagenized with 2-aminopurine and incubated at 39°C were treated withone of the three types of antisera (anti-VP123 complex, anti-gC, or anti-gAgB)and lysed by the addition of complement. Cells which survived immune cytolysiswere plated, and virus in the resulting plaques was isolated. Plaque isolates weretested for temperature sensitivity of growth and altered cytopathic effects in cellculture at 34°C (the permissive temperature) and 39°C. A total of 73 mutantswas isolated in this manner. Selection with glycoprotein-specific antisera resultedin a 2- to 16-fold enrichment for mutants compared with "mock"-selected mutantsusing normal rabbit serum. Phenotypically, 24 mutants were temperature sensi-tive for growth, 27 were partially temperature sensitive, and 22 were not temper-ature sensitive but exhibited markedly altered cytopathic effects at both permis-sive and nonpermissive temperatures. Nine mutants of each phenotype (temper-ature sensitive, partially temperature sensitive, and non-temperature sensitive)were selected at random for confinnatory immune cytolysis tests with the antiseraused in their selection. Cells infected with eight of the nine mutants were shownto be significantly more resistant to immune cytolysis at the nonpermissivetemperature than were the mock-selected mutants or the wild-type virus fromwhich they were derived.

Infection of cells with herpes simplex virus mine-labeled infected cell extracts, are found in(HSV) results in the synthesis of new, virus- purified membrane fractions of virus-infectedspecific glycoproteins (22, 25). These new gly- cells and in the envelopes of progeny virions (7,coproteins, which can readily be demonstrated 25, 26).by sodium dodecyl sulfate-polyacrylamide gel Although the precise functions of these gly-electrophoresis (SDS-PAGE) of ['4C]glucosa- coproteins in the biology of the virus have yet to

336

HSV MUTANTS RESISTANT TO IMMUNE CYTOLYSIS 337

be elucidated, one glycoprotein (gB) has recentlybeen shown to be essential for virion infectivity(19), whereas virus replication can apparentlyproceed in the absence of synthesis of another(gC) (10). The expression of the syncytial (syn)phenotype characteristic of cells infected withsome strains of HSV appears to involve at leasttwo viral glycoproteins (10) and three or more

distinct genetic loci (18; S. P. Little and P. A.Schaffer, manuscript in preparation). Glycopro-tein gE has been identified as the virus-inducedFc receptor on infected cell surfaces (1). In ad-dition, viral glycoproteins have been implicatedas the antigens which stimulate the humoralimmune response in the infected host (15). Vi-rus-infected cells are susceptible to lysis eitherby sensitized lymphocytes (14) or complement-dependent antibody (17). Glorioso and Smithand Glorioso et al. (5, 6) have shown that anti-body-dependent, complement-mediated cyto-toxicity is directed against components of themajor viral glycoprotein complex (VP123; aver-

age molecular weight, 123,000), observed in SDSpolyacrylamide gels. Recently, Norrild et al. con-firmed these findings and showed that gD (mo-lecular weight, 50,000 to 59,000) also functionsas a target for cytolysis (12). Additionally, theseinvestigators have demonstrated that all fourglycoprotein antigens (gA+gB, gC, gD) serve as

targets for cell-mediated immune cytolysis invitro (12). In addition to their role in the HSVreplicative cycle and in the induction of theimmune response, viral glycoproteins have beendetected on the surfaces of HSV-1 and HSV-2-transformed cells (2, 15, 16; R. J. Courtney,manuscript in preparation). That these glyco-proteins may play a role in modulating tumori-genicity is suggested by the observation thatanimals bearing HSV-induced tumors developneutralizing antibody to the virus (4, 9, 14).As major virion structural components, gly-

coproteins reach mimal rates of synthesis latein infection, consistent with their classificationin the y class of viral proteins (8). Five majorglycoproteins, gA, gB, gC, gD, and gE, derivedfrom four antigenically distinct polypeptides, aresynthesized (1, 24). Three of these polypeptides,gB, gC, and gD, are glycosylated in two stagesyielding partially glycosylated intermediates andfully glycosylated products. The final productsare found in the envelope of mature virions andin infected cell membranes. Glycopeptide gA,which is antigenically related to gB (R. Eberleand R. J. Courtney, personal communication; P.Spear, personal communication), is not exten-sively glycosylated, and little of it is incorporatedinto virions relative to glycoproteins gB, gC, andgD (24).To investigate the roles of individual glyco-

proteins in the biology of HSV, we made use ofthe involvement of viral glycoproteins in anti-body-dependent complement-mediated cytotox-icity.We describe here a procedure for the selective

enrichment of HSV mutants defective in theexpression of specific glycoproteins at the cellsurface, using antisera directed against individ-ual viral glycoproteins.

MATERIALS AND METHODSCells and cell culture. Monolayer cultures of hu-

man embryonic lung (HEL) fibroblasts (passages 8 to17) were used as target cells in immune cytolysisexperiments. Vero cell cultures were used for the prep-aration of virus stocks, for virus assay, and for theassay of infectious centers after treatment of infectedHEL cells with antibody and complement. Both HELand Vero cells were grown at 37°C in Eagle medium(Autopow, Flow LIaboratories, Inc., Rockville, Md.),supplemented with 10% fetal bovine serum containing0.075% NaHCO3 (for cultures in closed vessels) or0.225% NaHCO3 (for cultures in open vessels in a 5%CO2 atmosphere) (13).

Viruses and virus assays. Properties of the HSV-1 wild-type strain, KOS, and of the temperature-sen-sitive (ts) mutants, tsAl, tsE6, and tsJW2, have beendescribed (20, 22). Mutants tsAl, tsE6, and tsJ12 aremembers of complementation groups 1.-1, 1.-5, 1.-9,respectively, in the standard set of HSV-1 complemen-tation groups (21). Virus stocks were prepared asdescribed previously (13). Virus assays and infectiouscenter assays were performed in Vero cells with a 2%methylceliulose overlay in CO2 (5%) incubators(Wedco, Silver Spring, Md.) with temperature varia-tions of 0.2°C. The permissive temperature was 34°C,and the nonpermissive temperatures were 390C and400C.Immune cytolysis assay. The protocol employed

for the measurement of antibody-dependent, comple-ment-mediated cytotoxicity of virus-infected cells issummarized in Fig. 1. Confluent monolayers of HELcells (3 x 10' cells in 25-cm2 flasks) were infected withHSV-1 at a multiplicity of 2.5 to 5 PFU/cell. Afteradsorption for 1 h at 34°C, inocula were removed, andmonolayers were rinsed twice with 2 ml of Tris-buffered saline (pH 7.4) and overlaid with prewarnedmedium. Infected cultures were then incubated for theappropriate interval at either 340C or 390C. Afterincubation, monolayers were rinsed, and cells weredetached from the support medium by treatment with0.5% EDTA. Cells were pelieted by centrifugation at800 x g for 5 min and were suspended and diluted to105 cells per ml in serum-free medium. This cell sus-pension was then dispensed in 1-ml volumes, and cellswere pelleted a second time. Cells in paired tubes werethen suspended in either 0.1 ml of antiserum or 0.1 mlof control serum at the indicated dilutions and incu-bated for 60 min at 340C in a shaking water bath. Anequal volume of a 1:2 dilution of guinea pig comple-ment (GIBCO Laboratories, Grand Island, N.Y.) wasadded, and the suspension was incubated for an addi-tional 90 min in a shaking water bath at 34°C. Afterthe second incubation, infected cells were diluted in

VOL. 34, 1980

338 MACHTIGER ET AL.

medium containing 0.2% pooled human gamma glob-ulin (Merck Sharp & Dohme, West Point, Pa.) andplated with 3 x 106 indicator Vero cells at 34°C. (0.2%gamma globulin was sufficient to reduce the titer ofwild-type virus by greater than 99.9%.) When cells hadattached to petri dishes, they were overlaid with me-dium containing 2% methylcellulose. After 5 days ofincubation, neutral red was added to the overlay, andthe plaques, representing infectious centers derivedfrom cells which survived lysis by antibody and com-plement, were counted. The relative sensitivity of apopulation of infected cells to immune cytolysis wasdetermined by comparing the number of survivinginfectious centers in the treated population with thatin the population treated with normal serum.

This procedure was developed to measure the re-duction in infectious centers when a population ofvirus-infected cells was treated with antibody andcomplement before plating on indicator cells. Cellshaving viral antigens on their surfaces were lysed withantibody and complement under the conditions ofassay, and such cells did not form infectious centers.Cells not so treated or cells treated with antibody andcomplement before the appearance of viral antigenson their cell surface were not lysed and subsequentlygave rise to infectious centers upon plating. The plat-ing overlay contained immune gamma globulin to pre-vent plaque formation by cell-free, progeny, or unad-sorbed virus. Thus, the plaques which appeared prob-ably originated from virus in infected cells.

Infect HEL Cells with Virus(3X 106 Cells, Moi 5 PFU/Cel I)

Incubate 34° or 39°, 0-18 hrs.

Detoch and Pellet Cells

Reupend in Serum-Free MediumDilute to 105 CeIls/mI

Pellet Cells in I ml Cell Suspension

Resuspend in Antiserum Resuspend in Control Serum

Incubate 34°,60 min.

Add GP Complement

Incubate 34°,90 min

Dilute Infected Cells In Medium ContoiningHumon y Globulin and Plate with Indicator Cells, 34°

Overlay with Methylcellulosel I

Count Plaques after 5 Days, 340

FIG. 1. Immune cytolysis assay.

KOS-infected

Incubate 39°,18 hrs.

TS Al-infected Cells

Sensitivity of the immune cytolysis assay: re-construction experiments. Cells infected with mu-tants which do not express HSV-specific glycoproteinsor which express antigenically altered glycoproteinsshould be insensitive to immune cytolysis and, there-fore, should be recoverable from mixtures of cellsinfected with wild-type virus after treatment of thetotal population with antibody and complement. Theefficiency of recovery of mutant-infected cells wastested in the following way. Cells were infected eitherwith wild-type virus or with tsAl, a mutant defectivein glycoprotein synthesis at 390C (22). The two typesof cells were incubated at 390C for 18 h, mixed, treatedwith antibody and complement, plated at 34°C withindicator cells, and overlaid with medium containing2% methylcellulose (Fig. 2). Such cultures should yielda population of plaques enriched for those initiated bythe mutant.

Antisera. (i) Human antiserum to HSV. Serumwas obtained from an individual who experienced fre-quent, recurrent facial lesions caused by HSV-1. Thisserum, obtained 1 week after the appearance of themost recent recurrence, had a 50% plaque-reductionneutralization titer to HSV-1 of 1:512.

(ii) Antiserum to HSV-infected cells. The pro-cedure used for the preparation of antiserum to HSV-1-infected cells was that described by Sim and Watson(23). Extracts of HSV-1-infected rabbit kidney cellsmaintained in rabbit serum were used as the immu-nogen.

(iii) Antiserum to the major viral glycoproteincomplex, VP123 (average molecular weight,123,000). The preparation of polypeptide-specific an-tisera was carried out essentially as described byCourtney and Benyesh-Melnick (3). The viral glyco-protein complex (VP123) was purified from HSV-1-infected HEp-2 cells as described by R. Eberle and R.J. Courtney (manuscript in preparation). Briefly, cellswere infected at a multiplicity of 10 PFU/cell withHSV-1, strain KOS. Cells were harvested after 24 h,and glycoproteins were solubilized with 1% sodiumdeoxycholate and 1% Tween 40 for 1 h at 37°C. Theresulting extract was centrifuged at 100,000 x g for 1h, and the glycoprotein-containing supernatant fluidswere removed. The glycoprotein complex was sepa-rated by two cycles of preparative SDS-PAGE, andfractions containing the complex were pooled, concen-trated, and electrophoresed on cylindrical tube gels.The section of the tube gel containing glycoproteinsof molecular weight 115,000 to 130,000 was cut fromthe gel and frozen. Gel sections were later thawed andemulsified in Freund complete adjuvant. The emulsionwas used to immunize rabbits by the schedule de-scribed previously (3).

(iv) Antisera to viral glycoproteins gAgB andgC. The techniques used for the fractionation of virus-infected cells by preparative SDS-PAGE and SDS-

Pick PloquesMix-Treot with - Plote - ond Test/ Antibodyand at 34° for Temperature-

Complement Sensitivity

FIG. 2. Scheme for recovering tsAl-infected cells from an excess ofwild-type virus-infected cells by immunecytolysis using anti-HSV serum.

J. VIROL.

HSV MUTANTS RESISTANT TO IMMUNE CYTOLYSIS 339

hydroxylapatite column chromatography and thepreparation of antisera using the glycoproteins puri-fied in this way have been described (3; Eberle andCourtney, in preparation). Briefly, glycoproteins weresolubilized as described above and fractionated bySDS-hydroxylapatite chromatography by the proce-dure of Moss and Rosenblum (11). The two majorbands resolved were repurified twice by SDS-hydrox-ylapatite chromatography and were used for the prep-aration of antisera as described above. One band con-tained viral glycoprotein gC, and the other containedglycoproteins gA and gB. Thus, anti-gC antiserumprecipitated a single protein, and anti-gAgB precipi-tated two proteins, when immunoprecipitates of HSV-infected cells were examined by SDS-PAGE. No cross-reaction was demonstrated between anti-gC and anti-gAgB antisera (Eberle and Courtney, in preparation).Recent studies have demonstrated that the gA and gBcomponents are antigenically related (R. Eberle andR. J. Courtney, unpublished data; P. Spear, personalcommunication).

Mutagenesis. FourHEL cell cultures were infectedat a multiplicity of 1 PFU/ceil with HSV-1 strainKOS, using a virus stock which had been plaquepurified three times (Fig. 3). After adsorption, mono-layers were overlaid with 2 ml of maintenance mediumcontaining 700 yg of 2-aminopurine per ml. This con-centration of 2-aminopurine had been shown previ-ously to reduce the yield of HSV-1 by 80% after 48 h(J. Jofre, R. J. Courtney, and P. A. Schaffer, manu-script in preparation). Infected cultures were incu-bated for 48 h, at which time pronounced cytopathiceffects (CPE) were observed. Cultures were then fro-zen at -90°C. Later they were thawed, subjected totwo additional cycles of freezing and thawing, andclarified by low-speed centrifugation. Virus in super-

Infect 4 HEL Cell Cultureswith HSV-I, Moi PFU/Cell

Incubate 340, 48hr. in 700/Lg/ml2-Amino Purine

Harvest

i A1. e- O. 14.4 i 4 4

Infect 4 HEL Cell Cultures withMutogenized Stocks, Moi 2.5 PFU/Cell

I I I IIncubate 39° 12 hr.

Treat with Antiserum to:I. NRS 2.Anti-VP 123 Region 3.Anti-gAgB 4Anti-gC

and Add Complement

Plate Surviving Cells

1t 1 '1 1Pick Plaques, Test Isolates

for Temperature - Sensitivity,Altered CPE andResistonce to Immune Cytolysis

FIG. 3. Mutagenesis and selection of HSV-1 mu-tants resistant to immune cytolysis.

natant fluids from individual tubes was then used inthe mutant selection step. Multiple mutagenizedstocks were employed to avoid the selection of clonallyrelated mutants.Mutant selection. Cultures of HEL cells (2 x 106

to 3 x 106) were infected at a multiplicity of 2.5 PFU/cell with 2-aminopurine-mutagenized virus stocks(Fig. 3). Virus was adsorbed for 1 h at 34°C. Inoculawere then removed, and monolayers were rinsed threetimes with 2 ml of Tris-buffered saline, pH 7.4. Mono-layers were overlaid with maintenance medium whichhad been prewarmed to 39°C, and cultures were in-cubated for an additional 11 h at 39°C in a water bath.After incubation, monolayers were rinsed twice, andcells were detached with EDTA. Cells were washed,pelleted, and treated with antiserum as describedabove. The dilutions of antisera employed had beendetermined previously to yield an 80 to 90% decreasein infectious centers in cells infected with the wild-type virus. After incubation of the serum-cell mixturesfor 60 min at 34°C in a shaking water bath, an equalvolume of complement was added to each tube, andincubation was continued for 90 min. The cells werethen diluted and plated with indicator Vero cells sothat plates received approximately 10 to 1,000 infectedcells per dish. After 6 h of incubation at 34°C inimmune gamma globulin-containing medium, mediumwas removed from monolayers and replaced with me-dium containing immune globulin and methylcellu-lose. After 5 days of incubation at 34°C, neutral redwas added, and plaques were counted and picked thefollowing day. Plaque isolates were transferred to testtube cultures of HEL cells, and cultures were incu-bated until complete CPE was observed. The cultureswere then frozen and thawed three times to releasevirus; these virus stocks were frozen at -90°C. Theplaques isolated were then screened for temperaturesensitivity and altered CPE in the following manner.Duplicate 24-well cluster dishes were seeded with HELcells, and individual wells were infected with a 1:1,000dilution of each plaque isolate. After adsorption ofvirus, paired dishes were incubated at 34°C and 39°C,respectively, and observed daily for temperature sen-sitivity of growth and for the appearance of alteredCPE. Plaque isolates which exhibited markedly re-duced CPE at 39°C when maximal CPE had beenachieved in the well maintained at 34°C were consid-ered to be presumptive temperature-sensitive mu-tants. The temperature-sensitive phenotype was con-firmed by determining the efficiency of plating (EOP)of mutants at 39°C and 34°C. Those mutants demon-strating EOPs of 3 logs or greater were considered tohave been confirmed as temperature sensitive. Isolateswith EOPs of less than 3 logs as well as those whichexhibited equal EOPs at both temperatures but signif-icantly smaller plaques at 39°C compared with 34°Cwere designated partially temperature-sensitive (pts)mutants. Because temperature-sensitive and partiallytemperature-sensitive mutants exhibited extensiveleak at 39°C, the nonpermissive temperature used insubsequent tests was 40°C. Isolates which exhibitedequal EOPs and plaque sizes at 34°C and 39°C butmorphologically altered CPE were designated non-temperature-sensitive (nts) mutants. Altered CPE wasmost commonly evident as the tendency of cells to

VOL. 34, 1980

340 MACHTIGER ET AL.

form syncytia or any other readily recognizable alter-ation in social behavior, e.g., failure to induce "bal-looning" or lysis of cells. Mutants were plaque purifiedthree times, and stocks were prepared in Vero cells.

RESULTSThe immune cytolysis a8say. The number

of infectious centers produced in a population ofwild-type virus-infected HEL cells treated withanti-HSV serum and complement 18 h postin-fection at 340C was a linear function of serumdilution at a fixed number of input infected celLs(106) (Fig. 4, large graph). The inset graph showsthe reduction in infectious center formation asa function of serum dilution. In later experi-ments, we used 105 cells and a 1:8 dilution ofhuman anti-HSV serum, which resulted in areduction of approximately 80% of the numberof surviving infectious centers at 18 h postinfec-tion.Kinetics of antigen appearance. We next

examined the time at which cells infected withthe wild-type virus or with the glycoprotein-defective mutant, tsAl, were maximally suscep-tible to immune cytolysis (Fig. 5). Multiple HELcell cultures were infected either with wild-typevirus or tsAl. After virus adsorption and wash-ing, each set of cultures was divided into twogroups. One group was incubated at 340C, andthe other was incubated at 390C. At 2-h inter-vals, one culture incubated at each temperaturewas harvested, and the sensitivity to immunecytolysis of 105 cells in the culture was measuredusing a 1:8 dilution of human anti-HSV serum.

In cells infected with the wild-type virus, an-tigens involved in immune cytolysis appearedearlier at 390C (4 h) than at 340C (6 h), but aslightly greater total number of cells was lysed

v

,4 O&O

Q,-Kh4

' 4 8 16 32

Serum Dilution

FIG. 4. Effect of serum dilution on the survival ofHSV-1-infected cells after 18 h of incubation at 34°Cin immune cytolysis tests using anti-HSV serum.

when infected cells were maintained at 340Cthan when they were maintained at 390C. Thesedifferences in the kinetics of cytolysis are anal-ogous to differences in the growth curves ofwild-type virus at 340C and 390C: replication at 390Cprecedes that at 340C by approximately 2 h,through 12 h postinfection.

Virus-specific antigen synthesis as determinedby the chromium 51 (51Cr) release assay (5) wasincluded for comparison (Fig. 5, dashed line).The maximum number of cells lysed as detectedby 51Cr release was slightly greater than thatmeasured by reduction in infectious center for-mation. Although the 5'Cr release experimentswere conducted at 370C, the kinetics of antigenappearance as measured by either technique arestrikingly similar. In both cases, the time of firstlysis of infected cells was between 2 and 4 hpostinfection, and maximum lysis was observedat 10 to 12 h postinfection.At 340C, tsAl-infected cells were not lysed at

the same rate as wild-type virus-infected cells.Hence, the target antigens for immune cytolysiswere not expressed as fully as in wild-type virus-infected cells at this temperature. Such an ob-servation was not unexpected as temperaturesensitivity is a continuously variable function oftemperature, rather than an all-or-none phe-nomenon. At 390C, little or no cytolysis of tsAl-infected cells was evident during the 18-h incu-bation period. This observation is consistentwith the absence of glycoprotein synthesis char-acteristic of tsAl at 390C (22). When plated at340C after treatment, these cells did produceinfectious centers; therefore, the inhibition ofglycoprotein synthesis at 390C with this mutantis apparently reversible.Rescue oftsAl-infected cells. Experiments

were then conducted to determine whether cellsinfected with tsAl could be recovered from amixture of cells containing an excess ofwild-typevirus-infected cells using the immune cytolysistest. The procedure for this experiment is illus-trated in Fig. 2, and the results obtained aresummarzed in Table 1. We began with twopopulations of virus-infected cells, both ofwhichwere maintained at 390C for 18 h (Fig. 2). Smallnumbers of tsAl-infected celLs were mixed witha 100-fold excess or greater of wild-type virus-infected celLs. The mixed suspension was thentreated with human anti-HSV antibody andcomplement and plated. Plaques were picked,and plaque isolates were tested for temperaturesensitivity. All temperature-sensitive isolateswere screened for complementation with tsAland an unrelated mutant, tsE6 (20).The expected number of temperature-sensi-

tive isolates could be calculated from (i) the titerof the control after treatment with antibody and

J. VIROL.

HSV MUTANTS RESISTANT TO IMMUNE CYTOLYSIS

.-

INDQ

(3

6.5?

Time (hrs)FIG. 5. Kinetics of immune cytolysis of virus-infected ceUs using human anti-HSV serum. Srymbols: 0 and

0, the wild-type virus at 34-C and 39°C; A and A, glycoprotein-defective mutant tsAl at 34°C and 39WC; C,from Glorioso and Smith (5). This curve was derived from a 51Cr release assay after growth ofHSV-I-infectedcels at 37°C.

complement and (ii) the number of input tsAl-infected cells, knowing that the latter cellsshould not be sensitive to immune cytolysis. Theresults of the reconstruction experiment ares zed in Table 1. As seen in the last two

columns, temperature-sensitive isolates were re-covered from all three suspensions to whichmutant-infected cells had been added. More-over, all six temperature-sensitive mutant iso-lates failed to complement tsAl, but comple-mented tsE6 efficiently (Table 2), thus confirm-ing their identity as tsAl.Comparison of the actual percent recovery

with the expected percent recovery demon-strates that a two- to fivefold-higher percentageof infectious centers was initiated by the tem-perature-sensitive mutants than had been antic-ipated. This discrepancy may have been due toerror in the actual number of tsAl-infected cellsadded to the mixed suspension or to failure toinactivate all infectious progeny virus with hu-man gamma globulin during the plating proce-dure.

Nonetheless, these results demonstrate thattsAl-infected celLs were preferentially recoveredfrom a large excess of wild-type virus-infectedcelLs after immune cytolysis.Cytolysis mediated by glycoprotein-spe-

cific antisera. To isolate new glycoprotein-de-

TABLE 1. Reconstruction experiment: recovery oftsAl-infected cells from a mixture ofKOS- and

tsAl-infected ceUs

ExetdNo. of tempera-Input cell nO. No. of in- ture-sensitivein mixture: fectious % tsAl m- infectious cen-KOS tsAl centers centers ters recovered/

no. tested (%)io5 0 2.8 X 103 0.0 0/100 (<1.0)105 8 2.6 X 103 0.3 1/70 (1.4)105 20 2.4 X 103 0.8 2/100 (2.0)105 40 2.8 X 103 1.4 3/100 (3.0)

fective mutants, we next attempted to capitalizeupon the reduced sensitivity to cytolysis of cellsinfected by glycoprotein-defective mutants. An-tisera directed against combined HSV-specificglycoproteins gA and gB, and against gC, as wellas antiserum directed against the VP123 com-plex, were available for these studies. It wasreasoned that these antisera might facilitate theselection of mutants defective in the expressionof the glycoprotein antigens against which theantisera were prepared. It was first necessary todemonstrate that such antisera would indeedmediate antibody-dependent, complement-me-diated cytotoxicity (Fig. 6). Unlike normal rabbitserum, which lysed approximately 10% of wild-type virus-infected cells at a 1:2 dilution, rabbit

341VOL. 34, 1980

342 MACHTIGER ET AL.

TABLE 2. Complementation tests between tsA4,tsE6, and six putative reisolates of tsAI

Complementation index' intests with:

tsAl tsE6

Isolate 1 1.2 93Isolate 2 1.1 31Isolate 3 0.8 45Isolate 4 0.5 26Isolate 5 1.9 54Isolate 6 1.3 82

tsAl 0.6 75tsE6 47 1.6

a Complementation tests were performed as previ-ously described (19); indices of 2 or greater were con-sidered positive.

anti-HSV serum and antisera to glycoproteinsgAgB and gC and to the VP123 complex lysed80% or more of wild-type virus-infected cells atthis dilution. For all sera, the percent survivinginfectious centers increased as a function of se-rum dilution throughout the 1:2 to 1:16 range ofdilutions tested. The cytolytic activities of anti-gC and anti-VP123 complex sera (anti-gAgBgC)were similar to that of anti-HSV serum, whereasanti-gAgB serum was less active, particularly athigher dilutions. The ability of glycoprotein-spe-cific antisera to induce lysis of HSV-infectedcells confirmed earlier observations (5, 6) thatcytolytic antibody is directed against glycopro-teins in the VP123 complex and demonstratedthe potential usefulness of these sera for mutantselection.Mutagenesis and selection ofmutants re-

sistant to immune cytolysis. Since all threeglycoprotein-specific antisera were active in ourimmune cytolysis assay, these antisera were usedfor the selection of cytolysis-resistant mutantsderived from 2-aminopurine-mutagenizedstocks of HSV-1 strain KOS. Cells infected witheach of four mutagenized stocks and maintainedat 390C were treated with one of the following:nornal rabbit serum, anti-gAgB serum, anti-gCserum, or anti-VP123 serum (Fig. 3). Cells whichescaped immune cytolysis were plated at 340C,and plaques were picked. Selection by immunecytolysis was conducted at 390C in an attemptto isolate temperature-sensitive mutants; suchmutants might express viral glycoproteins oncell surfaces at 340C but not at 390C. In additionto classical temperature-sensitive mutants withdefects in essential replicative functions, mu-tants able to replicate at both temperatures butunable to express glycoproteins at 340C, 390C,or both temperatures might be selected. Suchmutants would be defective, by definition, in

J. VIROL.

Qo°r o normal rabbit serum

80k

~0\

anti-gAgB

60k

40k

20-

I2 4 8 16

ISerum Dilution

FIG. 6. Cytolysis mediated by antisera preparedagainst specific HSV-1 glycoproteins. HSV-1-in-fectedHEL cells were incubated at 39°C for 18 h andsubjected to immune cytolysis using various dilutionsofnormal rabbit serum, anti-VP123, anti-gC, or anti-gAgB.

functions not essential for replication or wouldsynthesize antigenically altered glycoproteins.Because glycoprotein-defective mutants mightbe expected to induce atypical CPE in cell cul-ture (e.g., syncytia formation), all plaque isolateswere screened for altered CPE, as well as fortemperature sensitivity. A total of 532 plaqueswere picked, and 73 mutants were isolated (Ta-ble 3). During plaque purification and stockpreparation, potential temperature-sensitivemutants exhibited extensive leak (3 to 5 logs) at390C. Consequently, in subsequent tests 400Cwas used as the nonpermissive temperature. At400C, leak was reduced to 2 to 4 logs. Threedistinct classes of mutants were identified: (i)mutants which produced plaques at 340C butnot at 400C (classical temperature-sensitive mu-tants with EOPs 400/340 23 logs), (ii) mutantswhich plated less well at 400C than at 340C(EOP 400/340<3 logs) or those which exhibitedequal titers at 340C and 400C but which pro-duced significantly smaller plaques at 400C, and(iii) mutants which plated equally well and pro-duced plaques of wild-type size at both temper-atures but which exhibited altered CPE. Allmutants in the latter class and some mutants inthe former two classes exhibited altered CPE atone or both temperatures. Mutants in the firstclass were designated temperature sensitive,those in the second class were designated par-tially temperature sensitive, and those in thethird class were designated non-temperaturesensitive.

Five mutants were identified from among 100plaque isolates derived from cells infected with

HSV MUTANTS RESISTANT TO IMMUNE CYTOLYSIS 343

TABLE 3. Isolation ofmutants resistant to inmune cytolysisNo. of mutants

Antiserum used for selec- % Survivors No. of mutant isolates/no. of Paray Non-tetion plaques picked (%) ure eni- tempera- peratureture 8e perature

tive ture snsi- sensitive

Normal rabbit serum 90.0 5/100 (5.0) 0 2 3

Anti-VP123 region 3.0 11/148 (7.4) 3 7 1

Anti-gAgB 0.7 30/192 (15.6) 12 10 8

Anti-gC 0.5 32/192 (16.6) 9 10 13a The quotient of the number of plaques after treatment with antiserum divided by the number of plaques

after treatment with normal rabbit serum multiplied by 100.

mutagenized virus and treated with nornal rab-bit serum. Mutants of two distinct types, par-tially temperature sensitive and non-tempera-ture sensitive, were identified. Because mutantsof each type could be clonally related, the fre-quency of occurrence of mutants in 2-aminopu-rine-mutagenized stocks was at least 2% but nogreater than 5%. No mutants were observedamong 100 plaques isolated from unmutagen-ized, unselected KOS (unpublished observa-tions).Although treatment of infected cells with an-

tiserum to the VP123 complex reduced the num-ber of surviving cells to only 3%, a higher per-centage of surviving cells contained mutantprogeny: compare 5% (normal rabbit serum) and7.4% (VP123 complex serum). Treatment of in-fected cells with anti-gAgB or anti-gC serumresulted in an even greater reduction in thenumber of infected cells. In both cases, approx-imately 16% of the virus from surviving cells wasmutant. Thus, treatment of cells infected withmutagenized virus with all three glycoprotein-specific antisera resulted in enrichment for mu-tant progeny.

Sensitivity of cells infected with new mu-tants to immune cytolysis. To examine theextent to which mock-selected and antiserum-selected mutants expressed antigens involved inimmune cytolysis compared with wild-type vi-rus, four mutants mock selected with normalrabbit serum (termed 0 mutants), one repre-sentative of each type of mutant (AB, C, and R,selected with anti-gAgB, anti-gC, and anti-VP123 complex sera, respectively, and whichhad been characterized as temperature sensitive,partially temperature sensitive, and non-temper-ature sensitive), and wild-type virus were testedin the immune cytolysis assay at 340C and 400C.The antisera used in these tests were anti-VP123complex serum for mock-selected mutants andanti-gAgB, anti-gC, and anti-VP123 complex

sera for mutants selected with these sera, re-spectively. The results ofimmune cytolysis testsare shown in Fig. 7. The four mutants isolatedafter treatment of mutagenized virus-infectedcells with normal antiserum were tested withanti-VP123 complex antiserum. This antiserumwas selected because, theoretically, it containedantibodies to all precursors, intermediates, andfully glycosylated products of the gA, gB, andgC glycoprotein species. Cells infected with eachof the four mock-selected mutants were approx-imately as sensitive to lysis as wild-type virus-infected cells at 400C, demonstrating no detect-able alteration in the expression of glycoproteinsgA, gB, and gC by these mutants (Fig. 7A). Incontrast, cells infected with nine antiserum-se-lected mutants demonstrated moderate (ntsl3O-AB, Fig. 7C) to greatly enhanced (pts63-R, Fig.7B; tslO8-AB, Fig. 7C; nts79-C, Fig. 7D) resist-ance to immune cytolysis at 400C when testswere conducted with the same antisera used formutant selection. tsAl and tsJW2, two glycopro-tein-defective mutants (20, 22; J. Jofre and S.Little, personal communication), also exhibitedenhanced resistance to immune cytolysis at 400Cin tests with all three antisera. These resultsthus demonstrate the effectiveness of the anti-serum selection procedure for the isolation ofmutants defective in immune cytolysis and, po-tentially, in the expression of viral glycoproteins.Because the immune cytolysis procedure used

in mutant selection was conducted at the non-permissive temperature, mutants might be ex-pected to exhibit either reduced or equal sensi-tivity to cytolysis at the permissive temperature.In fact, cells infected with some mutants werepartially temperature sensitive in the immunecytolysis test (e.g., pts63-R and tsll2-R, Fig. 7B;ts134-C and nts79-C, Fig. 7D) whereas otherswere equally resistant to cytolysis at 340C and400C (ptsllO-R, Fig. 7B; tslO8-AB and pts6l-AB, Fig. 70). Similarly, the glycoprotein-defec-

VOL. 34, 1980

344 MACHTIGER ET AL.

.12 1:2 1:4 1:2 1:2

KOS 56-0 54-0 55-0 5- KOS Al 63-R KOS Jlt it-f 110-1R KOS J12 Al l-A S-AS I30-AS KOS J12 Al 134-C 41-C 79-C(Pht) (at.) ("ts) ("to) (ts) (ph) Its) (to) (pls) (ts) (ts (Is) (.) (afs) (Is) Its Its) (phi) (at.)

Virus TestedFIG. 7. Immune cytolysis of cells infected with mutants selected for resistance to cytolysis. 105 HEL cells

were infected with the glycoprotein-defective mutants tsA1 and tsJW2, wild-type virus, or 13 newly isolatedmutants, randomly selected for testing. Infected cells were incubated for 18 h at 340C (U) and 400C (0) andsubjected to immune cytolysis using the antisera at the dilutions indicated.

tive mutants tsAl and tsJ12 exhibited tempera-ture-sensitive behavior in immune cytolysis testswith glycoprotein-specific antisera.

DISCUSSIONIn an effort to isolate mutants defective in

viral glycoprotein synthesis, we have developeda simple and effective procedure for the selectionof mutants which are resistant to immune cytol-ysis. The procedure is based upon the assump-tion that cells infected with mutants which ex-hibit decreased expression or aberrant synthesisof viral glycoproteins will survive treatment withanti-HSV serum and complement. This assump-tion was first tested and proven with a mutant,tsAl, which is both temperature-sensitive forglycoprotein synthesis and resistant to immunecytolysis at 390C.

All viral glycoprotein-specific antisera used inmutant selection were able to mediate cytolysis,indicating that the gAgB, gC, and possibly othercomponents ofthe major viral glycoprotein com-plex (VP123) can serve as targets for immunecytolysis and are inserted into the cell mem-brane. These observations are in agreement withthe findings of Glorioso and Smith and Gloriosoet al. (5, 6) and Norrild et al. (12) that viralglycoproteins within the VP123 complex serveas target antigens for immune cytolysis. We donot know the contribution of each HSV glyco-protein to complement-mediated cell lysis orwhether less extensively glycosylated precursorscan also serve as targets. Moreover, whether gEis involved in immune cytolysis remains to bedetermined. However, the isolation of mutantsresistant to immune cytolysis yet able to repli-cate as efficiently as the wild-type virus (non-temperature-sensitive mutants) demonstrates

that normal expression of target antigens is notessential for virus replication. This finding isconsistent with the observation that at least oneglycoprotein, the gC glycoprotein, is not essen-tial for virus replication (10). Thus, the densityof a given target antigen on the cell surface maybe too low to permit efficient immune cytolysis,but sufficiently high on the nuclear membraneto permit budding of particles which are able toinitiate infection. In this regard, accurate quan-titation of viral glycoprotein synthesis will beessential to an understanding of the roles ofglycoproteins in the viral replicative cycle.The 2- to 16-fold increase in mutant isolation

observed after selection with glycoprotein-spe-cific antisera is significant, demonstrating thatthe selection procedure enriches for progenywhich exhibit the desired' phenotype. In thefuture, it should be possible to include a secondcycle of antibody treatment in the selection pro-cedure which may further increase the propor-tion of lysis-resistant mutants.

In the present study, it is possible that someor all mutants within a given group (i.e., allmutants selected with a particular antiserum)could be clonally related because each groupwas selected from only one mutagenized stock.By the same token, however, mutants in differ-ent groups cannot be clonally related. Clonalrelationships should be detected by complemen-tation tests now in progress.Although the nature of the defects in individ-

ual mutants remains to be determined, one canpostulate a vanety of mutations which couldresult in resistance to immune cytolysis: muta-tions in structural or regulatory genes for indi-vidual glycoproteins, mutations in viral genesinvolved in glycosylation (if such genes exist),

J. VIROL.

HSV MUTANTS RESISTANT TO IMMUNE CYTOLYSIS 345

and mutations in genes encoding other viralmembrane proteins which can affect the abilityof normal glycoproteins to be inserted, to nameseveral.

Interestingly, preliminary tests of severallysis-resistant temperature-sensitive mutants in-dicate that they synthesize significant amountsof virus-specific glycoproteins at the nonpermis-sive temperature as assayed by SDS-PAGE. Thelocalization of these glycoproteins is presentlybeing investigated. It is possible that these mu-tants are either defective in the insertion ofglycoproteins into the cellular membrane or thatthe antigenic specificity of individual glycopro-teins has been altered so that glycoprotein-spe-cific antisera no longer recognize these targetantigens. In any event, the defects exhibited bythese temperature-sensitive mutants differmarkedly from that of tsAl, which exhibitsdrastically reduced expression of all viral glyco-proteins at the nonpermissive temperature, re-sulting in decreased expression of target antigensat the cell surface.Having isolated mutants which are resistant

to immune cytolysis and, thus, putatively alteredin the expression of individual viral glycopro-teins, we believe these mutants should proveuseful in defining (i) the kinetics of appearanceof virus-specific glycoproteins in cellular mem-branes; (ii) the mosaic of viral antigens on thecell surface; and (iii) the function of viral glyco-proteins in the HSV replicative cycle, in trans-fonnation by the virus, and in the induction ofhumoral and cellular immunity to HSV.

ACKNOWLEDGMENTSThe authors thank P. T. Gelep, P. A. Temple, and L. B.

Sandner for excellent technical assistance.This study was supported by Public Health Service re-

search grants CA 20,260, CA 24,564, and CA 25,000 andprogram project grant CA 21,082 from the National CancerInstitute. B.A.P. was the recipient of postdoctoral fellowshipCA 06,034 from the National Cancer Institute.

LITERATURE CITED1. Baucke, R. B., and P. G. Spear. 1979. Membrane pro-

teins specified by herpes simplex viruses. V. Identifica-tion of an Fc-binding glycoprotein. J. Virol. 32:779-789.

2. Camacho, A., and P. G. Spear. 1978. Transformation ofhamster embryo fibroblasts by a specific fragment ofthe herpes simplex virus genome. Cell 15:993-1002.

3. Courtney, R. J., and M. Benyesh-Melnick. 1974. Iso-lation and characterization of a large molecular weightpolypeptide of herpes simplex virus type 1. Virology 62:539-551.

4. Duff, R., and F. Rapp. 1971. Properties of hamsterembryo fibroblasts transformed in vitro after exposureto ultraviolet-irradiated herpes simplex virus type 2. J.Virol. 8:469-477.

5. Glorioso, J. C., and J. W. Smith. 1977. Immune inter-action with cells infected with herpes simplex virus:antibodies to radioiodinated surface antigens. J. Im-munol. 118:114-121.

6. Glorioso, J. C., LA. Wilson, T. W. Fenger, and J. W.Smith. 1978. Complement-mediated cytolysis of HSV-1 and HSV-2 infected cells; plasma membrane antigensreactive with type-specific and cross-reactive antibody.J. Gen. Virol. 40:443-454.

7. Heine, J. W., P. G. Spear, and B. Roizman. 1972.Proteins specified by herpes simplex virus. VI. Viralproteins in the plasma membrane. J. Virol. 9:431439.

8. Honess, R. W., and B. Roizman. 1974. Regulation ofherpes virus macromolecular synthesis. I. Cascade reg-ulation of the synthesis of three groups of viral proteins.J. Virol. 14:8-19.

9. Kimura, S., V. L Flannery, B. Levy, and P. A. Schaf-fer. 1975. Oncogenic transformation of primary hamstercells by herpes simplex virus type 2 (HSV-2) and anHSV-2 temperature-sensitive mutant. Int. J. Cancer 15:786-798.

10. Manservigi, R., P. G. Spear, and A. Buchan. 1977. Cellfusion induced by herpes simplex virus is promoted andsuppressed by different viral glycoproteins. Proc. Natl.Acad. Sci. U.S.A. 74:3913-3917.

11. Moss, B., and E. W. Rosenblum. 1972. Hydroxylapatitechromatography of protein-sodium dodecylsulfate com-plexes. A new method for the separation of polypeptidesubunits. J. Biol. Chem. 247:5194-5198.

12. Norrild, B., S. L Shore, and A. J. Nahmias. 1979.Herpes simplex virus glycoproteins: participation of in-dividual herpes simplex virus type 1 glycoprotein anti-gens in immunocytolysis and their correlation with pre-viously identified glycoproteins. J. Virol. 32:741-748.

13. Parris, D. S., R. J. Courtney, and P. A. Schaffer. 1978.Temperature-sensitive mutants of herpes simplex virustype 1 defective in transcriptional and post-transcrip-tional functions required for viral DNA synthesis. Vi-rology 90:177-186.

14. Pfizenmaier, K. H., H. Jung, A. Starzinski-Powitz,M. Rollinghoff, and H. Wagner. 1977. The role of Tcells in anti-herpes simplex virus immunity. I. Inductionof antigen-specific cytotoxic T lymphocytes. J. Immu-nol. 119:939-944.

15. Rawls, W. E. 1973. Herpes simplex virus, p. 302-303. InA. S. Kaplan (ed.), The herpes viruses. Academic PressInc., New York.

16. Reed, C. L, G. H. Cohen, and F. Rapp. 1975. Detectionof a virus-specific antigen on the surface of herpessimplex virus transformed cells. J. Virol. 15:668-670.

17. Roane, P. R., and B. Roizman. 1964. Studies of thedeterminant antigens of viable cells. II. Demonstrationof altered antigenic activity of HEp-2 cells infected withherpes simplex virus. Virology 22:1-8.

18. Ruyechan, W. T., L S. Morse, D. M. Knipe, and B.Roizman. 1979. Molecular genetics of herpes simplexvirus. II. Mapping of the major viral glycoproteins andof the genetic loci specifying the social behavior ofinfected cells. J. Virol. 29:677-697.

19. Sarmiento, M., M. Haffey, and P. G. Spear. 1979.Membrane proteins specified by herpes simplex viruses.III. Role of glycoprotein VP7 (B2) in virion infectivity.J. Virol. 29:1149-1158.

20. Schaffer, P. A., G. M. Aron, M. Biswal, and M. Ben-yesh-Melnick. 1973. Temperature-sensitive mutants ofherpes simplex virus type 1: isolation, complementation,and preliminary characterization. Virology 52:57-71.

21. Schaffer, P. A., V. C. Carter, and M. C. Timbury.1978. Collaborative complementation study of temper-ature-sensitive mutants of herpes simplex virus types 1and 2. J. Virol. 27:490-504.

22. Schaffer, P. A., R. J. Courtney, R.L McCombs, andM. Benyesh-Melnick. 1971. A temperature-sensitivemutant of herpes simplex virusdefective in glycoproteinsynthesis. Virology 46:356-368.

23. Sim, C., and D. H. Watson. 1973. The role of type-specific and cross-reacting structural antigens in the

VOL. 34, 1980

346 MACHTIGER ET AL.

neutralization of herpes simplex virus types 1 and 2. J.Gen. ViroL 19:217-233.

24. Spear, P. G. 1976. Membrane proteins specified by herpessimplex virus. Identification of four glycoprotein pre-cursors and their products in type-1 infected cells. J.Virol. 17:991-1008.

25. Spear, P. G., J. M. Keller, and B. Rolzman. 1970. The

protein specified by herpes simplex viru. II. Viral gly-coproteins amociated with cellular membranes. J. Virol.5:123-131.

26. Spear, P. G., and B. Roizman. 1972. Proteins specifiedby herpes simplex vir. V. Purification and structuralproteins of the herpes virion. J. Virol. 9:143-159.

J. VIROL.