Archives
Arch Virol (1990) 112:55-65
Vi rology © by Springer-Verlag 1990
Characterization of cytomegalovirus glycoproteins in a family of complexes designated gC-II with murine monoclonal antibodies B. Kari, R. Goertz, and R. Gehrz
Division of Biochemistry, Children's Biomedical Research Institute, St. Paul, Minnesota, U.S.A. Accepted January 24, 1990
Summary. Several murine monoclonal antibodies (MoAbs) were made to a family of human cytomegalovirus (HCMV) disulfide linked glycoprotein complexes designated gC-II. Characterization of these MoAbs by immunological methods showed that they could be divided into two groups recognizing different glycoproteins. Western blot analysis was done with immunoaffinity purified gC-II complexes. Under non-reducing conditions MoAbs from both groups recognized gC-II complexes with molecular weights of 67-93k and 130k to greater than 200 k. When purified gC-II complexes were reduced and individual glycoproteins separated by SDS-PAGE prior to Western blotting, Group 1 MoAbs reacted with glycoproteins having molecular weights of 4"7-63 k, while Group 2 MoAbs reacted with glycoproteins having molecular weights of 3948 k and 90 k to greater than 200 k. Thus, gC-II complexes contain glycoproteins recognized by both groups of MoAbs. By Coomassie blue staining and incorporation of [3H]Arg, Group 1 glycoproteins appeared to be minor components in the complexes relative to Group 2 glycoproteins. Surface labeling of extracellular virus with galactose oxidase and tritiated borohydride showed that gCII complexes of all molecular weights were on the surface of the virus. However, the most heavily labeled gC-II glycoproteins had molecular weights of 47-63 k. These data confirm our previous observations that the gC-II complexes of HCMV are comprised of a heterogeneous family of glycoproteins. Introduction Human cytomegalovirus (HCMV) is a member of the herpes family of viruses and is ubiquitous in humans. Infection with HCMV is usually asymptomatic and self-limiting in healthy individuals. However, it can cause congenital defects in the developing fetus if contracted in utero and can be devastating in im-
56
B. Kari et al.
munosuppressed patients [8]. We have demonstrated that in humans there are both T and B cell responses to HCMV envelope glycoproteins [12]. Thus, the surface glycoproteins of HCMV are likely to be important in host immune responses. In order to determine which responses are most important it will be necessary to obtain a thorough knowledge of the envelope glycoproteins of HCMV. It has been established that HCMV contains several disulfide linked glycoprotein complexes which can be extracted from extracellular virus using non-ionic detergents and separated from each other by anion exchange high performance liquid chromatography (HPLC) [9] or by sucrose gradient centrifugation [6]. A number of HCMV glycoprotein complexes could be immunoprecipitated with a single MoAb (9E10), and were therefore placed into a family, which we designated gC-II [6, 9]. The gC-II family consisted of disulfide linked glycoprotein complexes which varied in molecular weight, charge, density, and glycoprotein composition. Based on glycoprotein composition, two major types of complexes were detected when virus was labeled with [3H]GlcN. One of the these, which appeared to be less abundant, generated glycoproteins with molecular weights of 47-52 k (gp47-52) after reduction of disulfide bonds. A more abundant type of gC-II complex also generated gp4752 after reduction, but also contained several glycoproteins with molecular weights ranging from 90 k to greater than 200 k [6, 9]. It was determined that MoAb 9El0 recognized gp47-52 independent of other gC-II glycoproteins, and that gp47-52 had a much higher content of O-linked oligosaccharides as compared to the gC-II glycoproteins with molecular weights of 90-200k [10]. Subsequently it was shown that gp47-52 glycoproteins were encoded for in the HXLF gene family of HCMV [7]. When mRNA transcribed from the HXLF gene family was translated in vitro proteins with molecular weights of 20-30 k were obtained. The molecular weights of these precursor proteins were increased to 47-52 k by the addition of carbohydrates [7]. While these studies accounted for gp47-52 they did not explain the higher molecular weight glycoproteins found in the majority of the gC-II complexes. In order to better understand the composition of the gC-II family of complexes, additional gC-II specific MoAbs were made using partially purified gC-II as the immunizing antigen in mice. Several MoAbs reactive with purified gC-II in an enzyme linked immunosorbant assay (ELISA) were obtained and characterized. Based on their reactivity in Western blotting, these MoAbs could be placed into two groups. Group 1 MoAbs recognized gC-II glycoproteins with molecular weights of 4763 k and Group 2 MoAbs recognized gC-II glycoproteins with molecular weights of 39-48 k and from 90 k to 200 k. Thus, gC-II contained two distinct groups of glycoproteins. Materials and methods
Production and characterization of gC-H MoAbs
Adult BALB/cmice were immunizedwith 100 ~tg of whole Towne strain HCMV emulsified in complete Freund adjuvant and given a booster immunization with 100gg whole Towne
HCMV glycoproteins
57
HCMV at 3 weeks. After several weeks, a final boost was given using 20 txg of gC-II obtained by anion exchange HPLC as previously reported [9]. Three days after the final boost with gC-II, the mice were sacrificed and spleen cells fused wtih Sp2-O-Ag 14 myeloma cells. Cloning was done as previously described [9]. To identify positive hybridomas, culture medium was assayed for antibody against gC-II by ELISA. For this assay, purified gC-II was fixed in 96 well dishes. Hybridomas were sub-cloned twice by limiting dilution. All MoAbs used in these studies were found to have a single immunoglobulin subtype. Clonality was confirmed by agarose iso-electric focusing of purified MoAbs (data not shown).
SDS-PAGE and Western blot SDS-PAGE was done with either 7% or 10% polyacrylamide gels following the method of Laemmli [11]. Antigens used for Western blot included purified HCMV virions or gCII complexes immunoaffinitypurified as described below. Proteins in gels were electroblotted onto nitrocellulose paper, and the paper was then blocked with 3% gelatin in Tris-phosphate buffered saline [TBS, 20 mM Tris, 500 mM NaC1, phenylmethylsulfonyl fluoride (PMSF) 2 mM, pH 7.5]. Strips of the blocked paper were incubated overnight with MoAbs in ascites diluted t to 500, in TBS containing 1% gelatin. Strips were washed with TBS containing 0.05 % Tween 20, and then with TBS before incubation with phosphatase labeled goat antimouse IgG (Heavy and Light) (Kirkegaard and Perry, Gaithersburg, MD) diluted 1 to 1,000 with TBS containing 1% gelatin. After a 1 h incubation, the paper was washed and the substrate 5-bromo-4-chloro-3-indoyl phosphate in 0.1 M Tris buffer (Kirkegaard and Perry) was added. After visualization, the reaction was stopped by immersing the strips in water.
Radioactive labeling of HCMVproteins and immunoaffinity purification of gC-H Glycoproteins were radiolabeled with [3H]GlcN or [3H]Arg (New England Nuclear) as previously described [9]. Envelope proteins were extracted from extracellular virus using a Tris buffer [50mM Tris, pH 7.5, 150mM NaC1, 2mM phenylmethylsulfonyl fluoride (PMSF)] containing 1.0% NP-40. Insoluble material was removed by centrifugation. Monoclonal antibodies, biotinylated as previously described [4], were added to these extracts along with streptavidin agarose beads (BRL). These solutions were mixed for 1 h at room temperature and centrifuged to pellet the agarose beads. After several washes with PBS containing 0.1% NP-40, bound proteins were solubilized with SDS by heating at 100°C for 3 rain. After SDS-PAGE, radioactive proteins were detected by fluorography using Enhance (New England Nuclear).
Galactose oxidase labeling of HCMV Purified HCMV was labeled with tritiated borohydride after treatment with galactose oxidase following the methods of Gahmberg and Hakomori [3]. Briefly, virus in PBS pH 7.5 was treated with 10 units of galactose oxidase (Sigma) for 1 h at room temperature. Virus was then treated with 3.3 mCi of tritiated borohydride (600 mCi/mmol, New England Nuclear) for 30min. Excess borohydride was destroyed with 25% acetic acid (100 gl) at 0 °C. Virus was collected by centrifugation and washed with PBS.
Results
Western blot analysis of gC-II MoAbs M o A b s which reacted strongly with purified gC-II in an E L I S A assay were first analyzed to determine the m o l e c u l a r weights o f the proteins they recognized. This was d o n e by W e s t e r n blotting using whole T o w n e H C M V as the antigen.
58
B. Karl et al.
Fig. 1. Western Not analysisdone with whole Towne strain HCMV. HCMV was solubilized in SDS, urea and reduced with 2 mercaptoethanol prior to SDS-PAGE. 1 SP2 negative ascites control. 2-4 Group t MoAbs, 5-9 Group 2 MoAbs. Numbers to the left and right indicate molecular weights x 10 - 3
Whole HCMV was solubilized in SDS, urea, and reduced with beta-mercaptoethanol prior to SDS-PAGE and Western blotting. A mouse ascites (SP2), which did not contain HCMV antibody, was used as a negative control. This ascites did not react by Western blotting (Fig. 1, lane 1). Group 1 contained 3 MoAbs with different subtypes, 9El0 (IgG3), 8B4 (IgM), and 26E2 (IgG2a). Group 1 MoAbs reacted strongly with 47 k to 63 k molecular weight proteins and very weakly with higher molecular weight proteins (Fig. 1, lanes 2-4). Group 2 contained 7 MoAbs; 5 of these had an IgG2a subtype, while MoAb 15G5 was an IgG1 and MoAb 25C8 was an IgA. Group 2 MoAbs reacted strongly with proteins with molecular weights between 39,000 to 48 k, but they also reacted strongly with proteins from 90 k to greater than 200 k; representative data for 5 Group 2 MoAbs are shown in Fig. 1, lanes 5-9. The proteins with molecular weights of 39-48 k recognized by Group 2 MoAbs appear to be different from the 47-63 k molecular weight proteins recognized by Group 1 MoAbs since they did not completely overlap in molecular weight. Furthermore, Group 1 MoAbs showed little or no reactivity with the high molecular weight proteins while Group 2 MoAbs showed strong reactivity (Fig. 1). Thus, MoAbs in Group 1 recognize proteins different from those recognized by MoAbs in Group 2. To show that both groups of MoAbs recognized gC-II complexes,
HCMV glycoproteins
59
Fig. 2. gC-II complexes were immunoaffinity purified with Group 2 MoAb 15F9. Approximately 100-150 gg of gC-II protein was used to do Western blotting. Following SDSPAGE, proteins were transferred to nitrocellulosepaper and the paper cut into strips for Western blotting. Each strip contained approximately5-10 gg protein. A Western blotting done under non-reducing conditions, gC-II complexes were initially separated by SDSPAGE using a 7% polyacrylamidegel prior to transfer to nitrocellulosepaper for Western blotting. 1 SP2 negative control. 2 Group 1 MoAb 9El0; 3, 4 Group 2 MoAbs 12G9 and 15F9 respectively.B Same as A except that gC-II complexeswere reduced prior to SDSPAGE. SDS-PAGE was done in a 10% polyacrylamide gel. 1 SP2 negative control. 2 Group 1 MoAb 9El0; 3, 4 Group 2 MoAbs 12G9 and 15F9 respectively.Numbers to the fight indicate molecular weight x 10-3
Western blotting was done with gC-II complexes which had been immunoaffinity purified with Group 2 MoAb 15F9. Immunoaffinity purified gC-II complexes were separated by SDS-PAGE under non-reducing conditions prior to Western blotting. Under non-reducing conditions, MoAbs from either group reacted with complexes with molecular weights of 67-93 k and 130 k to greater than 200 k, representative data for Group 1 MoAb 9E10 and Group 2 MoAbs 12G9 and 15F9 are shown in Fig. 2A. Thus, gC-II complexes which varied in molecular weight were recognized by both groups of MoAbs. Immunoaffinity purified gC-II complexes were also reduced and individual gC-II glycoproteins separated by SDS-PAGE prior to Western blotting. After reduction of purified gC-II complexes, Group 1 MoAbs reacted with proteins having molecular weights of 47-63 k while Group 2 MoAbs reacted with proteins having molecular weights of 39-48 k and 90 k to greater than 200 k (Fig. 2B). Again after reduction the molecular weights of the proteins recognized were different when comparing the two groups of MoAbs. These data showed that gC-II complexes which had been immunoaffinity purified with Group 2 MoAb 15F9 contained both group 1 and 2 glycoproteins recognized by their respective MoAbs.
60
B. Kari et al.
Immunoprecipitation of gC-H glycoproteins To determine if the proteins detected by Western blotting were glycosylated, individual gC-II proteins labeled with radioactive glucosamine were immunoprecipitated with MoAbs from both groups. To do this gC-II complexes were labeled with radioactive glucosamine and immunoaffinity purified with Group 2 M o A b 15F9. The purified gC-II complexes were solubilized in SDS, urea, reduced with 20 m M dithiothreitol, and alkylated with 50raM iodoacetamide in 0.2 M Tris (pH 7.0). After dialysis to remove excess reagents, NP-40 was added so that the final NP-40 concentration was 1.0%. The reduced gC-II glycoproteins were serially immunoprecipitated with a negative control, followed by a Group 1 M o A b (26E2) and than with a Group 2 M o A b (15F9). Immunoprecipitated glycoproteins were examined by SDS-PAGE and fluorography. The total number of gtycoproteins in the starting material are shown in Fig. 3 lane 1. These included a broad band with molecular weights from 48 k -
Fig. 3. The gC-II complexes were immunoaffinity purified from extracellular virus which was labeled with radioactive glucosamine. Complexes were reduced and alkylated to obtain individual gC-II glycoproteins. This material was used for serial immunoprecipitations. The order of immunoprecipitation was a negative control followed by a Group 1 MoAb 26E2, and finally with Group 2 MoAb 15F9. 1 Fluorogram showing all the reduced gCII glycoproteins obtained from immunoaffinity purified gC-II complexes. 2 Ftuorogram of gC-II glycoproteins immunoprecipitated with a negative control SP2. 3 Fluorogram of glycoproteins immunoprecipitated with a Group 1 MoAb 26E2.4 Fluorogram of glycoproteins immunoprecipitated with a Group 2 MoAb 15F9. Numbers to the right of all lanes indicate molecular weights x 10 -3. T Top of the gel
HCMV glycoproteins
61
63 k and glycoproteins with molecular weights from 90 k to greater than 200 k. An SP2 ascites negative control did not immunoprecipitate any glycoproteins (Fig. 3, lane 2). After the negative control, Group 1 MoAb 26E2 was used. This MoAb immunoprecipitated glycproteins with molecular weights of 48-63 k (Fig. 3, lane 3). After using Group 1 MoAb 26E2 Group 2 MoAb 15F9 was used. This MoAb immunoprecipitated glycoproteins with molecular weights of 43-48 k, and 90 k to greater than 200 k (Fig. 3, lane 4). The order in which the MoAbs were used did not affect the result (data not shown). The glycoproteins immunoprecipitated by both groups of MoAbs appeared to account for all the glycoproteins detected in the starting material (Fig. 3, compare lane 1 with lanes 3 and 4). These data showed that most gC-II proteins were glycosylated.
Protein composition of the gC-H complexes The relative abundance of the various gC-II glycoproteins was examined by Coomassie blue staining and by incorporation of [3H]Arg. Group 2 MoAb 15F9 was used to immunoaffinity purify gC-II complexes which were reduced and examined by SDS-PAGE. By Coomassie blue staining glycoproteins with molecular weights of 48 k, and 90 k, to greater than 200 k appeared most abundant (Fig. 4, lane 1). With [3H]Arg labeling, the 48 k and 90 k, to greater than 200 k, were also most abundant (Fig. 4, lane 2). With either method glycoproteins
Fig. 4. gC-II complexes, radiolabeled with [3H]Arg or unlabeled, were immunoaffinity purified with Group 1 MoAb 15F9. These complexeswere reduced and proteins separated by SDS-PAGE in 10% polyacrylamidegels. Numbers to the right indicatemolecularweights x 10-3. 1 Coomassie blue stained gel showing reduced proteins obtained from immunoaffinity purified gC-II. 2 Fluorogram of reduced gC-II proteins labeled with radioactive arginine. T Top of the gel
62
B. Kari et al.
with molecular weights between 50-63 k appear to be minor components relative to the other glycoproteins. Most of the glycoproteins recognized by the Group 1 MoAbs were in this molecular weight range.
Surface labeling of extracellular virus The location of the gC-II glycoproteins in extracellular virus was determined by a surface labeling method. Since the glycoproteins of gC-II contain galactose and galactosamine [ 10], galactose oxidase treatment followed by reduction with tritiated borohydride was employed to surface label purified HCMV virions. After labeling, the virions were washed twice with PBS to remove excess reagents. Washed virions were extracted wtih 1.0% NP-40, insoluble material removed by centrifugation, and the labeled gC-II complexes immunoprecipitated from the extract with Group 2 MoAb 15F9. Immunoprecipitated complexes were examined by electrophoresis under reducing and non-reducing conditions followed by fluorography. Without prior treatment with galactose oxidase, no radioactivity was incorporated into gC-II glycoproteins showing the specificity
Fig. 5. Whole virions were treated with galactose oxidase followed by reduction with tritiated borohydride. As a negative control, virions which had not been treated with galactose oxidase were exposed to tritiated borohydride. Virions were pelleted and washed to remove reactants. After washing, the virions were extracted with 1.0% NP-40 and insoluble material removed by centrifugation. The extract was immunoprecipitated with Group 2 MoAb t 5F9 to obtain labeled gC-II complexes. Immunoprecipitated gC-II complexes were separated by SDS-PAGE under non-reducing conditions in a 7% polyacrylamide gel (1, 2) or under reducing conditions in a 10% polyacrylamide gel (3, 4). 1, 3 Controls done without galactose oxidase. 2 gC-II complexes; 4 gC-II glycoproteins labeled with galactose oxidase and tritiated borohydride treatment. Glycoproteins were detected by fluorography. Numbers to the right indicate molecular weights x 10 - 3
HCMV glycoproteins
63
of the reaction (Fig. 5, lanes 1 and 3). However, after galactose oxidase treatment gC-II complexes with molecular weights of 6 7 - 93 k and 130 k to greater than 200 k were heavily labeled (Fig. 5, lane 2). When these complexes were reduced gC-II glycoproteins with molecular weights from 48-63 k and 90 k to greater than 200 k were detected ( Fig. 5, lane 4). The most heavily labeled glycoproteins had molecular weights of 48-63 k. These data show that both high and low molecular weight gC-II glycoproteins were exposed on the surface of the virus. Discussion
The primary objective of the current study was to characterize the different glycoproteins in gC-II complexes. Our data clearly show that gC-II complexes contain two immunologically distinct groups of glycoproteins. Gtycoproteins with molecular weights of 47-63 k were recognized by 3 Group 1 MoAbs. Group 1 MoAb 9El 0 has been shown to recognize products of the HXLF gene family [7]. These glycoproteins appeared to be minor components when compared to other gC-II glycoproteins based on Coomassie blue staining of SDS-PAGE gels. For the purposes of discussion we previously designated these glycoproteins gp47-52 to simplify their description [7, 10]. Group 2 contained glycoproteins recognized by 7 MoAbs. These glycoproteins were much more diverse in their molecular weights. The glycoproteins with molecular weights greater than 90 k have been purified by gel-filtration HPLC. These purified glycoproteins were solubilized in SDS, urea, and reduced with beta-mercaptoethanol a second time prior to SDS-PAGE with no further reduction in their molecular weights [10]. Thus, the high molecular weights of these glycoproteins were not do to incomplete reduction of disulfide bonds. Group 2 also included glycoproteins with molecular weights of 39 k to 48 k. It is possible that some of the low molecular weight Group 2 glycoproteins were generated by proteolysis during the isolations. However, PMSF was used to prevent proteolysis. Moreover, whole virus was solubilized directly for SDS-PAGE and Western blotting to prevent proteolysis. When this was done the 39 k to 48 k molecular weight glycoproteins were still detected. The molecular weights and relative amounts of the Group 2 glycoproteins were also consistent between preparations. These data suggested that the diversity in molecular spear paper weights was not due to proteolysis. Therefore, our data indicate that, gC-II complexes are made during virus assembly which are composed of two distinct groups of heterogeneous glycoproteins. Group 1 glycoproteins are products of the HXLF gene family [7]. The glycoproteins in Group 2 may be products of different transcripts encoded by or alternative splicing the HXLF gene family or by a different gene. The heterogeneous composition of the gC-II complexes is different than other glycoprotein complexes in the envelope of HCMV. For example, the family of complexes variably designated gB, gp 116-52, or gC-I is derived from a glycoprotein encoded by a single gene. This glycoprotein is subjected to proteolysis during its biosynthesis to generate two glycoproteins which are disulfide linked to form these complexes [1, 2, 5].
64
B. Kari et al.
Surface labeling showed that the gC-II complexes and their high and low molecular weight glycoproteins are exposed on the surface of extracellular HCMV. However, while both high and low molecular weight glycoproteins were labeled by treatment with galactose oxidase and tritiated borohydride the 48 k to 63 k molecular weight glycoproteins were most heavily labeled by this method. Both groups of MoAbs recognized glycoproteins in this molecular weight range. However, these results were consistent with our previous analysis which showed that gC-II glycoproteins in this molecular weight range contained a higher content of O-linked oligosaccharides when compared to the higher molecular weight gC-II glycoproteins [10]. Thus, the high molecular weight gC-II glycoproteins are glycosylated differently from the low molecular weight glycoproteins. The data presented in this and previous reports [6, 7, 9, 10] show that the gC-II family of complexes are very heterogeneous. The reason for this heterogeneity is not clear. One possibility is that the gC-II complexes are biochemically heterogeneous because they have multiple functions. Nothing is known about the functions of the gC-II family of complexes. Thus, determining their functions will be the focus of our future investigations.
Acknowledgments This work was supported by NIH Grant HD19937-02 and by a grant from the Children's Biomedical Research Institute, St. Paul, Minnesota.
References 1. Britt WJ, Auger D (1986) Synthesis and processing of the envelope gp55-116 complex of human cytomegalovirus. J Virol 58:185-191 2. Britt WJ, Vugler LG (1989) Processing &the gp 55.-116 envelope glycoprotein complex (gB) of human cytomegalovirus. J Virol 63:403-410 3. Gahmberg CG, Hakamori S-I (1973) External labelling of cell surface galactose and galactosamine in glycolipid and glycoprotein of human erythrocytes. J Biot Chem 248: 4311-4317 4. Gretch DR, Suter M, Stinski MF (1987) The use of biotinylated monoclonal antibodies and streptavidin affinity chromatography to isolate herpesvirus hydrophobic proteins or glycoproteins. Anal Biochem t63:270-277 5. Gretch DR, Gehrz RC, Stinski MF (1988) Characterization of a human cytomegalovirus glycoprotein complex (gcI). J Gen Virol 69:1205-1215 6. Gretch DR, Karl B, Rasmussen L, Gehrz RC, Stinski MF (1988) Identification and characterization of three distinct families of gtycoprotein complexes in the envelope of human cytomegalovirus. J Virol 62:875-881 7. Gretch DR, Karl B, Gehrz RC, Stinski MF (1988) A multigene family encodes the human cytomegalovirus glycoprotein complex gCII (gp47-52 Complex). J Virol 62: 1956-1962 8. Ho M (1982) Cytomegalovirus biology and infection. In: Grennough WB III, Merigan TC (eds) Current topics in infectious disease. Plenum, New York 9. Karl B, Lugsenhop N, Goertz R, Wabuke-Bunoti M, Radeke R, Gehrz R (1986) Characterization of monoclonal antibodies reactive to several biochemically distinct human cytomegalovirus glycoprotein complexes. J Virol 60:345-352
HCMV glycoproteins
65
10. Kari B, Gehrz R (1988) Isolation and characterization of a human cytomegalovirus glycoprotein containing a high content of O-linked oligosaccharides. Arch Virol 98: 171-188 11. Laemmli UK (1970) Cleavage of structural proteins during assembly of the head of bacteriophage T4. Nature 227:680-684 12. Liu Y-N C, Kari B, Gehrz RC (1988) Human immune responses to major human cytomegalovirus glycoprotein complexes. J Virol 62:1066-1070 Authors' address: B. Kari, Children's Biomedical Research Institute, Division of Biochemistry, Room L-284, St. Paul, MN 55102, U.S,A. Received December 13, 1989
Arch Virol (1990) 112:55-65
Vi rology © by Springer-Verlag 1990
Characterization of cytomegalovirus glycoproteins in a family of complexes designated gC-II with murine monoclonal antibodies B. Kari, R. Goertz, and R. Gehrz
Division of Biochemistry, Children's Biomedical Research Institute, St. Paul, Minnesota, U.S.A. Accepted January 24, 1990
Summary. Several murine monoclonal antibodies (MoAbs) were made to a family of human cytomegalovirus (HCMV) disulfide linked glycoprotein complexes designated gC-II. Characterization of these MoAbs by immunological methods showed that they could be divided into two groups recognizing different glycoproteins. Western blot analysis was done with immunoaffinity purified gC-II complexes. Under non-reducing conditions MoAbs from both groups recognized gC-II complexes with molecular weights of 67-93k and 130k to greater than 200 k. When purified gC-II complexes were reduced and individual glycoproteins separated by SDS-PAGE prior to Western blotting, Group 1 MoAbs reacted with glycoproteins having molecular weights of 4"7-63 k, while Group 2 MoAbs reacted with glycoproteins having molecular weights of 3948 k and 90 k to greater than 200 k. Thus, gC-II complexes contain glycoproteins recognized by both groups of MoAbs. By Coomassie blue staining and incorporation of [3H]Arg, Group 1 glycoproteins appeared to be minor components in the complexes relative to Group 2 glycoproteins. Surface labeling of extracellular virus with galactose oxidase and tritiated borohydride showed that gCII complexes of all molecular weights were on the surface of the virus. However, the most heavily labeled gC-II glycoproteins had molecular weights of 47-63 k. These data confirm our previous observations that the gC-II complexes of HCMV are comprised of a heterogeneous family of glycoproteins. Introduction Human cytomegalovirus (HCMV) is a member of the herpes family of viruses and is ubiquitous in humans. Infection with HCMV is usually asymptomatic and self-limiting in healthy individuals. However, it can cause congenital defects in the developing fetus if contracted in utero and can be devastating in im-
56
B. Kari et al.
munosuppressed patients [8]. We have demonstrated that in humans there are both T and B cell responses to HCMV envelope glycoproteins [12]. Thus, the surface glycoproteins of HCMV are likely to be important in host immune responses. In order to determine which responses are most important it will be necessary to obtain a thorough knowledge of the envelope glycoproteins of HCMV. It has been established that HCMV contains several disulfide linked glycoprotein complexes which can be extracted from extracellular virus using non-ionic detergents and separated from each other by anion exchange high performance liquid chromatography (HPLC) [9] or by sucrose gradient centrifugation [6]. A number of HCMV glycoprotein complexes could be immunoprecipitated with a single MoAb (9E10), and were therefore placed into a family, which we designated gC-II [6, 9]. The gC-II family consisted of disulfide linked glycoprotein complexes which varied in molecular weight, charge, density, and glycoprotein composition. Based on glycoprotein composition, two major types of complexes were detected when virus was labeled with [3H]GlcN. One of the these, which appeared to be less abundant, generated glycoproteins with molecular weights of 47-52 k (gp47-52) after reduction of disulfide bonds. A more abundant type of gC-II complex also generated gp4752 after reduction, but also contained several glycoproteins with molecular weights ranging from 90 k to greater than 200 k [6, 9]. It was determined that MoAb 9El0 recognized gp47-52 independent of other gC-II glycoproteins, and that gp47-52 had a much higher content of O-linked oligosaccharides as compared to the gC-II glycoproteins with molecular weights of 90-200k [10]. Subsequently it was shown that gp47-52 glycoproteins were encoded for in the HXLF gene family of HCMV [7]. When mRNA transcribed from the HXLF gene family was translated in vitro proteins with molecular weights of 20-30 k were obtained. The molecular weights of these precursor proteins were increased to 47-52 k by the addition of carbohydrates [7]. While these studies accounted for gp47-52 they did not explain the higher molecular weight glycoproteins found in the majority of the gC-II complexes. In order to better understand the composition of the gC-II family of complexes, additional gC-II specific MoAbs were made using partially purified gC-II as the immunizing antigen in mice. Several MoAbs reactive with purified gC-II in an enzyme linked immunosorbant assay (ELISA) were obtained and characterized. Based on their reactivity in Western blotting, these MoAbs could be placed into two groups. Group 1 MoAbs recognized gC-II glycoproteins with molecular weights of 4763 k and Group 2 MoAbs recognized gC-II glycoproteins with molecular weights of 39-48 k and from 90 k to 200 k. Thus, gC-II contained two distinct groups of glycoproteins. Materials and methods
Production and characterization of gC-H MoAbs
Adult BALB/cmice were immunizedwith 100 ~tg of whole Towne strain HCMV emulsified in complete Freund adjuvant and given a booster immunization with 100gg whole Towne
HCMV glycoproteins
57
HCMV at 3 weeks. After several weeks, a final boost was given using 20 txg of gC-II obtained by anion exchange HPLC as previously reported [9]. Three days after the final boost with gC-II, the mice were sacrificed and spleen cells fused wtih Sp2-O-Ag 14 myeloma cells. Cloning was done as previously described [9]. To identify positive hybridomas, culture medium was assayed for antibody against gC-II by ELISA. For this assay, purified gC-II was fixed in 96 well dishes. Hybridomas were sub-cloned twice by limiting dilution. All MoAbs used in these studies were found to have a single immunoglobulin subtype. Clonality was confirmed by agarose iso-electric focusing of purified MoAbs (data not shown).
SDS-PAGE and Western blot SDS-PAGE was done with either 7% or 10% polyacrylamide gels following the method of Laemmli [11]. Antigens used for Western blot included purified HCMV virions or gCII complexes immunoaffinitypurified as described below. Proteins in gels were electroblotted onto nitrocellulose paper, and the paper was then blocked with 3% gelatin in Tris-phosphate buffered saline [TBS, 20 mM Tris, 500 mM NaC1, phenylmethylsulfonyl fluoride (PMSF) 2 mM, pH 7.5]. Strips of the blocked paper were incubated overnight with MoAbs in ascites diluted t to 500, in TBS containing 1% gelatin. Strips were washed with TBS containing 0.05 % Tween 20, and then with TBS before incubation with phosphatase labeled goat antimouse IgG (Heavy and Light) (Kirkegaard and Perry, Gaithersburg, MD) diluted 1 to 1,000 with TBS containing 1% gelatin. After a 1 h incubation, the paper was washed and the substrate 5-bromo-4-chloro-3-indoyl phosphate in 0.1 M Tris buffer (Kirkegaard and Perry) was added. After visualization, the reaction was stopped by immersing the strips in water.
Radioactive labeling of HCMVproteins and immunoaffinity purification of gC-H Glycoproteins were radiolabeled with [3H]GlcN or [3H]Arg (New England Nuclear) as previously described [9]. Envelope proteins were extracted from extracellular virus using a Tris buffer [50mM Tris, pH 7.5, 150mM NaC1, 2mM phenylmethylsulfonyl fluoride (PMSF)] containing 1.0% NP-40. Insoluble material was removed by centrifugation. Monoclonal antibodies, biotinylated as previously described [4], were added to these extracts along with streptavidin agarose beads (BRL). These solutions were mixed for 1 h at room temperature and centrifuged to pellet the agarose beads. After several washes with PBS containing 0.1% NP-40, bound proteins were solubilized with SDS by heating at 100°C for 3 rain. After SDS-PAGE, radioactive proteins were detected by fluorography using Enhance (New England Nuclear).
Galactose oxidase labeling of HCMV Purified HCMV was labeled with tritiated borohydride after treatment with galactose oxidase following the methods of Gahmberg and Hakomori [3]. Briefly, virus in PBS pH 7.5 was treated with 10 units of galactose oxidase (Sigma) for 1 h at room temperature. Virus was then treated with 3.3 mCi of tritiated borohydride (600 mCi/mmol, New England Nuclear) for 30min. Excess borohydride was destroyed with 25% acetic acid (100 gl) at 0 °C. Virus was collected by centrifugation and washed with PBS.
Results
Western blot analysis of gC-II MoAbs M o A b s which reacted strongly with purified gC-II in an E L I S A assay were first analyzed to determine the m o l e c u l a r weights o f the proteins they recognized. This was d o n e by W e s t e r n blotting using whole T o w n e H C M V as the antigen.
58
B. Karl et al.
Fig. 1. Western Not analysisdone with whole Towne strain HCMV. HCMV was solubilized in SDS, urea and reduced with 2 mercaptoethanol prior to SDS-PAGE. 1 SP2 negative ascites control. 2-4 Group t MoAbs, 5-9 Group 2 MoAbs. Numbers to the left and right indicate molecular weights x 10 - 3
Whole HCMV was solubilized in SDS, urea, and reduced with beta-mercaptoethanol prior to SDS-PAGE and Western blotting. A mouse ascites (SP2), which did not contain HCMV antibody, was used as a negative control. This ascites did not react by Western blotting (Fig. 1, lane 1). Group 1 contained 3 MoAbs with different subtypes, 9El0 (IgG3), 8B4 (IgM), and 26E2 (IgG2a). Group 1 MoAbs reacted strongly with 47 k to 63 k molecular weight proteins and very weakly with higher molecular weight proteins (Fig. 1, lanes 2-4). Group 2 contained 7 MoAbs; 5 of these had an IgG2a subtype, while MoAb 15G5 was an IgG1 and MoAb 25C8 was an IgA. Group 2 MoAbs reacted strongly with proteins with molecular weights between 39,000 to 48 k, but they also reacted strongly with proteins from 90 k to greater than 200 k; representative data for 5 Group 2 MoAbs are shown in Fig. 1, lanes 5-9. The proteins with molecular weights of 39-48 k recognized by Group 2 MoAbs appear to be different from the 47-63 k molecular weight proteins recognized by Group 1 MoAbs since they did not completely overlap in molecular weight. Furthermore, Group 1 MoAbs showed little or no reactivity with the high molecular weight proteins while Group 2 MoAbs showed strong reactivity (Fig. 1). Thus, MoAbs in Group 1 recognize proteins different from those recognized by MoAbs in Group 2. To show that both groups of MoAbs recognized gC-II complexes,
HCMV glycoproteins
59
Fig. 2. gC-II complexes were immunoaffinity purified with Group 2 MoAb 15F9. Approximately 100-150 gg of gC-II protein was used to do Western blotting. Following SDSPAGE, proteins were transferred to nitrocellulosepaper and the paper cut into strips for Western blotting. Each strip contained approximately5-10 gg protein. A Western blotting done under non-reducing conditions, gC-II complexes were initially separated by SDSPAGE using a 7% polyacrylamidegel prior to transfer to nitrocellulosepaper for Western blotting. 1 SP2 negative control. 2 Group 1 MoAb 9El0; 3, 4 Group 2 MoAbs 12G9 and 15F9 respectively.B Same as A except that gC-II complexeswere reduced prior to SDSPAGE. SDS-PAGE was done in a 10% polyacrylamide gel. 1 SP2 negative control. 2 Group 1 MoAb 9El0; 3, 4 Group 2 MoAbs 12G9 and 15F9 respectively.Numbers to the fight indicate molecular weight x 10-3
Western blotting was done with gC-II complexes which had been immunoaffinity purified with Group 2 MoAb 15F9. Immunoaffinity purified gC-II complexes were separated by SDS-PAGE under non-reducing conditions prior to Western blotting. Under non-reducing conditions, MoAbs from either group reacted with complexes with molecular weights of 67-93 k and 130 k to greater than 200 k, representative data for Group 1 MoAb 9E10 and Group 2 MoAbs 12G9 and 15F9 are shown in Fig. 2A. Thus, gC-II complexes which varied in molecular weight were recognized by both groups of MoAbs. Immunoaffinity purified gC-II complexes were also reduced and individual gC-II glycoproteins separated by SDS-PAGE prior to Western blotting. After reduction of purified gC-II complexes, Group 1 MoAbs reacted with proteins having molecular weights of 47-63 k while Group 2 MoAbs reacted with proteins having molecular weights of 39-48 k and 90 k to greater than 200 k (Fig. 2B). Again after reduction the molecular weights of the proteins recognized were different when comparing the two groups of MoAbs. These data showed that gC-II complexes which had been immunoaffinity purified with Group 2 MoAb 15F9 contained both group 1 and 2 glycoproteins recognized by their respective MoAbs.
60
B. Kari et al.
Immunoprecipitation of gC-H glycoproteins To determine if the proteins detected by Western blotting were glycosylated, individual gC-II proteins labeled with radioactive glucosamine were immunoprecipitated with MoAbs from both groups. To do this gC-II complexes were labeled with radioactive glucosamine and immunoaffinity purified with Group 2 M o A b 15F9. The purified gC-II complexes were solubilized in SDS, urea, reduced with 20 m M dithiothreitol, and alkylated with 50raM iodoacetamide in 0.2 M Tris (pH 7.0). After dialysis to remove excess reagents, NP-40 was added so that the final NP-40 concentration was 1.0%. The reduced gC-II glycoproteins were serially immunoprecipitated with a negative control, followed by a Group 1 M o A b (26E2) and than with a Group 2 M o A b (15F9). Immunoprecipitated glycoproteins were examined by SDS-PAGE and fluorography. The total number of gtycoproteins in the starting material are shown in Fig. 3 lane 1. These included a broad band with molecular weights from 48 k -
Fig. 3. The gC-II complexes were immunoaffinity purified from extracellular virus which was labeled with radioactive glucosamine. Complexes were reduced and alkylated to obtain individual gC-II glycoproteins. This material was used for serial immunoprecipitations. The order of immunoprecipitation was a negative control followed by a Group 1 MoAb 26E2, and finally with Group 2 MoAb 15F9. 1 Fluorogram showing all the reduced gCII glycoproteins obtained from immunoaffinity purified gC-II complexes. 2 Ftuorogram of gC-II glycoproteins immunoprecipitated with a negative control SP2. 3 Fluorogram of glycoproteins immunoprecipitated with a Group 1 MoAb 26E2.4 Fluorogram of glycoproteins immunoprecipitated with a Group 2 MoAb 15F9. Numbers to the right of all lanes indicate molecular weights x 10 -3. T Top of the gel
HCMV glycoproteins
61
63 k and glycoproteins with molecular weights from 90 k to greater than 200 k. An SP2 ascites negative control did not immunoprecipitate any glycoproteins (Fig. 3, lane 2). After the negative control, Group 1 MoAb 26E2 was used. This MoAb immunoprecipitated glycproteins with molecular weights of 48-63 k (Fig. 3, lane 3). After using Group 1 MoAb 26E2 Group 2 MoAb 15F9 was used. This MoAb immunoprecipitated glycoproteins with molecular weights of 43-48 k, and 90 k to greater than 200 k (Fig. 3, lane 4). The order in which the MoAbs were used did not affect the result (data not shown). The glycoproteins immunoprecipitated by both groups of MoAbs appeared to account for all the glycoproteins detected in the starting material (Fig. 3, compare lane 1 with lanes 3 and 4). These data showed that most gC-II proteins were glycosylated.
Protein composition of the gC-H complexes The relative abundance of the various gC-II glycoproteins was examined by Coomassie blue staining and by incorporation of [3H]Arg. Group 2 MoAb 15F9 was used to immunoaffinity purify gC-II complexes which were reduced and examined by SDS-PAGE. By Coomassie blue staining glycoproteins with molecular weights of 48 k, and 90 k, to greater than 200 k appeared most abundant (Fig. 4, lane 1). With [3H]Arg labeling, the 48 k and 90 k, to greater than 200 k, were also most abundant (Fig. 4, lane 2). With either method glycoproteins
Fig. 4. gC-II complexes, radiolabeled with [3H]Arg or unlabeled, were immunoaffinity purified with Group 1 MoAb 15F9. These complexeswere reduced and proteins separated by SDS-PAGE in 10% polyacrylamidegels. Numbers to the right indicatemolecularweights x 10-3. 1 Coomassie blue stained gel showing reduced proteins obtained from immunoaffinity purified gC-II. 2 Fluorogram of reduced gC-II proteins labeled with radioactive arginine. T Top of the gel
62
B. Kari et al.
with molecular weights between 50-63 k appear to be minor components relative to the other glycoproteins. Most of the glycoproteins recognized by the Group 1 MoAbs were in this molecular weight range.
Surface labeling of extracellular virus The location of the gC-II glycoproteins in extracellular virus was determined by a surface labeling method. Since the glycoproteins of gC-II contain galactose and galactosamine [ 10], galactose oxidase treatment followed by reduction with tritiated borohydride was employed to surface label purified HCMV virions. After labeling, the virions were washed twice with PBS to remove excess reagents. Washed virions were extracted wtih 1.0% NP-40, insoluble material removed by centrifugation, and the labeled gC-II complexes immunoprecipitated from the extract with Group 2 MoAb 15F9. Immunoprecipitated complexes were examined by electrophoresis under reducing and non-reducing conditions followed by fluorography. Without prior treatment with galactose oxidase, no radioactivity was incorporated into gC-II glycoproteins showing the specificity
Fig. 5. Whole virions were treated with galactose oxidase followed by reduction with tritiated borohydride. As a negative control, virions which had not been treated with galactose oxidase were exposed to tritiated borohydride. Virions were pelleted and washed to remove reactants. After washing, the virions were extracted with 1.0% NP-40 and insoluble material removed by centrifugation. The extract was immunoprecipitated with Group 2 MoAb t 5F9 to obtain labeled gC-II complexes. Immunoprecipitated gC-II complexes were separated by SDS-PAGE under non-reducing conditions in a 7% polyacrylamide gel (1, 2) or under reducing conditions in a 10% polyacrylamide gel (3, 4). 1, 3 Controls done without galactose oxidase. 2 gC-II complexes; 4 gC-II glycoproteins labeled with galactose oxidase and tritiated borohydride treatment. Glycoproteins were detected by fluorography. Numbers to the right indicate molecular weights x 10 - 3
HCMV glycoproteins
63
of the reaction (Fig. 5, lanes 1 and 3). However, after galactose oxidase treatment gC-II complexes with molecular weights of 6 7 - 93 k and 130 k to greater than 200 k were heavily labeled (Fig. 5, lane 2). When these complexes were reduced gC-II glycoproteins with molecular weights from 48-63 k and 90 k to greater than 200 k were detected ( Fig. 5, lane 4). The most heavily labeled glycoproteins had molecular weights of 48-63 k. These data show that both high and low molecular weight gC-II glycoproteins were exposed on the surface of the virus. Discussion
The primary objective of the current study was to characterize the different glycoproteins in gC-II complexes. Our data clearly show that gC-II complexes contain two immunologically distinct groups of glycoproteins. Gtycoproteins with molecular weights of 47-63 k were recognized by 3 Group 1 MoAbs. Group 1 MoAb 9El 0 has been shown to recognize products of the HXLF gene family [7]. These glycoproteins appeared to be minor components when compared to other gC-II glycoproteins based on Coomassie blue staining of SDS-PAGE gels. For the purposes of discussion we previously designated these glycoproteins gp47-52 to simplify their description [7, 10]. Group 2 contained glycoproteins recognized by 7 MoAbs. These glycoproteins were much more diverse in their molecular weights. The glycoproteins with molecular weights greater than 90 k have been purified by gel-filtration HPLC. These purified glycoproteins were solubilized in SDS, urea, and reduced with beta-mercaptoethanol a second time prior to SDS-PAGE with no further reduction in their molecular weights [10]. Thus, the high molecular weights of these glycoproteins were not do to incomplete reduction of disulfide bonds. Group 2 also included glycoproteins with molecular weights of 39 k to 48 k. It is possible that some of the low molecular weight Group 2 glycoproteins were generated by proteolysis during the isolations. However, PMSF was used to prevent proteolysis. Moreover, whole virus was solubilized directly for SDS-PAGE and Western blotting to prevent proteolysis. When this was done the 39 k to 48 k molecular weight glycoproteins were still detected. The molecular weights and relative amounts of the Group 2 glycoproteins were also consistent between preparations. These data suggested that the diversity in molecular spear paper weights was not due to proteolysis. Therefore, our data indicate that, gC-II complexes are made during virus assembly which are composed of two distinct groups of heterogeneous glycoproteins. Group 1 glycoproteins are products of the HXLF gene family [7]. The glycoproteins in Group 2 may be products of different transcripts encoded by or alternative splicing the HXLF gene family or by a different gene. The heterogeneous composition of the gC-II complexes is different than other glycoprotein complexes in the envelope of HCMV. For example, the family of complexes variably designated gB, gp 116-52, or gC-I is derived from a glycoprotein encoded by a single gene. This glycoprotein is subjected to proteolysis during its biosynthesis to generate two glycoproteins which are disulfide linked to form these complexes [1, 2, 5].
64
B. Kari et al.
Surface labeling showed that the gC-II complexes and their high and low molecular weight glycoproteins are exposed on the surface of extracellular HCMV. However, while both high and low molecular weight glycoproteins were labeled by treatment with galactose oxidase and tritiated borohydride the 48 k to 63 k molecular weight glycoproteins were most heavily labeled by this method. Both groups of MoAbs recognized glycoproteins in this molecular weight range. However, these results were consistent with our previous analysis which showed that gC-II glycoproteins in this molecular weight range contained a higher content of O-linked oligosaccharides when compared to the higher molecular weight gC-II glycoproteins [10]. Thus, the high molecular weight gC-II glycoproteins are glycosylated differently from the low molecular weight glycoproteins. The data presented in this and previous reports [6, 7, 9, 10] show that the gC-II family of complexes are very heterogeneous. The reason for this heterogeneity is not clear. One possibility is that the gC-II complexes are biochemically heterogeneous because they have multiple functions. Nothing is known about the functions of the gC-II family of complexes. Thus, determining their functions will be the focus of our future investigations.
Acknowledgments This work was supported by NIH Grant HD19937-02 and by a grant from the Children's Biomedical Research Institute, St. Paul, Minnesota.
References 1. Britt WJ, Auger D (1986) Synthesis and processing of the envelope gp55-116 complex of human cytomegalovirus. J Virol 58:185-191 2. Britt WJ, Vugler LG (1989) Processing &the gp 55.-116 envelope glycoprotein complex (gB) of human cytomegalovirus. J Virol 63:403-410 3. Gahmberg CG, Hakamori S-I (1973) External labelling of cell surface galactose and galactosamine in glycolipid and glycoprotein of human erythrocytes. J Biot Chem 248: 4311-4317 4. Gretch DR, Suter M, Stinski MF (1987) The use of biotinylated monoclonal antibodies and streptavidin affinity chromatography to isolate herpesvirus hydrophobic proteins or glycoproteins. Anal Biochem t63:270-277 5. Gretch DR, Gehrz RC, Stinski MF (1988) Characterization of a human cytomegalovirus glycoprotein complex (gcI). J Gen Virol 69:1205-1215 6. Gretch DR, Karl B, Rasmussen L, Gehrz RC, Stinski MF (1988) Identification and characterization of three distinct families of gtycoprotein complexes in the envelope of human cytomegalovirus. J Virol 62:875-881 7. Gretch DR, Karl B, Gehrz RC, Stinski MF (1988) A multigene family encodes the human cytomegalovirus glycoprotein complex gCII (gp47-52 Complex). J Virol 62: 1956-1962 8. Ho M (1982) Cytomegalovirus biology and infection. In: Grennough WB III, Merigan TC (eds) Current topics in infectious disease. Plenum, New York 9. Karl B, Lugsenhop N, Goertz R, Wabuke-Bunoti M, Radeke R, Gehrz R (1986) Characterization of monoclonal antibodies reactive to several biochemically distinct human cytomegalovirus glycoprotein complexes. J Virol 60:345-352
HCMV glycoproteins
65
10. Kari B, Gehrz R (1988) Isolation and characterization of a human cytomegalovirus glycoprotein containing a high content of O-linked oligosaccharides. Arch Virol 98: 171-188 11. Laemmli UK (1970) Cleavage of structural proteins during assembly of the head of bacteriophage T4. Nature 227:680-684 12. Liu Y-N C, Kari B, Gehrz RC (1988) Human immune responses to major human cytomegalovirus glycoprotein complexes. J Virol 62:1066-1070 Authors' address: B. Kari, Children's Biomedical Research Institute, Division of Biochemistry, Room L-284, St. Paul, MN 55102, U.S,A. Received December 13, 1989
