in the epitope recognized by JF186, as shown for P3HR-1. ... EBNA-5 was detected as a ladder of protein species of 20 to 130 ..... of as few as 104 IB4 cells.
Vol. 61, No. 12
JOURNAL OF VIROLOGY, Dec. 1987, p. 3870-3878
0022-538X/87/123870-09$02.00/0 Copyright C) 1987, American Society for Microbiology
Monoclonal and Polyclonal Antibodies against Epstein-Barr Virus Nuclear Antigen 5 (EBNA-5) Detect Multiple Protein Species in Burkitt's Lymphoma and Lymphoblastoid Cell Lines JURGEN FINKE,t* MARTIN ROWE, BENGT KALLIN, INGEMAR ERNBERG, ANDERS ROSEN, JOAKIM DILLNER, AND GEORGE KLEIN
Department of Tumor Biology, Karolinska Institute, S-104 01 Stockholm, Sweden Received 13 April 1987/Accepted 10 September 1987
The Epstein-Barr virus nuclear antigen 5 (EBNA-5) is encoded by highly spliced mRNA from the major IR1 (BamHI-W) repeat region of the virus genome. A mouse monoclonal antibody, JF186, has been raised against a synthetic 18-amino-acid peptide deduced from the EBNA-5 message of B95-8 and Raji cells. The antibody showed characteristic coarse nuclear granules by indirect immunofluorescence and revealed multiple EBNA-5 species by immunoblotting and immunoprecipitation. The B95-8 line itself and all B95-8 virus-carrying cells, whether lymphoblastoid cell lines or in vitro-converted sublines of Epstein-Barr virus (EBV)-negative Burkitt's lymphoma (BL) lines, were EBNA-5 positive. Among 36 cell lines carrying different EBV strains, only 10 expressed the B95-8-Raji-prototype EBNA-5 recognized by JF186; this was probably due to genetic variation in the epitope recognized by JF186, as shown for P3HR-1. Human antibodies, affinity purified against EBNA-5-JF186 immunoprecipitates, detected EBNA-5 in the majority of EBV-positive BL lines and in all lymphoblastoid cell lines containing the BL-derived viruses. Thus, EBNA-5 can be expressed by all virus isolates examined, but is down-regulated, together with other latent gene products, in a minority of BL lines which have a particular cellular phenotype. EBNA-5 was detected as a ladder of protein species of 20 to 130 kilodaltons (kDa), with a regular spacing of 6 to 8 kDa, consistent with the coding capacity of the combined BamHI-W 66- and 132-base-pair exons, together with shifts of 2 to 4 kDa, consistent with the size of the separate 66- and 132-base-pair exons. Multiple EBNA-5 proteins can be expressed by the single cell as shown by cloning of newly infected cells.
4) cell lines have led to the identification of highly spliced transcripts, derived from the W fragment repeats, containing different numbers of 66- and 132-base-pair (bp) tandem exons. These cDNA data suggest the existence of a highly repetitive, arginine- and proline-rich protein of variable size encoded in addition to EBNA-2 in the BamHI WYH region. This protein may be encoded either by a separate message (33) or by polycistronic messages that also encode other latent proteins of the EBNA family (32). The origin of the putative coding exons of EBNA-5 is summarized in Fig. 1. Using rabbit antisera raised to synthetic peptides corresponding to the W repeat exon, we have previously identified a protein of variable size which was designated EBNA-5 (7). Recently, Wang et al. (38) have confirmed these findings using an affinity-purified human serum reagent obtained by reaction with a bacterial fusion protein expressing part of the EBNA-5 sequence. They designated the corresponding viral protein the leader protein, EBNA-LP, since it was encoded by a leader sequence of a bicistronic mRNA that also encoded EBNA-2 (32). We continue to use the designation EBNA-5, because the possibility cannot be excluded that this protein is expressed independently of EBNA-2 and of other EBNAs. To study more precisely the biology of EBNA-5, we have now raised a mouse monoclonal antibody (designated JF186) against the peptide 186 that was used in our previous study (7). Because the 18-amino-acid peptide sequence (Pro-ArgGly-Asp-Arg-Ser-Glu-Gly-Pro-Gly-Pro-Thr-Arg-Pro-GlyPro-Pro-Gly) was deduced from the message corresponding to the BamHI-W repeats of B95-8 and Raji (3, 4), we refer to the corresponding protein as the B95-8/Raji (BR) prototype EBNA-5. Furthermore, EBNA-5-specific human antibodies
Epstein-Barr virus (EBV) is the causative agent of infectious mononucleosis (IM) and is closely associated with two malignancies, Burkitt's lymphoma (BL) and nasopharyngeal carcinoma. It is a highly efficient transforming agent for human B lymphocytes, leading to the establishment of continuously growing lymphoblastoid cell lines (LCLs). A family of EBV nuclear antigens (EBNAs) and a latent membrane antigen are expressed in growth-transformed cells (6a, 15). Their analysis is crucially important for the understanding of the transformation process and the role of EBV in malignancy. The EBNA complex was originally defined by anticomplement fluorescence (24) and is expressed in every EBV-transformed cell. Recent analyses revealed that EBNA is a family of at least five proteins. Latent transcripts from the BamHI K and WYH regions code for EBNAs 1 and 2, respectively, whereas latent transcripts from the BamHI E region encode EBNA-3, EBNA-4, and EBNA-6 (5, 6a, 10, 11, 14, 31, 35, 36; A. Ricksten, B. Kallin, H. Alexander, J. Dillner, R. F?ahreaus, G. Klein, R. Lerner, and L. Rymo, Proc. Natl. Acad. Sci. USA, in press). The BamHI WYH region is of particular interest since it is most probably involved in growth transformation; deletion of the EBNA-2-encoding part of the WYH region, as in the P3HR-1 and Daudi lines, is associated with the loss of transforming ability (13, 20, 23). Studies on cDNA clones derived from the Raji (3), JY (34), IB4 (32, 33), and B95-8 (2, * Corresponding author. t Present address: Abteilung Hamatologie Onkologie, Medizinische Universitatsklinik, D-7800 Freiburg, Federal Republic of
Germany. 3870
VOL. 61, 1987
MONOCLONAL ANTIBODY TO MULTIPLE EBNA-5 PROTEINS
1A) IRI
TR -
(B)
m
IR3
U2 _ r;;:=--a -
U1
_
_
_
--
U3
-
IR4
-W4
_
TR
(I=
WW
C
mRNA|..| WO
WI
W.W.W |
|,|
W W WW i||
||.|
SW Y.
.|.
|
Y'
W2
3
.n 13
H
Y2 Y3 *22
IC) Protein
FW01F|W2
pW
2>t|Y
n
FIG. 1. Schematic representation (not to scale) of the genetic origin of the EBNA-5 proteins. (A) Map of the EBV B95-8 genome, which is composed of largely unique (Ul to U5) and internal repeat (IR1 to IR4) regions. (B) BamHl restriction map of the U1-IR1-U2 region of EBV DNA, beneath which are indicated promoter sites (arrows) and the exons within the W and Y fragments which may be used in the generation of the EBNA-5 message (32, 33). The splicing together of a 28-bp noncoding WO exon (providing nucleotides AT) and a 66-bp Wl exon (providing G) generates the essential methionine initiation codon which replaces the first two amino acid codons of the intact Wl exon. Theoretically, the W01 exon may be generated in any of the W repeats, thus creating the possibility of several mRNA species with different numbers of subsequent W1-W2 exons. Only the first two exons within the Y region are coding exons. (C) The EBNA-5 proteins contain a constant aminoterminal W01-W2 sequence and a constant carboxy-terminal Y1-Y2 sequence. In between there may be variable numbers (0 to 10 for B95-8 virus) of W1-W2 repeat sequences.
were derived by affinity purification of human serum reactive with EBNA-5-JF186 immunoprecipitates in order to detect other epitopes of the antigen. The monoclonal antibody proved to be useful to study a unique feature of EBNA-5, that is, the expression of multiple proteins from the same W repeat region. MATERIALS AND METHODS Cell lines. Cells were grown in RPMI 1640 (GIBCO Laboratories, Grand Island, N.Y.) supplemented with 10% heat-inactivated fetal calf serum, 2 mM glutamine, 25 ,ug of gentamicin per ml, 200 ,ug of streptomycin per ml, and 60 ,ug of penicillin per ml and were kept in a moist atmosphere with 5% CO2 at 370C. IARC-prefixed BL cell lines were kindly provided by G. Lenoir (18), QIMR-prefixed lines were provided by D. Moss, and BL-postfixed cell lines and LCLs containing BL-derived EBV were provided by A. Rickinson (26, 27). Cherry and SPIM lines (9) are spontaneous LCLs from the blood of IM patients; B95-8 is a marmoset cell line carrying IM-derived EBV (19). PG N8652 was established from the blood of a patient with chronic lymphocytic leukemia by spontaneous outgrowth. The QIMR-WW1-LCL was an LCL immortalized by virus derived from the QIMR-WIL cell line (22), and the CBC-SEB lines were derived by in vitro transformation of cord blood lymphocytes with EBV from throat washings of IM patients (9). The BL lines presented in Table 1 have been characterized previously (9, 14, 27, 30, 30a). EBV-converted sublines of the EBVnegative lines BJAB, Ramos, BL41, and BL47 were established by in vitro infection with B95-8 virus, as were the fetal bone marrow cell line FEBM-15MnE95 and the cord blood lymphocyte line CBC-E95. The subline EHR-A-Ramos was established with P3HR-1 virus.
3871
Generation of monoclonal antibodies. A 6-week-old BALB/c mouse was immunized subcutaneously with 100 ,ug of peptide 186 coupled to keyhole limpet hemocyanin in Freund complete adjuvant (Difco Laboratories, Detroit, Mich.). After 3 and 6 weeks, the mouse was boosted with the same amount of peptide-keyhole limpet hemocyanin in Freund incomplete adjuvant. After 4 weeks, the mouse was injected intravenously with 100 ,ug of peptide, and it was sacrificed 4 days later for hybridoma production. The mouse serum had an antipeptide titer of >1:2,048 in enzyme-linked immunosorbent assay on the day of fusion. Spleen cells (n = 108) were fused with 25 x 106 cells of the mouse myeloma line Sp2/0 according to a modification of the Kohler and Milstein procedure (16). The cell suspension was distributed into seven 96-microwell tissue culture plates, together with mouse thymocyte feeder cells (4 x 104 cells per well) and cultured in RPMI medium supplemented with 15% fetal calf serum. After 24 h, half of the medium was exchanged for 2 x
TABLE 1. Expression of EBNA-5 by different EBV isolates Cell line
Origin of EBVa
B95-8 SPIM 119 SPIM 199 SPIM KK Cherry QIMR-WW1-LCL
IM
CBC-SEB-M7 CBC-SEM M9 CBC-SEB M13 CBC-API CBC-SEM-108 PG N8652
IM IM IM IM IM CLL
Raji Namalwa Daudi P3HR1 Jijoyeb
BL BL BL BL BL BL
Raelb Ag876b ELI-BLb
WAN-BLb MAK-BLb MUK-BLb OBA-BLb
KYU-BLb MWI-BLb
QIMR-WWlb
QIMR-WW2b IARC-BL16b IARC-BL18b
IARC-BL29b IARC-BL36b IARC-BL37b
IARC-BL60b IARC-BL72b IARC-BL74b
IM IM IM IM IM
BL BL BL BL BL BL
EBNA-5 detected by: JF186
+_ _
+
_
+
_
+
+ +
+ +
+
+ + + + + +
+ +
+
.
+
BL BL BL BL BL BL BL BL BL BL BL BL
Human anti-E5
+ + + + +
+
+ +
lymphocytic leukemia. For the BL lines indicated, virus was also rescued into normal B cells (see text). All these LCLs had EBNA-5 as detected with human anti-E5 (not a
b
Origin of virus: CLL, chronic
shown).
3872
FINKE ET AL.
hypoxanthine-aminopterin-thymidine (HAT) medium for the selection of hybrid cells. Cell culture supernatants were screened for antibody production by enzyme-linked immunosorbent assay against peptide 186 and by immunoblotting against cell extracts. Of several clones which recognized EBNA-5 in immunofluorescence (IF) and immunoblotting and which did not show any cross-reaction with extracts of six EBV-negative cell lines, one was selected and recloned three times. This hybridoma, JF186, produced protein Abinding antibodies of subclass immunoglobulin Gl (IgGl), as determined by immunodiffusion test according to Ouchterlony and Nilsson (21). The standard working dilution of JF186 culture supernatant was 1:20 for IF and 1:5 for immunoblotting. Enzyme-linked immunosorbent assay screening. Microtiter plates (96 wells; Dynatech, Zug, Switzerland) were coated with 100 pI of uncoupled peptide 186 (10 RgIml) in carbonate buffer, pH 9.4, at 4°C overnight. The plates were washed three times with phosphate-buffered saline (PBS) supplemented with 5% nonfat dry milk (PBSM). Each well was filled with 100 ,ul of hybridoma culture supernatant diluted 1:1 in PBSM. After overnight incubation at 4°C, the plates were washed three times with PBSM followed by 1 h of incubation at room temperature with 100 pul of anti-mouse IgG alkaline phosphatase conjugate (Sigma Chemical Co., St. Louis, Mo.), 1:1,000 in PBSM. After five washes with PBS-0.1% Tween 20, 100 ,ul of Sigma phosphatase substrate (1 mg/ml in 0.1 M Tris-glycine, pH 10.3; 1 mM MgCl2; 1 mM ZnCl2) was added and incubated for 1 h at 37°C. The reaction was measured at 405 nm in a Titertek Multiscan spectrophotometer. Values of five times above background (SP2/0 cell culture supernatant) were defined as positive. Immunoblotting. Whole-cell extracts were prepared by 10 s of sonication (Soniprep 150 sonifier; MSE Scientific Instruments, Crawley, Sussex, England) and subsequent boiling in electrophoresis sample buffer. Samples of 106 cells in 100 pul were loaded per lane for polyacrylamide gel electrophoresis according to Laemmli (17). Electrophoretic transfer to nitrocellulose (Hybond C; Amersham Corp., Arlington Heights, Ill.) was done essentially as described by Towbin et al. (37) for 4 h at 14°C at 8 V/cm. The filters were stained with Ponceau S (Sigma), and the positions of the molecular weight markers (Bio-Rad Laboratories, Richmond, Calif.) were noted. After preabsorption in PBSM, the filters were incubated with hybridoma cell culture supernatants diluted 1:5 in PBSM overnight at 4°C. After three washes with PBSM, the filters were incubated with rabbit anti-mouse IgG (DAKOPATTS, Copenhagen, Denmark), 1:2,000, for 1 h, washed with PBSM and subsequently incubated with alkaline phosphatase-conjugated goat anti-rabbit IgG (Sigma) 1:1,000 in PBSM for 2 h at room temperature. The filters were washed six times with PBS-0.5% Tween 20 and stained with 50 mM Tris hydrochloride (pH 8.8)-2 mM MgCl2-0.2 mg of a-naphthylphosphate per ml-0.5 mg of Fast Red Salt (Sigma) per ml. Immunoprecipitation. Cells (5 x 107) were disrupted by sonication (10 s on ice) in 1 ml of lysis buffer (0.5% sodium dodecyl sulfate 0.5% sodium desoxycholate, 0.5% Nonidet P-40, 150 mM NaCl, 50 mM Tris [pH 7.5], 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride). Lysates were incubated with 200 pul of monoclonal antibody hybridoma supernatant at 4°C by end-over-end rotation overnight before the addition of 50 pu1 of protein A-Sepharose CL-4B (Pharmacia, Uppsala, Sweden) for a further 4 h. The Sepharose beads were
were
washed twice with lysis buffer and twice with PBS and subsequently boiled in electrophoresis sample buffer.
J. VIROL.
Samples equivalent to material from 107 cells were used for immunoblotting as described above. Purification of EBNA-5-specific human antibodies. EBNA-5 immunoprecipitates from 400 x 106 IB-4 cells were loaded onto a 16-cm-wide Laemmli gel for sodium dodecyl sulfatepolyacrylamide gel electrophoresis and transferred to nitrocellulose as described above. Small strips from both edges of the nitrocellulose sheet were reacted with human serum from a chronic lymphocytic leukemia patient known to have high anti-EBNA titers in anticomplement IF (1:320). After immunostaining as described above but with goat antihuman IgG alkaline phosphatase-conjugated antibodies (Sigma), the developed strips were realigned with the original nitrocellulose sheet and a 1-cm-wide horizontal strip corresponding to the 42-kilodalton (kDa) EBNA-5 band of IB-4 was cut out and reacted with human serum, 1:20 in PBSM, for 2 h at room temperature. After four washes in PBS-0.5% Tween 20, the strip was placed in a plastic tube containing 1 ml of 500 mM diethylamine-150 mM NaCl, pH 11.5, for antibody elution (25). After 2 min of gentle mixing, the strip was withdrawn and washed in PBS for reuse. The antibody-containing solution was neutralized with 0.2 M phosphate buffer, pH 2.0, and diluted 1:1 in PBSM for use in blotting or in PBS-10% fetal calf serum for IF. IF. Air-dried cell smears were fixed with acetonemethanol 2:1 at -20°C for 5 min. After being washed in PBS, the smears were incubated with monoclonal antibody (1:20 dilution) for 1/2 h at room temperature in a moist chamber. Rabbit anti-mouse IgG fluorescein isothiocyanate conjugate (DAKOPATTS) diluted 1:10 in PBS was used as second antibody for 30 min at room temperature. Anticomplement IF, either with monoclonal antibody or with EBNA-reactive human sera, was done as described previously (24). Primary infection of B cells with B95-8 virus in vitro. High-density resting B cells were isolated from tonsils as described elsewhere (7). The EBV-negative BL line Ramos was fed 24 h before the harvesting of cells for infection with B95-8 virus. Cells (3 x 107) were incubated for 2 h at 37°C with 40 ml of B95-8 cell culture supernatant. Cells were pelleted and suspended in fresh medium and distributed equally into three culture flasks. One flask was harvested every 24 h to test for EBNA-5, EBNA-1, and EBNA-2 expression by immunoblotting and by IF. Cloning of B95-8-transformed B cells. Cloning was performed as described previously (8). Ficoll-Isopaque-isolated cord blood lymphocytes were infected with EBV B95-8 at a low multiplicity of infection and subsequently were seeded in 96-well microculture plates with an average of 10 cells per well. Cells were fed weekly. After 4 to 6 weeks, growing clones were picked and transferred to culture flasks. The assay for EBNA-5 was performed 8 weeks after establishment. RESULTS EBNA-5-specific monoclonal antibody JF186. The monoclonal antibody JF186 was used to detect EBNA-5 in indirect IF, immunoblotting, and immunoprecipitation. In indirect IF, EBNA-5 showed a characteristic pattern dominated by three to eight coarse, brilliant intranuclear granules, appearing against a fainter background of diffuse nuclear staining in 100% of cells. This was very different from the diffuse fine granular nuclear staining observed in anticomplement IF using polyvalent human sera directed to the other EBNAs (results not shown). Hybridoma culture supernatants showed a titer of 1:10,000 in IF on IB4 cells.
VOL. 61, 1987
MONOCLONAL ANTIBODY TO MULTIPLE EBNA-5 PROTEINS
The monoclonal antibody JF186 detected EBNA-5 in a proportion of LCLs and BL cell lines carrying different isolates of EBV as detailed below. Some of the EBNA-5positive lines were stringently virus nonproducers (e.g., IB4 and Namalwa; Table 1). The presence of EBNA-5 was therefore independent of the lytic cycle. By immunoblotting, JF186 detected EBNA-5 in extracts of as few as 104 IB4 cells. In routine immunoblotting using whole-cell extracts from 106 cells per track, EBNA-5 was characteristically detected as a ladder of bands of various sizes and intensities in different EBV-positive lines (e.g., IB4, 42, 40, 38, 32, and 28 kDa; and B95-8, 66, 60, 54, 48, and 42 kDa). Lower levels of expression could be detected more reproducibly when immunoprecipitated material, corresponding to 107 cells, was used for immunoblotting (e.g., Raji, 44 and 42 kDa) (Fig. 2). Certain lines (Rael, Jijoye, Daudi, P3HR-1; Fig. 2) were consistently nonreactive with JF186. By using JF186, EBNA-5-specific bands were demonstrated in 10 of 36 cell lines tested, each carrying a different isolate of EBV (Table 1). EBNA-5 expression in cell lines carrying different EBV isolates. To investigate EBNA-5 expression in the cell lines that do not express the BR prototype antigen, EBNA-5specific antibodies were purified from human serum. This human anti-E5 serum detected multiple EBNA-5 bands of various sizes in the majority of EBV-positive cell lines, including the lines which were nonreactive with the monoclonal antibody (Fig. 3). The typical EBNA-5 size variation among different virus isolates (IB4, 42 kDa; Namalwa, 84 and 80 kDa; Jijoye, 35 kDa) as well as among different cell lines containing the same virus (P3HR-1, 20 kDa; EHR-ARamos, 90 and 82 kDa) was demonstrated with human anti-E5 (Fig. 3). Table 1 summarizes the results obtained with JF186 and with human anti-E5 on 36 cell lines, each carrying a unique isolate of EBV. B95-8 and Raji cells did not react with human anti-E5, indicating a poorer sensitivity of this reagent. Taken together, the two reagents detected
d:l
a
a:
m
fi
92 66
< IgH 45
31
IgL m
FIG. 2. Immunoblot probed with JF186 monoclonal antibody. Cell lysates from 107 cells of seven EBV-positive lines were immunoprecipitated with JF186, separated on a 9% polyacrylamide gel, and subsequently transferred to nitrocellulose. In this procedure, in addition to EBNA-5, immunoglobulin heavy (IgH) and light (IgL) chains of the monoclonal antibody in the immune complex were detected as indicated. Numbers at left indicate molecular size in kilodaltons.
>-
0
cc
2O
ct:
-
*i
mO
z
w
W
C')
c
Jz
3873
0I
I 0
0
*@ I.. .,
.i
e
,.
92--
X
.
4_.~~~~~~ : di
66-
45-
31-
FIG. 3. Detection of EBNA-5 by using affinity-purified human antibodies. Whole-cell extracts (106 cells per lane) were separated on a 9% polyacrylamide gel and probed with human anti-E5 reagent. The EHR-A-Ramos line has been established by infecting the EBV-negative BL line Ramos with P3HR-1 virus in vitro. CB-M3Jijoye is a cord blood-derived LCL transformed with Jijoye virus in vitro. BJAB is an EBV-negative lymphoma line. Numbers at left indicate molecular size in kilodaltons.
EBNA-5 in all LCLs and in the majority (18 of 24) of BL lines. EBV-negative human sera did not react with EBNA5-JF186 immunoprecipitates nor with EBNA-5 characteristic bands in whole-cell extract immunoblots (data not shown). Neither JF186 nor human anti-E5 cross-reacted with wholecell extracts of any of six EBV genome-negative cell lines tested. To study the influence of the cellular phenotype on EBNA-5 expression, independent virus isolates carried by 20 BL cell lines were used to transform normal B cells as described previously (9, 26). These BL-LCL pairs proved useful for study of the influence of the cellular phenotype on latent viral gene expression. The LCLs thus established invariably expressed EBNA-1, -2, -3, and -4, and latent membrane antigen, even when the BL cells from which the transforming virus was derived lacked expression of EBNA2 and latent membrane antigen (29, 30a). Of 20 LCLs, 4, as well as their parental BL lines, reacted with JF186 (Table 1). However, all 20 LCLs were shown to be EBNA-5 positive by using human anti-E5 antibodies (data not shown), whereas 6 of 20 parental BL lines (Rael, Eli, WW2, IARCBL37, Mak, IARC-BL29; Table 1) lacked detectable EBNA5. These LCL controls rule out the possibility that the BL lines lack EBNA-5 expression because of genetic deletions. Figure 4 illustrates results obtained with BL-LCL pairs representing four BL viruses for which EBNA-5 was detected by JF186 (WAN-BL, IARC BL36, MUK-BL, and IARC BL72), together with an example of a BL virus for which EBNA-5 was detected only by human anti-E5 (IARC BL37). Multiple EBNA-5 protein expression in different cell lines carrying the B95-8 virus. The monoclonal antibody JF186 proved a useful reagent to study the complex pattern of EBNA-5 expression in cell lines of various origins containing
J. VIROL.
FINKE ET AL.
3874
(B)
(A) WAN
MUK BL72 BL37
BL36
4
.-1
_.4Q .40 jC)C
J cn.4 .4.
.4
.0 .40
c
92 -
.4 -j (
-i J m.4n
BL37
-m
-JoL
. . ,. _.
_
_of _ .... ..
S
66%4 545-,~..*._ ~
-j
To answer the question of whether each component of the EBNA-5 ladder was produced by a different cell within the general population, or whether a single cell was capable of producing several EBNA-5 species, cord blood cells were cloned by limiting dilution after transformation with B95-8 virus. The results shown in Fig. 7 demonstrate that different LCL clones differ in their predominant expression of particular EBNA-5-size species and that each clone contained multiple bands common to each other. These findings are also in line with the EBNA-5 ladder expression in the naturally monoclonal BL lines.
a.
_--
31 -
FIG. 4. Detection of EBNA-5 in pairs of BL lines and corresponding LCLs transformed with EBV from the tumor lines. Wholecell extracts (106 cells per track) were separated on a 9o polyacrylamide gel and probed for EBNA-5 with JF186 (A) and human anti-E5 (B). BL cells and LCLs containing EBV strains of WANBL, IARC-BL36, MUK-BL, and IARC-BL72 expressed the B958/Raji (BR) prototype EBNA-5. The BL-LCL pair of BL37 was negative with JF186 but the LCL expressed EBNA-5 as recognized by human anti-E5 (B). Numbers at left indicate molecular size in kilodaltons.
the same prototype transforming virus B95-8. All B95-8 virus-containing cell lines were found to express EBNA-5, usually as multiple bands in the form of a ladder, unlike the other EBNAs, which have a constant size for a given virus isolate (Fig. 5; Table 2). EBNA-5 proteins varied in size between 20 kDa (BL41/95) and 130 kDa (E95-D-Ramos). Not all proteins of the ladder were expressed in every line, and the size of the predominantly expressed band as well as the overall expression level varied between different lines, irrespective of the BL versus LCL origin of the line. By analysis of the EBNA-5 proteins produced in several B95-8-carrying cell lines, it is clear that the spacing between the different EBNA-5 bands is more complex than the 6 kDa previously reported (7, 28). The data shown in Table 2 are more consistent with the existence of three ladders, each having spacings of 6 to 8 kDa, superimposed upon each other. Freshly infected cells and newly established lines generally expressed considerably more peptide species than longer established lines. Thus, the expression of multiple EBNA-5 proteins was particularly prominent after primary infection of resting B cells and the EBV-negative Ramos BL line (Fig. 6). A 12-band EBNA-5 protein ladder, ranging between 20 and 119 kDa, appeared within 24 h of infection. The same ladder was expressed after primary infection both of normal B cells and of Ramos cells, but the signal intensity was much lower in Ramos cells, because a maximum of 10% of the Ramos cells became EBV positive, whereas the majority of the normal B cells became EBV positive. Parallel immunoblots probed with an EBNA-1- and EBNA-2reactive human serum showed that EBNA-5 appeared coincidentally with EBNA-1 and -2 after infection of Ramos cells, whereas in tonsillar B cells, EBNA-5 appeared coincidentally with EBNA-2, and about 24 h earlier than EBNA1 (data not shown).
DISCUSSION
This report describes the generation of a monoclonal antibody, JF186, which is specific for EBNA-5 and which has allowed the expression of the viral protein to be studied in a wide panel of lines each carrying one of several strains of EBV. Using JF186 together with an EBNA-5-specific affinity-purified human serum, we have shown that all different virus isolates tested are capable of expressing EBNA5 proteins, and we present evidence for EBNA-5 epitope variation among different EBV strains. In addition, the sensitive monoclonal antibody allowed the characterization of a much more complex and broader-ranging variation in the molecular size of the EBNA-5 proteins than was previously suspected (7, 28, 38). Unexpectedly, a ladder of different-sized EBNA-5 proteins was demonstrated in monoclonal cell lines. These findings help to explain possible splicing mechanisms involved in the generation of this highly unusual viral protein.