V irus I/iru.s Research, 31 (1994) 57-6.5

0 1994 Elsevier Science B.V.

All rights reserved 016%1702/94/$07.00 Research VIRUS 00953

Electrophoretic migration of adenovirus hexon under non-denaturing conditions

Eleni Fortsas a, Martin Petric a~b and Martha Brown a2* a Department of Microbiology, University of Toronto, 150 College St, Toronto, Ontario M5S IAS, Canada

and b The Hospital for Sick Children, 55.5 University Ave., Toronto, Ontario M5G 1X8, Canada

(Received 25 May 1993; revision received 27 August 1993; accepted 30 August 1993)

Summary

While SDS-PAGE/ immunoblotting is a valuable approach for the characteriza- tion of monoclonal antibodies, the denaturing conditions involved can compromise the recognition of conformational epitopes. This report demonstrates that a group specific epitope on adenovirus hexon can be recognized by immunoblotting follow- ing SDS-PAGE provided that samples are not boiled prior to electrophoresis. Under these conditions, multiple bands corresponding to native forms of hexon were detected above the position of the denatured hexon monomer. Among representative serotypes of subgroups A, B and F, two predominant bands, corresponding to hexon trimers and ‘group of nine’ hexons (GONS), were routinely observed. In contrast, higher order structures, in addition to trimers and GONs, were characteristic of subgroup C adenoviruses. These serotypic differences in stability of hexon structures may reflect differences in protein-protein interactions within the corresponding virions.

Adenovirus; Hexon; Monoclonal antibodies; Electrophoresis; SDS-PAGE

The studies described in this report stem from observations made during characterization of several monoclonal antibodies (mAbs) to human adenoviruses prepared in our laboratories. One of the mAbs in the panel, designated mAb GS,

* Corresponding author.

SSDZ 0168-1702(93)E0074-8

recognized multiple adenovirus serotypes when tested by immunofluorescent stain- ing of infected cells but consistently failed to react with proteins of the same adenovirus serotypes on a Western blot. The serotype specificity of the mAb, coupled with its negative reaction with blotted antigens, suggested that the mAb is specific for a conformational epitope on adenovirus hexon. Conditions for sample

preparation and electrophoresis were then modified to facilitate detection of the epitope on immunoblots.

The virus strains used in this study included prototype strains of Ad40 (Dugan), Ad41 (Tak) and Ad7 which were obtained from the American Type Culture Collection (ATCC). Ad5 was obtained from Dr. F. Graham, McMaster University, Hamilton, Ontario, Canada. Prototype strains of Ad2 and Ad3 were obtained from Dr. N. Singh-Naz, Children’s Hospital, National Medical Center, Washington, DC. Stool specimens containing Ad31 and Ad41 were obtained from patients with gastroenteritis at the Hospital for Sick Children, Toronto, Canada. All adenovirus

serotypes were propagated in 293 cells, a continuous line of transformed human embryonic kidney cells (Graham et al., 1977) as previously described (Brown et al., 1984). The 293 cell line was acquired from Dr. F. Graham, McMaster University,

Hamilton, Ontario, Canada. All hybridoma cells were propagated in minimal essential medium (MEM)

supplemented with 15% FCS, 0.5% glucose, penicillin (500 units/ml) and strepto- mycin (250 pg/ml). Cells were subcultured twice weekly at a split ratio of 1: 3 or 1: 5. Hybridoma cells producing mAb 2Hx-2 (Cepko et al., 1981, 1983) were obtained from ATCC. The hybridoma cells producing mAb GS were prepared by standard techniques (Harlow and Lane, 1988) using NS-1 myeloma cells (obtained from ATCC) fused with spleen cells from BALB/c mice which had been immu-

nized with Ad5. For use as an immunogen, Ad5 virions were propagated in 293 cells, purified by cesium chloride equilibrium density gradient centrifugation (Brown, 1985) and inactivated by irradiation with a portable UV light for 15 min at a distance of l-2 cm. Hybridomas which produced antibody were subcloned twice by limiting dilution then amplified and stored in liquid nitrogen (lo6 cells/ml in culture medium supplemented with 10% dimethyl sulfoxide).

Virus to be used for analysis of proteins by SDS-PAGE and Western blotting was obtained by infection of subconfluent 293 cells at an input MO1 < 1 using virus stocks below passage level 4. For preparation of labelled virus, Tran ?S-Label (70% methionine, 15% cysteine) (ICN) was added to the culture fluid 12 h post-infection (final concentration lo-15 &i/ml). To harvest progeny virus, cells from cultures exhibiting maximal cytopathic effect were collected by centrifugation at 300 x g for 10 min, then resuspended in a minimal volume of culture medium and subjected to 5 cycles of freezing and thawing. The lysate was clarified by centrifugation at 300 x g for 10 min, layered on a preformed CsCl gradient (1.2-1.4 g/ml CsCl in 50 mM Tris-HCI pH 8.1), and centrifuged at 120,000 X g for 1 h. Alternatively, the clarified cell lysate was adjusted to a density of 1.32 g/ml with solid CsCl and centrifuged at 120,000 X g for 20 h. The virus band (~1.34) was collected and dialyzed against 50 mM Tris-HCI (pH 8.1) containing 1 mM phenyl- methylsulfonylfluoride (PMSF) (Boehringer Mannheim).

59

Proteins were separated by polyacrylamide gel electrophoresis (PAGE) using either conventional gels or concave gradient gels. Conventional gels consisted of a resolving gel containing 12.5% acrylamide (acrylamide : bisacrylamide 30 : 0.24) and a stacking gel containing 5% acrylamide (acrylamide : bisacrylamide 30 : 0.8) (Maize& 1971). Concave gradient gels containing 4-15% acrylamide (acryl- amide : bisacrylamide 30 : 0.8) were used for improved separation of high molecular weight proteins (Hames, 1990). All gels were prepared in 0.375 M Tris-HCI (pH 8.9) and contained 0.1% (w/v) SDS. Gels were polymerized with TEMED and ammonium persulfate (APS) as follows: 0.03% (v/v) TEMED and 0.05% (w/v) APS for the 12.5% resolving gel; 0.05% (v/v) TEMED and 0.1% (w/v> APS for the 5% stacking gel as well as the 4-15% concave gradient gel (Hames, 1990; Maize& 1971). Samples were mixed with an equal volume of 2 X sample buffer containing 100 mM Tris-HCl (pH 6.8), 2% (w/v> SDS, 20% (v/v) glycerol, 2% (v/v> 2-mercaptoethanol(2-ME), and phenol red as the tracking dye. For complete denaturation of proteins, samples were boiled in sample buffer for 3 min prior to electrophoresis. All gels were run at constant current using Tris-glycine running buffer (5 mM Tris, 38 mM glycine, 0.1% SDS) until the tracking dye (phenol red) reached the bottom of the gel (Maizel, 1971).

Electrophoretic transfer of proteins from gels onto nitrocellulose or polyvinyli- dine difluoride (PVDF) membranes (BIO-RAD) was carried out at 200 mA for 18 h at 4°C using the Trans-Blot@ Electrophoretic Transfer Cell (BIO-RAD). The transfer buffer contained 49.6 mM Tris, 384 mM glycine, and 20% methanol. For transfer from conventional gels to nitrocellulose membranes, SDS (0.01%) was included in the transfer buffer. Immunostaining was carried out using the Immun- Blot Assay Kit (BIO-RAD) according to the manufacturer’s specifications, except that Tween-20 was left out of the blocking solution. The primary antibody was undiluted hybridoma culture supernatant or ascites fluid diluted (lo-*) in ITBS (20 mM Tris (pH 7.51, 500 mM NaCl, 0.05% Tween-20, and 1% gelatin). Binding of the secondary antibody (goat anti-mouse IgG conjugated with alkaline phos- phatase) was detected calorimetrically using 5-bromo-4-chloro-3-indolyl phosphate (BCIP) as the substrate for alkaline phosphatase and nitroblue tetrazolium (NBT) as the chromogen. When proteins were 35S-labelled, the stained membranes were exposed to Kodak XARJ film for autoradiography.

When virion proteins were separated using the conditions described by Tsuru- dome et al. (1989) for the detection of conformational epitopes, mAb GS reacted with slowly migrating bands of Ad40, Ad41 and Ad5 proteins (Fig. 1). Initially, mercaptoethanol (ME) was eliminated from the sample buffer and samples were not boiled prior to electrophoresis, but were instead kept at 4°C for 1 h. As SDS is insoluble at 4”C, lithium dodecyl sulfate (LDS) was substituted for SDS in the gel, running buffer and sample buffer (Tsurudome et al., 1989). It was subsequently demonstrated that the crucial factor in preserving the reactivity of the epitope was the absence of boiling prior to electrophoresis. The presence of ME in the sample buffer did not appear to affect reactivity of the epitope since reactivity was preserved even when ME was included in the sample buffer provided that samples were not boiled. Reactivity was lost, however, when proteins were denatured by

60

I

-116250 -97400

~.. 66200

-31000

-21500

-14400

Fig. 1. Western blot of virion proteins reacted with monoclonal antibody GS. 35S-Labelled virion

proteins were separated by LDS-PAGE (12.5% a&amide) and transferred to a nitrocellulose mem-

brane for reaction with mAb GS. The serotype is indicated above each lane. The sample in the first

lane (*) was denatured by boiling for 3 min. The samples in the next three lanes were not boiled

(minimal denaturing conditions). Molecular weight markers stained with colloidal gold are indicated on

the far right. The immunostained membrane is on the right and the autoradiograph of the stained membrane showing the total protein pattern is on the left.

boiling whether or not ME was included in the sample buffer. Furthermore, the presence of SDS did not appear to compromise the reactivity of mAb GS with the slowly migrating bands (results not shown). Consequently, in subsequent experi- ments, LDS was replaced with SDS and electrophoresis was done at room temperature. Non-denaturing conditions were achieved by not boiling samples prior to electrophoresis.

61

Hexon protein was the only virion protein whose migration in SDS-PAGE was altered when samples were not boiled. Other virion proteins migrated to the same position irrespective of boiling. In contrast, the single band corresponding to denatured hexon monomer, which was seen when samples were boiled, was replaced with a group of slowly migrating bands when samples were not boiled. Although some hexon protein migrated as the denatured monomer, even when samples were not boiled (Fig. 11, this amount was variable and represented only a small proportion of the total amount of hexon protein.

As shown in Fig. 1, mAb GS reacted with a subset of these slowly migrating bands. For Ad40 and Ad41, only the lower band (of two) reacted with the mAb whereas for Ad5, only the lower two bands (of four) reacted. The identity of these reactive bands was determined using mAb 2Hx-2 which has been shown, by immunoprecipitation studies, to be specific for hexon trimers and ‘group of nine’ hexons (GON) (Cepko et al., 1981). Hexon trimers consist of three hexon polypep- tides held together by non-covalent interactions and represent the morphological subunits (hexons) of the adenovirus capsid (Roberts et al., 1986). When virions are disrupted with SDS, interactions between the capsomers of a given facet are sufficiently strong to hold these nine hexons together in a complex referred to as a ‘group of nine’ hexons (GON) (Smith et al., 1965). Fig. 2 illustrates the results of immunoblotting virion proteins from Ad40, Ad41, Ad31 and Ad5 with mAb 2Hx-2. Duplicate samples were analyzed for each serotype; one of the samples was boiled prior to electrophoresis so that the position of the hexon monomer (labelled as

5 3-l 31 si 41 do 40 5il 31 4: 41 & 40

Fig. 2. Recognition of hexon trimers and GONs by monoclonal antibody 2Hx-2. Virion proteins of 35S-labelled, gradient purified Ad40, Ad41, Ad31 and Ad5 were separated by SDS-PAGE using a concave gradient gel (4-15% acrylamide) and transferred along with biotinylated molecular weight markers to a PVDF membrane for reaction with mAb 2Hx-2. The serotype is indicated above each lane. Samples indicated by a dot (0) over the appropriate lanes were boiled for 3 min prior to electrophoresis (complete denaturation). Remaining samples were not boiled (partial denaturation). The denatured hexon monomer is labelled as protein II. The slowly migrating bands are indicated (-). The band recognized by mAb 2Hx-2 and corresponding to hexon trimers is identified by an arrow. (A)

Stained membrane. (B) Autoradiograph of the stained membrane.

62

protein II) could be identified by autoradiography of the stained membrane. Fig. 2B shows the four well-separated, slowly migrating bands of 35S-labelled Ad5 protein. In contrast, there are only two comparable bands for Ad31, Ad41 and Ad40 proteins, with the lower band being predominant. In the immunoblot shown

in Fig. 2A, mAb 2Hx-2 recognized all but the uppermost slowly migrating band of Ad5 protein. Although the four bands appeared to be of comparable intensity on the autoradiograph (Fig. 2B1, a stronger signal was generated by reaction of mAb 2Hx-2 with the two lower bands, as expected for bands representing hexon trimers and GONs. For the remaining serotypes (Ad31, Ad41 and Ad40), mAb 2Hx-2 reacted strongly with the predominant lower band whose migration was consistent with that of hexon trimers. A faint signal corresponding to the upper band of Ad31 and Ad41 hexon species was also seen. For these serotypes (including Ad40), the intensity of the calorimetric signal from the upper and lower bands (Fig. 2A) can be correlated with the amount of protein present in each band as seen in the autoradiograph (Fig. 2B). The uppermost band in the Ad31 lane is considered to be an artifact which resulted from streaking from the Ad5 lane.

Immunoblotting experiments repeated with mAb GS, following SDS-PAGE of Ad40, Ad41, Ad31 and Ad5 proteins through a concave gradient gel, gave results comparable to those obtained with mAb 2Hx-2 (Fig. 3A). As shown for mAb 2Hx-2 in Fig. 2A, mAb GS reacted strongly with the lower slowly migrating band of Ad40

r-=1 40&o 414-i 31 3-l 55

A ‘.

2HX-2

5

Fig. 3. Recognition of hexon trimers and GONs by monoclonal antibody GS. Virion proteins of ._ “S-labelled, gradient purified Ad40, Ad41, Ad31 and Ad5 were separated by SDS-PAGE using a

concave gradient gel (4-15% acrylamide) and transferred along with biotinylated molecular weight markers to a PVDF membrane for reaction with mAb GS or mAb 2Hx-2, as indicated. The serotype is

indicated above each lane. Samples indicated by a dot (0) over the appropriate lanes were boiled for 3

min prior to electrophoresis. Remaining samples were not boiled (partial denaturation). The denatured

hexon monomer is labelled as protein II. Hexon trimers and GONs (-) are indicated. (A) Stained

membrane. (B) Autoradiograph of the stained membrane.

63

and Ad41 (Fig. 3A). The amount of Ad31 protein on the membrane was likely not sufficient to generate a visible signal in this particular experiment, but other experiments have repeatedly shown a strong reaction between mAb GS and the lower slowly migrating band of Ad31 protein (results not shown). In reaction with Ad5 proteins, both mAb GS and mAb 2Hx-2 recognized the lower two of the four slowly migrating bands (Fig. 3).

Results with both mAb GS and mAb 2Hx-2 clearly show that conformational epitopes on adenovirus hexon can be maintained during SDS-PAGE provided that samples are not boiled prior to electrophoresis. Similar results have been reported by Khilko et al. (1990) for human adenoviruses from subgroups A to E as well as simian, avian and bovine adenoviruses. In that report, polyclonal antiserum, prepared against complete virions, recognized native hexon species in multiple slowly migrating bands, some of which were shown by electron microscopy to consist of hexon trimers.

MAb GS, isolated in this laboratory following immunization with Ad5 virions, and mAb 2Hx-2, isolated by Cepko et al. (1981, 1983) following immunization with extracts of HeLa cells infected with Ad2, have comparable specificities and may recognize the same epitope. Both mAbs recognize hexon trimers and GONs but not denatured hexon polypeptides (Figs. 2 and 3). MAb GS, like mAb 2Hx-2, is a group specific antibody recognizing representative serotypes of each subgroup (A-F) of human adenoviruses as well as canine adenovirus (Cepko et al., 1983; results not shown). However, mAb GS did not react with intact virions in immunogold labelling studies done in our laboratory (results not shown) suggesting that the binding site, like that of mAb 2Hx-2, is not exposed on the surface of intact virions (Cepko et al., 1981, 1983). Further studies (e.g., competitive binding assays, epitope mapping) would be necessary to determine whether the two mAbs do, in fact, recognize the same epitope.

It is apparent from Figs. 2 and 3 that the non-denatured hexon species of enteric adenoviruses (Ad40, Ad41, Ad31) have a migration pattern in gradient gels which is different from the pattern seen for the non-denatured hexon species of Ad5 (a serotype associated with respiratory disease). For Ad40, Ad41 and Ad31, hexon trimers were the predominant species, while for Ad5, GONs and higher order structures were seen in addition to trimers.

To determine whether the migration pattern of non-denatured forms of hexon reflected structural differences between enteric and non-enteric adenoviruses, migration patterns were examined for other non-enteric adenoviruses, specifically Ad2 (subgroup 0, Ad3 and Ad7 (both subgroup B). Ad41 virions purified directly from stool were included in this comparison to ensure that the difference in hexon migration pattern between enteric and non-enteric serotypes was not related to properties of enteric adenoviruses produced in cell culture (293 cells) as opposed to natural host cells. As shown in Fig. 4, each adenovirus serotype, with the exception of Ad2, displayed the same pattern of hexon migration previously observed for Ad40, Ad41 and Ad31 (Figs. 2 and 3) in which hexon trimers are predominant. Ad2, like Ad5, displayed GONs and higher order structures in addition to trimers. Based on these results and those reported by Khilko et al.

64

116.25

Fig. 4. Serotypic differences in migration of non-denatured forms of hexon. Virion proteins of gradient

purified Ad3, Ad7, Ad2 and Ad41 were separated by SDS-PAGE using a concave gradient gel (4-15%

acrylamide) and visualized by silver staining (Morrissey et al., 1981). The serotype is indicated above

each lane. Samples indicated by a dot (0) over the appropriate lanes were boiled for 3 min prior to

electrophoresis (complete denaturation). Remaining samples were not boiled (partial denaturation).

The Ad41 samples in the two lanes next to Ad2 represent virus purified directly from a stool specimen;

the last two lanes of Ad41 represent the same virus strain propagated in cells inoculated with the

specimen and subsequently purified by gradient centrifugation. The denatured hexon monomer is

labelled as protein II. The slowly migrating bands of Ad2 proteins are indicated (-1. Molecular weight

markers are indicated on the left.

(19901, it appears that the pattern of hexon migration seen for the enteric adenoviruses is representative of adenoviruses in general with the exception of human adenoviruses of subgroup C. Although the bands corresponding to non-de- natured forms of hexon were not well separated in the study reported by Khilko et al. (19901, it appears that similar patterns were observed for bovine, simian and avian adenoviruses, as well as human adenoviruses belonging to subgroups A, B, D and E. Both studies show a unique migration pattern for non-denatured hexon species of subgroup C adenoviruses (Khilko et al., 1990; Fig. 4). This unique pattern may reflect the nature of specific protein-protein interactions within the capsid of subgroup C adenoviruses. It may not be appropriate, therefore, to consider the well-characterized serotype 2 (subgroup C) (Stewart and Burnett, 1993; Stewart et al., 1993) as totally representative of all adenoviruses.

Acknowledgements

The technical assistance of Sumita Fleming and William Mallari, in preparation of the monoclonal antibody, is greatly appreciated. Development of the mono-

65

clonal antibody was supported by funds from The Hospital for Sick Children Research and Development Limited Partnership. The remaining work was sup- ported by Grants MA 8656 and MA 11333 from the Medical Research Council of Canada.

References

Brown, M. (1985) Selection of nonfastidious adenovirus species in 293 cells inoculated with stool specimens containing adenovirus 40. J. Clin. Microbial. 22, 205-209.

Brown, M., Petric, M. and Middleton, P.J. (1984) Diagnosis of fastidious enteric adenoviruses 40 and 41 in stool specimens. J. Clin. Microbial. 20, 334-338.

Cepko, C., Changelian, P.S. and Sharp, P.A. (1981) Immuno-precipitation with two-dimensional pools as a hybridoma screening technique: production and characterization of monoclonal antibodies against adenovirus 2 proteins. Virology 110, 385-401.

Cepko, C., Whetstone, C.A. and Sharp, P.A. (1983) Adenovirus hexon monoclonal antibody that is group specific and potentially useful as a diagnostic reagent. J. Clin. Microbial. 17, 360-364.

Ginsberg, H.S. (1979) Adenovirus structural proteins. In: H. Fraenkel-Conrat and R.R. Wagner (Eds.), Comprehensive Virology, pp. 409-457. Plenum Press, New York.

Graham, F.L., Smiley, J., Russell, W.C. and Nairn, R. (1977) Characteristics of a human cell line transformed by DNA from human adenovirus type 5. J. Gen. Virol. 37, 59-72.

Hames, B.D. (1990) One-dimensional polyactylamide gel electrophoresis. In: B.D. Hames and D. Rickwood (Eds.), Gel Electrophoresis of Proteins-A Practical Approach, 2nd edn., pp. 116-124. Oxford University Press, Oxford.

Harlow, E. and Lane, D. (1988) Monoclonal antibodies and growing hybridomas. In: E. Harlow and D. Lane (Eds.), Antibodies - A Laboratory Manual, pp. 139-281. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

Khilko, S.N., Kirasova, M.A., Kiseleva, E.K. and Tikchonenko, T.I. (1990) Comparisons of adenoviral hexon polypeptides (monomers) and of native hexons (trimers) by SDS-polyacrylamide gel elec- trophoresis. Acta Microbial. Hung. 37, 233-245.

Maizel, J.V. (1971) Polyacrylamide gel electrophoresis of viral proteins. Methods Virol. 5, 179-246. Morrissey, J.H. (1981) Silver stains for proteins in polyacrylamide gels: a modified procedure with

enhanced uniform sensitivity. Anal. Biochem. 117, 307-310. Pettersson, U. and Wadell, G. (1985) In: M.H.V. van Regenmortel and A.R. Neurath (Eds.), Immuno-

chemistry of Viruses, pp. 295-321. Elsevier, Amsterdam. Roberts, M.M., White, J.L., Grutter, M.O. and Burnett, R.M. (1986) Three-dimensional structure of

the adenovirus major coat protein hexon. Science 232, 1148-1151. Smith, K.O., Gehle, W.D. and Trousdale, M.D. (1965) Architecture of the adenovirus capsid. J.

Bacterial. 90, 254-261. Stewart, P.L. and Burnett, R.M. (1993) Adenovirus structure as revealed by X-ray crystallography,

electron microscopy and difference imaging. Jap. J. Appl. Physics 32, 1342-1347. Stewart, P.L., Fuller, S.D. and Burnett, R.M. (1993) Difference imaging of adenovirus: bridging the

resolution gap between X-ray crystallography and electron microscopy. EMBO J. 12, 2589-2599. Tsurudome, M., Nishio, M., Komada, Bando, H. and Ito, Y. (1989) Extensive antigenic diversity among

human parainfluenza type 2 virus isolates and immunological relationships among paramyxoviruses revealed by monoclonal antibodies. Virology 171, 38-48.