Preparation and uses of immunoabsorbent monolayers in the purification of virus proteins and separation of cells on the basis of their cell surface antigens

Journal of Immunological Methods, 60 (1983) 147-165

147

Elsevier Biomedical Press

Preparation and Uses of Immunoabsorbent Monolayers in the Purification of Virus Proteins and Separation of Cells on the Basis of their Cell Surface Antigens R.E. Randall Division of Virology, National Institute for Medical Research, Mill Hill, London NW7 1AA, U.K.

(Received 19 July 1982, accepted 11 December 1982)

A method is described for forming monolayers of Staphylococcus aureus Cowan strain A on plastic tissue culture plates (A plates). Bacteria remain bound to such plates even under strong protein denaturing conditions. The specific binding of antibody to A plates provides a rapid method for immune precipitation on a solid matrix. Antibody can be covalently crosslinked to staphylococcus monolayers by fixation with paraformaldehyde without significant loss of specific antigen binding capacity. Furthermore antigen-antibody complexes may be efficiently disrupted with 9 M urea and non-ionic detergent and both antibody and antigen renatured by removal of urea, antibody regaining the ability to bind fresh antigen. In the presence of 9 M urea and 1% Nonidet P-40 fixed antibody remains bound to plates while antigen is released into the urea solution, providing a method for the immunological purification of proteins. Plates with such fixed antibody may be used multiple times to bind antigen. The use of this method is illustrated by the purification of 2 adenovirus structural proteins and a herpes simplex virus glycoprotein, by means of specific monoclonal antibodies to these proteins crosslinked to A plates. A method is also described to enrich for functional antibody from immune precipitates by dissociating antibody-antigen complexes with 9 M urea and 1% Nonidet P-40 and isolating the antigen free antibody on Staphylococcus aureus Cowan strain A. A plates have also been used to separate suspension cells on the basis of their cell surface antigens (e.g., lymphocyte subpopulations or cells expressing virus antigens on their surface). Cells were either (1) reacted with specific antisera to cell surface determinants and panned over A plates, cells with antibody on their surface binding to the monolayer, or (2) panned over A plates to which specific antibody had been coupled. Separated cells may be probed directly for certain properties, such as their ability to support virus replication. The staphylococcus monolayer does not prevent the growth of tissue culture cells on the plastic surface and the selected cell population can be examined for cells giving rise to infectious centres by the subsequent addition of permissive cells. Key words: immunoabsorbent monolayers - - protein purification - specific antibody enrichment - - cell separation - - cell surface antigen

Introduction It has been Staphylococcus

well documented t h a t a cell w a l l c o m p o n e n t o f m o s t s t r a i n s termed protein A, has a high affinity for the Fc portion

aureus,

0022-1759/83/0000-0000/$03.00 © 1983 Elsevier Science Publishers

of of

148 certain subclasses of IgG immunoglobulins of a number of animal species (reviewed by Goding, 1978). Staphylococcus aureus Cowan strain A has a large amount of protein A on the cell surface, about 80,000 molecules per organism (Kronvall et al., 1970). As the binding of antibody to protein A does not interfere with the antigen binding capacity of the antibody, these bacteria have been used widely as solid-phase absorbents for the isolation of antigen-antibody complexes, antibody being bound to the bacteria either before or after interaction with antigen. MacSween and Eastwood (1981) reported that antibody to bovine serum albumin (BSA) could be crosslinked to the staphylococcus absorbent by treatment with 0.5% paraformaldehyde for 45 min at 37°C. This treatment did not significantly reduce the binding of BSA to the absorbent complexed with anti-BSA. The requirement in our laboratory for an efficient and economical method (1) to separate small numbers of cells on the basis of their cell surface antigens with the ability to examine separated cells directly for certain properties (e.g., presence of virus genomes) and (2) to immunoprecipitate virus antigens, has led to the development of a technique, described in this paper, that uses immunoabsorbent monolayers. The basis of the technique is to form monolayers of Staphylococcus aureus Cowan strain A, to which antigen (either in soluble form or on the surface of cells) will bind after interaction with specific antibodies. A further adaptation of this technique of making immunoabsorbent monolayers has led to a general immunological panning method for the purification of proteins.

Materials and Methods Cells and viruses

Human peripheral blood lymphocytes were isolated from heparinised blood of healthy donors by equilibrium centrifugation on Ficoll-Paque gradients (Pharmacia Fine Chemicals, Uppsala). Mouse splenocytes were prepared from B A L B / c mice by homogenising the spleens and collecting the suspension cells after lysis of red blood cells with ammonium chloride (Weir, 1978). HeLa $3 and KB suspension cells were grown in minimum essential medium with Earle's salts and 7% calf serum, at 37°C with continuous stirring. Vero cells were propagated as monolayers at 37°C in Dulbecco's modification of Eagle's medium containing 10% calf serum. Human adenovirus type 5 (Ad 5) was grown in KB cells as described by Russell et al., (1967). Herpes simplex virus type 1 strain H F E M / S T H 2 (Honess et al., 1980) and type 2 strain HG52 (Timbury, 1971) and the attenuated derivative of herpesvirus saimiri strain 11 (HVS(11Att), Schaffer et al., 1975) were also used in these studies. A n tigens

Soluble antigens were obtained from adenovirus-infected KB cells after fluorocarbon extraction and caesium chloride gradient centrifugation (Russell et al., 1967). The material above the virus band, containing the soluble antigens, was dialysed against phosphate-buffered saline and centrifuged at 36,000 rpm for 1 h before use. Monolayers of Vero cell cultures in 80 oz. Winchester bottles were infected with

t49 herpes simplex virus (HSV) or herpesvirus saimiri (HVS) at high multiplicities of infection (2-10 p.f.u, per cell). The virus was allowed to absorb to the cells for 1-2 h at 37°C, and the inoculum was then decanted and fresh culture medium added. At various times after infection (see Results) the culture medium was replaced with [35S]methionine labelling medium, consisting of Dulbecco's modification of Eagle's medium with 1/10 the normal concentration of methionine, 2 #Ci/ml L[35S]methionine (500 Ci/mmol; Amersham International) and 2% calf serum. At times after infection when 90-100% of the cells showed a c.p.e. (24 h for cells infected with HSV and 48 h for cells infected with HVS), the cells were scraped into the culture medium which was then clarified by centrifugation (3000 rpm for 15 min) to sediment cells. The cells were resuspended in phosphate-buffered saline (approximately 1 × 108 per 2 ml), disrupted with an ultrasonic probe (total cell antigen fraction), centrifuged at 36,000 rpm for 1 h in an MSE 6 x 5.5 swing out rotor to remove virus and other particulate material and the supernatant (soluble antigen fraction) collected. A ntisera General anti-HSV antiserum (a gift from D.H. Watson and R.A. Killington, University of Leeds, Leeds) was prepared by repeated immunization of rabbits with rabbit kidney (RK13) cells infected with herpes simplex virus strain HFEM (Watson et al., 1966). Rabbit antiserum to an infected cell polypeptide (precursor mol. wt. 30,000; product mol. wt. 28,000) of herpesvirus saimiri infected ceils was prepared as described by Randall et al. (1982). Rabbit serum reactive with the Fab' fragment of mouse immunoglobulins was a gift of R.M.E. Parkhouse (National Institute for Medical Research, Mill Hill, London). A 7S fraction of rabbit serum to human immunoglobulins was purchased from Nordic Immunological Laboratories (Maidenhead). Monoclonal antibodies to adenovirus hexon (Ad5Hx-4) and fibre (Ad5Fb-1) were provided as ascites fluids by W.C. Russell (for preparation and properties see Russell et al., 1981). Similarly, LP2 monoclonal antibody was provided as ascites fluid by A. Minson (University of Cambridge, Cambridge), who showed that the antibody was of subclass IgG2a and reacted with a glycoprotein of about 50,000 in both herpes simplex virus type 1 and type 2 infected cells. Purified mouse OKT3 monoclonal antibody (subclass IgG2a; 25 /~g/ml) which reacts with human peripheral blood T lymphocytes (Reinherz and Schlossman, 1980) was purchased from Ortho Diagnostic Systems (High Wycombe). Preparation of Staphylococcus aureus Cowan strain A Heat inactivated and formalin fixed preparations of Staphylococcus aureus Cowan strain A were prepared by the method of Kessler (1975) and stored as a 10% (w/v) suspension in phosphate buffered saline at - 7 0 ° C before use. Polyacrylamide gel electrophoresis Samples, including those in 9 M urea and 1% Nonidet P-40, were disrupted in 2% SDS, 5% mercaptoethanol and 0.05% Tris-HCl, pH 7.0, by heating at 80°C for 5 min. Polyacrylamide electrophoresis was carried out using a discontinuous buffer

150 system modified by the inclusion of SDS and the polyacrylamide slab gels were crosslinked with N,N'-diallyltartardiamide (Heine et al., 1974). The gels were stained with Coomassie brilliant blue, destained in acetic acid and methanol and dried. Analysis of radioactive proteins was undertaken by autoradiography with Fuji X-ray film.

Results

Preparation and properties of monolayers of Staphylococcus aureus Cowan strain A Preliminary observations showed that Staphylococcus aureus Cowan strain A bound stably to certain plastic surfaces. Culture plates were flooded with a 1% (w/v) suspension of bacteria in phosphate-buffered saline and left for a minimum of 4 h at 4°C. Unbound bacteria were removed by washing the plates with buffer, leaving a confluent monolayer of bacteria bound to the plastic surfaces (termed A plates). Studies on the binding of staphylococcus to different plastic culture plates showed that certain types of plastic bound bacteria more strongly than others. For example, bacteria bound stably to 25 cm 2 and 75 cm 2 Nunc Delta (tissue culture) flasks (cat. no. 1-63371 and 1-53732), but less stably to either Nunc or Flow, 24-well tissue culture dishes (cat. no. 143982 and 76-033-05 respectively). In situations where binding of bacteria to plastic was weak, vigorous washing resulted in the continuous removal of bacteria from these surfaces. However, the binding of staphylococcus to all plastic surfaces could be made stable by drying the monolayers before use. Monolayers were washed with distilled H 2 0 , drained and dried under vacuum without freezing. This procedure did not result in a significant decrease in the capacity of the staphylococci to bind immunoglobulins. Once the bacteria were stably bound to plastic surfaces they could not be removed by vigorous washing or by a variety of protein denaturing solutions (e.g., 2% SDS, 5% mercaptoethanol for 5 min at 100°C; 9 M urea and detergent; 4 M NaC1 or 3.5 M MgC12). When serum or ascites fluids were incubated with A plates immunoglobulins were specifically absorbed by virtue of the protein A on the staphylococcus; plates with antibody absorbed have been termed Ab plates. Fig. 1 demonstrates the specific binding of immunoglobulins from rabbit, human and common marmoset monkey (Callithrixjacchus) sera to A plates (compare slots 1-3 with slots 6-4), in comparison to serum proteins which bind non-specifically to untreated culture plates (slot 7). The amount of rabbit immunoglobulin binding to a 75 cm 2 A plate was estimated to be between 60-100 ttg. The binding of immunoglobulin to A plates occurred rapidly; 25 cm 2 plates were saturated with immunoglobulins by incubating the plate for 15-30 min at 37°C with 1 ml of rabbit serum, diluted 1/25 in phosphate-buffered saline. As illustrated below, monolayers of Staphylococcus aureus may be used in a wide variety of immunological procedures including the purification of proteins and separation of cells on the basis of their cell surface antigens. For a number of these purposes, e.g., purification of proteins, methods had to be established by which antibody could be covalently linked to the monolayers without loss of specific

151

QI

Fig. 1. Analysis of total polypeptides (Coomassie brilliant blue-stained gel film), separated by electrophoresis through a 10% polyacrylamide slab gel from rabbit serum (slot 1), human serum (slot 2) and common marmoset serum (slot 3). These sera were diluted in phosphate-buffered saline (1 in 20), incubated with A plates for 30 min at 37°C and the proteins which bound were eluted with disruption buffer (slot 4, common marmoset Ab plate; slot 5, human Ab plate; slot 6, rabbit Ab plate). Rabbit serum proteins that bound non-specifically to plastic culture plates are shown in slot 7.

antigen binding capacity, and a n t i g e n - a n t i b o d y complexes dissociated without loss of the antigenic integrity of the proteins. In these studies antibodies to herpes simplex virus proteins were crosslinked to the bacterial monolayer by a n u m b e r of methods and the antigen binding capacity of these anti-HSV-Ab-crosslinked plates was examined. In agreement with the results of MacSween and Eastwood (1981), optimal crosslinking of antibody to bacteria, without significant loss in the antigen binding capacity of the antibody, was obtained with 0.5% paraformaldehyde for 45 min at 37°C. This procedure prevented the removal of antibody from the plates by 2% SDS, 5% mercaptoethanol at 100°C for 2 min, with less than a 20% decrease in the antigen binding capacity of the crosslinked antibody (Table I). Concentrations of glutaraldehyde which crosslinked antibody to bacteria significantly reduced the capacity of antibody to bind antigen (data not shown). Once optimal conditions for crosslinking antibody to bacteria had been achieved, various denaturing solutions were examined for their ability to disrupt immune complexes (Table I). The results clearly showed that in the absence of detergent, 6 M or 9 M urea did not dissociate a n t i g e n - a n t i b o d y complexes but in the presence of ionic (SDS) or non-ionic

152 TABLE I A N T I G E N B I N D I N G CAPACITY O F C R O S S L I N K E D A N T I B O D Y Normal rabbit Ab plates or anti-HSV-Ab plates were prepared by rocking a 1/20 dilution of the respective serum over 25 cm 2 A plates for 30 min at 37°C. The absorbed antibodies were or were not 'fixed' to the bacteria by irrigating the plates with 0.5% paraformaldehyde in phosphate-buffered saline for 45 min at 37°C and washed with 0.1 M glycine for 30 min at 37°C. A soluble antigen preparation of herpes simplex virus infected Vero cells, labelled with [35S]methionine from 5 h to 24 h after infection, was incubated with the plates for 1 h at 4°C, the plates were then thoroughly washed with phosphatebuffered saline and 1% Nonidet P-40 and immune complexes dissociated with various denaturing solutions. The efficiency with which the immune complexes were disrupted or released from the plates was estimated by monitoring the amount of radioactivity released into the respective solutions. It was assumed that disruption buffer (1 ml of 2% SDS, 5% mercaptoethanol, in 0.05 M Tris-HCl, pH 7.0, for 5 min at 100°C) would release all HSV polypeptides from a plate. Other plates were treated with 1 ml of the denaturing solution for 20 rain at 37°C. The results are presented as the amount of radioactivity released by the denaturing solutions expressed as a percentage of the amount of radioactivity released by disruption buffer. The figures in brackets indicate the actual number of counts per minute present in 100 /zl of disruption buffer as estimated by counting in a Beckman LS7000 scintillation counter. Antigens removed by

% of bound antigen removed

Disruption buffer (7038) Disruption buffer (36290) 9 M urea 9 M u r e a + 1% NP40

100% 100% 21% 98%

Disruption buffer (2632) Disruption buffer (30008) 6 M urea 6 M urea+0.1% NP40 6 M u r e a + 1.0% NP40 9 M urea 9 M urea+0.1% NP40 9 M u r e a + 1% NP40 9 M urea + 0.1% SDS 9 M urea + 1.0% SDS 9 M urea + 5% mercaptoethanol 0.1 M HCl-glycine, p H 1.0 0.1 M HCl-glycine, pH 2.0 0.1 M HCl-glycine, pH 2.4 0.1 M HCl-glycine, pH 2.8 0.1 M HCl-glycine, pH 3.2 3.5 M MgCI 2 4 M NaCI 0.1 M N a O H 1.0 M N a O H

100% 100% 4% 54% 57% 10% 48% 82% 41% 51% 15% 6% 10% 8% 6% 3% 2.8% 2% 33% 42%

A b plates not crosslinked with para formaldehyde

Normal rabbit Ab plate Anti-HSV-Ab plate Anti-HSV-Ab plate Anti-HSV-Ab plate A b plates crosslinked with para formaldehyde

Normal rabbit Ab plate Anti-HSV-Ab plate Anti-HSV-Ab plate Anti-HSV-Ab plate Anti-HSV-Ab plate Anti-HSV-Ab plate Anti-HSV-Ab plate Anti-HSV-Ab plate Anti-HSV-Ab plate Anti-HSV-Ab plate Anti-HSV-Ab plate Anti-HSV-Ab plate Anti-HSV-Ab plate Anti-HSV-Ab plate Anti-HSV-Ab plate Anti-HSV-Ab plate Anti-HSV-Ab plate Anti°HSV-Ab plate Anti-HSV-Ab plate Anti-HSV-Ab plate

153

(Nonidet P-40) detergent these agents were very efficient at disrupting immune complexes. Low pH buffers (0.1 M HCl-glycine, pH 1.0-3.2) and high molarity salts (4 M NaC1 and 3.5 M MgC12) were very inefficient in removing antigen from these plates. Although antigen could be eluted from the plates with sodium hydroxide, this procedure weakened the interaction of bacteria with plastic surfaces and was likely to irreversibly denature proteins. HSV antigens that were eluted with 9 M urea and 1% Nonidet P-40 could be renatured by removal of the urea by dialysis against phosphate-buffered saline. At least 70% of the antigen treated in this manner could be specifically rebound to anti-HSV-Ab plates. Furthermore, antibody that had been crosslinked to such plates regained the ability to bind fresh antigen by being renatured in phosphate-buffered saline (see below). Antibody on anti-HSV-Ab plates that had been dried under vacuum could also be renatured in phosphatebuffered saline; these plates regaining their original capacity to bind antigen.

Uses of monolayers of staphylococcus in various immunological procedures (I) Purification of proteins. The use of monoclonal antibodies to purify virus I - para.

I

+ para.

I

hexonpenton
fiber



3 4 5 6

7 8

9 10 11 12

Fig. 2. Polyacrylamide gel analysis of the total protein (Coomassie brilliant blue-stained gel film) in a soluble antigen preparation of adenovirus type 5 infected KB cells (slot 1). The soluble antigen preparation (1 ml) was 'panned', for 1 h at 4°C, over A plates to which monoclonal antibodies (Ad5Hx-4, slots 2 and 5; LP2, slot 3 and 6; Ad5Fb-1, slot 4, 7, 8 and 9) had or had not been irreversibly fixed with paraformaldehyde ( + or - p a r a ) . U n b o u n d proteins were removed by washing the plates with phosphate-buffered saline and immune complexes dissociated with 9 M urea and 1% Nonidet P-40. Prior to this experiment 1 plate had been previously treated once (slot 8) and 1 plate 6 times (slot 9) with 9 M urea and 1% Nonidet P-40. In between each treatment with 9 M urea and 1% Nonidet P-40 the plates were washed with phosphate-buffered saline and left for a m i n i m u m of 6 h with buffer at 4°C. Eluted proteins were separated by electrophoresis through a 10% polyacrylamide gel. Monoclonal antibodies were bound to the A plates by diluting the respective ascites fluids (LP2, slot 10; Ad5Fb-l, slot 11 and Ad5Hx-4, slot 12) 1 in 20 in phosphate-buffered saline and ' p a n n i n g ' the fluid over the plates for 30 min at 37°C. Antibody was fixed to the plates with paraformaldehyde as previously described.

154

a) +

para. - +

+ [Mot.wt. 3 xlO"

b)

para. 1-

+

-

+

+l

5

6 7

155

19H

48

26

IgL 1

2

3

4

5

6

7

1 7

34

Fig. 3. Analysis of (a) total polypeptides (Coomassie brilliant blue-stained gel film) and (b) the

[ 33S]methionine.labelled polypeptides (autoradiogram) separated by eleetrophoresis through a 10% polyacrylamide slab gel from a lysate of herpes simplex virus type 2 infected Vero cells (slot 1) and a soluble antigen preparation of these cells (slot 2). The soluble antigen preparation was 'panned' for 1 h at 4°C, over A plates to which monoelonal antibodies (LP2, slots 3, 4; Ad5Fb-1, slot 7) had or had not been fixed with paraformaldehyde (+ or - para). Unbound proteins were removed by washing the plates with phosphate-buffered saline and immune complexes dissociated with 9 M urea and 1% Nonidet-P40. The LP2 Ab plates were washed with phosphate-buffered saline, left for 6 h at 4°C, and used to bind fresh antigen (slots 5 and 6 respectively),

polypeptides was d e m o n s t r a t e d by using 2 m o n o c l o n a l antibodies (Ad5Hx-4 a n d A d 5 F b - 1 ) with different specificities to adenovirus structural proteins a n d one m o n o c l o n a l a n t i b o d y (LP2) to a glycoprotein of herpes simplex virus. A n t i b o d i e s were absorbed to A plates and were or were not fixed to the plates by treatment with paraformaldehyde. A soluble antigen p r e p a r a t i o n derived from KB cells infected with adenovirus type 5 was p a n n e d over these plates for 1 h at 4°C; the u n b o u n d material was then removed easily and rapidly b y washing the surface of the plate with phosphate-buffered saline. B o u n d antigen was removed from the plates with 9 M urea a n d 1% N o n i d e t P-40 a n d samples electrophoresed through a polyacrylamide gel (Fig. 2). Results indicate that A d 5 F b - I a n t i b o d y b o u n d both the p e n t o n base and fibre polypeptides, the p e n t o n base a n d fibre occurring as a complex in the

155

o) toter proteins

Mot.w~ x 10-3

b) autorad iocjram

150 110 85 albumin IgH actir 32 28 A 1

2

3

L

5

1

2

3

4

5

Fig. 4. Purification and properties of specific antibody molecules from polyclonal rabbit'antiserum to a herpesvirus saimiri polypeptide (product mol. wt. 28,000 (28 K); precursor mol. wt. 30 K) analysed by electrophoresis through a 10% polyacrylamide slab gel. Anti-28-30 K antiserum (1 ml) was reacted for 18 h at 4°C with a soluble antigen preparation of herpesvirus saimiri infected cells (not radioactively labelled). The immune precipitate that formed was sedimented by centrifugation (3000 rpm for 15 min), washed with cold phosphate-buffered saline and resedimented. The immune complexes were disrupted by sonication in 9 M urea and 1% Nonidet P-40 (2 ml) and left for 30 min at 37°C. This material was then added to an excess of 'fixed' Staphylococcus aureus Cowan strain A resuspended in a large volume of phosphate-buffered saline (20 ml of a 1% w / v suspension) and left at room temperature for 10 min. The bacteria were then pelleted by centrifugation (3000 rpm for 10 min), washed once with phosphate-buffered saline, and resedimented. The bacteria were resuspended in 2 ml of 9 M urea and 1% Nonidet P-40 for 20 min at 37°C to dissociate antibody from the bacteria which were then removed by centrifugation (3000 rpm for 10 min). The supernatant (slot A), in the presence of 1 mg/ml of carrier bovine serum albumin, was dialysed against decreasing concentrations of urea and finally against phosphate-buffered saline. Antibody in this preparation was bound to an A plate and the capacity of this plate (slot 5) to bind herpesvirus saimiri polypeptides was compared to A plates that had bound antibody from the original serum (slots 4) or from normal rabbit serum (slots 3). A soluble antigen fraction (slots 2) of Vero cells that had been infected with HVS (slots 1) and radioactively labelled with [35S]methionine from 24 h to 48 h after infection, was 'panned' over these plates (4°C for 1 h). Unbound material was removed by washing the plates with buffer and the immune complexes dissociated in 9 M urea and 1% Nonidet P-40 (20 min at 37°C).

s o l u b l e a n t i g e n p r e p a r a t i o n ( R u s s e l l et al., 1981). T h e A d 5 H x - 4 a n t i b o d y s p e c i f i c a l l y s e l e c t e d t h e h e x o n p o l y p e p t i d e , w h i l e L P 2 a n t i b o d i e s f a i l e d to p r e c i p i t a t e a n y o f t h e polypeptides present in the adenovirus soluble antigen preparation. Crosslinking of these monoclonal antibodies to the bacteria with paraformaldehyde did not alter t h e i r a b i l i t y t o b i n d a n t i g e n (Fig. 2, c o m p a r e s l o t s 1 - 3 w i t h s l o t s 4 - 6 ) a n d as b o u n d a n t i g e n c o u l d b e d i s s o c i a t e d f r o m t h e s e p l a t e s b y 9 M u r e a a n d 1% N o n i d e t P - 4 0

156

without the removal of antibodies, virus proteins could be purified by this method. These results also show that antibodies which remain bound to these plates after treatment with 9 M urea and 1% Nonidet P-40 could be renatured by washing the plates with phosphate-buffered saline, regaining their original capacity to bind fresh antigen (Fig. 2, slot 8). This process could be repeated multiple times, for example a Ad5Fb-1 Ab-crosslinked plate has been used 6 times to precipitate adenovirus fibre and penton base polypeptides (Fig. 2, slot 9). The amount of fibre bound by a 25 cm 2 Ad5Fb-l-Ab plate was 20-30 #g, similar to the amount of antibody bound (Fig. 2, slot 4) whereas 5-10 /~g of hexon polypeptide bound to 25 cm 2 Ad5Hx-4-Ab plate (Fig. 2, slot 2). Less Ad5Hx-4 antibody bound to A plates than Ad5Fb-1 and LP2 antibodies (Fig. 4, compare IgH slots 2-4). It seemed likely that this was because Ad5Hx-4 antibodies were of mouse subclass IgG1 while clone LP2 antibodies were of subclass IgG2a and that antibodies of subclass IgG1 bind less efficiently to protein A than antibodies of subclass IgG2a.

Scheme for separation of mouse splenocytes as shown in Fig. 5. Mouse splenocytes

f

A plates

f

Non-specifically bound cells

(45 rain, 4°C)

Unbound cells (5 X 107 cells)"j J + anti-mous~ Fab' serum (15 ~1, 30 min, 4°C)

~(5

X 107 ceils) + normal rabbit serum (15 ~1, 30 rain, 4°C)

I (PBS wash) A plate

(PBS wash) A ~late

bound(Fig. 5A)Cells (PBS wash) _ (45 rain, 4°C)

(PBS wash)

(45 min, 4°C)

bound ceils (panel B) Unbound cells

Unbound cells I

normal rabbit Ab plate (45 min, 4°C)

anti-mouse Fab' Ab plate (45 rain, 4°C)

F

normal rabbit Ab plate (45 min, 4°C)

7

anti-mouse Fab ~Ab plate (45 min, 4°C)

i

(PBS wash) Bound ceils (Fig. 5C)

l (PBS wash) I Bound cells (Fig. 5E)

(PBS wash) Bound cells (Fig. 51))

(PBS wash) Bound ceils (Fig. 5F)

157

LP2 monoclonal antibody specifically precipitated a virus polypeptide, with an estimated mol. wt. of 48,000 and a minor polypeptide, mol. wt. 26,000, from a soluble antigen preparation of Vero cells infected with herpes simplex virus type 2. Crosslinking of the antibody to bacteria, with paraformaldehyde did not alter the ability of LP2 antibody to precipitate these 2 virus polypeptides. Such plates with fixed antibody can be used multiple times to precipitate antigen (Fig. 3). Small amounts of certain polypeptides bound non-specifically to these plates, including the major capsid protein of herpes simplex virus (155 K). This non-specific binding

Fig. 5. Binding of mouse spleen 'B lymphocytes' to monolayers of staphylococcus either by (1) reacting splenocytes with rabbit anti-mouse Fab' serum followed by panning over A plates, or (2) panning splenocytes over anti-mouse Fab' Ab plates. The percentages of cells which did not bind to A plates are shown. See opposite page for experimental scheme.

158 could be reduced by preabsorption of the soluble antigen fraction with excess fixed Staphylococcus aureus Cowan strain A. In addition plates could be incubated with soluble antigen fractions to absorb out non-specific sites of protein binding on the bacteria. Antibody would then be bound to these plates and paraformaldehyde used to fix both antibody and non-specifically bound proteins. (II) Enrichment of specific antibodies in a polyclonal serum. In polyclonal antiserum the majority of antibody molecules will not be directed against the protein(s) to which the antiserum was raised. Consequently when purifying proteins with polyclonal antiserum the relatively small amount of specific antibody may be a disadvantage. However, results presented in Fig. 4 demonstrate how to enrich for specific antibody to a particular protein in a polyclonal antiserum. A rabbit antiserum raised against a herpesvirus saimiri polypeptide, with a tool. wt. of 28,000 (28 K; precursor polypeptide 30 K), was used to immunoprecipitate these polypeptides from a soluble antigen preparation of herpesvirus saimiri infected cells. The immunoprecipitate formed was sedimented by centrifugation, washed and the ira-

Fig. 6. Panel A demonstrates the binding of Staphylococcus aureus to plastic culture dishes, in the form of 0.5 cm diameter discs. Human peripheral blood lymphocyteswere reacted with normal rabbit serum, Ortho OKT3 monoclonalantibody and rabbit anti-human serum before being 'panned' over these discs and cells which bound are demonstrated in panels B, C and D respectively (for experimental details see text).

159

mune complexes dissociated by sonication in 9 M urea and 1% Nonidet-P40. Antibody was purified from antigen by diluting the dissociated complexes in a large volume of excess Staphylococcus aureus Cowan strain A, conditions which allowed antibody to bind to bacteria but prevented reassociation of antigen antibody complexes. The bacteria were sedimented and antibody removed by 9 M urea and 1% Nonidet P-40. Antibody in this preparation was bound to an A plate and the capacity of this plate to precipitate antigen was compared to A plates that had antibody bound from the original antiserum or from normal rabbit serum (Fig. 4). These results clearly show that while a similar amount of antibody was bound to all 3 plates, the plate to which purified anti-30 K antibody had been bound precipitated approximately 5 times more antigen than a plate with antibody bound from the original anti 30 K antiserum. This method of purifying antibody from immune complexes was also used to enrich for antibodies to a specific subset of herpes simplex virus proteins (glycopro-

TABLE 1I B I N D I N G OF HERPES SIMPLEX VIRUS (HSV) INFECTED HeLa (SUSPENSION) CELLS TO IMMUNOABSORBENT MONOLAYERS AFTER THEIR REACTION WITH ANTISERUM RAISED AGAINST HSV ANTIGENS 2 × 107 HeLa suspension cells were infected with herpes simplex virus, by shaking the cells with virus, for 30 min at 37°C, at a multiplicity of 500 p.f.u./cell or 50 p.f.u./cell and either in the presence or absence of 50/Lg/ml of cycloheximide. The cells were then sedimented by centrifugation (1000 rpm for 5 min), washed twice with phosphate-buffered saline and resuspended in 10 ml of culture medium, with or without 50 p,g/ml of cycloheximide, and incubated at 37°C. At various times after infection 2 × 106 cells were pelleted by centrifugation (1000 rpm for 5 min), resuspended in 0.1 ml of phosphate-buffered saline and reacted with 60/tl of anti HSV antiserum on ice for 20 min. The cells were pelleted, washed once with phosphate-buffered saline, and allowed to settle on a 25 cm 2 A plate for 30 min at 4°C. The plates were examined under a microscope and the number of cells in 10 random fields counted. Unbound cells were then removed by carefully washing the immunoabsorbent monolayer 3 times with phosphate-buffered saline. The number of cells remaining bound to the plate was estimated as described above and the results are expressed as a percentage of the total number of cells prior to washing. Time after infection

Mock infected a/HSV

HSV infected 500 p.f.u, cell

50 p.f.u./cell

(h) Pre-I c

a/HSV b

+ 50 ktg cycloheximide a/HSV

Pre-I

a/HSV

+50 vg cycloheximide a/HSV

19% 6% 20% 55% 99% a

18%

<1%

12% a

<1%

28% 22% 18% 66% 90%

29% 25% 30% 20% 14%

1

< 1%

< 1%

3 6

-

-

11

-

23

-

-

< 1% a

a Cells binding to 'A plates' shown in Fig. 7. b a / H S V , anti-HSV antiserum. c Pre-l, pre-immune serum.

160

teins) from polyprecipitin anti-HSV antiserum. Previously it had been shown that only a few soluble virus proteins were released into the culture medium of baby hamster kidney cells infected with HSV. These proteins were glycoproteins and contained all the antigenic sites involved in the neutralization of virus infectivity (Randall et al., 1980). Antibodies to HSV glycoproteins were purified as above from the immunoprecipitate formed when polyprecipitin anti-HSV antiserum was reacted with the culture medium of infected cells (data not shown).

(III) Separation of suspension cells by virtue of their cell surface markers. (a) Lymphocyte subpopulations. Lymphocyte subpopulations can be separated on monolayers of staphylococcus using specific antisera to cell surface antigens either by (1) reacting the lymphocytes with antisera, washing the cells and then panning over A plates, cells with immunoglobulin on their surface will bind to the plate; or (2) by absorbing specific antibody to an A plate over which lymphocytes are then panned. Fig. 5 demonstrates the use of both these methods in separating mouse spleen B lymphocytes from other splenocytes. Mouse splenocytes were panned over an A plate to remove the small percentage (approximately 5%) of cells which bind to the staphylococcus in the absence of added antibody. The unbound lymphocytes were reacted with either normal rabbit serum or with rabbit anti-mouse Fab'

~

~...... i~ '~Hii'i~ ¸~i~ ~i~!i!~ ~ i ~

Fig. 7. H e L a cells w e r e i n f e c t e d with h e r p e s simplex virus at 50 p . f . u . / c e l l in the p r e s e n c e o r a b s e n c e of c y c l o h e x i m i d e ( + or - c), 23 h a f t e r i n f e c t i o n the cells w e r e r e a c t e d w i t h either n o r m a l r a b b i t s e r u m ( p a n e l A) o r r a b b i t a n t i - H S V a n t i s e r u m ( p a n e l s B a n d C) a n d p a n n e d o v e r A p l a t e s ( f o r e x p e r i m e n t a l details see legend to T a b l e II).

161

Fig. 8.5 × 105 Vero cells were plated on a 25 c m 2 A plate and left to settle and grow for 18 h at 37°C. The plates were then fixed with 10% formolsaline and stained with 0.1% crystal violet. The staphylococciform the granular monolayeron which the darkly staining Vero cells have plated out, removingthe surrounding bacteria and forming small plaques in the bacterial monolayer. antiserum, washed and then panned over A plates. Approximately 60% of the splenocytes which had been reacted with anti-mouse Fab' antiserum bound to A plates while less than 1% of the splenocytes bound to A plates after reaction with normal rabbit serum. The unbound cells from these plates were collected, divided into equal portions and panned over either anti-mouse Fab' Ab plates or normal rabbit Ab plates. Less than 1% of unbound cells from the original reaction with anti-mouse Fab' serum adhered to either of these Ab plates whereas unbound cells from the reaction with normal rabbit serum bound to anti-mouse Fab' Ab plates. Immunofluorescence studies confirmed that about 99% of lymphocytes with immunoglobulins on their surface adhered to A plates after reaction with rabbit anti-mouse Fab' serum (R.M.E. Parkhouse, personal communication). Most immunological methods to separate lymphocyte subpopulations require fairly large numbers of cells to work with. Fluorescence activated cell sorting requires moderate cell numbers, but is expensive and time-consuming and most existing equipment cannot safely be applied to sorting cultures containing hazardous micro-organisms. However, by using staphylococcus bound to small areas of plastic, termed A discs, small numbers of cells can be very rapidly separated from large numbers of samples under conditions of high containment if required. This method has been illustrated by the separation of human peripheral blood B and T lymphocytes. A discs were prepared by placing 15 /~1 drops of a 2% (w/v) suspension of Staphylococcus aureus on Nunc 100 mm tissue culture petri dishes (cat. no. N1415), which were then left for 12 h at 4°C; plates were inverted for 5 min before unbound bacteria were removed by washing and the discs were dried under vacuum (Fig. 6A).

162

Human peripheral blood lymphocytes (1 × 106 in 100 /~1 of phosphate-buffered saline) were reacted for 20 min on ice, with normal rabbit serum (10 ~tl), rabbit anti-human immunoglobulin (10/~1), or with a monoclonal antibody (Ortho OKT3, 10 t~l of a 1/25 dilution i.e. 1 /~g/ml) that reacts with T cells but not other lymphocytes. Lymphocytes were then washed with phosphate-buffered saline, pelleted by centrifugation, and resuspended in 50/zl of buffer. Fifteen microlitres of the cell suspensions (3 × l0 s cells) were added to duplicate discs and the cells allowed to settle for 45 min at 4°C. Unbound cells were removed by carefully washing the discs with phosphate-buffered saline. The ratio of cells bound to the A discs after treatment with normal rabbit serum, rabbit anti-human immunoglobulin serum or with Ortho OKT3 monoclonal antibody was 1 : 2 : 6.7 (Fig. 6). In a similar experiment we have examined the ability of Ortho OKT3 antibody to bind human and common marmoset monkey peripheral blood lymphocytes to A discs and could not show any reaction of the OKT3 antibody with marmoset lymphocytes by this method. (b) Separation of cells expressing novel (virus) proteins on their surface. Suspension HeLa cells were infected with herpes simplex virus, either in the presence or absence of cycloheximide, a drug which prevents protein synthesis. At various times after infection cells were reacted with either normal rabbit serum or with anti-HSV antiserum, washed and panned over A plates. Unbound cells were removed by carefully washing the plate with phosphate-buffered saline and the percentage of cells which remained bound calculated (Table II and Fig. 7). A proportion of cells (approximately 20%) bound to A plates immediately after their infection with virus either in the presence or absence of cycloheximide. The proportion of cells binding to these plates increased at later times after infection in the absence of cycloheximide but not in the presence of cycloheximide (Table II). Thus these results show that infected cells could be bound to A plates via virus antigen on their cell surface, acquired either from the inoculum used to infect the cells or by de novo synthesis of virus antigens after infection. Once bound to A plates cells can be assayed directly for a number of properties, for example their ability to support virus replication. Tissue culture cells grow readily to form a monolayer on A plates thus allowing an infectious centre assay to be performed on bound cells. Tissue culture cells remove the bacteria from the plates, forming plaques in the staphylococcus monolayer when seeded at low densities (Fig. 8).

Discussion

A method is described for making monolayers of Staphylococcus aureus Cowan strain A and it is shown that such monolayers can be used in a wide variety of immunological procedures, including the purification of proteins and the separation of cells on the basis of their cell surface antigens. Also described is a method by which functionally active specific antibody can be purified from immunoprecipitates. These methods rely on 3 main findings: (1) that Staphylococcus aureus Cowan

163

strain A remains stably bound to certain plastic surfaces in the presence of strong denaturing solutions, (2) antibody can be covalently crosslinked to bacteria with paraformaldehyde without significant loss of the antigen binding capacity of the antibody and (3) immune complexes can be dissociated with either 6 M or 9 M urea in the presence of non-ionic detergent and that antibody and antigen can be renatured by the removal of the urea. The use of A plates for immunoprecipitation studies has a number of advantages over the conventional method in which suspension Staphylococcus aureus are used as solid-phase absorbents. The latter method is time consuming, requiring bacteria-antibody-antigen complexes to be pelleted multiple times, by centrifugation, during the washing of the complex to remove unbound antigen. Immune complexes bound to A plates can be washed rapidly and easily with a large excess of buffer. Also antigen can be bound from dilute solutions by continuously irrigating the plate with antigen. The ability to precipitate antigen onto a solid matrix may have a number of further applications. For example, in monitoring water for micro-organisms specific antibody may be bound to a staphylococcus monolayer, over which the water is then passed, organisms that bind being detected by probing the plates with a second radioactively labelled antibody preparation in a radioimmune assay. The use of A discs in radioimmune assays also seems a further obvious application of the general method. For example, to detect hepatitis B virus surface antigen (HBsAg) in the serum of patients, specific antibody to HBsAg could be bound to A discs, before incubation with patients' serum. Any HBsAg in patients' serum would bind to the discs and could be detected by addition of radioactive antibody to HBsAg. In any situation in which a second labelled antibody was used to probe for bound antigen prior to the addition of that antibody all the protein A binding sites would have to be occupied, or the antibody or antibody fragment must not bind to protein A. The ability to dry Ab plates while retaining the antigen binding capacity of the antibody means that these plates may be made prior to immunoprecipitation. Although our observations have shown that such plates may be stored for a number of days without loss of the antigen binding capacity no studies on longer term storage have been performed. The results presented here which show that immune complexes can be dissociated by treatment with either 6 M or 9 M urea in the presence of detergent and that proteins may be renatured by removal of the urea is an important finding. Thus for example, specific antibody purified from immune complexes, can be crosslinked to an A plate with paraformaldehyde and reused multiple times to bind fresh antigen. Although the amount of protein that can be purified from an Ab plate (60-100 /~g/75 cm z plate) may be insufficient for certain purposes, the ability to covalently crosslink functional antibody to a solid matrix and to recover renatured antibody from antigen-antibody complexes is obviously of importance when larger scale preparations of proteins are required, e.g., by antibody affinity column chromatography. Suspension cells can be bound to A plates after their interaction with antibody to cell surface antigens. The method is extremely economical in reagents, especially when cells are reacted with antibody before being panned over A plates; these plates

164 also have a high capacity to bind cells. In agreement with the results of Ghetie and SjOquist (1975) (who also bound mouse spleen B lymphocytes to monolayers of Staphylococcus aureus) a small percentage of lymphocytes bind to A plates without reacting with antibody. Their results showed that cells which adhere directly to bacterial monolayers would appear to be equal for macrophages and lymphocytes. However, it has been shown that a proportion of B lymphocytes expressing IgG on their surface interact with protein A (Ghetie et al., 1974) and thus a proportion of those cells that bound directly to A plates would have been B lymphocytes. Observations on the direct binding of common marmoset splenocytes to A plates have shown that the number of cells which bind to the plates can be inhibited much more efficiently with purified marmoset immunoglobulins than with the same concentration of bovine serum albumin (data not shown). Presumably the direct binding of the majority of cells in this situation at least, is mediated through antibody on the cell surface. Also the relatively large percentage of human peripheral blood lymphocytes (ca. 11%) that bind directly to A plates may in large be made up of B cells expressing IgG on their surface. The amount of direct binding to A plates can be decreased by coupling antibody to the plates, in amounts that saturate the protein A antibody binding sites, followed by panning of lymphocytes over the plates, rather than by reacting the lymphocytes with antisera prior to panning over A plates. In addition, prior to specific cell separation those cells that bind directly to monolayers of staphylococcus may be removed by incubation on A plates. The binding of a proportion of cells directly to staphylococcus monolayers was not a property of all suspension cells. Less than 0.1% of RPMI 1788 cells (a human B cell line that secretes IgM lambda chains, isolated by Huang and Moore (1969)) and HeLa suspension cells bound directly to A plates. However 6% of B95-8 cells (a cotton top marmoset B cell line, Miller and Lipman (1973)) bound directly to A plates, possibly through IgG expression on these cells. HeLa suspension cells can be bound to A plates after infection with herpes simplex virus and reaction with antibody to herpes simplex virus infected cell antigens. These cells, while remaining bound to the plates, may be examined directly for their ability to support virus replication by an infectious centre assay. Cells bound to monolayers may also be probed directly for other properties. For example, we are currently developing methods to directly examine separated lymphocyte subpopulations, from common marmoset monkeys that have been infected with herpesvirus saimiri, for the presence of virus antigens by immunofluorescence techniques. In addition, the binding of cells to A plates via specific antibody could provide a method for isolating specific cell surface receptors. Antibody to a particular cell surface determinant would be crosslinked with paraformaldehyde to A plates and reactive cells that bound to this monolayer would then be lysed with detergent, leaving the receptor complex attached to the monolayer.

Acknowledgements I am indebted to the Cancer Research Campaign for fellowship support and would like to thank Drs. R.A. Killington, A. Minson, R.M.E. Parkhouse, W.C.

165

Russell and D.H. Watson for gifts of antisera. The technical assistance of Mrs. C. Newman and the secretarial help of Mrs. V. Rogers are gratefully acknowledged.

References Ghe~ie, V. and J. Sjrquist, 1975, J. Immunol. 115, 659. Ghe~ie, V., H.A. Fabricius, K. Nilsson and J. Sjrquist, 1974, Immunology 26, 1081. Goding, J.W., 1978, J. Immunol. Methods 20, 241. Heine, J.W., R.W. Honess, E. Cassai and B. Roizman, 1974, J. Virol. 14, 640. Honess, R.W., A. Buchan, I.W. Halliburton and D.H. Watson, 1980, J. Virol. 34, 716. Huang, C.C. and G. Moore, 1969, J. Nat. Cancer Inst. 43, 1119. Kessler, S.W., 1975, J. Immunol. 115, 1617. Kronvall, G., P.G. Quie and R.C. Williams Jr., 1970, J. Immunol. 104, 273. MacSween, J.M. and S.L. Eastwood, 1981, Methods Enzymol. 73, 459. Miller, G. and M. Lipman, 1973, J. Exp. Med. 138, 1398. Randall, R.E., R.A. Killington and D.H. Watson, 1980, J. Gen Virol. 48, 297. Randall, R.E., R.W. Honess and P. O'Hare 1982, J. Gen. Virol. 64, 19. Reinherz, E.L. and S.F. Schlossman, 1980, Cell 19, 821. Russell, W.C., K. Hayashi, P.J. Sanderson and H.G. Pereira, 1967, J. Gen. Virol. 1,495. Russell, W.C., G. Patel, B. Precious, I. Sharp and P.S. Gardner, 1981, J. Gen. Virol. 56, 393. Schaffer, P.A., L.A. Falk and F. Deinhardt, 1975, J. Nat. Cancer Inst. 55, 1243. Timbury, M.C., 1971, J. Gen. Virol. 13, 373. Watson, D.H., W.I.H. Shedden, A. Elliot, T. Tetsuka, P. Wildy, D. Bourgaux, Ramoisy and E. Gold, 1966, Immunology 11,399. Weir, D.M. (ed.), 1978, Handbook of Experimental Immunology (Blackwell Scientific Publications, Oxford).