HYBRIDOMA Volume 11, Number 2, 1992 Mary Ann Liebert, Inc., Publishers

Monoclonal Antibodies Against Bacterial Glucosamine 6-Phosphate Synthase: Production and Use for Structural Studies OLIVIER

COCHET,1

BERNARD

BADET,1

and JEAN-LUC

TEILLAUD2

'Laboratoire de Bioorganique UA CNRS 1389, Ecole Nationale Supérieure de Chimie de Paris, 75231 PARIS Cedex 05, FRANCE et Clinique, Unité INSERM 255, Institut Curie, 75231 PARIS Cedex 05, FRANCE

2Laboratoire d'Immunologie Cellulaire

ABSTRACT Fifteen

mouse

rat

hybridoma cell

lines

producing rat monoclonal antibodies (MAbs) directed

Escherichia coli Glucosamine 6-P Synthase (GlmS)

were established and characterized. Most of them (13/15) are IgG2a while 2 were typed as IgGl. Their Kaff ranged from 1.5 IO6 to 9.6 IO8 M-1 as determined by Beatty et al. (1). The epitopes recognized by these MAbs were assigned to one of the two catalytical domains of the enzyme (CTi and CT2) as demonstrated both by ELISA and Western-blotting using purified GlmS proteolytic fragments. The binding of the MAbs on either the native or denatured forms of GlmS, CTi and CT2 was further analyzed by competitive immunoassay and most of the MAbs were found to bind preferentially to the denatured

to

proteins. The study of the antigenic topography of GlmS by competitive radioimmunoassay demonstrated the existence of at least 10 independent epitopes on GlmS, divided into three groups. The first one (3/15) includes MAbs whose binding was not inhibited by any of the other MAbs. The second group (9/15) is comprised of MAbs that exhibit reciprocal binding inhibitory activity while the third group includes MAbs (3/15) presenting asymmetric inhibitory activity. Finally, since most of the isolated antibodies (10/15) bind to the 27 kDa amino-terminal glutamine binding domain (CT2), the capacity of these MAb to interfere with the associated glutaminase activity was analyzed. INTRODUCTION

D-Glucosamine-6P

Synthase (EC 2.6.1.16) is the enzyme responsible glucosamine-6P (GlcN-6P) according to the following equation: D-Fructose-6P + L-Glutamine

—>

L-Glutamate

The

+

for the

production

of

D-Glucosamine-6P.

synthase reaction involves the binding of fructose-6P then glutamine, followed by the sequential release of glutamate and glucosamine-6P (2). GlcN-6P can only be synthesized through this pathway and is a precursor of UDP-N-Acetyl-Glucosamine which leads to the formation of bacterial peptidoglycan, fungal chitin and proteoglycans in eucaryotic cells. The formation of an intermediate -glutamyl thiolester with the aminoterminal cysteine residue of the enzyme is, for this particular amidotransferase, the only way to provide ammonia to the acceptor, fructose-6P, which

225

undergo a keto/aldose isomerisation to produce 2-amino 2-deoxy D-glucose-6P. In addition, can function as glutaminase in the absence of fructose-6P. We have previously described (3) the organization of the native Escherichia coli GlmS in two domains which could be individually characterized following limited proteolysis with achymotrypsin. The amino-terminal part (CT2) responsible for glutamine binding and glutaminase activity is linked through Tyr 240 to the larger carboxyterminal domain (CTi) responsible for sugar binding and catalytic transformation (4). However, little is known about the antigenic topography of the enzyme as well as on the precise location of its catalytic sites. So far, the polyclonal antibodies can

GlmS

raised in rabbits have been only used to detect the presence of the enzyme in crude cell extracts. In the present study, we developed a set of rat monoclonal antibodies and determined the antigenic topography of the GlmS. Fifteen MAbs were raised and characterized for their properties (IgG subclass typing and affinity) and for their ability to bind native or/and denatured GlmS. Their reactivities towards CTi and CT2 were analyzed by direct ELISA, Western-blotting and competitive enzyme-immunoassay. Using a radiocompetitive binding assay, we also defined groups of antibodies specific for either CTi or CT2. At least, 3 independent antigenic determinants were characterized on each catalytic domain of GlmS. However, these inhibition binding assays also revealed complex patterns of steric hindrance between antibodies and indicated that some MAbs bind to overlapping epitopes. Finally, we studied the effects of anti-CT2 MAbs on glutaminase activity of the amino-terminal domain. Two MAbs were able to completely inhibit catalysis while other MAbs interfered only weakly with enzymatic activity and one MAb slightly stimulated the hydrolysis of the substrate by CT2. MATERIALS & METHODS Purification of Glucosamine-6P Synthase and

Preparation of Chymotryptic Fragments

Glucosamine-6P Synthase from E. coli was purified from an overproducing strain as described by Dukta et al. (5). The proteolytic domains CT2 and CTi encompassing residues 1-240 and 241608 respectively were purified from GlmS after treatment with 1% (w/w) -chymotrypsin and followed by gel filtration chromatography (G-75 Superfine) as previously reported (3). ELISA

Analysis

performed as previously described (6). Briefly, polystyrene plates (Nunc Maxisorp, Roskilde, Denmark) were coated overnight at 4°C with purified GlmS (50 µg/ml in PBS, 50 µ /well). After washing (PBS-0.05% Tween), non-specific binding sites were blocked with 1% BSA (Sigma, Saint Louis, USA) in PBS (37°C, 90 min). After further washings, 50 µ of two-fold serial dilutions of sera or culture supernatants were incubated at 37°C, 90 min. A polyclonal alkaline phosphatase labeled goat anti-rat IgG (H+L) (Immunotech, Marseille, France), was used (0.2 µg/ml) as revealing reagent. After extensive washing, p-nitrophenyl phosphate (1 mg/ml in 10% diethanolamine (v/v), 0.1 mg/ml MgCl2 6H20, 0.02 % NaN3, pH 9.8) was added to each well. Absorbances were read at 405 nm using a microplate ELISA reader (Titertek Multiscan Plus MKII, Labsystems, France). ELISA

was

Production of Monoclonal Antibodies 3 female LEWIS rats were immunized on day 0 by intraperitoneal injection of 250 µg purified GlmS in a 1:1 PBS-complete Freund adjuvant mixture (1 ml). Two similar injections using incomplete Freund adjuvant were performed on days 15 and 30. Antibody responses to GlmS were assessed by ELISA before boosting was performed. On day 48, three days after the last boost (performed in PBS), Sp2/0 cells were fused according to Fazekas and Scheidegger (7) with spleen lymphocytes from the immunized rat giving the highest serum titer. Briefly, Sp2/0 hybridomas and splenocytes (ratio 1:2) were fused with PEG 1500 (Sigma). Fused cells were plated (105 Sp2/0 per

226

(Gibco, Paisley, Scotland) supplemented with Hepes (25 mM), glutamine (2mM), penicillin (100 ug/ml), streptomycin (100 µg/ml), 20% fetal bovine serum (Boehringer Mannheim, FRG) and HAT (Hypoxanthine 100 µ , Aminopterin 0.4 µ , Thymidine 16 µ ) in 96-well culture plates (Falcon, New Jersey, USA). The plates were seeded 3 days before fusion with feeders cells (fibroblasts derived from rats embryos or mouse peritoneal macrophages, IO4 cells/well). The supernatants were screened by ELISA for their anti-GlmS antibodies content. Of 16 plates, 1400 wells presented growing clones (93%) and ELISA screening demonstrated the presence of IgG anti-GlmS in 97 wells (6.9%). Cells from positive wells were transferred in 24-well plates (Falcon) and cloned in soft-agar (Sea-plaque, FMC, Rockland, USA) as previously described (8). Individual clones were picked and grown in liquid culture HT medium. Supernatants were retested by ELISA for their anti-GlmS rat IgG contents. Positive hybridomas were then frozen and stored in liquid nitrogen. Hybridomas cell lines were also cultured in HT medium in 75 ml culture flasks (Falcon). After centrifugation (10,000 g, 10 min) and filtration (0.22 urn, Millipore, Molsheim, France) of the supernatants, immunoglobulins were then purified by protein-G affinity chromatography (Mab-Trap G, Pharmacia, Uppsala, Sweden), adjusted to 0.5 mg/ml in PBS-azide (0.02%) and stored at 4°C. They were checked for purity by SDS-PAGE (9). Immunoglobulin concentration (mg/ml) was determined by absorbance at 280 nm using a correction factor of 0.74. well) in

RPMI 1640 medium

Determination of Rat Monoclonal Antibodies

Isotypes

Polystyrene microtiter plates were coated with different mouse monoclonal anti-rat IgG isotypes (Zymed, San Francisco, USA) at 10 µg/ml in PBS. Microplates were washed and saturated as described above. Anti-GlmS MAbs to be typed and control rat monoclonal immunoglobulins (IgGi and IgG2bfrom Zymed, IgG2aand IgG2c from PharMingen, Clinisciences, Paris, France) were incubated at ^g/ml for 1.5 h at 37°C. After washing, a monoclonal alkaline-phosphatase labeled mouse antibody to rat light chain (PharMingen, ^g/ml final conjugate concentration) was used as revealing reagent (1.5 h at 37°C). ELISA

Competition with Native Forms of GlmS. CT¡^ and CT£

of the MAbs preferentially bind to the native form of GlmS, CTi or performed (10). 100 µ of each MAb (0.2 µg in PBS) were incubated for 15 min at room temperature with 100 µ of 10-fold molar excess of the protein solution (GlmS, 10 µ^ ; CTi, 6 µg/ml ; CT2, 5 µg/ml) or with 100 µ of PBS used as control. 50 µ of each mixture were then added to a GlmS-coated microtiter plate. ELISA plates were then processed as described above. The assays were performed in triplicate. The percentage of inhibition of the binding on immobilized GlmS of each MAb by native GlmS, CTi or CT2 was then calculated. The binding of each MAb preincubated with PBS alone was used as reference (100%). To determine whether

CT2, competitive ELISA

Determination of MAbs

some was

Affinity Constants

The affinity constant (Kaff) of each MAb for GlmS coated onto the plate was determined using the protocol described by Beatty et al. (1). These authors demonstrated that mathematical processing of the mass action law leads to a simple expression of Kaff :

(n-1)

Kaff-2(n[Ab2]-[Abi]) where [Abi] and [Ab2] represent the respective MAb concentrations required to reach 50% of the maximum absorbances obtained at two different concentrations of coated antigen ([Ag]i n[Ag]2) and is the dilution factor between the two concentrations of antigen used. In our experiments, these two MAbs concentrations were determined by plotting the optical densities =

227

obtained with the two different concentrations of coated-GlmS as a function of the logarithm of the corresponding MAbs concentrations ( see Fig. 1). Microtiter plates were coated with four concentrations (400, 200, 100 and 50 ng/ml) of GlmS in PBS. The final concentration of protein in each well was maintained at 5 µg/ml with BSA during coating. Serial-two fold dilutions (from 106 down to 10~9 M) of MAbs in PBS-Tween were then performed in duplicate for each concentration of coated GlmS. The subsequent steps were identical to those described above except that the conjugate was also diluted in PBS-Tween.

Mapping Epitopes in the Two Domains of GlmS The pattern of

reactivity

of each monoclonal

(CTi and CT2) of the native enzyme

was

antibody toward the two proteolytic fragments analyzed by three different and complementary

approaches : analysis : 50 ug of GlmS, CTi and CT2 were mixed and submitted to electrophoresis (Mini Protean II, BioRad). Proteins were transferred onto nitrocellulose (Amersham Hybond C-extra, 0.45 µ ) for 1 h (100 Volts, Mini Trans-blot, BioRad). The nitrocellulose was then saturated with Tris Buffer Saline (TBS)-3% gelatin for 1.5 h at room temperature, washed (TBS-0.05% Tween) and cut in strips which were soaked for two hours with the different MAbs to be tested (0.5 µg/ml) in TBS-1% gelatin. After washing (TBS-0.05% Tween), the nitrocellulose strips were incubated for 1.5 h at 37°C with a goat antiserum (0.15 µg/ml) to rat IgG (H+L) alkaline-phosphatase conjugated (Immunotech) in TBS-1% gelatin. Staining was then performed with Bromo-chloro indoyl phosphate (BCIP) and Nitroblue tetrazolium (NBT), (KPL, Gaithersburg, USA). Western-blot

-

Indirect ELISA : Polystyrene microtiter plates coated (1.5 h, 37°C) with 30 µg/ml of either GlmS, CTi or CT2 were saturated with PBS-1%BSA and incubated with two-fold dilutions of MAbs (from 1 µg/ml down to 7.8 ng/ml). The ELISA was performed as described above. -

Competitive radioimmunoassay : 8 strips, composed of 12 cleavable flat-bottom wells, were assembled into holders to form a 96-well plate (Removawell, Dynatech, Marnes-la-Coquette, France). Plates were coated (1.5 h, 37°C) with GlmS (200 ng/ml in PBS), washed and saturated by an overnight incubation at 4°C with PBS-1%BSA. Each row was then incubated for 1 h at room temperature with 100 µ of two-fold serial dilutions (from 80 µg/ml down to 80 ng/ml) of a competing unlabeled MAb in PBS-0.2% BSA and washed before addition of the radiolabeled MAb. Iodination of antibodies was performed by the chloramine method (11). 400,000 cpm of radiolabeled MAb (in 50 µ ) were added to each well of the microtiter plates which were then incubated 2 h at room temperature before washings. Each well was cleaved and evaluated for 125IMAb binding using a multigamma counter (LKB 1260, Pharmacia-LKB, St Quentin-en-Yvelines, -

France).

Inhibition of the Glutaminase

Activity

The inhibition of the hydrolysis of -glutamyl-p-nitroanilide (a chromogenic substrate of CT2, releasing p-nitroaniline upon hydrolysis) by the different MAbs against CT2 was investigated. The enzymatic reaction was conducted in a microtiter plate using a final volume of 250 µ . Each MAb

incubated in five-fold molar excess (0.5 µ ) with CT2 (0.1 µ ) for 15 min in PBS-0.2% BSA temperature, -glutamyl-p-nitroanilide (Sigma) was then added at a final concentration of 0.15 mM and the A405 was immediately recorded. The plates were then incubated at 37°C for 4 h before a new A405 reading. The percentage of the substrate hydrolyzed in presence of MAb was determined by comparison with the hydrolysis of -GpNA in a mixture containing no antibody was

at room

(100%).

228

RESULTS and DISCUSSION Production of Monoclonal Antibodies

High antibody titers (1:16000) were obtained when rat sera were assayed by ELISA for their binding to GlmS-coated plates (data not shown). The fusion of rat splenocytes with Sp2/0 cells resulted in the production of a large number of hybridomas of interest since about 7% of the wells (97 wells) contained IgG directed to GlmS as tested by ELISA. 15 stable different hybridoma cell lines, secreting rat monoclonal anti-GlmS IgG, were finally isolated. The IgG subclasses of these MAbs were determined using an indirect ELISA (see Materials and Methods). Most of them (13 of 15) were IgG2a whereas only two MAbs (510.2 and 602.2) belonged to the IgGi subclass. This is in agreement with previous reports demonstrating that most of the rat IgG directed against thymodependent antigens belong to the IgG2a and IgGi subclass (12).

Affinity Constant Measurement The relative affinity constants of the MAbs for GlmS were calculated and compared. Instead of using a method based on the formation of antibody-antigen complexes in a soluble phase (13, 14), we performed the assay described by Beatty et al. (1) which allows the measurement of Kaff by solid-phase EIA, using different antigen concentrations for the plate coating (Fig 1). Although, the exact amount of the antigen adsorbed onto the well is unknown, the method has been found to compare favorably with classical RIA (15). The simplicity of the technique allowed us to classify easily the 15 MAbs with regards to their Kaff. Since 4 concentrations of GlmS were used for coating (400, 200, 100 and 50 ng/ml), six affinity constants (3 for n=2, 2 for n=4 and 1 for n=8) were calculated for each MAb and the mean Kaff values were then determined (Table 1). The affinity of MAb 104.2 (9.6 108 M"1) is almost 1000 greater than the affinity of MAb 504.1 (1.5 IO6 M"1) which exhibited the lowest affinity. These Kaff values reflect the relative affinity of the MAbs for immobilized GlmS. Their determination could be biaised due to partial enzyme denaturation leading to conformational epitope modification which could occur upon polystyrene binding (10, 16). We, therefore, also

-9,6 -9,2 -8,8 -8,4 -8,0 -7,6 -7,2 -6,8 -6,0 -6,4 FIGURE 1. Experimental Kaff determination by non-competitive enzyme immunoassay for MAb 410.1 at different GlmS coating concentrations (ng/ml): 1: 400, 2: 200, 3: 100 and 4: 50. The antibody concentrations (nM) at OD-50 were: 1: 15.8, 2: 17.8, 3: 25.1 and 4: 28.1. Calculated

Kaffwasl.8±0.8xl07M-1·

229

TABLE 1.

Kaff Determination of anti-GlmS Monoclonal Antibodies. MAb MAb Kaff" (M-i) 415.3 9.6 ±2.9x108 104.2 1.5 ±0.3 107 421.1 105.1 5.0 424.3 ±2.2 IO6 108.3 504.1 3.8 ± 1.8 106 113.3 115.1 1.1 ± 0.3 IO8 505.1 510.2 409.2 3.8 ± 1.7 IO7 522.2 410.1 ± 1.8 0.8 IO7 8.3 ±2.1 IO8 602.2 a

KaffMM-l) 3.3 ± 1.0 6.8 ± 3.6 2.0 ± 0.7 1.5 ± 0.8 2.4 ± 1.4 4.6 ± 1.4

4.2 ± 0.3

IO6 IO7 IO8 IO6 IO6 IO7 IO8

Mean of 6 values.

analyzed whether the MAbs bound more efficiently to the immobilized GlmS, CTj and CT2 fragments or to the corresponding native components in the soluble phase. This was performed by an ELISA competition test (14) where the MAbs and a ten-fold molar excess of the native GlmS, CTi or CT2 were first incubated in solution for 15 min. In this assay, MAbs which rapidly bind to the native form of GlmS (or native proteolytic fragments) added in excess, were prevented from binding to GlmS-coated plate. By contrast, all the MAbs against denatured antigenic determinants of GlmS remained free in solution and were quantified by ELISA using a GlmS-coated plate. It should be stressed that an antibody binding to a denatured protein may nevertheless still bind to the native form of this protein but with a decreased affinity (17). In our assay, an antibody was defined as anti-native when only less than 15% of the MAb amount bound to immobilized GlmS after antibody-antigen complexes formation during preincubation with excess of antigen. The results of this analysis, summarized in Table 2, show that 4 MAbs (indicated in bold type) were strongly reactive with the native GlmS. MAbs 104.2, 409.2 and 505.1 exhibited a strong binding to CT2, which was clearly confirmed by other tests (see below). Moreover, this assay was the only method able to assign to MAb 602.2 a specificity to the CTi domain. (See Fig. 2 and 3 for ELISA and Western-blotting analysis). The corresponding epitope was very labile since, although present on coated-GlmS, it was no longer detectable on immobilized CTj or following electrophoretic transfer of this fragment onto nitrocellulose. Finally, the Kaff for MAbs 104.2, 409.2, 505.1 and 602.2 TABLE 2.

ELISA

Competition with Native Forms of GlmS, CTi

Native protein used for complexes formation:

GlmS CTI CT2

and

CT2a

% of binding of the following MAbs

104.2 105.1 108.3 113.3 115.1 409.2 410.1 415.3 421.1 424.3 504.1 505.1 510.2 522.2 602.2 2 83 84 98 100 5 53 90 93 96 95 52 9 94 2 46 100 100 100 44 100 100 100 100 100 95 92 82 2 7 15 93 93 97 95 66 85 100 96 3 100 86 100

a

Competitive binding assay. MAbs (2µ^ 1) were first incubated for 15 min at room temperature with ten-fold molar excess of GlmS (10 µ^ ), CT] (6 µg/ml) or CT2 (5 µg/ml). Each mixture was then allowed to adsorb on GlmS-coated plate. Bound antibodies were delected by ELISA. * Results are expressed as the percentage of the binding of the MAb to GlmS-coated plate after GlmSMAb, CTi-MAb or CT2-MAb complexes formation in solution. The binding of each MAb preincubated with PBS alone was used as reference (100%). MAbs which bind native protein are typed in bold. a

230

OD 405 nm

32000

250

32000

250

32000

250

FIGURE 2. Reactivity of MAbs 104.2, 421.1 and 602.2 by ELISA to GlmS (— "), CT and CT2 ( ~~0~ ) (10 µg/ml) coated on ELISA microtiter plates. Curves are expressed as OD 405 nm vs dilution factor (1/250 to 1/32000) of MAb 104.2,421.1 and 602.2 (0.1 µg to0.8 ng/well).

(-A—)

determined by the technique described by Beatty et al. (1), that uses denatured GlmS, were probably underestimated compared to their Kaff for the soluble protein, as these MAbs preferentially bind to native GlmS.

Specificity of MAbs for the Two Domains

of GlmS

Since Glucosamine 6-P synthase is organized in two domains around a hinge structure at Tyr240, specificity of each MAb for each purified domain (3) was analyzed in detail. Two approaches were used for this purpose. An indirect ELISA on GlmS, CTi and CT2 coated-plates was first performed using two-fold serial dilutions of the MAbs (Fig. 2). The specificity of the different MAbs for either CTi or CT2 was also analyzed by Western-blotting (Fig. 3). The data obtained with these two techniques are compared to those obtained with the ELISA competition test in Table 3. This table indicates that : (i) all the purified MAbs bind to the immobilized enzyme as expected since this assay was used for the first screening of growing the

FIGURE 3. Reactivity of the MAbs on GlmS, CTi and CT2 as detected by Western blotting. Mixture of GlmS, CTi and CT2 (50 ug of each protein) was submitted to electrophoresis on a 12% polyacrylamide slab gel and then transferred onto niuocellulose. The niuocellulosc strips were incubated with MAbs (0.5 ug/ml), rat anti-serum to GlmS (Conuol +) or with non-immune rat serum (Control -) and then stained with alkaline-phosphatase labeled goat anti-rat IgG.

231

TABLE 3.

Comparison of the binding profile of the 15 MAbs to GlmS, CTI Western-blotting (W.B.) and ELISA competition test (Comp.). Tested by: MAbs

and CT2

Preferentially

Tested on:

W.B.6

ELISA" 409.2 104.2

GlmS CTI CT2

505.1

GlmS CTI CT2

105.1 108.3 113.3 415.3 504.1 522.2

GlmS CTI CT2

410.1

GlmS CTI

by ELISA, binds

to:

Comp. +

+/-

native CT2

native CT2

"denatured"

CT2

"denatured"

CT2

+/-

CT2

115.1 421.1 424.3 510.2

602.2

a

Results of indirect ELISA

GlmS CTI CT2

"denatured" CTi

GlmS CTI CT2

native CTi

using alkaline-phosphatase

labeled

polyclonal goat anti-rat IgG

as

revealing

reagent.

Western-blotting on GlmS, CTi and CT2. See Figure 3. ELISA competition test: MAbs (2mg/ml) were first incubated for 15 min at room temperature with GlmS or CTj or CT2 (10 mg/ml). Each mixture was then allowed to adsorb on GlmS-coated plate. Bound antibodies were detected by ELISA. (+) indicates a binding lower than 15% of the maximal binding of MAbs on GlmS-coated plate after GlmS-MAb complexes formation in solution. (+/-) indicates a binding of about 50% of the maximal binding of the MAbs. "

c

hybridomas. (ii) 13 MAbs bound to either CTi (4) or CT2 (9). (iii) Two MAbs (505.1 and 602.2) recognized only GlmS when tested by ELISA. We first hypothesized that they bound to an epitope disappearing after the proteolytic cleavage of GlmS. However, the MAb 505.1 exhibited a strict CT2 specificity by immunoblot analysis. ELISA competition test confirmed this reactivity of 505.1 to native CT2. Thus, partial denaturation or epitope disappearance during the coating of CT2 on polystyrene plates could account for the discrepancy between ELISA and Western-blotting. The MAb 602.2 was also found to be negative in Western-blotting with GlmS, CTi and CT2 fragments. However, as reported in the previous section, we demonstrated that 602.2 recognized an epitope of the native form of CTi by the ELISA competition test.

Antigenic Topography of GlmS Competitive radioimmunoassays (RIA) (18, 19) were used to determine the antigenic topography of GlmS using our set of MAbs. It should be noted that monoclonal antibodies recognizing the same epitope usually compete with each other in a reciprocal fashion for the binding to this epitope. However, the relationships between MAbs binding to different antigenic determinants may be complex, including non-reciprocal inhibition of binding and even enhanced binding. Basically, each MAb directed to GlmS was tested in a competitive RIA where coated-GlmS was first reacted with the different unlabeled MAbs and then with one given ,25I-labeled MAb. Binding of 125I-labeled MAb on GlmS-coated plate in the absence of competing MAbs was used as the 232

FIGURE 4. MAbs groups defined by competitive radioimmunoassay. Hatched areas defined MAbs recognizing the same epitope or close antigenic determinants as shown by reciprocal inhibition. Cross-hatched areas delineated epitopes recognized by a single MAb. Open areas indicate non-reciprocal interactions between antibodies (for instance, the binding of labeled 409.2 was prevented by unlabeled 415.3 but labeled 415.3 could bind to GlmS-coated plate preincubated with 409.2). The assay also confirmed the epitope mapping on either CT] or CT2 determined by ELISA and western-blotting.

reference

(100%). It reached between 2.5 and 10% of the total cpm (400,000) added in each well, depending on the MAb used. In this assay, a marked decrease in bound cpm reflected the binding to the same determinant or to very close determinants of the MAbs tested. Partial inhibition of binding (which range with these MAbs from 50 to 80%) represented significant steric hindrance between the MAbs, although more complex situations can exist as detailed below. In our experimental conditions, the binding of one 125I-labeled MAb was always inhibited (>80%) by the same

unlabeled MAb. Our results indicate that the monoclonal antibodies against GlmS could be divided into three groups as shown in Fig. 4. One group (hatched areas) comprises MAbs for which reciprocal binding inhibition (above 80%) was observed (for example MAb 104.2 and 522.2 or 421.1, 424. 3 and 510.2) The antibodies of this group were specific for the same epitope or topologically close antigenic determinants. Antibodies 113.3, 602.2 and 115.1 which define the second group (cross-hatched areas), did not present inhibitory effects with each other or with any other MAbs. These antibodies delineate independent epitopes on GlmS. The third group includes MAbs which competed with one another in a non-reciprocal fashion. For instance, unlabeled MAb 415.3 was able to prevent binding of labeled MAb 409.2 to GlmS but 409.2 did not inhibit the binding of labeled MAb 415.3. The asymmetric reciprocal competitions observed could result from differences in the affinities of competing antibodies for the same epitope or from allosteric alterations of the enzyme. It could induce the structural alteration of one antigenic determinant upon binding of a MAb to another epitope. This hypothesis has been previously suggested by Lübeck and Gerhard (20) with the PR8 virus hemagglutinin. The interactions of antibodies of the third group with those of the first group may be due to steric constraints resulting from the size of the competing antibody or because of structural alterations of antigenic sites upon binding. In this context, MAb 504.1, which belongs to this third group, is of particular interest. The binding of this MAb, recognizing an epitope clearly mapped on CT2 by ELISA and Western-blotting, was prevented by 522.2 that bound to a CT2 epitope and by 510.2 which was specific for the CTi domain. This result suggests that the binding of one antibody to GlmS could indeed allosterically affect the binding of another antibody to a topologically distant site. The competition data obtained with 505.1 are puzzling : all the other MAbs against the enzyme prevented its binding, although, as expected, it competed with itself for the binding to GlmS.

233

However, it was unable to inhibit the binding of the 14 remaining MAbs. One can hypothesize that

505.1 could recognize a conformational epitope which is lost upon binding of any other MAbs to GlmS. The competitive RIA made it possible to determine the relative binding site location of each MAb on GlmS and to confirm the mapping of the epitopes on either CTi or CT2. It is noteworthy that 10 of the 15 MAbs were specific for the first third of the amino-terminal region of GlmS. This apparent immunodominance of the CT2 domain may be due to a particular folding of the GlmS that allows preferential accessibility of B-cells to the N-terminal region of the enzyme as reported for human papillomavirus 18 E7 open reading frame protein (21). Finally, with the fifteen MAbs, at least ten topologically distinct epitopes have been mapped on the two domains of GlmS. Inhibition of Glutaminase Activity on the effects of MAbs on the enzymatic activity of GlmS were the by measuring hydrolytic activity of GlmS for glutamine which has been also observed for amidotransferases all previously investigated (22, 23). This catalytic activity, although about two hundred times less efficient than glucosamine-6-P synthesis, is shared by the native protein (4.2 min1) and, to a lesser extent, by the CT2 amino-terminal proteolytic fragment (0.3 min-1). The analysis of this catalytic activity has been improved using -glutamyi p-nitroanilide ( GpNA) as an alternate substrate (3) which is hydrolyzed by native protein with the same efficiency (4.4 min-1) as glutamine and by the CT2 domain with a much higher turnover (2.4 min-1) than the normal substrate. This hydrolysis could be further increased (up to 2.5 fold) by addition of 0.2% BSA to the incubation buffer (24).

Preliminary investigations

conducted

The ten CHVspecific MAbs were, therefore, tested for their ability to inhibit this reaction. The results shown in Fig. 5 are expressed as the percentage of release of p-nitroaniline upon incubation of MAbs with CT2 3 by comparison with the release obtained in the same conditions but without any MAb (100%). Four different patterns of interactions were observed : antibodies 113.3 and 409.2 had no influence on CT2 catalytic activity. The low affinity of the MAb 113.3 for the soluble native form of GlmS could explain its lack of interference with the enzyme activity. In contrast, 409.2 recognizes the native glutamine binding domain (CT2) as well as the native form of the enzyme (Table 3). Thus, one can conclude that this particular MAb binds to a conformational epitope away from the glutamine active site. Five antibodies (104.2, 108.3, 410.1, 415.3 and 504.1) led to a 50% inhibition of glutaminase activity and might induce a slight conformational change or induce a weak steric hindrance resulting in a decreased hydrolytic rate.

Two antibodies (105.1 and 505.1) directly interfered with the hydrolysis of -GpNA on CT2. Since they bound to two different epitopes as shown by the competitive binding experiments described above, this suggests that two antigenic determinants are related to the glutamine binding site. MAb 410.1, which inhibits the binding of 105.1 and whose binding to GlmS is prevented by the latter, exhibited only a slight inhibitory effect. Thus, it probably recognizes an antigenic determinant different from that recognized by 105.1. However, some MAbs (105.1, 108.3, 415.3, 410.2 and 504.1) which have a slight or a marked inhibitory effect on enzymatic activity bind preferentially to denatured Cp2 as determined by ELISA competition test (Table 2 and Table 3). Since glutaminase activity is only exerted by native GlmS and native CT2, one can conclude that these MAbs bind to native CT2 but with a much lower affinity than to denatured CT2. Thus, although ELISA competition test favors anti-denatured CT2 binding, these MAbs bind nevertheless native GlmS, inhibiting glutaminase activity. to

Finally, we observed that MAb 522.2 was able to slightly increase the efficiency of catalysis up experiments, range : 109% to 135%) although it did not induce the catalysis of -GpNA

118% (5

234

in absence of CT2. The binding of this MAb to the CT2 domain probably alters the conformation of the catalytic site, allowing a better accessibility of the substrate to its binding site. Such potentiating effects on enzymatic activity have been previously reported with polyclonal antibodies against defective ß-galactosidase (25). It should be noted that MAb 104.2, which inhibits the binding of 522.2, did not induce the same enhancing effect on the catalytic activity. As discussed above for MAbs 410.1 and 105.1, these two MAbs could delineate two distinct but topologically related epitopes on CT2.

140 & hydrolysis

FIGURE 5. Percentage hydrolysis of -Glutamyl-p-Nitro-Anilide ( -GpNA) by CT2 domain in presence of anti-CT2 MAbs. Antibodies ( . µ ) were incubated for 15 min with CT2 (0.5 µ ) before adding -GpNA (0.15 mM). Hydrolysis of -GpNA was measured at 405 nm after 4 h incubation at 37°C. Percentages represent the mean value ± SE (bars) of at least 3 different experiments, a : incubation of CT2 without antibody ; b : spontaneous hydrolysis of -GpNA ; c : incubation of CT2 with an irrelevant rat IgG MAb ; d : incubation of MAb 522.2 and -GpNA without CT2.

From the analysis described above, a restricted number of antigenic structures are directly involved in GlmS hydrolytic activity. The influence of the corresponding monoclonal antibodies on the Glucosamine-6-P synthesizing activity together with the fine mapping of linear epitopes sequence is currently under investigation.

ACK NO WLEDGMENTS

This work was supported in part by a grant from Institut CURIE (contrat coopératif n° 90-31). O.C. is a PhD student supported by the French Ministère de la Recherche et de la Technologie. J.L.T. is Directeur de Recherche (DR II) at the INSERM and . . is Directeur de Recherche (DR II) at the CNRS. The authors thanks Dr W.H. Fridman for valuable discussion and constant support, Mrs J. Moncuit for technical assistance, Dr A. Saul for reviewing english, Mrs Bussière for preparation of glossy prints and Professor Le Goffic for his hospitality.

235

REFERENCES 1.

BEATTY, J.D., BEATTY, B.G.

AND

VLAHOS, W.G. (1987) Measurement of monoclonal

antibody affinity by non-competitive enzyme immunoassay. J. Immunol. Methods. 100, 173179. 2.

BADET, ., VERMOOTE, P., HAUMONT, P.Y., LEDERER, F. AND LE GOFFIC, F. (1987) Glucosamine synthetase from Escherichia coli : purification, properties and glutamine utilizing site location. Biochemistry 26, 1940-1948.

3.

DENISOT, MA., LE GOFFIC, F.

AND Badet, . (1991) Glucosamine-6-phosphate synthase yield two proteins upon limited proteolysis: identification of the glutamine amidohydrolase and 2R ketose/aldose isomerase-bearing domains based on their biochemical properties. Archivs Biochem. Biophys. 288, 225-230.

from Escherichia coli

4.

GOLINELLI-PIMPANEAU, B.

AND

BADET, . (1991) Possible involvement of lysine 603 of

Escherichia coli glucosamine-6-phosphate phosphate. Eur. J. Biochem. 201, 175-182. 5.

overexpression 287-290.

6.

Badet, . (1988) Molecular cloning and glucosamine synthetase gene from Escherichia coli. Biochimie. 70,

DUTKA-MALEN, S., MAZODIER, P. of the

synthase in the binding of its substrate fructose-6-

and

THIBAUT, E., AMIGORENA, S., MONCUIT, J., FRIDMAN, W.H. ANDTEILLAUD, J.L. (1987) Software for the quantitative evaluation of in vitro monoclonal ELISA data. J. Immunol. Methods. 104, 15-24.

antibody production

from

7.

FAZEKAS DE St Groth, S. AND SCHEIDEGGER, D. (1980) Production of monoclonal antibodies, strategy and tactics. J. Immunol. Methods. 35, 1-21.

8.

COFFINO, P., BAUMAL, R., LASKOV, R.

AND

SCHARFF, M.D. (1972) Cloning of

myeloma cells and detection of rare variants. Cell. 9.

LAEMMLI, U.K. (1970) Cleavage of structural bacteriophage T4. Nature. 227, 1940-1948.

10.

FRIGUET, B., DJAVADI-OHANIANCE, antibodies raised with native Immunol. 21, 673-679.

protein

L. AND

bind

mouse

429-440.

Physiol. 79, proteins during

the

assembly of the

head of

GOLDBERG, M.E. (1984) Some monoclonal

preferentially

to

the denatured

antigen.

11.

MCCONAHEY, P.J. AND DlXON, F.J. (1980) Radioiodination of chloramine T. Meth. Enzymol. 70, 210-213.

12.

DER BALIAN, G.P., SLACK, J., CLEVINGER, B.L., BAZIN, H. AND DAVIE, J.M. (1980) Subclass restriction of murine antibodies. Ill Antigens that stimulate IgG3 in mice stimulate IgG2c in rats. J. Exp. Med. 152, 209-218.

13.

ROE, R., ROBBINS, R.A, LAXTON, R.R. AND BALDWIN, R.W. (1985) Kinetics of divalent monoclonal antibody binding to tumour cell surface antigens using flow cytometry standardization and mathematical analysis. Mol. Immunol. 22, 11-21.

14.

FRIGUET, B., DJAVADI-OHANIANCE, L.

AND

protein by

the

Mol.

use

of

:

GOLDBERG, M.E. (1989) Immunochemical : a practical approach (Edited by T.E.

analysis of protein conformation. In Protein structure Creighton) p. 287. 15.

AND SHIVELY, J.E. (1983) Monoclonal antibodies for carcinoembryonic antigen and related antigens as a model system : determination of affinities and specificities of monoclonal antibodies by using biotin-labeled

WAGENER, C, CLARK, B.R., RlCKARD, K.J

236

antibodies and avidin 2302-2310.

precipitating agent

in soluble

phase immunoassay. J.

Immunol.

130,

AND Houe D.W. (1986) A comparison of ELISA screening methods for the production of monoclonal antibodies against soluble protein antigens. /. Immunol. Methods. 93, 9-18. 17. JEMMERSON, R. (1987) Antigenicity and native structure of globular proteins : low frequency of peptide reactive antibodies. Proc. Nati Acad. Sci. U.S.A. 84, 9180-9184. 18. SMITH-GILL, S.J., LA VOIE, T.B. AND MAINHART, C.R. (1984) Antigenic regions defined by monoclonal antibodies correspond to structural domains of avian lysozyme. J. Immunol. 133,

16.

BRENNAND, D.M., DansON, M.!.

384-393. 19. 20. 21.

22.

and Rodriguez, r. (1990) Epitope mapping of the major allergen from yellow mustard seeds. Moi Immunol. 27, 143-150. LÜBECK, M.D. AND GERHARD, W. (1981) Topological mapping of antigenic sites on the Influenza A/PR/8/34 virus hemagglutinin using monoclonal antibodies. Virology 113, 73-85.

Menendez-Arias, L, Domínguez, J., moneo, .

SELVEY, L.A., TINDLE, R.W., GEYSEN, H.M., HALLER, C.J., SMITH, J.A AND FRAZER, I.H. (1990) Identification of B-epitopes in the human papillomavirus 18 E7 open reading frame. J. Immunol. 145,3105-3110. BUCHANAN, J.M. (1973) The amidotransferases. Adv. Enzymol. Relat. Areas. Moi Biol. 39, 91-183.

23.

ZALKIN, H. (1985) Glutamine amidotransferases. Meth. Enzymol. 113, 263-305.

24.

WINTER, H.C.

AND

peanut leaves. Plant. 25.

DEKKER, E.E. (1991) 4-Methyleneglutamine Amidohydrolase from Physiol. 95, 206-212.

AND CELADA, F (1968) Antibody-mediated activation of a defective ßgalactosidase extracted from an Escherichia coli mutant. Proc. Nati Acad. Sci USA. 60, 660-

ROTMAN, B.M. 667.

Address reprint requests: Dr Bernard BADET Laboratoire de Bioorganique et Biotechnologies Ecole Nationale Supérieure de Chimie 11, rue Pierre et Marie Curie 75231 Paris Cedex 05 France

237

Monoclonal antibodies against bacterial glucosamine 6-phosphate synthase: production and use for structural studies.

Fifteen mouse x rat hybridoma cell lines producing rat monoclonal antibodies (MAbs) directed to Escherichia coli Glucosamine 6-P Synthase (GlmS) were ...
3MB Sizes 0 Downloads 0 Views