Journal of Immunological Methodv, 154 (1992169-76

69

© 1992 Elsevier Science Publishers B.V. All rights reserved 0022-1759/92/$05.110

JIM (16419

Use of defined oligosaccharide epitopes in an ELISA for group B streptococcus R o b e r t J. C h a l i f o u r a, F r a n c i s M i c h o n b, H a r o l d J. J e n n i n g s b a n d M a r t i a l Lacroix a "BioChem hnmunoSystems, 5.71 Boulecard des Prairies, Lm "al, Quebec H7V I BZ Canada, and i, Dirision of Biological Sciences, National Research Council of Canada, Ottawa, Ontario KIA OR6, Canada

(Received 11 March 1992,accepted 17 April 19921 A single-site ELISA for group B streptococcal polysaccharide based on a monoclonal antibody against an immunodominant trirhamnoside epitope was inhibited at high capture antibody coating densities. The inhibition was eliminated when less antibody was coated or when high antigen concentrations were tested. This antigen is polyvalent with respect to the terminal trirhamnoside epitope and therefore it appears that closely spaced capture antibodies bound the epitope completely, leaving no sites for attachment of the enzyme-labeled second antibody with the same specificity. To make use of the trirhamnoside epitope feasible, a two-site ELISA was evaluated with this monoclonai antibody and a polyclonal antibody isolated by affinity chromatography. ELISA inhibition studies using oligosaccharides derived from the group B polysaccharide were used to evaluate the specificity of the polyclonal antibody. This showed that the antibody recognized both a-L-rhamnose(l --, 3)-D-galactose and a-L-rhamnose(l --* 3)-D-glucitoi side chains, which together represent 30 potential binding sites per antigen molecule. A two-site ELISA with the anti-trirhamnoside monoclonai antibody to capture the antigen and the polyclonal antibody as enzyme conjugated second antibody reacted with only group B streptococci when tested against a panel of bacteria representative of the vaginal flora and was able to detect 3 x 10 4 c f u / t e s t of group B streptococci. This two-site ELISA, based on well defined oligosaccharide epitopes had the sensitivity and specificity necessary to identify women at risk of infecting their newborns with group B streptococcus. Key words: Streptococcus agalactiae; Group B streptococci; Affinity purification; Oligosaccharideepitope; Trirhamnose; ELISA

Introduction Group B streptococcus causes serious neonatal disease when tcansmitted from mother to child

Correspondence to: R.J. Chalifour, BioChem lmmunoSysterns, 531 Boulevard des Prairies, Laval, Quebec H7V IB7, Canada. Tel.: (5141-687-49111;Fax: (5141-687-3521. Abbreciations: BSA, bovine serum albumin; GBS, group B streptococcus; GBSP, group B streptococcal polysaccharide antigen; MAb, mono~'!onalantibody; PBS. phosphate-buffered saline; PBST, phosphate-bufferedsaline containingTween 211.

dulihg bh ;.h. aucn inlcctic~ls can be prevented if women who are carriers of GBS are identified and treated with antibiot ~:s prior to parturition (Morales and Lira, 19871. Rapid immunodiagnostic testing is preferable to t;ae method of overnight culture in this situation d~e to the unpredictability of the time of delivery. Because of the significance of neonatal streptococcal disease, we have sought to determine whether a recently identified immunodominant trirhamnoside epitope of the GBS group-specific polysaccharide might be useful in a rapid screening test for this bacterium. A

number of studies have found that tests for GBS antigen, both of the ELISA and latex agglutination types, lack sensitivity (Hoppe et al., 1989; Yawn et al., 1990; Greenspoon et al., 1991; Skoll et al,, 1991) and that crossreaction with non-GBS bacterial antigens may occur (Costa-Cruz et al., 1990). Immunodiagnostic assays for GBS detect the group-specific polysaccharide synthesized by all serotypes of this bacterium. This antigen has a complex multian~c~mary skr~cture previously described in detail (Michon et al., 19881 c~mtaining a number of immunodominant epitopes. Terminal rhamnose i~ one such cpitope (Curtis and Krause, 1964; loleidelbergc~ et al., 19671; however, due to its frequent appearance on other bacterial poIysaccharides, this epitope is of limited diagnostic use ,~,,.~d may be a cause of test non-specificity. The trisaccharide structure, a-LR h a p ( l ~ 2~a-L-Rhap(! ~ 2 ) a - L - R h a p , was identified as another immunodominant epitope of the group-specific antigen (Michon et al., 19911. A monoclonal antibody raised against GBS bacteria was previously shown by cytochemical methods to be useful for detecting GBS bacteria in necropsy specimens (Feldman et al., 19861 and recent studies have shown that this antibody recognizes the trh,:_mnoside epitope (Michon et al., 19911. Here we l~ave used this monoclonal antibody to evaluate t~c utility of the first structurally defined epitopes of the GBS antigen in an ELISA for detection of this bacterium.

Materials and methods

Preparation of antisera to group B streptococcus Polyclonal antiserum was raised in a Suffolk sheep immunized intradermally with 4.5 ml of an emulsion of two parts Freund's complete adjuvant (Gibco Laboratories, G r a n d Island, USA) and one part PBS (0.05 M N a z H P O 4 / N a H 2 P O 4, pH 7.4, (I.15 M NaCI), containing 0.5 mg of GBS polysaccharide-tetanus toxoid conjugate. Blood was collected after 3 months, serum separated and stored at - 2 0 ° C until purification as described below. The mouse lgG3 monoclonal antibody directed against GBS polysaccharide, designated G B S l / 1 8 : 6 / D I , (Feldman et al., 19861 was produced in ascites fluid and purified by protein A affinity chromatography.

Oligosaccharides and protein conjugates of oligosaccharides These were obtained from the native polysaccharide or by synthetic procedures as previously described (Michon et al., 19911.

Preparation of affinity colamns An affinity gel containing a-L-Rhap(I ~ 2)at.-Rhap(l ~ 2)a-L-Rhap units was prepared as previously described (Michon et al., 19911. An affinity gel containing rhamnose units was prepared using a published procedure (Fornstedt and Porath, 19751 with the following modifications: Sepharose CL-4B (Pharmacia (Canada),

TABLE I AFFINITY PURIFICATION OF ANTI-MONORHAMNOSE POLYCLONAL ANTIBODY FROM SHEEP SERUM Affinity column ligand

Fraction

Protvin (mg)

Trirhamnose

Applied Flowthrough Applied Bound Applied FIowthrough

nd " nd nd 556 55(} 4~)1

Monorhamnose Trirhamnose Not determined.

Antimonorhamnose bound (~) Ill{) 36.4 36.4 nd nd 28.2

Antitrirhamnose bound (%) I(X) 6.4 6.4 nd nd 0.1

71 Dorval, Canada) was used and after activation of this gel with divinylsulphone, a 20% w / v solution of L-rhamnose (Aldrich Chemical Co., Milwaukee, USA) in (}.5 M Na2CO3/NaHCO 3 buffer, pHI(} was added.

town, Canada) by the periodate method (Wil.~m and Nakane, 1978). Conjugates containing 2 mol enzyme per tool antibody were prepared and stored as a 50% glycerol .solution in PBS at - 20°C at a concentration of I mg/ml.

Affinity purification of the polyclonal antibody

ELISA procedures

Two affinity columns used in succession were employed to purify a monorhamnose binding/trirhamnoside non-binding polyclonal antibody fraction from immune sheep serum. Antibody binding to trirhamnoside was first removed by passage of 250 ml of the immune sheep serum through a 3.5 ml column containing trirhamnoside units, equilibrated in 0.05 M NazHPO 4, pH 7.4, I).30 M NaCI and I).01% NaN 3. The non-binding fraction was then passed through a column containing rhamnose monosaccharide units equilibrated in the same buffer. An antibody fraction which bound to the second column was eluted in buffer containing 0.1 M rhamnose and dialysed to eliminate all free rhamnose. A small amount of residual trirhamnoside binding antibody was removed from the dialysed fraction by passing it a second time through the trirhamnoside gel. The flowthrough fractions of the second trirhamnoside column were combined, adjusted to 50% glycerol and stored at -20°C. ELISA measurements using trirhamnoside-BSA and rhamnoside-BSA coated microtiter plates were used to determine the recovery of antibodies from the affinity columns (Table !).

Microtitcr plates (Immulon 4, Dynat~ch Laboratories, Chantilly, USA) were coated with antibody or antigen in tl.10 M NazCO.~/NaHCO.+ buffer, pH 9.6, Gvernight at 4°C. The plates were then blocked fi~r I h at 37°C with PBS containing 0.05% v / v Twecn 20 (PBST), washed three times with PBST followed by three washes with PBS and dried at 37°C for I h. To determine antibody binding to antigen, IIH} p.I aliquots of a 2 / ~ g / m l .solution of polysaccharidc-BSA conjugate were coated. Antibody diluted into PBST (l(X}/zl) was pipetted into each well and after 3(1 min the wells were washed five times with PBST. Rabbit anti-sheep lgG (Ill) ttl) Fc specific (Jackson Immunoresearch Laboratories, West Grove, USA) was added and after 30 min the plates were again washed five times with PBST. A hydrogen peroxide/tetramethylbenzidine substrate for peroxidase, I(X}p.I (Bos et al., 1981) was added to each well and the absorbance at 650 nm was read after 15 min. To test oligosaccharide inhibitors of antibody binding, the antibody (ILK}/.d) and inhibitor (100 /zl) in PBST were premixed in a low binding type microtiter plate (Linbro/Titertek EIA, Flow Laboratories, McLean, USA) for ! h at room temperature. A 10(} ttl volume of this inhibitor/ antibody solution was then transferred to GBSPBSA conjugate coated microtiter wells and incubated for l h. Peroxidase conjugated rabbit antisheep and enzyme substrate were then added as described above. To detect GBSP, microtiter plates were coated with 2l•} ~l of a 5/.tg/ml .solution of the trirhamnosidc recognizing monoclonal antibody. Antigen in 200 p.! of PBST was applied to each well for 31) min then the wells were washed five times with PBST. Pcroxidase-conjugated second antibody diluted l/lll,IHI0 in PBST was added to each well and after 30 min the wells were again washed five times with PBST. Enzyme substrate was added and the absorbance read as described above.

Gel filtration chromatography A gel filtration column 11.6× !111) cm) of Sephacryl S-3110 superfine (Pharmacia (Canada), Dorval, Canada) equilibrated in 0.05 M Hepcs buffer pH 7.4, containing 11.311M NaCI and 0.ill% NaN 3 was calibrated using gel filtration standards (Bio-Rad Laboratories, Mississauga, Canada). Purified antibody, 13.8 mg, was passed through the gel at 5 m l / h , 4°C, and the elution volume of the protein, determined by reading the absorbance at 281) nm, was used to estimate its molecular weight.

Conjugation of antibodies to peroxidase Antibodies were conjugated to horseradish peroxidase (Diagnostic Chemicals. Charlotte-

To detect group B streptococci, the bacteria were first nitrous acid extracted in the wells of a low binding microtiter plate (Linbro/Titertek EIA), Bacteria in 200 p,I of 0.15 M NaCI, 25 p,i of 5.6 N sodium nitrite and 25 ~! of 1.4 N acetic acid were combined for 15 rain and t',en neutralized with 25 #,l of 1.4 N Tris chloride, pH 10. A 200/zi aliquot of this extract was transferred to antibody-coated microtiter plates to detect group B polysaccharide as described above.

Preparation of bacteria A panel of bacteria including representative species of the vaginal flora (Bartlett and Polk, 1984) (see Table l i d were purchased from the lnstitut Armand-Frappier, Laval, Canada. Dilutions of 10~ cfu/ml in 0.15 M NaCI were prepared by the Macfarland optical density method. Dilutions of group B streptococcus at l09 c f u / m l were prepared by the colony count method. The bacteria were killed by addition of 2% final concentration of formalin and stored at - 80°C until used. The serotypes of GBS prepared were; Streptococcus agalactiae la (ATCC 12400), Ib (ATCC 12401), ll (ATCC 12973), III (ATCC 12403).

Results

Characteristics of a single-site ELISA based on the trirhamnoside epitope. Detailed structura~ studies of the group B streptococcal polysaccharide antigen found it polyvalent for a terminally located trirhamnoside (Michon et al., 1988). A double antibody sandwich ELISA was set up using a monocional antibody (MAb) specific for the trirhamnoside as both capture and peroxidase labeled second antibody. Varying the concentration of the MAb in the solution used to coat microtiter plates revealed that this ELISA responded in a biphasJc manner to the density of antil~ ~dy on the solid phase (Fig. 1). The EL1SA response increased proportionally to the concentration of coating MAb up to 160 ng antibody/100/~i/wei|, beyond which the response decreased, rather than exhibiting the usual plateau that results from the microtiter plate beconting saturated. At a concen-

E 1. 1.2 0 o 0.8

f

~

c 0.6

0.2 0.0

0

100 200 300 400 500 Antibody (ng/well)

Fig. 1. ELISAfo, -fion of group B-specificpolysaccharide antigen with the trh. .moside-specificmonoclonal antibody as both capture antibody and peroxidase-labeledsecond antibody. The amount of antibody used to coat each microtiter well is in6icated and the amount of group-specificpolysaccharide antigen added per well was: o, 0 ng; a. 0.37 ng; v, 1.11 ng; v, 3.33 ng, D, 10ng; II,100 rig. tration of 500 ng capture antibody/100 /.H/well no ELISA response was observable with the clinically significant antigen concentration of 11 n g / m l (Rench et al., 1984) or less. This biphasic response of the ELISA to capture antibody coating was likely due to closely spaced capture antibodies binding all of the trirhamnoside epitopes of each polysaccharide antigen. Higher antigen concentrations were found to overcome the loss of ELISA response (Fig. 1) probably by creating a situation of antigen excess. We concluded that the terminal trirhamnoside of GBSP is not ideal for use as a target epitope in a single-site ELISA in spite of it being present in four copies per GBSP molecule (Michon et al., 1988).

Isolation by affinity chromatography of an antibody recognizing monorhamnoside epitopes We tested whether the trisaccharide might be more effectively used in a ELISA, paired with a second distinct Terminal monorhamnose was chosen as ond site because a-L-rhamnopyranoside

epitope two-site epitope. this secresidues

are known to be important sites o f antibody recognition of this polysaccharide (Curtis and Krause, 1964; Heidelberger et al., 1967). To p r e p a r e an antibody specifically directed against these terminally located r h a m n o s e epitopes immune s h e e p serum was passed through an affinity column containing trirhamnoside units which removed 94% o f the anti-trirhamnoside binding activity but only 64% o f the anti-monor h a m n o s e binding activity (Table !). Subsequent passage o f trirhamnoside absorbed serum through a m o n o r h a m n o s e affinity column absorbed antibody which was later eluted by addition o f 0.1 M L-rhamnose to the buffer. Repassage of the dialysed eluate o f the m o n o r h a m n o s e column through the trirhamnoside column r e d u c e d the residual anti-trirhamnoside activity to 0.1% from 6.4% (Table !). T h e final polyclonal antibody thus interacts with m o n o r h a m n o s e but not with the trirhamnoside epitopes. A n inhibition E L I S A technique was employed to f u r t h e r define the

specificity o f the m o n o r h a m n o s c binding polyclonai antibody (Fig. 2). T h e most effective inhibitors of its binding were the oligosaccharides derived from G B S P which contained the a-LR h a p ( l -~ 3)D-glueitol moiety (Fig. 2, nos. !, 5, 6 and 9 and those containing the a-L-Rhap(I---, 3)a-D-Galp moiety (Fig. 2, nos. 5 and 7). Inhibition by the trirhamnoside, dirhamnoside, monorhamnoside or methyl glycoside o f rhamnose (Fig. 2, nos. 3, 4, IO and I l) was about two orders o f magnitude less effective than the oligosaccharides containing a - L - R h a p ( l - , 3 ) D - g l u c i t o l and a-LR h a p ( l --, 3 ) a - D - G a l p moieties. T h e inability o f this antibody to be retained by the column containing trirhamnos¢ residues further illustrates its poor recognition of the trirhamnoside epitope.

Characteristics of a two-site ELISA based on trirhamnoside and monorhamnoside epitopes A two-site E L I S A set up with the monoclonai an'i-trirhamnosid¢ and the polyclonal anti-mono-

TABLE II STRUCTURE OF OLIGOSACCHARIDE INHIBITORS (I) (2) (3) (4) (5)

aq.-Rhap I ~ 2 a-L-Rhap 1 -~ 2 a-L-Rhap I ~ I' I)-glucitol3'-1 vt-L-Rhap aq.-Rhap I -, 2 ol-L-Rhap I --, 2 Ct-L-Rhap 1 --, I' D-glucitoI a-L-Rhap l --, 2 a-L-Rhap I -, 2 a-l.-Rhap OMe a-t.-Rhap I ~ 2 a-t.-Rhap aq.-Rhap l -'-,2 a-L-Rhap I --*2 a-L-Rhap I ~ l' l)-glucitol 3'-I a-L-Rhap 4 T I

(6)

(7) (8) (9) (lid (11) (12)

/3q~-GlcpNAc 3 T I a-D-Galp 3 T l aq.-Rhap a-L-Rhap I --~2 a-L-Rhap I --, I' D-glucitol3'-I aq,-Rhap 4 I flq)-GlcpNAc 3 T I a-D-Galp aq.-Rhap I ~ 3 a-D-Galp 1 ~ 3/3-D-GIcpNAc I --, 4 a-I.-Rhap OMe otq.-Galp I ~ 3/3-D-GIcpNAc 1 -, 4 aq,-Rhap OMe D-glucitol3'-1 a-t.-Rhap L-Rhap a-L-Rha OMe D-glucost~

100

5 _.4m4

9

1

11

.;2o

.0//;7 ~5

2.5

20 ~ , ~ ~ ,

o

60

in,, ~l--'-Bj

v / v - v ~ 1.0

20

0.5

.v.

,EL

12 0.0

0

1

2

3

4

i

5

Fig. 2. Inhibition of the binding of an affinity purified sheep antibody to group B-specific polysaccharide. The structures of the oligosaccharides are shown in Table II. rhamnoside antibodies exhibited a normal response to capture antibody with a plateau o f maximum activity corresponding to the highest possible capture antibody coating (Fig. 3). Terminal m o n o r h a m n o s i d e is a structural feature o f many bacterial polysaecharides and so the polyclonal antibody against this epitope was not used as the capture antibody in o r d e r to retain specificity for group B streptococcus in the capture step. F u r t h e r studies t h e r e f o r e focused on the two-site E L I S A with the monoclonal anti-trirhamnoside as capture agent and the polyclonal a n t i - m o n o r h a m n o s i d e as probe antibody. T h e two-site E L I S A for group B streptococci bacteria was sensitive enough to detect 3 × l04 cfu (Fig. 4). W o m e n presenting this n u m b e r o f bacteria p e r cervical/vaginal swab are in a high risk group for intrapartum infection o f their infants with GBS (Morales and Lira, 1987). A n evaluation o f the specificity of the two-site E L I S A was carried out using a library o f bacteria representative o f the vaginal tiara (Bartlett and Polk, 1984). W h e n bacteria were tested at a concentration o f 2.4 x 107 c f u / t e s t , which is a 1000 fold excess over the clinically relevant n u m b e r for group B streptococcus (Morales and Lim, 1987), only group B streptococcus gave a positive reaction (Table Ill). Various serotypes o f GBS, Strep-

,

0

Oligosaccharide (log uM)

I

,

100

I

,

200

I

,

300

I

,

400

I

500

Antibody (ng/well) Fig. 3. ELISA lot detection of group B-specific polysaccharide antigen with the trirhamnoside-specific monoelonal antibody as capture antibody and the affinity purified monorhamnosespecific polyclonal antibody labeled with peroxidase used as the second antibody. Other details are as described in Fig. 1.

E t:::

o

g

0.6

0.4

o

,~

0.2

.

0.0 .

0.5 .

. 1.0

.

1.5

2.0

i

25

i

30

GBS (,10 s c f u per t e s t )

Fig. 4. Sensitivity of the two-site ELISA employing the trirhamnoside-specific monoclonal antibody as capture antibody. The serotype Ill group B streptococcus cells were extracted as described in the materials and methods section and transferred to microtiter wells coated with the trirhamnose specific monoclonal antibody.

TABLE III SPECIFICITY OF THE TWO-SITE ELISA Bacteria

Source

Candida albicans Escherichia coli Eubacterium lentum Gardnerella ¿'aginalis Klebsiellapneumoniae Lactobacillus fermentum Neisseria gonorrhoeae Peptostreptococcus anaerobus P*'eudomonas aeruginosa Staphylococcus epidermis Streptococcus A Streptococcus B Streptoccccus equi Streptococcus pneumoniae

ATCC 111231" K-12 (C-6001 LSPQ h ATCC 141118 ATCC 13883 LSPQ ATCC 19424 LSPO ATCC 27853 ATCC 12228 ATCC 19615 ATCC 12403 LSPQ ATCC6305

Absorbance 16511nm) 0.11111 0.1118 (1.1X18 I).11113 1).01141 {}.{lt)I 11.[1112 (I.II~X) 0.046 I).U00 ().(X)I 1.5111) 11.001 0.(1~1~)

~' American Type Culture Collection number. h Laboratoire de Sant~ Publique du Quebec, Ste-Anne-deBellevue, Quebec, Canada. Bacteria were extracted as described in the materials and methods section. An aliquot equivalent to 2.4× I[17 cfu of each extract (lUll /.d), was transferred to microtiter wells coated with 1.11p.g of the trirhamnoside-specificmonoelonal antibody. The peroxidase-conjugatedsecond antibody was the monorhamnose epitope-specificpolyclonal antibody.

tococcus agalactiae la, Ib, 11 and 111 were tested

and each produced a positive response (results not shown).

Discussion The group B streptococcal polysaccharide consists of a complex, multiantennary structure expressing serologically dominant epitopes corresponding to the terminal trirhamnoside of each branch and to terminal monorhamnoside residues of various side chains (Michon et ai., 1991). The presence of this antigen in clinical samples is indicative of infection by this organism. When assaying for such a multi-epitoped analyte it is possible to use a single-site ELISA which requires only a single antibody specificity, thus simplifying develor~ment and reducing productio~ costs. A single-site ELISA based on the trirhamnoside epitope of GBSP was found in this study not to be feasible as it was adversely affected by

high capture antibody coating densities. It appears that as the monoclonal anti-trirhamnoside antibody was coated at increasing density onto the plastic surface of the microtiter plates, it eventually became closely enough packed that capture of virtually all trirhamnoside epitopes took place. This resulted in a loss of response because the peroxidase-conjugated second antibody, of the same specificity, had few or no sites at which to attach itself. A single-site ELISA could conceivably still be setup with this trirhamnoside epitope, however, antibody coating would require strict control to maintain activity and to avoid large test result variation. A more ~ r i o u s disadvantage of the trirhamnoside-based singlesite ELISA is that the capture antibody could not be coated to maximum levels and therefore a reduced test sensitivity would result. Based on these observations, when analysing for GBSP using antibodies against the trirhamnoside epitope, a two-site ELISA is a better test design. In this design the multisite capture of GBSP, a detriment to the single-site ELISA, becomes an advan*9~e as a result of the increase in functional affinity of the capture step which would result (Hornick and Karush, 1972; Devey and Steward, 19881. The choice of specificity of the second antibody for use in tandem with the anti-trirhamnoside antibody was made to maximize the amount of enzyme-conjugated second antibody that can attach to each molecule of GBSP captured by the solid phase. An abundant structural feature of GBSP is the terminally located monorhamnose residue, whi~.h . :urs in three types on GBSP (Michon et al., 1988) a-L-Rhap(l --* 3)a-D-Galp, a-L-Rhap(! ~ 3)D-glucitoi and a-L-Rhap(I --* 2)a-L-Rha. Antibodies against the latter structure, found as part of the trisaccharide recognized by the capture monoclonal antibody, were removed from the polyclonal antibody preparation by passage of the immune sheep serum through a column containing trirhamnoside units. Antibodies directed against the remaining two types of terminal rhamnose structures were retained by a column containing monorhamnose residues. Even though free rhamnose was able to displace these antibodies from the affinity column, the analysis of their specificity summarized

in Fig. 2, showed that r h a m n o s e was in fact a relatively p o o r inhibitor of their binding to the whole polysaccharide. T h e most potent inhibitor was the oligosaccharide containing both types of terminal r h a m n o s c , t ~ - L - R h a p ( 1 - , 3 ) a - D - G a l p and a-L-Rhap(l--*3)D-glucitol (Fig. 2, no. 5), The disaccharide a - L - R h a p ( l --, 3)D-glueitol (Fig. 2, no. 9) was however a very effective inhibitor as well, being about 100-fold m o r e effective than the free rhamnose. This suggests that the actual epitope which was used as the second site in the two-site E L I S A is at least the disaccharide terminating in r h a m n o s e r a t h e r than the m o n o r h a m nose itself. This being the case the second antibody may be more specific for G B S than would an antibody recognizing only terminally located r h a m n o s e residues. T h e use of antibodies with high specificity for GBS as both p r o b e and capture antibody would increase the overall tegt specificity. This study d e m o n s t r a t e s the utility of the recently defined trirhamnoside epitope of the g r o u p B streptococcal antigen for diagnosing infection by g r o u p B streptococci.

Acknowledgements We thank Dr. R, Feldman for providing the hybridoma for the antibody against the a-L-(1 2)-trirhamnopyranoside epitope on the g r o u p specific polysaccharide of g r o u p B streptococci.

References Bartlett, J.G. and Polk, B.F. (1984) Bacterial flora of the vagina: Quantitative study. Rev. Infect. Dis. 6, $67. Bos, E.S., Van der Doel~ n A.A.. Van Rooy, N. and Schuurs, A.H,W.M. 11981) 3,3",'~,5'-Tetramethylbenzidine as an Ames test negative chromogen for horse-radish peroxidase in enzyme immunoassay.J. Immunoassay2, 187. Costa-Cruz. O.. Sesso, A.M.. Shrikrishna, M. and Frank, E. 11990) Pasteurella multocida meningitis. NJ Med. 87, 127. Curtis. S.N. and Krause, R.M. 11964) Antigenic relationships between groups B and G streptococci. J. Exp. Med. 120, 629.

Devey, M.E. and Steward, M.W. 11988) The role of antibody affinity in the performance of solid phase assays. In: D.M. Kemeny and S.J. Challacombe (Eds.), ELISA and Other Solid Phase lmmunoassays.Wiley, New York. p. 135. Feldman, R.G., Law, S.M. and Salisbury, J.R. 11986) Detection of group B streptococcal antigen in necropsy specimens using monoclonal antibody and immunoperoxidase staining. J. Clin. PathoL 39, 223, Fornstedt, N. and Porath, J. 11975) Characterization studies on a new lectin found in seeds of Vicia errilia. FEBS Lett. 57, 187. Greenspoon, J.S., Fishman, A.. Wilcox, J.G., Greenspoon, R.L. and Lewis, W. (1991) Comparison of culture for group B streptococcus versus enzyme immunoassay and latex agglutination rapid tests: results in 250 patients during labor. Obstet, Gynecol. 77, 97. Heidelberger, M., Davie, J.M. and Krause, R.M. (1967) Crossreactions of the group-specific polysaccharides of streptococcal groups B and G in anti-pneumococcal sera with special reference to type XX]I1 and its determinants. J. lmmunol. 99, 794. Hoppe, J.E., Lindenau, C. and Hffler. W. (1989) Rapid detection of group B streptococci in vaginal swabs of parturients by latex particle agglutination. Zbl. Bakt. Hyg. A 270, 379. Hornick, C.L. and Karush. F. (1972) Antibody affinity. Ill. The role of multivalence. Immunochemistry9, 325. Michon, F., Brisson, J.-R., Dell, A., Kasper, D.L. and Jennings, H.J. (1988) Multiantennary group-specific polysaccharide of group B streptococcus. Biochemistry, 27, 5341. Michon, F., Chalifour, R., Feldman, R., Wessels, M., Kasper, D.L., Gamian, A., Pozsgay, "~'. and Jennings, H.J. (1991) The a-L-(l ~ 2)-trirhamnopyranoside epitope on the group-specific polysaccharide of group B streptococci. Infect. Immun. 59, 1690. Morales, W.J. and Lira, D. 11987) Reduction of group B streptococcal maternal and neonatal infections in preterm pregnancies with premature rupture of membranes through a rapid identification test. Am. J. Obstet. Gynecol. 157,13. Rench, M.A., Metzger, T.G. and Baker, C.J. (1984) Detection of group B streptococcal antigen in body fluids by a latex-coupled monoclonal antibody test. J. Clin. Microbiol. 21l, 852. Skoll, M.A., Mercer, B.M., Baselski, V.. Gray, J.P., Ryan, G. and Sibai, B.M. 11991) Evaluation of two rapid group B streptococcal antigen tests in labor and delivery patients. Obstet. Gynecol. 77, 322. Wilson, M.B. and Nakane, P.K. 11978) In: W. Knapp, H. Holubar and G. Wick (Eds.), Immunofluorescence and Related Techniques. Elsevier, Amsterdam, p. 215. Yawn, B.P., Yawn, R.A. and Henning, T. 11990) Evaluation in rural practice of a rapid group B streptococcus screening test. Faro. Med. 22, 122.