Research in Veterinary ,Science; 1992, 53, 300-308

Characterisation of monoclonal antibodies against a fimbrial structure of Salmonella enteritidis and certain other serogroup D salmonellae and their application as serotyping reagents C. J. THORNS, M. G. SOJKA, I. M. MCLAREN, M. DIBB-FULLER, Ministry of Agriculture, Fisheries and Food, Central Veterinary Laboratory, Weybridge, Surrey KT15 3NB

A panel of 13 monoclonal antibodies from different hybridomas was produced against a novel salmonella fimbrial antigen expressed predominantly by Salmonella enteritidis strains. The specificity of the monoclonal antibodies to this antigen (SEF14) was confirmed by enzyme-linked immunosorbent assay (ELISA)using purified SEFI4, immune electron microscopy and, with 11 monoclonal antibodies, the identification of a repeating protein subunit (14,300kDa) on the antigen. Blocking-ELISAwith the monoclonal antibodies identified epitopes in at least three, non-overlapping clusters which appeared evenly distributed on SEF14 in immune electron microscopy. The use of the monoclonal antibodies in direct-binding ELISA on a range of salmonella serotypes suggested that the epitopes on SEF14 are highly conserved and were expressed by all the S enteritidis strains examined; some strains of S dublin and the only strain of S moscow available were the only other serotypes that expressed SEF14. A latex agglutination reagent based on a monoclonal antibody was developed and used to test for SEF14 on 280 strains (representing 120 serotypes in 24 serogroups of salmonellae) that had been grown on Sensitest agar for 18 hours at 37°C. All S enteritidis strains (64) and most S dublin strains (28 of 33) produced SEF14 as did the two strains representing S blegdam and S moscow. SEF14 was not detected in any other strains of serotypes from serogroup D or from any other serogroup examined.

molecular weight of 14,300 (Thorns et al 1990). Mt~ller et al (1991) refer to this fimbrial structure as SEF14 and the same term will be used herein. So far SEF14 has been detected only on S enteritidis strains (58 out of 58), S dublin strains (12 out of 36) and on a single strain of S moscow, that is serotypes within serogroup D. To date, 169 isolates from 17 other salmonella serogroups have not been found to express this antigen (Thorns et al 1990). S enteritidis has become the dominant serotype isolated from cases of food poisoning in the United Kingdom and the United States, and this phenomenon is generally thought to be associated with an increase in poultry products infected with this serotype (O'Brien 1988, St Louis et al 1988). The discovery of the fimbrial structure, which is expressed predominantly by S enteritidis, has important implications in understanding the pathogenesis of these organisms and might in turn lead not only to the design of novel oral vaccines against salmonelloses, but also to the development of rapid assays for detecting speciesspecific antigen and, or, antibody. This paper describes the production and characterisation of monoclonal antibodies directed towards SEF14 fimbriae and their evaluation in latex agglutination tests for the rapid identification of cultured S enteritidis.

A NOVEL fimbrial structure on the surface of Salmonella enteritidis has recently been described (Thorns et al 1990, Mt~ller et al 1991). The structure comprises thin, filamentous organelles, less than 5 nm in diameter, which carry a major protein that consists of repeating subunits with a

Materials and methods

Bacteria and media The serotype and number of Salmonella strains used to characterise the monoclonal antibodies or for testing by latex agglutination are listed in

300

Monoclonal antibodies against S enteritidis TABLE 1: Salmonella strains examined by direct EUSA or latex agglutination with monoclonal antibodies

Serogroup

Serotype (number of strains tested)

S agama (1) S agona (3) S bredeney (2) S califomia (1) S chester(I) S coeln (1) S derby (2) S heidelberg (4) S indiana (3) S massenya ( 1) S reading (2) S saint paul (1) S san diego ( 1) S schwarzengrund (1) S stanley (1) S stanleyville (1) S typhimurium(23) C ........ S amersfoort (1) S bareilly (1) S brandenberg (1) S infantis (3) S hartford (1) S lille (1) S fivingstone (2) S mbandaka (2) S montevideo (2) S oakland (1) S ohio (2) S oranienburg (2) S oslo(1) S singapore ( 1) S tennessee (6) S thompson (1) S virchow (2) C2 ........ Sbovismorbificans(1) S goldcoast (1) S hadar (3) S kottbus (1) S manhattan (1) S meunchen (1) S nagoya (1) S newport (2) C3 ........ Salbany(1) S bardo (1) S emek (1) S haardt (1) S kentucky (1) S molade (1) S tado (1) D1 ........ S berta (6) S blegdam (1) S canastel (1) S dublin (33) S durban (1) S eastboume (1) S enteritidis (64) Sfresno(1) S gallinarum (3) S kapemba (1) S miami (1)

Serogroup

B .........

Total

=

280 strains

E1 ........

E2 ........

E3 ........ E4 ........

F .........

G1 ........ G2 ........

H ......... I .........

K .........

M ......... N ......... 0 .........

Q .........

R ......... S ......... T ......... X ......... Y .........

Serotype (number of strains tested) S moscow (1) S napofi (1) S ouakam (1) S pullorum (1) S wangata (1) S amsterdam (1) S anatum (3) S butantan (1) S falkensee (1) S lexington (1) S london (1) S meleagridis (2) S meunster (1) S nchanga ( 1) S orion (3) S regent (1) S uganda (1) S vejle (1) S weltevreden (1) S westhampton (1) S binza (2) S drypool (1) S manila (1) S newington (1) S wildwood(I) S liverpool (1) S Ilandoff (1) S senftenberg (3) S taksony (2) S bullbay(1) S chandans (1) S telhashomer (1) S havana (1) S poona (1) Sajiobo(1) S eubana (1) S idikan (1) S kedougou (2) S fischerkeitz(1) S chameleon (1) S gaminara (1) S tees (1) Scerro(1) Spomona (1) S godesberg (1) S urbana (1) S adelaide (1) S alachua (1) Sealing (1) S widemarsh (1) S anfo (1) S wandsworth (1) Sjohannesberg (1) S millesi ( 1) S omifisan (1) S offa (1) Sgera(1) S toricada (1) S bergen (1) S marina (1)

301

Table 1 and were obtained from the reference collection at the Central Veterinary Laboratory, Weybridge. Recent isolates of Citrobacter diversus, Escherichia coli, E hermannii, Proteus vulgaris and Yersinia ruckerii, all from the field, were also examined. All bacterial strains were stored on Dorset egg slopes. To characterise the monoclonal antibodies the bacteria were grown in peptone water (Oxoid, CM10) for 18 hours at 37°C. In addition, strains of S enteritidis were grown in different media to determine the optimum medium for expression of the SEF14 antigen. The liquid media used to grow strains were: enriched E broth (Francis et al 1982), heart infusion broth (Oxoid, CM225), Minca broth (Guinee et a11976), peptone water pH 6.0 and 7.2, Slanetz broth (Ness 1983) and Vogel Bonner medium. The following solid media were also used: bismuth sulphite agar (Difco, 0073), brilliant green agar (Oxoid, CM263), desoxycholate citrate agar (Oxoid, CM35), MacConkey agar (Oxoid, CM7), nutrient agar (Oxoid CM3), Salmonella Shigella agar (Oxoid, CM99), Sensitest and Isosensitest agar (Oxoid, CM409, CM471), 5 per cent sheep blood agar (Difco, 0045), and xylose lysine desoxycholate agar (Difco, 0788). Strains were cultured in liquid or solid medium for 18 hours at 37°C.

Antigen purification Fimbrial antigens were prepared from S enteritidis phage type 4 strain 1246/89 which has previously been identified as expressing large quantities of the fimbriae (Thorns et al 1990). The protocols for the isolation and purification of the fimbriae will be described in detail elsewhere (M. G. Sojka and C. J. Thorns, unpublished data). Briefly, the organisms were grown on Sensitest agar overnight at 37°C. Bacteria were sedimented and.suspended in phosphate buffered saline (PBS, pH 6.8). The fimbriae were then removed from the surface of the bacteria by heating the suspension at 60°C for 30 minutes. The cell-flee supernatant (crude SEF 14) was first purified by anion exchange chromatography with DEAE-sepharose (semi-pure SEF14), followed by size exclusion using high-pressure liquid chromatography (pure SEF 14). The purity of the respective SEF14 preparations was determined by SDSPAGE using 12.5 per cent gels.

302

C.J. Thorns, M. G. Sojka, L M. McLaren, M. Dibb-Fuller

TABLE 2: Properties of the 13 monoclonal antibodies (MAb) specific to the SEF14 fimbriae of Salmonella species Immunising antigens

MAb

lsotype

Immunoblot of the 14300-molecular weight SEF14

tEM*

Relative affinity1"

Epitope cluster

Whole S enteritidis strain 1246/89 cells

SEF14-1 SEF14-2 SEF14-3

IgG1 lgG~ IgG 1

+ + +

+ + +

++ ++ ++

1 1 1

Crude and semi-pure SEF14

SEF14-4 SEF14-5 SEF14-6

IgA IgA IgA

+ + +

+ + +

++++ ++++ ++++

3 3 3

Semi-pure and pure SEF14

SEF14-7 SEF14-8 SEF14-9 SEF14-10 SEF14-11

IgG 3 IgG 3 IgG 3 IgA IgA

+ + + + +

+ + + + +

++ ++ +++ +++ +++

2 2 2 2 2

SEF14-12 SEF14-13

IgM IgM

? ?

+ +

1 1

Semi-pure SEF14

o

* For details see text IEM Immune electron microscopy 1 Elution of monoclonal antibodies from SEF14 with 1M or less ammonium thiocyanate (NH4SCN) +; 1M NH4SCN ++; 2M NH4SCN +++; 3M NH4SCN ++++

Production of antibodies

New Zealand White rabbits were injected subcutaneously at two separate sites with 50 gg of purified SEF 14 emulsified in Freund's incomplete adjuvant. The injections were repeated seven and 21 days later and blood was collected 10 days after the final injection. The specificity of the antisera was checked by enzyme-linked immunosorbent assay (ELISA)and immune electron microscopy. A monoclonal antibody 69/25 (SEF 14-1) directed towards an epitope on SEF 14 has been described (Thorns et al 1990). Production of monoclonal antibodies

Female BALB/C mice were injected intraperitoneally with 50 gg of crude, semi-pure or pure SEF14 (Table 2). Booster injections were given four and eight weeks after the first injection and the spleens of the mice were removed three days after the final injection. Hybridomas were produced from fusion of the murine splenocytes with the BALB/Cmyeloma cell line NS1 as described previously (Morris et al 1985). Following fusion, the hybridomas were distributed into 96-well microplates (Nunclon, Roskilde) and regularly tested by ELISA(as described below) for antibody production. Antibody-secreting hybridomas were expanded, cloned and stored following protocols detailed elsewhere (Thorns et al 1990). Monoclonal antibodies from the cell lines, together with monoclonal antibody SEF 14-1, were pro-

duced from murine ascitic fluid and concentrated tissue culture supernatant. The antibodies were partly purified by precipitation with ammonium sulphate (40 per cent saturation) and dialysed against 0.01M PBSpH 7-2. The class and subclass of the antibodies were determined by immunodiffusion. Direct-binding ELISA

The initial screening of the hybridomas was carried out as previously described (Thorns et al 1990). Briefly, wells of polystyrene microtitre plates (Nunc) were coated with 25 ng of pure SEF14 antigen (diluted in 100 gl of 0.1 M phosphate buffer, pH 4.5) by incubation overnight at 37°C. The plates were then washed and blocked before the addition of the culture supernatants. To detect binding of the monoclonal antibody to the antigen, a peroxidase labelled anti-mouse immunoglobulin (Cappel Laboratories, Dynatech) followed by the chromogenic substrate tetramethylbenzidine (TMB) (Cambridge Veterinary Sciences) were added; the enzyme reaction was stopped with sulphuric acid and the optical densities of the reaction were recorded at 450 nm. Monoclonal antibodies from cloned hybridomas and rabbit antiserum against SEF14 (RaSEF14) were tested by ELISA for binding to a variety of Salmonella species and other bacteria (Table 3). Organisms grown in peptone water overnight were centrifuged at 3000 g for 10 rain-

Monoclonal antibodies against S enteritidis

303

TABLE 3: Direct binding of monoclonal antibodies to strains of salmonellae Serotype

Number of strains examined 1

2

3

4

5

Monoclonal antibodies 6 7 8

9

10

11

12

13

S enteritidis

10

69* 69 68 69 80 61 63 55 78 71 70 68 52 (40-89) (32-89) (33-88) (33-94)(43-100)(32-89)(28-94) (33-79)(41-100)(42-90)(43-93)(34-87)(28-73)

S dublin

5

24 21 24 25 33 22 18 23 30 29 32 30 17 (13-31) (12-30)(13-31) (13-33) (18-42) (12-29) (9-24) (12-33) (12-40) (16-38) (16-39) (12-40) (7-25)

S dublin

2

0

0

0

0

0

0

0

0

0

0

0

0

0

S moscow

1

49

47

44

46

57

41

37

39

52

53

49

42

31

Other salmonel[aet

9

0

0

0

0

0

0

0

0

0

0

0

0

0

Other bacteria:l:

5

0

0

0

0

0

0

0

0

0

0

0

0

0

* Binding of monoclonal antibodies relative to binding to 'high' control (see text) 1" Included: a strain of each of S berta, canastel, durban, gallinarum, ouakam, panama, typhimurium and wangata 1: See text for details 0 Range of binding

utes, and then suspended in pBs (pH 7.2) to give an absorbance of 1.25 at 400 nm. The wells of polystyrene microtitration plates were coated (100 gl per well) with the antigens in 0-1 M carbonate buffer (pH 9.6) by incubation overnight at 37°C. Optimum concentrations ofmonoclonal antibodies were then tested for binding to the bacterial antigens using the procedures described above. The findings were expressed as the percentage ofmonoclonal antibody binding to the test strains relative to the binding in the 'high' control, in which normal mouse serum (Miles Laboratories) was used in place of antigen. The extent of binding by RaSEF14 was detected by the addition of a biotinylated anti-rabbit immunoglobulin and biotinylated streptavidin-peroxidase complex (Amersham International).

times in PBST. Antibody binding was detected by adding TM~.

Immunoblot analysis Antigens containing SEF14 were transferred from an SDS-PAGE gel to a nitrocellulose membrane and reacted with the monoclonal antibodies following methods described previously (Thorns et al 1990).

Immune electron microscopy The binding of monoclonal antibodies and polyclonal RaSEF14 to strains of salmonellae was visualised by adding gold-labelled antiglobulins and viewed by the electron microscope as previously described (Thorns et al 1990).

Direct-blocking £LISA Monoclonal antibodies and RaSEF14 were conjugated to horseradish peroxidase (Sigma) by standard methods (Wilson and Nakane 1978). Wells of microtitre plates were coated with SEF 14 antigen, blocked and washed as described above. The concentration of all monoclonal antibodies was adjusted to twice the amount needed to saturate the antigen, and serial twofold dilutions Of monoclonal antibodies performed in PBS containing 0.05 per cent (v/v) Tween 20 (PBST), incubated for 30 minutes at 37°C and the plates washed six times in PBST. Optimum dilutions of the monoclonal antibody or RaSEF 14 conjugates were then added (100 gl per well) and incubated for 30 minutes at 37°C and washed a further six

Thiocyanate elution Elution ofmonoclonal antibodies from SEF 14coated wells by increasing concentrations of chaotropic thiocyanate ions (SCN-), was determined as a measurement of relative affinity (Macdonald et al 1988). Monoclonal antibodies were added to SEF14-coated wells (100 gl per well), incubated for 40 minutes at 37°C and then washed six times in PBST. Various molarities of ammonium thiocyanate were added (100 gl per well) to the wells and incubated for 15 minutes at room temperature and then washed six times in PBST. The effect of NH4SCN on monoclonal antibody binding was detected by adding goat anti-mouse immunoglobulin peroxidase conju-

304

C. J. Thorns, M. G. Sojka, I. M. McLaren, M. Dibb-FuIler

gate (Cappel Laboratories) for 30 minutes at 37°C and then TMB. The results are expressed as the lowest molarity of NH4SCN causing a 50 per cent reduction in binding ofmonoclonal antibody to SEF14 antigens.

(under 1M to 5M thiocyanate). In general, however, monoclonal antibodies with similar affinities originated from the same fusion. The specificity of all the monoclonal antibodies except SEF1412 and SEF 14-13 was confirmed by reacting them with the 14,300 SEF14 antigen in Western blots.

Latex agglutination Direct-binding ELISA

The monoclonal and polyclonal antibodies The binding of monoclonal antibodies to were coated on to latex particles using standard methods (Hechemy and Michaelson 1984). strains of salmonellae is summarised in Table 3. Briefly, an optimum concentration of antigen The 13 monoclonal antibodies bound strongly was added to a 10 per cent (w/v) suspension of to all the strains of S enteritidis and the single 0.8 ~m diameter blue latex particles (K080, strain of S moscow tested. Five Strains of S dublin Estapor, Rhone-Poulenc) and 0.1M glycine bound weakly to all the monoclonal antibodies, buffered saline (GBS,pH 8-2) in a ratio of approx- and two failed to react with any of the monoclonal imately 1:30:120 and incubated for two hours at antibodies used in this study. The monoclonal 37°C with constant, gentle rocking. The coated antibodies did not bind to any other strains repparticles were then washed and suspended in GBS resenting other Salmonella serotypes or to strains containing 0.1 per cent of fatty acids-free bovine of bacteria of related genera. Polyclonal serum albumin (Sigma) to a final concentration RaSEF14 reacted identically to the monoclonal of 0.25 per cent (w/v). The latex reagents were antibodies in the direct-binding ELISA. stored at 4°C. Control latex reagents were prepared by replacing antibodies with normal mouse Direct-blocking ELISA or rabbit serum. Tests were carried out by mixing Each monoclonal antibody was tested for its equal volumes (50 gl) of latex reagent and heavy, ability to block in serial twofold dilutions, the smooth suspensions of organisms (in GBS pH binding of peroxidase-conjugated monoclonal 8-2) prepared from bacteria grown on solid mediantibodies to epitopes on the SEF14 antigen. um or directly from overnight broth culture, on The results are expressed as the logarithm of the a disposable white plastic-coated card, rocking reciprocal of the highest dilution showing 50 per gently for up to four minutes and looking for cent blocking of the reaction compared with conany macroscopic agglutination. Autoagglutinajugated monoclonal antibody alone (Table 4). tion of the test suspensions was checked for by Monoclonal antibodies which showed reciprocal replacing the antibody-coated latex with the con- blocking were regarded as binding to identical trol latex. The performance of the reagents was or overlapping epitopes, while monoclonal antimonitored regularly using positive control prepa- bodies that did not block one another were rations of antigen in place of the test organisms. assumed to identify different, non-overlapping epitopes. Conclusions were not drawn on nonreciprocal blocking reactions since they may be Results caused by (i) affinity differences between two Monoclonal antibodies monoclonal antibodies; (ii) steric alteration of Twenty-seven cloned hybridomas from four monoclonal antibody-bound antigen affecting separate fusions secreted monoclonal antibodies binding of a second monoclonal antibody; or (iii) changes in the binding properties of conjuthat bound to the purified SEF14 antigen. Thirteen were selected for further study. Table gated monoclonal antibodies (Van Zijderveld et 2 gives details of their characteristics and the al 1989, Van Regenmortel 1990). Using the above criteria from 13 different monimmunising antigen used in their production. The relative affinities of the monoclonal antibodies oclonal antibodies, it can be concluded that they varied considerably, as indicated by the range indicate three distinct epitope groups or clusters of thiocyanate molarities capable of eluting the (Table 4). Single monoclonal antibodies identimonoclonal antibodies from SEF14 antigen fying the three clusters partly blocked the binding

Monoclonal antibodies against S enteritidis

305

TABLE 4: Direct-blocking ELISA of conjugated monoclonal antibodies ( M A b ) reacting with S E F 1 4 with all 13 monoclonal antibodies of the panel MAb number Conjugated MAb number

1

2

3

12

13

7

8

9

10 .

11 .

4 .

1

++*

++

++

++

++

++

++

.

2

++

++

++

++

++

-

-

++

3

++

++

++

++

++

.

.

.

.

.

.

.

. .

.

.

5

.

.

.

.

12

+

++

++

++

++

.

.

.

.

.

.

.

13

+

++

++

++

++

.

.

.

.

.

.

.

7

.

.

.

.

8

.

.

.

.

9

-

10

.

.

.

.

.

11

.

.

.

.

.

++

-

. .

-

++

-

6

.

.

++

+

++

++

++

-

-

-

++

++

++

++

++

-

-

-

++

++

++

++

++

++

-

-

++

++

++

++

++

++

-

++

++

++

++

++

-

-

4

.

.

.

.

.

.

.

.

++

++

++

++

++

5

.

.

.

.

.

.

.

.

+

+

++

++

++

6

.

++

++

++

.

.

.

.

.

.

It

.

.

. 2

3

* ++ Titre 0.48 or more; + Titre 0.18 or more; - Titre 0 1" E p i t o p e c l u s t e r n u m b e r

of polyclonal RaSEF14 to the antigen (45 per cent or less). When the monoclonal antibody that indicates cluster 3 was combined with monoclonal antibodies from cluster 1 or cluster 2 there was increased blocking of RaSEF-14 (45 to 55 per cent). Monoclonal antibodies representing the three cluster groups blocked the RaSEF14 by 70 per cent.

Immune electron microscopy studies Specific, immunogold labelling of SEF14 occurred with all the monoclonal antibodies (Table 2) and RaSEF14. No difference in intensity or distribution of gold particles labelling SEF 14 was apparent when monoclonal antibodies from different epitope groups were tested; the gold was distributed evenly throughout the SEF14 antigen in all cases and was similar to the labelling of the fimbrial antigen with RaSEF14.

Antibody-coated latex reagents These reagents were tested for their ability to agglutinate with cell-free SEF 14 antigen and with a panel of salmonellae and other related bacteria. The latex reagent coated with SEF14-9 monoclonal antibody was specific and the most sensitive (data not shown). It gave identical results to latex coated with RaSEF14 antibodies and was used for all further studies.

Use of latex particle agglutination to measure effect of growth media on expression of SEF14 The expression of SEF 14 antigen by S enteritidis grown in different culture media is detailed in Table 5. The six S enteritidis strains used were selected to represen L respectively, high and low producers of the antigen when the organisms were grown in peptone water (pH 7.2) at 37°C. Peptone water (pH 7.2) and enriched E broth were the only liquid media in which SEF14 antigen was detectably expressed by all six strains (Table 5). However, the strains grown in peptone water were agglutinated more strongly than when grown in enriched E broth. Conversely, when the strains were grown in Minca medium, Vogel Bonner and Heart infusion broth very little SEF14 could be detected, as evidenced by there being little or no agglutination with the latex reagent (Table 5). The addition of 0.1 per cent (w/v) glucose to all the liquid media reduced considerably the production of SEF14 antigen by the strains All six strains ofS enteritidis grown on nutrient agar and 5 per cent sheep blood agar agglutinated with the monoclonal antibody-coated latex, but the strains grown on Sensitest or Isosensitest agar were agglutinated most strongly (Table 5). When strains were cultured on media commonly used for isolation and selection of salmonellae the expression of SEF 14 antigen was either reduced or in the case of brilliant green and bismuth sulphite agars completely inhibited (Table 5).

306

C. J. Thorns, M. G. Sojka, I. M. McLaren, M. Dibb-Fuller

TABLE 5: Production of SEF14 fimbrial antigen by strains of Salmonella enteritidis grown on different media and tested by latex reagent Strain G r o w t h medium Liquid Enriched E broth Heart infusion broth M i n c a broth P e p t o n e water pH 7.2 P e p t o n e w a t e r pH 6.0 Slanetz Vogel B o n n e r Solid BNliant green Bismuth sulphite D e s o x y c h o l a t e citrate MacConkey Nutrient Salmonella Shigella Sensitest (Isosensitest) S h e e p blood Xylose lysine d e s o x y c h o l a t e

486/86

1925/86

+ ++ ++ ++ ++ +

+ + . ++ ++ ++ .

. . ++ ++ ++ ++ +++ ++ ++

. .

2531/86

+ .

.

.

.

562/89

+ +

+ +

++ ++ +

++ ++ +

++ ++ ++ ++ +++ ++ ++

++ ++ ++ ++ +++ ++ ++

. ++

. . . ++ + ++ ++ +++ ++ ++

3739/86

+ -

++ ++ +

. . ++ ++ ++ ++ +++ ++ ++

3583/86

. . .

. . .

+ + ++ ++ +++ ++ +

+ Agglutinated 3 to 4 minutes; ++ Agglutinated 1 to 3 minutes; + + + Agglutinated 1 minute or less

Detection of SEF14 antigen on strains of salmonellae

Two-hundred-and-eighty strains representing 120 serotypes from 24 serogroups were grown on Sensitest agar for 18 hours at 37°C, and examined by latex agglutination for the production of SEF14 antigen (Table 6). All the S enteritidis (64) and most S dublin strains (28 of 33) tested agglutinated the latex reagent. The single strains representing S blegdam and Smoscow also agglutinated the reagent. No other strains from serotypes within serogroup D or any other serogroup examined agglutinated the latex. Discussion

A panel of 13 monoclonal antibodies has been produced that are directed against the SEF14 fimbriae first described by Thorns et al (1990). TABLE 6: Detection of SEF14 fimbrial antigen on Salmonella strains by the latex agglutination test Serotype

S S S S

enteritidis dublin blegdam moscow

O t h e r salmonellae*

N u m b e r of strains examined

L a t e x agglutination test + -

64 33

64 28

O 5

1 1

1 1

O O

181

0

* T h e s e serotypes are listed in Table 1

181

Direct-binding ELISA to the purified antigen, Western blotting or immune electron microscopy confirmed the specificity of the monoclonal antibodies used in this study. SEF14 antigen was originally identified with monoclonal antibody SEF 14-1 which was shown to be directed against an epitope on the fimbria that appeared to be highly conserved when expressed by S enteritidis strains. However, not all S dublin strains expressed the same epitope. To test for epitope polymorphism within SEF-14, the monoclonal antibodies and polyclonal RaSEF14 were examined in direct-binding ELISA with a variety of serotypes of salmonellae and closely related bacteria. Individual monoclonal antibodies and RaSEF14 reacted identically suggesting that the fimbria contains a number of highly conserved epitopes. The results of the blocking ELISAindicate that the monoclonal antibodies produced to SEF14 antigen recognise three distinct clusters of epitopes. These may comprise individual or groups of overlapping epitopes; the large size of monoclonal antibodies compared with individual epitopes precludes further interpretation at present. Furthermore, combinations of monoclonal antibodies from the three clusters blocked the RaSEF 14 more effectively than monoclonal antibodies alone, confirming the existence of more than one cluster. Immune electron microscopy studies revealed that the clusters were distributed

Monoclonal antibodies against S enteritidis

evenly along the SEF 14 fimbria, with.no obvious difference in the number of repeats. Labelling with polyclonal RaSEF 14 produced similar numbers of gold particles associated with SEF14, suggesting that the size of the rabbit antibodies and gold particles inhibited their binding to closely oriented epitopes. The fact that most of the monoclonal antibodies reacted in Western blots indicates that the SEF14 antigen subunits contain a number of linear or continuous epitopes. The two monoclonal antibodies (SEF14-12 and SEF14-13) which failed to react with SEF14 in Western blots were able to block reciprocally monoclonal antibodies directed against continuous epitopes suggesting that they, too, identify them. These two monoclonal antibodies had the lowest affinity towards SEF14, which might account for their lack of reactivity in Western blots. The SEF14 fimbrial antigen is highly specific to a few serotypes within group D salmonellae including S enteritidis (Thorns et al 1990, Mtiller et al 1991). The DNA sequence encoding the SEF14 fimbriae has recently been identified and also shown to be highly specific within serogroup D salmonellae only (C. Turcotte, personal communication). It was these findings which prompted the investigation of the use of these monoclonal antibodies in rapidly identifying and serotyping isolates of S enteritidis and S dublin. Specific latex reagents have been successfully used to detect fimbrial antigens of cultured, enterotoxigenic Escherichia eoli (Thorns et al 1989). Using similar methods a monoclonal antibody-coated latex reagent that can rapidly detect the SEF 14 antigen on cultured salmonella organisms was produced. Although a variety of fimbriae have been described on salmonellae (Duguid and Gillies 1958, Clegg and Gerlach 1987) very little information exists regarding the optimum medium for their expression. The importance of composition of the growth medium on the production of many fimbriae expressed by enterotoxigenic Escherichia coli is well recognised (Parry and Rooke 1985). Similarly, latex agglutination to detect SEF14 on salmonellae indicates that its expression is also very dependent on the nature of the medium. Indeed, one explanation of why this antigen has only recently been described might be that many media used routinely for the growth and isolation of salmonellae completely inhibit the expression of the SEF14 antigen. Of

307

the liquid and solid media tested in this study, peptone water (pH 7.2) and Sensitest or Isosensitest (Oxoid) were the media of choice. When the Salmonella strains were grown on Sensitest (Oxoid) agar for 18 hours at 37°C SEF14 was detected on all the strains of S enteritidis and the majority of strains of S dublin. Single isolates from only two other serotypes, S blegdam and S moscow produced SEF14. Both serotypes, which are very closely related antigenically to S enteritidis, are extremely rare worldwide. No example has been encountered in the reference laboratory of the Central Veterinary Laboratory since the Zoonoses Order (1975) started in the United Kingdom in 1976. The detection of strains expressing SEF14 by latex agglutination is therefore an indication of S enteritidis or S dublin, and on isolates originating from poultry products can be regarded as a presumptive identification of S enteritidis. Detection of SEF 14 antigen should at least complement conventional culture methods by rapidly distinguishing S enteritidis from other salmonellae isolated from poultry products. A latex agglutination kit for the detection of SEF14 is now available (McLaren et al 1992). References CLEGG, S, & GERLACI-I, G. F. (1987) Enterobacterial fimbriae. Journal of Bacteriology 169, 934-938 DUGUID, J. P. & GILLIES, R.R. (1958) Fimbriae and haemagglutinating activity in Salmonella, Klebsiella, Proteus and Chromobacterium. Journal of Pathology and Bacteriology 75, 519520 FRANCIS, D. H., REMMERS, G. A. & DEZEEUW, P. S. (1982) Production of K88, K99 and 987P antigens by Escherichia coli cultered on synthetic and complex media. Journal of Clinical Microbiology 15, 181-183 GUINEE, P. A. M., JANSEN, W. H. & AGTERBERG, C. M. (1976) Detection of the K99 antigen by means of agglutination and immuno-electrophoresis on Escherichia coliisolates from calves and its correlation with enterotoxigenicity. Infection and Immunity 13, 1369-1377 HECHEMY, K. E. & MICHAELSON, M. A. (1984) Latex particle assays in laboratory medicine, part 1. Laboratory Management 22, 27-40 MACDONALD, R. A., HOSKING, C. S. & JONES, C. L. (1988) The measurement of relative antibody affinity by ELISAusing thiocyanate elution. Journal of Immunological Methods 106, 191-194 McLAREN, I. M., SOJKA, M. G., THORNS, C. J. & WRAY, C (1992) An interlaboratory trial of a latex agglutination kit for rapid identification of Salmonella enteritidis. Veterinary Record 131,235236 MORRIS, J. A., THORNS, C. J. & WOOLLEY, J. C. (1985) The identification of antigenic determinants on Mycobacterium boris using monoclonal antibodies, Journal ofGeneralMicrobiology 131, 1825-1831 MULLER, K-H., COLLINGSON, K. S., TRUST, T. J. & KAY, W. W. (1991) Type 1 f'trnbriae of Salmonella enteritidis. Journal of Bacteriology 173, 4765-4772

308

C. J. Thorns, M . G. S o j k a , I. M . M c L a r e n , M . D i b b - F u l l e r

NESS, E. (1983) The detection of the 987P antigen in Escherichia coli isolated from piglets with diarrhoea. Acta Veterinaria Scandinavica 24, 521-523 O'BRIEN, J.D.P. (1988) Salmonella enteritidis infection in broiler chickens. Veterinary Record 122, 214 PARRY, S. H. & ROOKE, D. M. (1985) Adhesins and colonisation factors of Eseheriehia coli. Special Publications for the Society of General Microbiology: The Virulence Markers of Escheriehia coli. Ed M Sussman. London, New York, Academic Press. pp79-155 ST LOUIS, M. E., MORSE, D. L., POTTER, M. E., DEMELFI, T. M., GUZEWICK, J. J., TAUXE, R. V. & BLAKE, B. A. (1988) The emergence of grade A eggs as a major source of Salmonella enteritidis infections. New implications for the control of salmonellosis. Journal of the American Medical Association 259, 2103-2107 THORNS, C. J., SOJKA, M. G. & CHASEY, D. (1990) Detection of a novel fimbrial structure on the surface of Salmonella enteritidis by using a monoclonal antibody. Journal of Clinical Microbiology 28, 2409-2414

THORNS, C. J., SOJKA, M. G. & ROEDER, P. L. (1989) Detection of fimbrial adhesins of ETECusing monoclonal antibody-based latex reagents. Veterinary Record 125, 91-92 VAN REGENMORTEL, M. H. V. (1990) Structure of viral B-cell epitopes. Research in Microbiology 141,747-756 VAN ZIJDERVELD, F. G., WESTENBRINK, F., ANAKOTTA, J., BROUWERS, R. A. M. & VAN ZIJDERVELD, A. M. (1989) Characterisation of the F41 fimbrial antigen of enterotoxigenic Eseheriehia coli by using monoclonal antibodies. Infection and Immunity 57, 1192-1199 WILSON, M. B. & NAKANE, P. K. (1978) Recent developments in the periodate method of conjugating horseradish peroxidase (HRVO)to antibodies. Immunofluorescenceand related staining techniques. Ed W. Knapp, K. Holubar and G. Wicks. Amsterdam, Elsevier/North Holland Biomedical Press. pp215-224

Received November 1, 1991 Accepted March 16, 1992

Characterisation of monoclonal antibodies against a fimbrial structure of Salmonella enteritidis and certain other serogroup D salmonellae and their application as serotyping reagents.

A panel of 13 monoclonal antibodies from different hybridomas was produced against a novel salmonella fimbrial antigen expressed predominantly by Salm...
727KB Sizes 0 Downloads 0 Views