The protective efficacy of cloned Moraxella bovis pili in monovalent and multivalent vaccine formulations against experimentally induced infectious bovine keratoconjunctivitis (IBK)

veterinary microbiology Veterinary Microbiology 45 (1995) 129-138

The protective efficacy of cloned MoruxeZZabovis pili in monovalent and multivalent vaccine formulations against experimentally induced infectious bovine keratoconjunctivitis (IBK) A.W.D. Lepper a, J.L. Atwell b, P.R. Lehrbach ‘, C.L. Schwartzkoff ‘, J.R. Egerton d, J.M. Tennent a,* aCSIRO Division of Animal Health, Animal Health Research Laboratory, Private Bag No.1 Parkville. Vie. 3052 Australia b CUR0 Division of Biomolecular Engineering, 343 Royal Parade Parkville, Vie. 3052 Australia ’ Arihur Webster Pty. Ltd., 23 Victoria Avenue Castle Hill, NSW 2154 Australia d University of Sydney, Department of Animal Health, Werombi Road Camden, NSW 2570 Australia

Received 22 July 1993; accepted 6 October 1994

Abstract Calves were vaccinated with cloned Moraxella bovis pili of serogroup C (experiment 1) or B (experiment 2) either as a monovalent formulation or as part of a multivalent preparation with pili of six other serogroups. Within 4 weeks of the second vaccine dose vaccinated calves and nonvaccinated controls were challenged via the ocular route with either virulent M. bovis strain Dal2d (serogroup C) or M. bovis strain 3WO7 (serogroup B) in experiments 1 and 2, respectively. Calves vaccinated with multivalent vaccines had significantly lower antibody titres than those vaccinated with monovalent preparations. Nevertheless, the levels of protection against infectious bovine keratoconjunctivitis (IBK) achieved with multivalent vaccines were 72% and 83% for the groups challenged with M. bouis strains of serogroups B and C, respectively. The serogroup C monovalent vaccine gave 100% protection against experimentally induced IBK and M. bouis isolates cultured from the eyes 6 days post-challenge were identified as belonging solely to serogroup C. Unexpectedly, only 25% protection was achieved against homologous strain challenge of calves that received the monovalent serogroup B vaccine. Furthermore, the majority of M. bovis isolates recovered from calves in this group belonged to serogroup C, as did half of those isolates cultured from the multivalent vaccinates. The remaining bacterial isolates from the latter group, together with all isolates from the non-vaccinated controls, belonged to serogroup B. Results are consistent with the hypothesis that derivatives of the serogroup B challenge inoculum had expressed serogroup C pilus antigen within 6 days of the challenge, possibly as a result of pilus gene inversion occurring in response to the presence of specific antibody in eye tissues and tears. * Corresponding author. Phone 61-3-342-9700 Fax 61-3-347-4042 E-mail [email protected] 0378-l 135/95/$09.50

0 1995

SSDIO378-1135(94)00123-5

Elsevier Science

B.V. All rights reserved

130

A. W.D. L.epper et al. /Veterinary

Keywords: Morarella bouis; Pili; Vaccine; Recombinant

Microbiology

45 (1995) 129-138

DNA

1. Introduction Moraxella bovis is recognised as the primary cause of infectious bovine keratoconjunctivitis (IBK) . Seven different serogroups of pili associated with strains of M. bovis from Australia and Great Britain have been characterised to date (Moore and Lepper, 1991) . The capacity of pili from one organism to protect against infection by another organism of the same serogroup is due to the possession of identical or closely-related serogroup-specific protective pilus epitopes (Lepper et al., 1992). In order to provide broad spectrum protection, any pilus-based M. bovis vaccine must contain a sufficient quantity of pili of each of the seven known serogroups to produce a suitable sustained immunological response. This vaccine must be capable of protecting bovine cornea1 epithelium against colonisation and damage by the majority of M. bouis strains likely to be encountered in the field. Despite being major immunogens, we have found that the pili of M. bovis are not readily produced by this organism in liquid culture. However, the pilin structural genes from M. bovis can be expressed by recombinant DNA technology in Pseudomonas aeruginosa (Elleman et al., 1990) and cloned pili have been shown to protect cattle against IBK induced by challenge with the homologous virulent strain (Lepper et al., 1993). This paper describes two trials that tested the protective effect of cloned pili antigens of two different serogroups, either in monovalent vaccine formulation or as components of a multivalent preparation that included similar doses of authentic pili prepared from M. bouis strains representative of five or six other pilus serogroups.

2. Materials and methods 2.1. Preparation of vaccine antigens and vaccine formulations M. bovis strains S276R, 3WO7, R593L, TAT849,218R and FL462 (representing serogroups A, B, D, E, F and G, respectively) were used to prepare authentic pili for vaccine production. Pili were prepared after at least two subcultures of the strains on 5% bovine blood agar plates to enrich for piliated colony variants. Pili were sheared from M. bouis cells harvested from this solid medium and purified by MgCl,-precipitation and centrifugation steps as previously described (Lepper and Hermans, 1986). P. aeruginosa recombinants containing the cloned M. bovis pilin gene of serogroups B or C were used to produce cloned pili antigen according to a previously published method (Lepper et al., 1993). Pili were formulated into monovalent or multivalent vaccine preparations in oil adjuvant as described elsewhere (Lepper et al., 1993; Schwartzkoff et al., 1993). 2.2. Vaccination schedules and challenge A total of fifty-seven individually identifiable, Hereford steer and heifer calves were used in the two separate experiments reported in this study. The animals were between 5 and 7

A. W.D. tipper

et al. / Veterinary Microbiology

45 (1995) 129-138

131

months of age and were assigned to three groups of seven calves (experiment 1) or three groups of 12 calves (experiment 2) (animals in each group having been matched with regard to weight, age and origin). Animals in each experiment were allowed to run together at pasture until vaccination had been completed, and then were housed together in one large covered yard, open to the weather on all sides wherein they were fed hay and pellets with water ad-libitum. All calves were known to be free of detectable M. bovis infection as a result of bacteriological examination of the conjunctival sac prior to vaccination. Vaccinated calves received a primary dose of vaccine (VI ) and a booster dose (V2), each of 2m1, injected subcutaneously on opposite sides of the neck at an interval of six (experiment 1) or four weeks (experiment 2). The challenge inoculum of M. bovis strain Dal2d or strain 3WO7 was prepared and administered as previously described (Lepper et al., 1993). Both eyes of all calves were challenged four weeks after V2 (experiment 1) or three weeks after V2. (experiment 2).

2.3. Assessment of protective efficacy and antibody response

Bacteriological examination of the conjunctival sacs of both eyes of all calves was performed 14 days before vaccination and on day 6 after challenge. Colonies resembling M. bovis were further identified by Gram-stained smear, direct fluorescent antibody (FA) test (Lepper and Barton, 1987), slide agglutination (Moore and Rutter, 1989) and microplate agglutination (MPA; Lepper et al., 1993). Calves were bled before Vl, at V2 and then at 2,4,6 and 12 weeks (experiment 1) or 3 weeks (experiment 2) after V2. Sera were tested by MPA for antibody to serogroup-specific pili antigen of M. bovis. Antibody titres were log transformed and the responses of each group compared by one way analysis of variance. Clinical examinations were carried out daily to determine the number of calves in each experimental group that had developed comeal ulceration. The eyes of any calves that developed either unilateral or bilateral keratitis involving two thirds or more of the comeal surface, were topically treated with benzathene cloxacillin eye ointment (Orbenin, Beecham Australia) within 24 hr of this observation in order to avoid undue suffering. Experiments were terminated by similar topical treatment. The percentage protection of vaccinates versus non-vaccinated controls, together with the mean lesion scores per group, were determined according to previously described methods (Lepper et al., 1992). Statistical analysis of clinical results was performed using Fisher’s exact test (two-tailed).

2.4. DNA manipulation procedures

Methods for the preparation of genomic DNA from M. bovis and for the generation of PCR fragments using primers based on the M. bovis pilin gene sequence have been described elsewhere (Elleman et al., 1990; Atwell et al., 1994). Standard methods (Sambrook et al., 1989) were employed for agarose gel electrophoresis, visualisation, size estimation and photography of DNA.

132

A. W.D. Lepper et al. /Veterinary

Microbiology

45 (1995) 129-138

3. Results Two separate experiments were conducted to evaluate the protective effect of cloned M. bouis pili of serogroups B and C when used as a monovalent vaccine or as part of a multivalent vaccine formulation against experimentally-induced IBK. 3.1. Experiment

1

Three separate groups of seven calves were used. Group 1 was vaccinated with a monovalent preparation consisting of the cloned serogroup C pili of M. bovis strain Dal2d. Group 2 was vaccinated with a multivalent preparation containing a similar dose of cloned serogroup C pili together with equal doses of authentic pili derived from M. bovis strains S276R, 3WO7, R593L, TAT849, 218R and FL462. Group 3 comprised non-vaccinated calves. Four weeks after V2, both eyes of all calves were challenged with lo9 colony forming units (CFU) of virulent M. bovis strain Dal2d via the conjunctival sac. Blood samples were collected for serological analysis according to the schedule in Materials and Methods. Clinical examinations were carried out daily to determine the number of calves in each group that had developed comeal ulceration and eyes were swabbed 6 days after challenge to determine the number of calves infected with M. bouis. The serological identity of isolates recovered from two animals in each group was checked by slide agglutination with antisera raised in rabbit and goat that had been immunised with M. buds serogroup C pili. The percentage indices of protection against IBK achieved with serogroup C pili of M. bovis strain Dal2d compared with non-vaccinated control animals were 100% for calves vaccinated with the monovalent cloned pili and 72% for calves vaccinated with multivalent vaccine (Table 1). The number of animals protected against IBK in Group 2 was not significantly different from that in Group 1. The mean lesion score of calves in Group 2 was 0.3 compared with 1.2 in non-vaccinated controls. Anamnestic agglutinating antibody responses were recorded for animals in both Groups 1 and 2 (Table 2), however, the geometric mean antibody titre at V2 + 2 1 days was markedly higher for calves that received the monovalent vaccine than for those vaccinated with the multivalent formulation. This difference was statistically significant (P < 0.01). M. bouis isolates recovered from two animals in each of the Groups 1, 2 and 3 were all found to belong to serogroup C and were presumed to be identical to the challenge strain. 3.2. Experiment 2 Three separate groups of 12 calves were used. Group 1 was vaccinated with a monovalent preparation consisting of the cloned serogroup B pili of M. bovis strain 3WO7. Group 2 was vaccinated with a multivalent preparation containing the same dose of cloned serogroup B and serogroup C pili together with equal doses of authentic pili derived from M. bovis strains S276R, R593L, TAT849, 218R and FI.462. Group 3 comprised non-vaccinated calves that were housed in a covered yard with the vaccinated calves for the duration of the challenge experiment. Groups 1,2 and 3 were challenged with lo9 CFU of virulent M. bovis strain 3WO7 three weeks after V2. Clinical examinations. assessment of cornea1 ulceration,

A. W.D. Lepper et al. /Veterinary

133

Microbiology 45 (1995) 129-138

Table 1 The degree of protection achieved against IBK in calves vaccinated with monovalent or multivalent vaccines comprising cloned and authentic pili and challenged with virulent strains of M. bovis representing serogroups C andB NO. Animals challenged

No. eyes challenged

No. animals with IBK

No. eyes with IBK

Mean lesion score

No. lesions requiring treatment”

% protection index

Dal2d

7

14

0

0

0

0

100

C Dal2d

7

14

2

2

0.30

1

72

C Dal2d

7

14

7

8

1.20

4

_

C 3wo7

12

24

8

10

1.00

6

25

B 3wo7

12

24

1

1

0.08

0

83

B 3wo7

12

24

11

15

1.25

10

_

Experiment no.

Group and Challenge strain and vaccine serogroup serogroup

1

1 C 2 A to G 3 NIL 1 B 2 A to G 3 NIL

B

“Within 5 days of challenge.

Table 2 Geometric mean agglutinating antibody response with 95% confidence intervals for calves vaccinated with two doses (V 1 and V2) of monovalent or multivalent vaccines in experiments 1 and 2 and tested with killed, piliated cells of the respective Moruxella bovis challenge strain

Experiment, test antigen and serogroup

1

Sample time and vaccine group V1 v2

3wo7 B

“Experiment 1. bExperiment 2. ND not done

V2 + 42 DAYS

MONO

MULTI

MONO

MULTI

MONO

MULTI

MONO

MULTI

<20

<20

706

40

10,240

780

11,306

707

(1699)

<20

(2381522) 320

(5,20520,142) 2,665

(336 1808) 958

(6,45819,793) ND _

(3101611) ND

(1231,827)

(44-261)

(1,6544,295)

(644-129)

Dal2d C 2

V2 + 21” or 2gb. DAYS

<20

107

134

Table 3 Serogrouping

A. W.D. Lepper et al. /Veterinary

Microbiology 45 (1995) 129-138

of M. bovis isolates recovered from animals in experiment

Group, vaccine and serogroup

Challenge strain and serogroup

No. of M. bovis isolates

2 Serogroup B

identity of isolates

C

A, Q JL F 0rG

1 Monovalent B 2 Multivalent A to G 3 Nil

3wo7 B

17

1

16

0

3wo7 B

13

6

7

0

3wo7 B

16

16

0

0

blood collection, treatment regimes and the calculation of percentage indices of protection and mean lesion scores were performed as described for Experiment 1. Eyes were swabbed 6 days after challenge to determine the number of calves infected with M. bouis. The serological identity of all confirmed M. bovis isolates was checked by MPA using an antisera panel derived from rabbits that had been individually immunised with one of the seven pilus serogroups. The number of isolates recovered from each group with the same or different serological identity to that of the challenge inoculum were compared by contingency table analysis using Fisher’s exact test (two-tailed). Only 25% of calves in Group 1 that had been vaccinated with monovalent cloned pili of serogroup B were protected against IBK induced by the homologous M. bouis strain 3WO7 (Table 1) . In contrast, 83% of animals in Group 2 that had received the same antigen in a multivalent formulation together with pili of serogroups A, C, D, E, F and G, were protected against IBK. Furthermore, the mean lesion score for animals in Group 1 was almost as high as that for the non-vaccinated calves in Group 3, whereas the score for Group 2 animals was insignificant by comparison (Table 1) . Anamnestic antibody responses were recorded for calves in both Group 1 and 2 (Table 2). Consistent with the results in experiment 1, it was found that the geometric mean titre to serogroup B pili antigens was significantly less in calves that had received the multivalent vaccine compared with the titre from those calves vaccinated with the monovalent formulation (P
A. W.D. Lepper et al. f ~~ter~~~ ~~cr~bioio~~ 45 (1995) 129-138

135

4. Discussion This study has demons~ated that despite a significant reduction in antibody response to pilus antigens compared with those levels achieved following monovalent vaccination, an effective level of protection against IBK can be achieved by multivalent vaccination. It is most probable that the reduction in antibody titre using the multivalent vaccine was due to the phenomenon of antigenie competition. This term describes what occurs when proteins (Adler, 1964), haptenated co-polymers (Werdelin, 1982) or amino acid co-polymers (Rock and Benacerraf, 1983) compete for presentation to T lymphocytes by the same MHC molecule ( IJnanue and Cerottini, 1989). The antigenic competition observed in our work parallels the findings of Schwartzkoff et al. ( 1993) and Raadsma et al. ( 1994), who have studied the interaction of pilus antigens belonging to multiple serogroups in ovine footrot vaccines. In these studies, antibody titres induced by multivalent Dichelobacter nodosus pilus antigen preparations provided significant protection against disease though it was noted that these titres were always reduced compared with those achieved following vaccination with monovalent vaccines. As the D.nodosus pilus antigens used by Schwartzkoff et al. ( 1993) were approximately 90% pure, these workers concluded that the observed antigenic competition was due to the presence of a family of ~tige~c~ly-related pilus proteins rather than to the interference of any extraneous material. Since similar criteria for the evaluation of pilus antigen purity have been adopted for the M. bovis pili preparations described in this study, we propose that the antigenic competition observed in the IBK vaccine trials has occurred for the same reason. The level of protection achieved with the monovalent vaccine comprising cloned serogroup C pili of M. bovis strain Dal2d (Experiment 1) was as high as that recently reported for a similar formulation (Lepper et al., 1993). However, the failure of the monovalent vaccine comprising cloned serogroup B pili of M. boois strain 3WO7 to protect a majority of calves (Experiment 2)) despite the presence in these animals of what has been considered in the past to be an adequate geometric mean antibody titre (Lepper, 1988: Lepper et al., 1993), was contrary to expectations. The rapid (within 6 days) and almost universal ( 16 out of 17) switch in pili antigenicity from serogroup B to C for M. bouis isolates recovered from the eyes of Group I calves has not previously been observed in any of our vaccine experiments. As the protection data reflects, this switching rendered the monovalent vaccine ineffective. We have dete~ined that no serogroup C antibody existed in those calves vaccinated with the monovalent cloned serogroup B pili vaccine (data not shown) thus rendering these animals fully susceptible to infection caused by organisms that expressed serogroup C pili. A similar serological switch (B to C) of approximately half of the M. boais isolates recovered from the eyes of Group 2 calves that had received the multiv~ent vaccine was also observed. In this case, it is interesting to recall the effect of antigenic competition in that these animals had a significantly lowered geometric mean titre of antibody to serogroup B pili antigen (Table 2) although they did possess an antibody titre to serogroup C pili antigen that had been stimulated by the presence of that antigen in the multivalent vaccine. At least two possible explanations for the observations in Experiment 2 can be considered. The first is that one or more of the calves in Group I may have had an M. bovis serogroup C infection prior to challenge with the serogroup B strain 3WO7. In this scenario, it is

136

A. W.D. Lepper et al. /Veterinary

Microbiology 45 (1995) 129-138

possible that sufficient antibody to serogroup B pili was generated within 4 to 6 days as a result of vaccination that selectively prevented the adherence to cornea1 cells of the piliated challenge organisms but did not interfere with the adherence of any resident serogroup C bacteria. However, M. bovis was not detected in the eyes of any calves (either vaccinates or non-vaccinates) prior to challenge nor was there any M. bovis of serogroup C recovered from the eyes of the non-vaccinated control group after challenge despite their being housed in the same enclosure as calves of Groups 1 and 2. A second more likely explanation could be that a specific alteration to the pilin gene, and thus a change in the adherence-promoting pilus antigen (Moore and Rutter, 1989), may have occurred in progeny of the 3WO7 challenge strain. The pilin gene of the American strain of M. bovis Epp63 was the first such determinant of this species to be cloned and sequenced (Marrs et al., 1985). Subsequent studies (Marrs et al., 1988; Ruehl et al., 1988; Fulks et al., 1990) demonstrated the existence of more than one pilin gene in this strain. M. bouis Epp63 possesses both a complete and an incomplete pilin gene that can undergo pilin gene phase variation as a result of the inversion of a 2.1 kilobase segment of genomic DNA. It was shown that inversion resulted in two antigenically distinct pilin proteins, of either quick or slow electrophoretic mobility, being produced by this strain in vitro and that a switch from one type to the other occurred at low frequency. Furthermore, by the use of specific DNA probes for each of the complete or incomplete pilin genes it was shown that two patterns of pilin gene organisation existed, each one associated with the production of a particular pilin protein. Significantly, specific pilin gene probes have now been used to show that Australian strains of M. bouis, including Dal2d and 3WO7, contain both a complete and an incomplete pilin gene (Elleman et al., 1990; Atwell et al., 1994). The results of Experiment 2 suggest that a change in pilus antigenicity (in this case from serogroup B to serogroup C) may have occurred for the challenge organism M. bouis 3WO7. Preliminary genetic analysis of the various 3WO7 isolates suggest that this “switch” has arisen by way of a genomic inversion event that places distinct pilin genes in the pilin gene expression locus. Several pieces of data support this hypothesis and specifically argue against a superinfection with a serogroup C M. bouis strain being responsible for the experimental observations. The first of these is that the plasmid profile of the serogroup B challenge strain 3WO7 was found to be identical to that of the serogroup C organisms recovered, and different from that of the M. bouis serogroup C prototype strain Dal2d, supporting their common derivation (data not shown). Secondly, the genomic organisation of pilin genes in all M. bouis isolates recovered from calves in Experiment 2 (irrespective of their serogroup) was identical to that of the serogroup B strain 3WO7 and different to that of the serogroup strain Dal2d (data not shown). Finally, it has been possible to design a specific primer based on the highly conserved 5’ coding region of M. bovis pilin genes (Atwell et al., 1994) for use in a PCR analysis of genomic DNA prepared from various M. bovis strains (Fig. 1) . This analysis showed that the complete and incomplete pilin genes are located 2.9 kilobase pairs (kbp) apart in the genome of the serogroup B challenge strain 3WO7 (lane 1) . The same size PCR fragment was generated from the serogroup B 3WO7 strain recovered from a multivaccinate (lane 2) and also from both antigenically-altered 3WO7 (ie: serogroup C) strains recovered from either a monovaccinate (lane 3) or a multivaccinate (lane 4). In contrast, the complete and

A. W.D. Lepperet al. I Veterinarymicrobiology45 (1995) 12%I38

137

- 3.59 - 2.81

- 1.81 Fig. 1.PCR pmducts from M. bouisstrains. Agarose gel electrophoresis ( I .5% f wtivol] ) of PCR products from genomic DNA of M. bouisstrains associated with Experiment 2. Lanes: 1, serogroup B M. boaischallenge strain 3WO7; 2, serogroup B M. bouisstrain recovered from a multivaccinate; 3, serogroup C M. bouisstrain recovered from a monovaccinate: 4, serogroup C M. bouisstrain recovered from a multivaccinate: 5, serogroup C prototypic M. butk strain Dal2d; 6, DNA molecular weight markers (sizes indicated in kilobase pairs).

incomplete pilin genes are located 1.9 kbp apart in the genome of the prototypic serogroup C strain Dal2d (lane 5). We propose that the ability of M. bo~is strain 3WO7 to express pili of altered antigenicity results from the vaccine-induced selective pressure of circulating antibody. Clinically this was manifest as a breakdown of protection from infection in cattle that received only the monovalent vaccine. Antigenic variation of strain 3WO7 occurred in just over half of those isolates recovered from calves that received the multivalent vaccine although it was reassuring to note that a significant majority of these animals were protected from IBK since the variant serogroup was included in the multivalent formulation. The involvement of selective antibody pressure in antigenic variation is supported by the observation that none of the bacteria isolated from non-vaccinated calves exhibited any change in pilus antigenicity. It can only be speculated that the remaining strains recovered from multivaccinates that did not express altered antigenicity were subject to a lesser selective antibody pressure possibly related to individual animal responses, different microbial physiological states or some effect of the environment on either of these. To date it has not been possible to demonstrate any in vivo or in vitro alteration in pilus antigenicity expressed by hrl. bock strains Dal2d despite attempts to do so by routine laboratory sub-culture, by culture in the presence of homologous antibody or by passage in ocular tissues of mice or cattle (Lepper, 1988; Lepper et al., 1993; Experiment 1, this study). It may be that strain Dal2d is “switch-locked” or alternatively variation could have gone unnoticed if the spectrum of serogroups tested for was incomplete or if the variant serogroup remained as serogroup C. Further work in this area will be necessary to define the extent to which altered ~tigenicity is possible via pilin gene inversion in order to formulate a comprehensive multivalent pilus-based vaccine to combat IBK. Acknowledgements This work was carried out under Meat Research Corporation Grant No. CS 096. We are grateful to Mrs S. Spiess, Mr A. Matheson and Mrs N. Franks for skilled technical assistance.

138

A. W.D. tipper

et al. /

VeterinaryMicrobiology 45 (1995) 129-138

References Adler, F.L., 1964. Competition of antigens. Progr. Allergy., 8: 41-57. Atwell, J.L., Tennent, J.M., Lepper, A.W.D. and Elleman, T.C., 1994. Characterization of pilin genes from seven serologically defined prototype strains of MoraxeZZa bouis. J. Bacterial. 176: 4875-4882. Elleman, T.C., Hoyne, P.A. and Lepper, A.W.D., 1990. Characterization of the pilin gene of Moraxella bovis Dalton 2d and expression of pili from M. bouis in Pseudomonas aerqinosa. Infect. Immun., 58: 1678-1684. Fulks, K.A., Marrs, C.F., Stevens, S.P. andGreen, M.R., 1990. Sequenceanalysisof the inversionregioncontaining the pilin genes of Moraxella bouis. J. Bacterial., 172: 310-316. Lepper, A.W.D., 1988. Vaccination against infectious bovine keratoconjunctivitis: protective efficacy and antibody response induced by pili of homologous and heterologous strains of Moraxella bouis. Aust. Vet. J., 65: 3 lO316. Lepper, A.W.D. and Barton, I.J., 1987. Infectious bovine keratoconjunctivitis: seasonal variation in cultural, biochemical and immunoreactive properties of MoruxeZZa bouis isolated from the eyes of cattle. Aust. Vet. J., 64: 33-39. Lepper, A.W.D. and Hermans, L.R., 1986. Characterisation and quantitation of pilus antigens of Moraxella bouis by ELISA. Aust. Vet. J., 63: 401405. Lepper, A.W.D., Moore, L.J., Atwell, J.L. and Tennent, J.M., 1992. The protective efficacy of phi from different strains of Moraxella bouis within the same serogroup against infectious bovine keratoconjunctivitis. Vet. Microbial., 32: 177-187. Lepper. A.W.D., Elleman, T.C., Hoyne, P.A., Lehrbach, P.R., Atwell, J.L., Egerton, J.R., Schwartzkoff, L.C. and Tennent, J.M., 1993. A Morarellu bouis pili vaccine produced by recombinant DNA technology for the prevention of infectious bovine keratoconjunctivitis. Vet. Microbial., 36: 175-183. Marrs, C.F., Schoolnik, G., Koomey, J.M., Hardy, J., Rothbard, J. and Falkow, S., 1985. Cloning and sequencing of a Moraxella bouis pilin gene. J. Bacterial., 163: 132-139. Marrs, C.F., Ruehl, W.W., Schoolnik, G.K. and Falkow, S., 1988. Pilin gene phase variation of Moraxella bovis is caused by an inversion of pilin genes. J. Bacterial., 170: 3032-3039. Moore, L.J. and Lepper, A.W.D., 1991. A unified serotyping scheme for Moraxella bouis. Vet. Microbial., 29: 75-83. Moore, L.J. and Rutter, J.M., 1989. Attachment ofMoraxella bouis to calf comeal cells and inhibition by antiserum. Aust. Vet. J., 66: 39-42. Raadsma, H.W., O’Meara, T.J., Egerton, J.R., Lehrbach, P.R. and Schwartzkoff, CL., 1994. Protective antibody titres and antigenic competition in multivalent Dichelobacter nodosus Fimbrial vaccines using characterised rDNA antigens. Vet. Immunol. Immunopathol., 40: 253-274. Rock, K.L. and Benacetraf, B., 1983. Inhibition of antigen-specific T lymphocyte activation by structurally related Ir gene-controlled polymers. J. Exp. Med., 157: 1618-1634. Ruehl, W.W., Marrs, C.F., Femandez, R., Falkow, S. and Schoolnik, G.K., 1988. Purification, characterization, and pathogenic&y of Moraxeh bouis pili. J. Exp. Med., 168: 983-1002. Sambrook, J., Fritsch, E.F. and Maniatis, T., 1989. Molecular cloning: a laboratory manual, 2nd edn. New York, Cold Spring Harbor Laboratory. Schwartzkoff, C.L., Egerton, J.R., Stewart, D.J., Lehrbach, P.R., Elleman, T.C. andHoyne, P.A., 1993. The effects of antigenic competition on the efficacy of multivalent footrot vaccines. Aust. Vet. J., 70: 123-126. Unanue, E.R. and Cerottini, J-C., 1989. Antigen presentation. FASEB J., 3: 2496-2502. Werdelin, O., 1982. Chemically related antigens compete for presentation by accessory cells to T cells. J. Immunol., 129: 1883-1891.