Vaccinal properties of Salmonella abortusovis mutants for streptomycin: screening with an ovine model P Pardon, J Marly, F Lantier, R Sanchis To cite this version: P Pardon, J Marly, F Lantier, R Sanchis. Vaccinal properties of Salmonella abortusovis mutants for streptomycin: screening with an ovine model. Annales de Recherches Vétérinaires, INRA Editions, 1990, 21 (1), pp.57-67. <hal-00901921> HAL Id: hal-00901921 https://hal.archives-ouvertes.fr/hal-00901921 Submitted on 1 Jan 1990 HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
Original article Vaccinal properties of Salmonella abortusovis mutants for streptomycin: screening with an ovine model P Pardon J Marly F Lantier R Sanchis 1 INRA, Station de pathologie de la reproduction, Unité de pathologie infectieuse ovine, Tours-Nouzilly, 37380 Nouzilly 2 Ministère de l agriculture, Services vétérinaires, Laboratoire national de pathologie des petits ruminants, 63, avenue des Arènes, 06051 Nice Cedex, France (Received 27 October 1988; accepted 7 June 1989) Summary ― Living attenuated vaccines may help control infection with Salmonella enterica subspecies enterica ser abortusovis in sheep without constraints incompatible with husbandry. Virulence and immunogenicity of 2 Abortusovis streptomycin-dependent strains and 3 reverse mutants were compared in sheep, the virulent parental strain Abortusovis 15/5 being used as reference. Reduction of virulence with conservation of some immunogenicity was demonstrated for the 5 mutant strains after intravenous or subcutaneous inoculation. A 5-day bacteriemia was detectable after intravenous inoculation with parental strain, whereas none was shown with mutant strains. Mutant reverse strains survived in loco-regional lymph nodes for at least 12 days. The dependent strain D5 and reverse strain Rv6 were selected for subcutaneous vaccination of 2 groups of 18 ewes 32 days before fecundation. Ten pregnant ewes per group and 10 unvaccinated ewes were randomly selected among pregnant animals and were challenged subcutaneously at mid-gestation. Both vaccines conferred protection, but Rv6 was more efficient than D5 and was chosen as the vaccinal strain. vaccine / Salmonella abortusovis / ovine / mutant strain Résumé ― Propriétés vaccinales de souches du sérotype Salmonella abortusovis mutantes vis-à-vis de la streptomycine : criblage avec un modèle ovin. Un vaccin vivant atténué peut faciliter le contrôle de l infection ovine par Salmonella enterica subspecies enterica ser abortusovis sans imposer des contraintes incompatibles avec l élevage. La virulence et l immunogénicité de 2 souches mutantes streptomycino-dépendantes de ce sérotype et de 3 mutants réverses ont été comparées sur ovins, la souche sauvage parentale servant de contrôle. La réduction de la virulence avec conservation d une immunogénicité a été démontrée pour ces 5 souches mutantes après inoculation intraveineuse ou sous-cutanée. Une bactériémie était détectable pendant les 5 premiers jours après inoculation intraveineuse de la souche sauvage, tandis qu aucune bactériémie des souches mutantes n était détectable. Les bactéries réverses survivaient au moins 12 jours dans les ganglions lymphatiques loco-régionaux après injection sous-cutanée. Une souche dépendante (D5) et une souche réverse (Rv6) ont été retenues pour la vaccination sous-cutanée de 2 lots de 18 brebis 32 jours avant fécondation; 10 brebis par lot et f0 brebis non vaccinées ont été tirées au sort parmi les brebis présumées gravides et ont été éprouvées par voie sous-cutanée à mi-gestation. Les 2 souches augmentaient la résistance des brebis à l épreuve; la souche Rv6 étant plus protectrice, cette souche a été retenue comme souche vaccinale. vaccin /Salmonella abortusovis / ovin / souche mutante
INTRODUCTION Abortion and vaginal bacterial excretion are the main characteristics of endemic infection of sheep due to Salmonella enterica subsp enterica ser abortusovis (called Abortusovis; LeMinor and Popoff, 1987). Collective and repeated antibiotic treatment is expensive and usually unsatisfactory. Inactivated vaccines require 2 or 3 injections per gestation and give irregular results. Living attenuated vaccines may help control Abortusovis infection without constraints hardly compatible with sheep husbandry. Analogy with the Brucella melitensis strain Rev 1 vaccine (Elberg, 1981) has directed research to a living vaccine (Pardon et al, 1980). Eight streptomycindependent and 6 reverse mutants were selected (Lantier et al, 1981 Screening of the vaccinal strain among these 14 mutant strains was carried out in vivo, in mice and then in sheep. Virulence and immunogenicity studies in a murine model (Pardon and Marly, 1979) allowed us first to exclude 9 out of the 14 mutant strains (Lantier et al, 1981). The final choice among the 5 remaining strains was made at the end of 3 successive experiments in sheep (Pardon et al, 1983; Sanchis and Pardon, 1984). The first 2 experiments led to only 2 strains being retained, 1 per category of mutants, according to virulence and immunogenicity in non-gravid sheep. The last experiment, a vaccinationchallenge test in gravid ewes, led to the choice of the vaccinal strain Rv6. MATERIALS AND METHODS Bacterial strains and growth conditions Abortusovis 15/5 was originally isolated by Plommet from the placenta of an aborting ewe (Pardon and Marly, 1979). This streptomycinsensitive (Str s) strain was used to obtain the resistant mutant, 15/5 Str, the dependent mutants, Dl and D5 and the reverse mutants, Rvl, Rv4 and Rv6 (Lantier et al, 1981 Conditions of strain preservation, diluent and preparation of the inoculum have been described elsewhere (Pardon and Marly, 1979). Strains were plated on Trypticase Soy Agar (TSA; BioM6rieux, Lyon) or Salmonella-Shigella medium (SS; Bio- Mérieux, Lyon). Mutant strains differed from the parent wild strain 15/5 Str s by their growth capacity. In the presence of streptomycin (500 pg/ ml; Specia, Paris) added to the culture medium, growth was observed only with the resistant mutant 15/5 Str (colonies of at least 1 mm in diameter in less than 24 h of incubation) and the reverse mutants Rvl, Rv4 and Rv6 (colonies of at least 1 mm in diameter in over 24 h). The parent strain Abortusovis 15/5 Str s was used as a reference for virulence, and the Abortusovis strain 15/5 Str as a challenge strain. Animals Sheep of both sexes or male lambs (7 to 8 months old), Préalpes-Lacaune reared in an isolated flock free of contagious abortions, were distributed randomly in experimental groups and placed in a specialized building 1 week before inoculation. Ewes were fertilized in a controlled covering after hormonal heat synchronization. Three successive experiments with different aims were made between August and October 1978. Multiplication and persistency Nine adult sheep (7 ewes, 2 rams) were distributed in 3 experimental groups. Each group received in the jugular vein 5 ml of bacterial suspension containing 6 x 10! to 7 x 10! CFU of 1 of these strains: wild 15/5 Str s, dependent D5 or reverse Rv6. Blood samples for serological tests were taken each day for 2 weeks, then twice during the 3rd week, and once a week until slaughter. Blood samples for bacteriological examination were taken at 30 min and 6 h after inoculation, then twice a day for 4 days, and each morning for the next 4 days. Rectal tempera-
tures were recorded at the time of inoculation and then each morning for 6 days. At slaughter, 42 days after inoculation, samples were harvested: supramammary, precrural, prescapular, medial retropharyngeal, submaxillary, parotid, hepatic, mediastinal and distal jejunal lymph nodes; portions of liver, spleen, lung, kidneys, udder or testis; gall-bladder; contents of rumen and intestine at different levels; bile and heart clot. Ewes (55) were distributed in 3 experimental groups. Two out of the 3 groups received as a vaccination an sc injection behind the left shoulder of 2 ml with 7 x 108 Abortusovis strain D5 or Rv6. Thirty-four days later, ewes were fertilized in a controlled covering after hormonal heat synchronization. About 2 months after fecundation, each initial group was reduced to 10 ewes presumed to be gravid. From each group, the 8 remaining ewes were isolated as unchallenged unvaccinated controls. The 3 groups of 10 ewes received as challenge a sc injection behind the right shoulder of 2 ml with 8 x 109 Abortusovis 15/5 Str. At each lambing, cultures were made from colostrum and from vaginal discharge collected with a vaginal swab. All aborted foetuses and dead lambs were necropsied, and cultures were made from colostrum and from spleen, lung, stomach contents and brain. The 28 proven gravid ewes in challenged groups were slaughtered and necropsied between 2.5 and 3 months after abortion or normal lambing. At slaughter, samples for Salmonella detection were: prescapular, hepatic, ileo-caecal and distal jejunal lymph nodes; gall-blader and uterus. Invasiveness and immunogenicity Young rams (18; 7-8 months old) were distributed in 6 experimental groups. Each group received subcutaneously (sc) behind the right shoulder 2 ml of bacterial suspension containing between 5 x 108 to 1 x 10! viable bacteria of 1 of these strains: wild 15/5 Str s, dependent D1 or D5, reverse Rv1, Rv4 or Rv6. Rectal temperatures were recorded twice a day for 3 days after inoculation and each morning for the following 7 days. Blood samples for serological tests were collected at the time of inoculation and at days 2, 5, 7, 9 and 12 after inoculation. Skin thickness at the site of inoculation was measured each morning from the day of inoculation to day 9 after inoculation. At slaughter, 12 days after inoculation, samples were harvested: prescapular, distal jejunal and hepatic lymph nodes, spleen, liver, gall-bladder, kidneys, testis, epididymis and seminal vesicles. Protection against abortion Clinical examination Animals were observed once or twice daily. Skin thickness at the site of sc injection was measured with a caliper. Serological tests Venous blood samples were placed in tubes without anticoagulant (Vacutainer, Becton Dickinson, Meylan). Sera were subjected to agglutination tests with a microtechnique described previously (Pardon et al, 1983; Sanchis et al, 1985). Bacteriologic tests Venous blood samples were placed in tubes with ethylene diaminetetraacetic acid (Vacutainer, EDTA). Bile and gall-bladder, fcetal stomach contents and intestinal contents were individually and directly placed in sterile pots. Other samples, entirely or partially removed, were dissected, flamed and individually placed in sterile pots. Liquid samples and swabs were directly plated on SS medium. Blood was also enriched in Trypticase Soy Broth (BioM6rieux, Lyon) before plating on SS medium. Uteruses were ground separately in a homogenizer (Kenwood, Woking, UK) with a small volume of diluent. Other organs were fragmented with a stomacher (Colworth Stomacher, Cofralab, Bordeaux, France) in a sterile plastic bag after addition of a small volume of sterile diluent. Intestinal contents were diluted with a small volume of diluent and crushed with a stomacher. One aliquot of each ground or crushed sample was seeded onto 2
plates, 1 of SS medium and the other of SS medium with added streptomycin (500 1g/ml). For the numeration, tissues were weighed, diluted (1/5 weight/volume), crushed and aliquots were plated. After 48 h at 37 C, the identity of dubious colonies was checked by agglutination with sera against 0 or H Salmonella antigens (Diagnostics Pasteur, Marnes-La-Coquette, France). Definitions A lambing took place before 140 days of gestation. A sample was noted as infected when at least 1 colony of Abortusovis was isolated. A lambing was noted as infected when Abortusovis was isolated from at least 1 of the samples taken from the ewe or from products of this lambing. was noted as an abortion when it Presentation of results and statistics Mean (m) and standard errors (SE) were calculated from CFU of all animals per experimental group. Samples from which no Abortusovis was isolated were noted as containing 1 viable bacteria. The level of infection of a sample was also noted semi-quantitatively, the score being scaled from 0 to 4 (1, 1 to 5 colonies; 2, 6 to 25; 3, 26 to 125; 4, >125). Titres of sera in agglutinating antibodies were expressed in log 2 of reciprocal of dilutions. Comparisons between experimental groups were statistically evaluated using a t-test or an F-test with the appropriate number of degrees of freedom of error (df), the level of significance between means being established by comparison with the calculated value of the least significant differences (isd). RESULTS Systemic multiplication and persistency The intravenous route was used to explore capacity for survival and systemic multiplication of the bacterial strains. The group inoculated with the parent wild strain presented irregularly detectable bacteremic phases (table I); all positive blood samples were taken during the first 5 days after inoculation. The number of Abortusovis in the blood was low: on 11 positive blood samples, direct plating without enrichment was successful once. Samples taken at slaughter, 42 days after injection, were not infected, including those from the group inoculated with the wild strain. Differences between evening and morning rectal temperatures did not contribute significantly to the observed variations (results not presented); a mean between the 2 daily individual records was used to calculate the evolution of rectal temperatures in each group. The moments of rise in rectal temperature and those later in serologic response were not different according to the groups, but the peaks were lower and the febrile phase shorter in groups injected with the mutant strains (table I). Animal No 9 exhibited the lowest hyperthermic response and did not react serologically to injection of the Rv6 strain; the 2 other animals in this group presented a serological reaction similar to that of animals injected with the D5 strain. A local reaction at the injection site was noted but not measured.
The serological response was lower after injection of the mutant strains. Beside the short in vivo survival of wild or mutant strains, this first experiment demonstrated a reduction in immunogenicity and in the systemic dissemination capacity of the 2 mutant strains compared with the parent wild strain. Invaslveness and immunogenicity In the second experiment, bacteria were injected sc to screen strains according to their invasiveness and their capacity for stimulation of an immunologic response after an sc vaccination. Screening was extended to all 5 mutant strains, the parent wild strain being used as reference. Groups of 3 rams were inoculated sc with a dose about 15 times higher than that used for intravenous inoculation. Twelve days after injection, the 9 examined samples other than the sites of local reactions and prescapular lymph nodes of the inoculated side were uninfected in all groups. The Rv6 strain was first of the mutant strains, in frequence and level of local and regional persistence of bacteria. The serologic response was higher after injection of reverse strains than after injection of dependent strains (table II). All selected criteria evidenced minimal reactions after injection of the dependent mutant strains (table II). The intensity of cutaneous reactions to Rv4 or Rv6 injections was of the same order as that to the parent strain injection, but regional lymphadenitis and fever remained lower. Local
cutaneous reactions caused by injection of dependent strains were undetectable at the time of slaughter and consequently no examination of local infection was carried out. The low capacity for systemic colonization of all strains, and the reduced virulence of mutant strains were confirmed. One strain per category of mutants was retained: D5 because of higher local and regional reactions, and Rv6 because of higher scores of local and regional infection (table II). Protection against abortion Ewes were challenged sc at mid-gestation to test and compare protections induced by vaccination before gestation with about 7 x 10 8 Abortusovis strain D5 or Rv6. As in the previous experiment, local postvaccinal reaction measured by increase in skin thickness was higher and more persistent after Rv6 injection than after D5 injection: at day 7 after vaccination, skin thickness was 23.0 + 1.0, 11.2 + 0.8 or 2.5 ±0.1 mm (mean ± SE, n = 18) in Rv6, D5 or the unvaccinated group respectively. One abscess discharching pus through the skin was observed in 6 out of 18 ewes vaccinated with the Rv6 strain. In 4 weeks, local reaction was not longer visible and hardly palpable except in animals with transiently discharging abscesses, skin thickness being 6.1 ± 1.0, 2.8 ± 0.1 or 2.5 ± 0.1 mm (mean ± SE, n = 18) in Rv6, D5 or unvaccinated group respectively.
Anti-O agglutinating titers measured at day 7 after vaccination were higher in animals vaccinated with the Rv6 strain (Rv6 group: 7.9 ± 0.2; D5 group: 6.6 ± 0.2; unvaccinated group: 2.0 ± 0.2; mean ± SE, n = 18). Then the titers in the vaccinated groups decreased. At the time of challenge, 102 days after vaccination, agglutinating titers remained slightly higher after Rv6 vaccination (Rv6 group: 4.7 ± 0.16; D5 group: 4.1 ± 0.13; unvaccinated group: 2.3 ± 0.18; mean ± SE, n = 18). In unchallenged groups of 8 ewes, no vaccinal Salmonella was isolated at lambing or from samples taken at slaughter. Lambings of 28 among the 30 challenged ewes were observed. Eight out of 9 unvaccinated ewes and 7 out of 19 vaccinated ewes aborted after challenge; 8 out of 9 unvaccinated ewes and 9 out of 19 vaccinated ewes excreted the challenge strain (table III; fig 2). Local cutaneous reaction to the challenge injection resulted in an abscess discharging pus through the skin in 4 or 5 animals depending on the
group; the cutaneous reaction decreased more rapidly in Rv6-vaccinated animals (fig 2a). Febrile reaction was shortened in vaccinated groups (fig 2b). In D5- and Rv6-vaccinated groups, gestation lengths were not directly related to maximal values of antibody titers or of local skin reactions after vaccinal injections. Individual serological titers remained at the same level during the 2 successive weekly examinations the day of a normal or infected lambing. According to clinical and serological results, the protective effect of the Rv6 strain was higher than that of the D5 strain (table IV). Bacteriological examination at lambing detected on the whole 16 vaginal excretions associated with an infected foetus, but 1 infected foetus in the Rv6-vaccinated group which was expelled with uncontaminated lochia. Sampling of mammary secretions was possible from ewes lambing at term or near term; 1 challenge bacteria excretion was detected in colostrum from 1 out of the 2 infected lambings at normal term in the Rv6-vaccinated group. From samples taken at slaughter, bacteria of the challenge strain were detected only from the right prescapular lymph node of 1 animal in the unvaccinated and challenged group; no vaccinal bacteria were isolated from any challenge animals. DISCUSSION Attenuation of virulence in these 5 dependent or reverse mutant strains has been established in the animal species naturally susceptible to the infectious disease. The Rv6 strain was finally chosen as the vaccinal strain because of its immunogenicity and protective activity. The absence of bacteriemia detectable during the days following intravenous inoculation of the mutant strains was a first indication of their low virulence, inasmuch as bacteriemia is a necessary pathogenic step towards abortion and vaginal excretion of bacteria. However, in view of results obtained in mice (Lantier et al, 1981) as well as in sheep in the first 2 experiments, it was feared that residual virulence was too low considering the pursued objective: ie to
find a vaccine efficient with 1 injection over a reproduction period if not during the economical life of a ewe. Considering that the in vivo persistence of bacteria and the duration of host reactions to infection are related to the duration of immunity, the strains inducing relatively high biological reactions and long local persistence of bacteria were selected after the first 2 experiments (tables I and II). However, the small differences between strains with the same type of mutation would have required more animals per group and smaller variations between inoculated doses for a choice based on statistical grounds in the 2 first experiments. Evident pragmatic reasons account for the small size of groups, as well as for the use of non-gravid animals and the substitution of kinetics of numbers of viable bacteria in organs (Pardon and Marly, 1979; Lantier, 1987) for kinetics of bacteriemia and timepoint observation of bacterial colonization. In the 3rd experiment, the use of gravid ewes permitted a better simulation of the natural disease. Beside measures of indirect effects of infection such as fever, local reaction and serological response, 2 criteria essential to test a vaccine against a contagious abortive disease were taken into account: clinical issue of gestation and bacterial excretion. Housing in a closed ventilated building with a metallic slippery floor, with 1 small compartment for each group, may be factors explaining the low number of viable lambs (table III). The sc route of vaccination was chosen in view of a polyvalent vaccine (Plommet et al, 1987) including the Brucella melitensis strain Rev 1 vaccine, which is at present injected sc (Alton and Elberg, 1967; Elberg, 1981 An sc challenge and a high dose of virulent bacteria injected at mid-gestation regularly induce abortion in non-immunized animals, allowing a reduction of the number of animals in experimental groups (Pardon et al, 1983; Sanchis et al, 1985). But the sc route of challenge by-passes initial phases of the natural infectious process, and the particular site of sc injection is exposed to the risk of non-uniform behavior of different cutaneous regions in terms of immunological reaction and of bacterial implantation or dissemination. Twelve days after a virulent challenge regularly inducing placental colonization in gravid ewes, the wild strain was not detectable in samples other than at the site of challenge injection and the prescapular lymph node draining this site. This absence or undetectable level of systemic colonization is explainable by the low level and short duration of bacteriemia and by a maximum of systemic colonization observed early, about 1 week after inoculation (Lantier, 1987). Evolution of rectal temperature indicates systemic effects even more transient after inoculation of mutant strains. Among the dependent strains, the D5 strain was preferred to the D1 strain because of its greater biological effects (table II). Bacterial colonization of the inoculated region and serological response are generally more intense with reverse strains than with dependent strains (table II). Injection of the Rv6 strain caused relatively high local and regional bacterial scores (table II), so this strain was retained as the best choice among the reverse strains. In comparing Rv6- and D5-vaccination regarding protection against abortion, the higher number of animals observed during the period preceding the challenge revealed the regularity of the postvaccinal serological response and confirmed its level as slightly more elevated (about 1 dilution) in the Rv6-vaccinated ewes. In this group, the local reaction was also more intense. The vaccination-challenge test demonstrated the protective activity of the 2
mutant strains against a severe challenge. The Rv6 strain appears to be more protective in sheep and in mice (table IV; Lantier et al, 1981 Parameters measured before challenge do not provide information allowing replacement of the virulent inoculation to test the increased capacity of resistance to the infection. The low persistence of virulent bacteria in tissues precludes a comparison based on the localization and intensity of colonization of samples at slaughter and necropsy of ewes a few weeks after lambing; such parameters may be utilized in Brucellosis (Fensterbank et al, 1982). Available genetic and physiological knowledge on the dependent and reverse mutations for streptomycin (Hancock, 1981) indicates that phenotypic variability observed between mutants of the same type corresponds to several possible sites of mutation. Reverse mutants originate from at least 1 mutation added to 1 of those leading to dependent strains from streptomycin-susceptible strains. In order to vaccinate man or various animal species, dependent mutant strains were obtained from several bacterial species (Reitman, 1967; DuPont et al, 1970; Anon, 1972; Meyer et al, 1973; Vladoianu et al, 1975; Wei and Carter, 1978; Chengappa et al, 1979, 1980; De Alwis et al, 1980; Kucera et al, 1981 ). In mice orally infected by a Typhimurium serotype, Vladoianu and Dubini (1975) confirmed the stability of the reduction of virulence in dependent strains. But among reverse mutants that can be obtained in vitro from dependent strains of Enteritidis or Typhimurium serotypes, about 1 % presents a reversion to virulence of the wild parental strain; the frequency of this type of reversion from a dependent strain is 10-9 to 10-1! (Vladoianu and Dubini, 1975). Such a risk of reversion should be very low if a reverse strain is chosen directly as a vaccinal strain. The Rv6 strain was further tested with a a lower lo- reduced vaccinal dose inducing cal reaction (Pardon and Marly, unpublished results). Present knowledge on the epidemiology of the disease does not allow a precise forecast of an abortive outbreak in flocks located in endemic areas; experiments aimed at the demonstration of protection against natural contamination should cover a large number of flocks over several years (Sanchis and Pardon, 1984). Protection was confirmed with the reduced vaccinal dose in a vaccination-challenge experiment with pregnant ewes maintained in other environmental conditions, and was favourably compared with 2 injections of commercial dead vaccines (Sanchis and Pardon, 1981). ACKNOWLEDGMENTS This work was supported by a contract between INRA and the Federation Nationale des Groupements de Defense Sanitaire du B6tail and by a grant from the EEC programme of research on Mediterranean agriculture. We thank M Plommet as coordinator of the project of multivalent living vaccine against sheep abortion, and C Le Louedec, B Poutrel and M Pépin for helpful discussions. We are indebted to P L6chopier and G Bourgy for organization of experimental facilities and to R Delaunay, D Musset and W Piemont for the care and maintenance of the animals. REFERENCES Alton GG, Elberg SS (1967) Rev. 1. Brucella melitensis vaccine. A review of ten years of study. Vet Bull 37, 793-800 Anon. (1972) Les vaccins buccaux contre les entérobactéries. Rapport technique n 500. World Health Organisation, Geneva Chengappa MM, Carter GR, Chang TS (1979) A streptomycin-dependent live Pasteurella multocida type-3 vaccine for the prevention of fowl cholera in turkeys. Avian Dis 23, 57-61
with Chengappa MM, Myers RC, Carter GR (1980) A streptomycin-dependent live Pasteurella multocida vaccine for the prevention of rabbit pasteurellosis. Lab Anim Sci 30, 515-518 de Alwis MCL, Carter GR, Chengappa MM (1980) Production and characterization of streptomycin dependent mutants of Pasteurella multocida from bovine haemorrhagic septicaemia. Can J Comp Med 44, 418-422 DuPont HL, Hornick RB, Snyder MJ, Libonati JP, Woodward TE (1970) Immunity in typhoid fever: evaluation of live streptomycindependent vaccine. Antimicrob Ag Chemother 236-239 Elberg SS (1981) Rev 1 Brucella melitensis vaccine. Part 11. 1968-1980. Vet Bull 51, 67-73 Fensterbank R, Pardon P, Marly J (1982) Comparison between subcutaneous and conjunctival route of vaccination with Rev 1 strain against Brucella melitensis infection in ewes. Ann Rech Vet 13, 295-301 Hancock REW (1981) Aminoglycoside uptake - and mode of action special reference to streptomycin and gentamycin. I. Antagonists and mutants. J Antimicrob Chemother 8, 249-276 Kucera CJ, Wong JCS, Eis RL (1981) Development of a chemically altered Pasteurella multocida vaccinal strain. Am J Vet Res 42, 1389-1394 Lantier F (1987) Kinetics of experimental Salmonella abortus ovis infection in ewes. Ann Rech Vet 18, 393-396 Lantier F, Pardon P, Marly J (1981) Vaccinal properties of Salmonella abortus ovis mutants for streptomycin: screening with a murine model. Infect Immun 34, 492-497 Lantier F, Pardon P, Marly J (1983) Immunogenicity of a low virulence vaccinal strain against Salmonella abortus ovis infection in mice. Infect Immun 40, 601-607 LeMinor L, Popoff MV (1987) Designation of Salmonella enterica sp. nov, nom. rev., as the type and only species of the genus Salmonella. Request for an opinion. lnt J Syst Bacteriol37, 465-468 Meyer H, Hartmann H, Steinbach G, Koch H, Kittlick M, Stelzner A, Linde K (1973) Untersuchungen zur Salmonellose des Kalbes. 2. Mitteilung: lmmunisierungversuche mit attenuierten S.-dublin-Keimen. Archi Exp Veterinaermed 27, 301-320 Pardon P, Lantier F, Marly J, Sanchis R (1980) Mise au point d un vaccin contre la Salmonellose abortive ovine. Bull Soc Vét Prat Fr 64, 465-469 Pardon P, Marly J (1979) Experimental Salmonella abortus ovis infection of normal or Ann Immunol primo-infected CD1 mice. 130B, 21-28 Pardon P, Marly J, Sanchis R, Fensterbank R (1983) Influence des voies et doses d inoculation avec Salmonella abortus ovis sur 1 effet abortif et la réponse sérologique des brebis. Ann Rech Vét 14, 129-139 Plommet M, Bosseray N, Lantier F, Bernard F, Pardon P, Rodolakis A (1987) Simultaneous vaccination by three living attenuated strains of Brucella Salmonella and Chlamydia in mice. Vaccine 5, 27-32 Reitman M (1967) Infectivity and antigenicity of streptomycin-dependent Salmonella typhosa. J Infect Dis 117, 101-107 Sanchis R, Pardon P (1981) Comparaison de I activit6 de trois vaccins par une methode expérimentale d avortement Salmonella abortus ovis chez la brebis. Recl Méd Vdt 157, 501-505 Sanchis R, Pardon P (1984) Essai en milieu contamin6 d un vaccin atténué [yophilis6 contre I avortement de la brebis do S abortus ovis. Ann Rech Vét 15, 381-386 Sanchis R, Polveroni G, Pardon P (1985) S6rodiagnostic de la Salmonellose abortive des brebis. Bull Lab V6t (19/20), 45-51 Vladoianu IR, Dubini F (1975) Experimental model of oral antityphoid vaccination with live streptomycin-dependent Salmonella typhimurium in C57BL/6 mice. J Hyg 75, 215-218 Vladoianu IR, Dubini F, Bolloli A (1975) Contribution to the study of live streptomycindependent Salmonella vaccines: the problem of reversion to a virulent form. J Hyg 75, 203-214 Wei BD, Carter GR (1978) Live streptomycindependent Pasteurella multocida vaccine for the prevention of hemorrhagic septicemia. Am J Vet Res 391, 534-1537