Japan. J. Med. Sci. Biol., 20, 205-212, 1967 STUDIES ON THE ASAKUSA GROUP OF ENTEROBACTERIACEAE (EDWARDSIELLA TARDA) RIICHI SAKAZAKI Department of Bacteriology I, National Institute of Health, Tokyo (Received: February 16th, 1967) SUMMARY: A group of the Enterobacteriaceae has been studied using 256 cultures which were isolated mainly from snakes. Although the term gasakusa h group was first used as the designation for the organisms by Sakazaki and Murata (1962), a scientific name Edwardsiella tardy was recently proposed by Ewing and his co-workers. The members of the group are closely related to the Salmonella group in hydrogen sulfide production and lysine decarboxylation, but differ in their indol production, and mannitol-, arabinose-, xylose-, and trehalose-non-fermentable characteristics. Within the 256 cultures of the group, seventeen O groups and eleven H antigens were established, and an antigenic schema was set up for 18 serotypes of the group. It was considered that the organisms are normal intestinal inhabitants of reptiles. Several cultures were isolated from human pathological materials, but no conclusive results on pathogenicity have been obtained in this study. INTRODUCTION An unusual biotype of Enterobacteriaceae has been recovered from snakes while performing a survey to determine the prevalence and distribution of Salmonella and Arizona in reptiles. The organisms were first recovered in 1959 and reported in 1962 by Sakazaki and Murata according to the morphology, physiology, and biochemical characteristics of 153 isolates possessing similar properties. A complete search of the literature failed to produce a description of these organisms, therefore, these authors suggested the name gasakusa hand the organisms to be placed in the family Enterobacteriaceae. One hundred and three additional cultures were subsequently added to bring the total number of isolates examined to 256. Ewing, McWhorter, Escobar, and Lubin (1964), in independent studies, examined 37 isolates submitted to the Communicable Dssease Center, Atlanta, Georgia by various laboratories. The biotype of these organisms was labeled 1483-59. In 1965 this group suggested the term gedwardsiella has the generic name and gtarda has the specific name for this organism which they incorporated within the family Enterobacteriaceae. King and Adler (1954) likewise reported a strain of organisms with the same characteristics and proposed the designation of gbarthoromew hgroup. Thus, the purpose of this paper is to report the biochemical and serological studies of the gasakusa hgroup, although a summary was presented by the present author and published in the International Bulletin of Bacterial Nomenclature and Taxonomy 15: 45-47, 1965.
206 SAKAZAKI Vol. 20 MATERIALS AND METHODS Specimen source: Of the 256 cultures employed here, 248 were originally isolated from snake feces, 2 from seal intestinal contents obtained at autopsy following a febrile course and death, and 5 from human patients with acute gastroenteritis. Morphology: Morphological and tinctorial examinations were conducted on organisms grown on trypticase soy agar at 37 C for 18 to 24 hr. The flagellation was determined by electron micrographs. Motility was determined in SIM medium. Biochemical examination: The oxidase test, Gapy and Hadley (1957); indol production in tryptone broth; hydrogen sulfide production in triple sugar iron agar; utilization of organic acids including citrate, d-tartrate, and mucate, Kauffmann and Petersen (1956); utilization of glucose, citrate, acetate, and alginate as a sole carbon source in 7 days on ammonium agar (Simmons' agar base); citrate utilization on Christensen's agar; urea decomposition in 4 days, Christensen (1946); gelatin liquefaction with denatured gelatin strip in 10 days, Lautrop (1956); malonate utilization in 2 days, Ewing, Davis, and Reavis (1957); decarboxylation of lysine, arginine, and ornithine in 4 days, MƒÓller (1955); lipolysis of tributylin, corn oil, and triacetin, Davis and Ewing (1964); KCN test, MƒÓller (1954); sugar fermentation reaction in 30 days using tripticase broth with Andrade's indicator and 1 % carbohydrates for the following-arabinose, cellobiose, glucose, lactose, maltose, melezitose, melibiose, raffinose, rhamnose, sucrose, trehalose, xylose, adonitol, dulcitol, erthritol, inositol, mannitol, sorbitol, and salicin; Voges-Proskauer using Barritt's method after 2 days at 22 C; methyl red after 4 days at 37 C; nitrate reduction after 12 hr. Antiserum production: The antigens used for the production of the O antisera were prepared by heating 18-hr brain heart infusion broth cultures at 100 C for 2 1/2 hr. Rabbits received four intravenous injections in the amounts of 0.5, 1.0, 2.0, and 4.0 ml of the antigens. Since H antigenicity was poor in many cultures, serial passage through semisolid medium was necessary to insure active motility. It was found that a medium ph range of 6.6 to 6.8 produced the best H antigens. One loopful of the semisolid medium cultures was inoculated into brain heart infusion broth (ph 6.6) which was then incubated overnight at 37 C and subsequently added 0.3 % formalin for preservation. The inoculation schedule was as stated above for O antiserum production. Agglutination test: Tube agglutination tests were employed for the determination of O and H antigens. For the O agglutination test antigen suspensions were prepared from growth on brain heart infusion agar at 37 C for 18 hr and heated at 100 C for 1 hr. The suspensions were standardized to McFarland nephelometric standard No. 3 with 0.5 % saline. Antigens for H agglutination test were similar to those for H antiserum production, but 0.1 % merthiolate was used instead of formalin. Cross reaction was tested by reciprocal agglutination tests after 2 and 18 hr in a 50 C water bath for the H and O agglutinations, respectively. Absorption tests: Absorption tests were carried out when antigenic relationships were indicated by cross reaction in dilution of 1 : 50 and higher in O tube tests and in dilution of 1 : 100 and higher in H tube tests.
1967 ASAKUSA GROUP OF ENTEROBACTERIACEAE 207 RESULTS 1. Morphological Characteristics. All the cultures were composed of Gram-negative rods showing peritrichous flagellation. Neither capsule nor spore was seen. 2. Physiological and Biochemical Characteristics. All the 256 cultures studied here grew in ordinary media. In broth the organisms produced clouding. When the organisms were planted on ordinary agar plates, they formed moist, smooth, circular and translucent colonies bearing a striking resemblance to those of shigellae. They grew on MacConkey, SS, and brilliant green agar. On Table 1. The biochemical reactions given by the 256 cultures of the Asakusa group Remarks: + =positive, - =negative, (+) =weakly positive. Fermentation; + = positive within 24 hr, - =negative within 30 days' incubation.
208 SAKAZAZI Vol. 20 bismuth sulfite agar plate colonies of the cultures were black in appearance. The biochemical characteristics of the 256 cultures studied are summarized in Table 1. The 256 cultures behaved as follows: Nitrate was reduced to nitrite; hydrogen sulfide was produced strongly in butt of triple sugar iron agar; Voges-Proskauer was negative and methyl red was positive; citrate was utilized on Christensen's agar but not on Simmons' agar; urease was not produced; gelatin was not liquefied; phenylalanine was not deaminated; lysine and ornithine were decarboxylated but arginine was not; glucose, citrate, acetate, and alginate were not utilized on ammonium agar, as a sole carbon source; citrate and d-tartrate were utilized in Kauffmann-Petersen's broth, but mucate, i-tartrate, and 1-tartrate were not in the broth; lipase to tributylin, corn oil, and triacetin were not produced; esculin was not hydrolyzed; beta-d-galactosidase was not produced; and KCN broth did not allow to grow. Fermentation of carbohydrates was characteristic of the cultures as shown in the table. None of the 256 cultures fermented arabinose, cellobiose, lactose, melezitose, melibiose, raffinose, rhamnose, sucrose, trehalose, xylose, adonitol, dulcitol, erthritol, inositol, mannitol, sorbitol, and salicin. Of the 18 carbohydrates tested, only two, glucose and maltose, were attacked promptly with gas production. Although Ewing et al. stated that 3 % of their 37 cultures fermented arabinose, none of ours did so. 3. Serological Characteristics. By reciprocal agglutination and agglutinin absorption tests, 17 O antigen groups were demonstrated within the 256 cultures. The results of O tube agglutination tests using the representative strains for each O antigen group are indicated in Table 2. As shown Table 2. O antigenic relationships Remarks: The sign minus indicates negative in dilution of 1 : 50 and higher.
1967 ASAKUSA GROUP OF ENTEROBACTERIACEAE 209 in the table, unilateral reactions were observed in O antisera 1, 3, 5, 8, and 12, although no cross reactions with the other O groups were recognized in any of the 17 O antisera. Therefore, the determination of the O antigen could be accomplished by employing appropriately diluted unabsorbed antisera. In a preliminary examination, H agglutination tests were carred out using H antigens prepared in a manner similar to those described for Salmonella cultures, but the result was not satisfactory. Thus, conditions for the preparation of H antigens to yield a satisfactory H agglutination were investigated. Cultures passed successively through semisolid medium were inoculated into 4 tubes containing brain heart infusion broth adjusted to ph 6.6, 7.0, 7.4, and 7.8, respectively. After overnight incubation at 37 C, each. broth culture was divided into two parts, and one was preserved with 0.3 % formalin and the other with 0.1 % merthiolate. The results of H agglutination tests with these antigens are indicated in Table 3. Table 3. Influences of preservative and ph of broth to H agglutination As shown in the table, a ph over 7.0 was unsatisfactory for preparation of H antigens, and the formalized antigen also reduced its agglutinability. From these reasons, to obtain satisfactory H antigens, the actively motile organisms were inoculated into brain heart infusion broth adjusted to ph 6.6, incubated overnight at 37 C and then added 0.1 % merthiolate. H antigenic relationships among the 256 cultures employed here were studied using the antigen described above, and 11 H antigens were detected in the cultures. Cross agglutination pattern of the 11 H antigens revealed a reciprocal reaction between
210 SAKAZAKI Vol. 20 the H antisera 6 and 7, and a unilateral reaction in antiserum 4. However, no reactions with other H antisera were given by H antigens 1, 2, 3, 5, 8, 9, 10, or 11.T o determine if phase variation occurred in the H antigens, tests were performed b y using the modified technic of Edwards and Bruner (1942). All 256 cultures were inoculated into semisolid medium containing homologous H antiserum from which O antibody was removed by absorption with a suspension of the homologous culture boiled at 100 C for 2 1/2 hr, and then incubated at 37 C for 10 days. All cultures became immobile in the semisolid medium. From the results, it was concluded that the H antigens of these organisms were monophasic. All the 256 cultures were tested for K antgen by means of of tube agglutinatio n t est using both living and boiled broth cultures with corresponding O antiser a. It appeared that living cultures, as well as boiled cultures, were agglutinated in a homologous titer of the serum in all cases. To observe O and H antigenic relationships to other groups of the family Entero - bacteriaceae, each O and H antigens of the organisms were tested with 45 O and 56 Table 4. H antigenic relationships Remarks: The sign minus indicates negative in dilution of 1 : 100 a nd higher. Table 5. Antigenic scheme of the Asakusa group
1967 ASAKUSA GROUP OF ENTEROBACTERIACEAE 211 H antisera of Salmonella, 146 O and 41 H antisera of Escherichia, 32 O and 71 H antisera of Citrobacter, 49 O and 17 H antisera of Proteus, 53 O and 32 H antisera of Cloaca (Enterobacter cloacae), and 29 O and 23 H antisera of Hafnia. However, no significant reactions were obtained. From the results obtained with the O and H antigens of the cultures employed here, it is possible to establish an antigenic schema for this group. As shown in Table 4, the schema consists of 17 O groups, 11 H antigens, and 18 serotypes. DISCUSSION From their biochemical characteristics, it is obvious that the organisms employed here are involved into a definite group of bacteria within the family Enterobacteriaceae. Based on the morphological, physiological, and biochemical characteristics of the organisms described, a classification with the Adansonian concept has been accomplished to connect these organisms with members of the genera Salmonella, Escherichia, Citrobacter, and Proteus (P. mirabilis and P. vulgaris). In the classification, similarity value between the organisms and members of the other genera was indicated as under 50 %. In the preliminary paper, Sakazaki and Muata (1962) suggested the name gasakusa h, after the name of a ward of Tokyo where the original strain was isolated, for the group of organisms, although no generic or specific name was given. Later, Ewing and his co-workers (1965) suggested the generic term gedwardsiella hand the specific epithet g tarda hfor the organisms. The author has no objection to the nomenclature of Edwardsiella tarda by Ewing et al., because gasakusa hused the vernacular name and thus has no standing in the nomenclature. The organisms of the group are similar to members of the genus Escherichia in the IMViC reaction but they produce abundant hydrogen sulfide. The organisms are related to the genus Salmonella with regard to the hydrogen silfide production, lysine and ornithine decarboxylation, the lack of malonate utilization, urea decomposition, phenyllanine deamination and the fermentation of lactose, sucrose, adonitol, and salicin. However, they are distinguished from the Salmonella in their indol production, and mannitol, arabinose, xylose, and trehalose-non-fermentable characteristics. They also resemble members of the genus Proteus, since they are incapable of fermenting a number of carbohydrates, especially mannitol, but may be differentiated from this group due to negative phenylalanine deamination, urea decomposition, and gelatin liquefaction. Although the majority of the cultures studied by Ewing and his co-workers originated from human sources, it may be very rarely found in warm-blood animals including humans. It should be noted, on the other hand, the almost all of the cultures in the author's collection originated in snakes, and that there were only 7 of the 256 cultures isolated from human and animal sources. It is probable that the organisms are intestinal inhabitants of reptiles. King and Alder (1964) isolated this organism from the blood of a patient suffering from a febrile disease, and they considered the organism to be the causative agent in this case. Ewing and his co-workers stated 34 cultures from human sources, 12 originating in pathological materials such as diarrheal stool, blood, urine, and wound exudates. In the author's collection, the 5 cultures from human sources were isolates from patient's stools with gastroenteritis. However, no conclusive results on pathogenicity of the group have been obtained as yet.
212 SAKAZAKI Vol. 20 ADDENDUM After contribution of this article, the author was told, by Dr. W. H. Ewing, Communicable D isease Center, Atlanta, Georgia, U. S. A., that a brief summary of a study titled ga provisional antigenic scheme far Edwardsiella tarda is presented hwas presented by McWhorter. C., W. H. Ewing, and R. Sakazaki at the meeting of AMS in New York City in M ay, 1967. The antigenic symbols described in this paper are independent from theirs. Therefore, the symbols should be adjusted to those of McWhorter et al. in future to avoid a confusion., A REFERENCES CHRISTENSEN, W. B. (1946): Urea decomposition as a means of differentiating Proteus and paracolon cultures from each other and from Salmonella and Shigella. J. Bacteriol., 52, 461-467. DAVIS, B. R. AND EWING, W. H. (1964): Lipolytic, pectolytic, and alginolytic activities of Enterobacteriaceae. J. Bacteriol., 88, 16-19. EWING, W. H., DAVIS, B. R. AND REAVIS, R. W. (1957): Phenylalanine and malonate media and their use in enteric bacteriology. Public Health Lab., 15, 153-160. EWING, W. H., MCWHORTER, A. C., ESCOBAR, M. R. AND LUBIN, A. H. (1965): Edwardsiella, a new genus of Enterobacteriaceae Bact. Nomen. Taxon., 15, 33-38. based on a new species, E. tarda. Internat. Bull. GABY, W. L. AND HADLEY, C. (1957): Practical Laboratory test for the identification of Pseudomonas aeruginosa. J. Bacteriol., 74, 356-358. KAUFFMANN, F. AND PETERSEN, A. (1956): The biochemical group and type differentiation of Enterobacteriaceae by organic acids. Acta Path. Microbiol. Scand., 38, 481-491. KING, B. M. AND ADLER, D. L. (1964): A previously undescribed group of Enterobacteriaceae. Amer. J. Clin. Pathol., 41, 230-232. LAUTROP, H. (1956): A modified Kohn's test for the demonstration of bacterial gelatin lique - faction. Acta Path. Microbiol. Scand., 39, 357-384. MOLLER, V. (1954): Diagnostic use of the Braun KCN test within the Enterobacteriaceae. Acta Path. Microbiol. Scand., 34, 115-126. SAKAZAKI, R. AND MURATA, Y. (1962): The new group of the Enterobacteriaceae. The Asakusa group. Japan. J. Bacteriol., 17, 616-617 (with Japanese text). SAKAZAKI, R. (1965): A proposed group of the family Enterobacteriaceae nternat. Bull. Bact. Nomen. Taxon., 15, 45-48., the Asakusa group. I