Immunochemical Comparisons among Lipopolysaccharides. Bacteria Isolated from Several Luminous Marine Animals

Transcription

1 MicrobioI. Immunol. Vol. 26 (12), , 1982 Immunochemical Comparisons among Lipopolysaccharides from Symbiotic Luminous Bacteria Isolated from Several Luminous Marine Animals Toyoyasu KUWAE,* Shigeki FUKASAWA, Takeji SASAKI, and Munetsugu KURATA *Department ofphysiological Chemistry, Faculty ofpharmaceutical Sciences, and Department of Biochemistry, Faculty ofsciences, ]osai University, Keyakidai, Sakado, Saitama , and Research Centerfor Veterinary Science, The Kitasato Institute, Kashiwa, Chiba 277 (Accepted lor publication, October 25, 1982) The luminous marine fish may be divided into those that emit visible light by their own ability and those whose light is produced with the aid of symbiotic luminous bacteria or luminous substances of ingested crustaceans in their light organs (5). The symbiotic luminous bacteria inhabiting the light organs of their hosts are of great interest in regard to whether the relationship between the bacteria and the host is specific or not. Fitzgerald (3) and Reichelt et al (10) reported that certain species of luminous fish possess specific species of luminous bacteria in their light organs. On the other hand, Imai (6) reported that the luminous bacteria isolated from different fishes within the same species are not identical immunochemically. However, more detailed relationships between the hosts and the symbionts have not been clarified. Lipopolysaccharides (LPS) located in the cell walls of gram-negative bacteria not only have diverse biological effects, but also possess an a-antigenic region specific for the bacterial strain. In the present study, we investigated the relationship between the hosts and the symbionts by immunochemically studying the LPS from symbiotic luminous bacteria obtained from several different hosts. In this study, we used symbiotic luminous bacteria, facultative anaerobic enterobacteria, isolated from the light organs of five species (five genera) of luminous fish (Physiculus japonicus [PJ-I], Coelorhynchus kishinouyei [CK-I], and Ventrifossa garmani [VG-I] in the order Gadida; Chlorophihalmus albatrossis [CA-I] in the order Clupeida: and Acropoma japonicum [AJ-I b] in the order Percida) and from the light organ ofa luminous squid tdoryteuthis kensaki [DK-I]). For the name of the luminous bacterium isolated from each host, the abbreviation given in brackets ([]) is used hereafter. The morphological aspects and the derivations of these symbionts have been reported by Fukasawa (4). We identified luminous bacteria PJ-I, CK-I, VG-I, and CA-I as Photobacterium phosphoreum from the results obtained by comparison with the criteria of Bergey's manual (8th edition) (I) and with standard 1181

2 1182 T. KUWAE ET AI. strains of the photobacterial genera, Photobacterium phosphoreum ATCC 11040, Photobacterium leiognathi ATCC 25521, Vibrio.fischeri ATCC E7744, and Vibrio harveyi ATCC Luminous bacteria DK-I and AJ-Ib were species of Photobacterium other than P. phosphoreum but remain unidentified. This taxonomic study is in progress (manuscript in preparation). Four strains (~J-I, CK-I, VG-I, and CA-I) were cultured with shaking at 20 C for 20 to 24 hr in a medium containing NaCI 30 g, MgS0 4 7H g, Bacto peptone (Difco) 5.0 g, yeast extract (Difco) 3.0 g, glycerol 3.0 g, and 0.2 M phosphate buffer (ph 7.2) 10 ml in 1,000 rnl of distilled water and adjusted to ph 7.2 with I N NaOH. DK-I and AJ-Ib were similarly cultured at 25 C for 20 to 24 hr. LPS from these symbiotic luminous bacteria were extracted from formalin-treated cells by the method of Westphal and J ann (II) and purified by repeated ultracentrifugations (105,000 X g for 2 hr). The LPS preparations obtained had only a low 2-keto 3-deoxyoctonate (KDO) content (0.2%" w/w) or lacked it and consisted of 28 to 56% carbohydrate, 5 to 13% glucosamine, 3 to 5% phosphorus, and II to 27% fatty acid. The preparations also contained 2 to 8% nucleic acid and 0.4 to 1.6% protein except that VG-I LPS fell out ofthe range with 18.8%. Ofthese LPS, we reported the chemical and biological properties of PJ-I LPS in a previous paper (7). In order to show whether there is any similarity among these LPS or not, we investigated these LPS by means of immunochemical analysis. First, we performed comparative double immunodiffusion tests with these LPS and the antisera against the above bacteria according to the method of Ouchterlony (9). Antisera against LPS from these symbiotic luminous bacteria were prepared by injecting heat-killed (100 C, 2 hr) cells at a density of 10 5 or 10 9 cells per ml into rabbits as described previously (7). As shown in Fig. I, LPS from PJ-I, DK-I, AJ-I b, and CK-I formed one or two precipitin lines in the gel plate only when they were reacted with the homologous antiserum, for example, PJ-I LPS against anti-pj-i, but not ~J-I LPS against anti-dk-l. On the other hand, CA-I LPS and VG-I LPS formed two or more precipitin lines even when they were reacted with the heterologous antiserum (CA-I LPS against anti-vg-i or VG-I LPS against anti CA-I) as well as with the homologous antiserum. These results show that LPS from CA-I and VG-I possess a common antigen and that at least one of these precipitin lines formed by reaction of the LPS and homologous or heterologous antiserum is a line of complete coalescence. Further, absorption tests were performed with CA-I LPS and VG-I LPS. As shown in Fig. 2, anti-ca-i or anti-vg-i antibody was completely absorbed by both the homologous and the heterologous LPS. The precipitin lines seen between peripheral wells (for example, between well a and well a") are the results of reaction between the antisera and excess LPS during the preparation of the absorbed antisera. These results show that the LPS from CA-I and VG-I are identical immunochemically and that the other LPS, from PJ-I, DK-I, AJ-Ib, and CK-I, do not have any similarity with each other. The electrophoretic mobilities of these LPS preparations were investigated by immunoelectrophoresis in gel plates. As shown in Fig. 3, the immunoelectrophoretic patterns of LPS from both CA-I and VG-I were almost the same and

3 NOTES 1183 Fig. I. Double immunodiffusion tests among LPS from various symbiotic luminous bacteria. The double immunodiffusion tests for these LPS and their antisera were performed according to the method of Ouchterlony (9). The gel plate was made with 1% agarose (A-37: Nakarai Chemical, Ltd.) in 0.07 M barbital buffer (ph 8.6). All LPS were dissolved with physiological saline containing 0.5% sodium deoxycholate (DOC) at I mg per ml. Central wells a, b, c, d, e, and f were filled with anti-ca-i, anti-vg-i, anti-pj-i, anti-dk-i, anti-aj-ib, and anti-ck-i sera, respectively. Peripheral wells 1,2,3,4,5, and 6 were filled with CA-I LPS, VG-I LPS, PJ-I LPS, CK-I LPS, DK-I LPS, and AJ-Ib LPS, respectively. The gel plate was kept in a humid chamber at room temperature. The results 24 hr later are shown. Fig. 2. Absorption of anti-ca-i and anti-vg-l sera with homologous LPS or heterologous LPS. One mg of CA-I LPS was added to I ml of homologous anti-ca-i or heterologous anti-vg-i serum. Similarly, I mg of VG-I LPS was added to I ml of homologous anti VG-I or heterologous anti-ca-i serum. These mixtures were incubated at 37 C for I hr and then overnight at 4 C. After incubation, the mixtures were centrifuged at 1,500 <s for 10 min at 4 C, and the supernates were analyzed by double immunodiffusion tests as described for Fig. I. The results 48 hr later are shown. Central wells I and 2 were filled with CA-I LPS and VG-I LPS, respectively. Peripheral wells a and b were filled with anti CA-I and anti-vg-i sera, a" and b" were filled with anti-ca-i andanti-vg-i sera, each absorbed with the homologous LPS, and a' and b' were filled with anti-ca-i and anti VG-I sera, each absorbed with the heterologous LPS, respectively.

4 1184 T. KUWAE ET AL Fig. 3. Immunoelectrophoretic patterns of LPS from various luminous bacteria. Eight-ul quantities of LPS solutions (I mg per ml in physiological saline containing 0.5% DOC) were placed in wells (3 mm in diameter) in the gel plate, and the plate was run at 2 rna per em of gel width in 0.07 M barbital buffer (ph 8.6). The electrophoresis was stopped when rat serum albumin stained with Evans blue as marker had migrated approximately 3.5 em into the gels, and then each trench in the gel plate was filled with the corresponding antiserum. The plate was kept in a humid chamber at room temperature. Figure 3 shows the results 24 hr later. Wells a, b, c, d, e, and f were original points of VG-I LPS, CA-I LPS, PJ-I LPS, CK-I LPS, DK-I LPS, and AJ Ib LPS, respectively. Trenches 1,2,3, 4,5, and 6 were filled with anti-ca-i, anti-vg-i, anti-pj-i, anti-ck-i, anti DK-I, and anti-aj-ib sera, respectively. consisted of two bands with slow electrophoretic mobility and of one obscure band with fast electrophoretic mobility but the bands of these two LPS slightly differed from each other in their electrophoretic mobility. The immunoelectrophoretic patterns of the other LPS exhibited one or two clear bands with fast electrophoretic mobility and were specific for each LPS. These results also show that LPS from ~J-1, CA-l, and VG-l are heterogeneous LPS consisting of two or three fractions. Further, we investigated by the quantitative precipitation test whether there is a difference in the affinity to certain antisera among these LPS or not. The results are expressed as the amount of nitrogen in precipitates formed by reaction between anti-ca-l antibody and homologous or heterologous LPS. As shown in Fig. 4, the precipitation curves obtained by reacting a constant amount of anti CA-l antiserum with varying concentrations of CA-l LPS or VG-l LPS nearly overlapped, while other LPS formed hardly any precipitate, antigen-antibody complex, with anti-ca-l serum. The maximal reactions of anti-ca-l serum with CA-l LPS or VG-l LPS were obtained with 2 mg of these LPS per ml. These results also support the results obtained by the double immunodiffusion tests with LPS from these bacteria. However, in the case of CA-l LPS, the values obtained were due only to antibody in the antigen-antibody complex since the protein content of this LPS is low (1.4%, wjw), while in the case of VG-l LPS, the values obtained were expressed as the total amount of antibody nitrogen plus antigen nitrogen in the complexes formed, since this LPS includes approximately 19% (wjw) protein. Therefore, we consider that the precipitation curve obtained by reaction

5 .. E 0.6 ::l Ql!II Ẹ... ~ Ql.. " 'is. 0.3 u Ql ii 0.2 z ~ 0.1 NOTES ", II..~ LPS added (mg/ml) Fig. 4. Quantitative precipitation curves of reaction between anti-ca-i serum and homologous LPS and heterologous LPS. Each LPS was dissolved in physiological saline to the desired concentration. One-tenth ml quantities of various concentrations of each LPS were added to test tubes containing the same volume of anti-ca-i serum. These mixtures were incubated at 37 C for I hr and additionally overnight at 4 C. The mixtures were then centrifuged, and the precipitates were washed three times with cold saline at 1,500 Xg for 10 min at 4 C. Each precipitate was dissolved in 0.2 ml of 0.1 N NaOH and the protein content of the precipitates was determined by the method of Lowry et al (8) with bovine serum albumin as the standard, and the values obtained were converted into the amounts of nitrogen (2). The results are expressed as the amount of nitrogen in the precipitates per ml of anti-ca-i serum. Symbols:., CA-I LPS; 0, VG-I LPS; o, CK-I LPS; \1, PJ-I LPS; 0, DK-I LPS; and 6, AJ-Ib LPS. between anti-ca-i antibody and heterologous VG-I LPS is actually lower than that obtained by reaction between anti-ca-i antibody and the homologous CA-I LPS. This result appears to suggest that there is more or less difference in the affinity of the common antigen between LPS from CA-I and VG-I for anti-ca-l. However, this question should be further examined by using protein-free VG-I LPS since protein comprises a large part of its composition (approximately 19%, w/w). In this investigation, we showed, by comparative immunochemical studies among the LPS preparations from symbiotic luminous bacteria and the antisera against these bacteria, that, except for the LPS from CA-I and VG-I, the LPS are specific for the bacteria isolated from different luminous animals; in other words, the relationship between these symbiotic luminous bacteria and their hosts is a specific one. The relationship between CA-I or VG-I and its host is also specific, since the two bacteria are different from each other morphologically (4). However, in the present study, the questions of whether hosts belonging to the same species always carry the same strain of symbiotic luminous bacterium and whether an individual host possesses only a single strain ofsymbiotic luminous bacterium were not answered. In order to answer these questions, further study is in progress. An interesting finding is that the two fish serving as hosts for CA-I (rod shape) and VG-I (spherical shape) belong to the order Clupeida and Gadida, respectively,

6 1186 T. KUWAE ET AL which are themselves very different as fish and whose bacteria also differ morphologically (4), yet the two bacteria are very closely related in terms of LPS. On the other hand, the fish that are hosts for PJ-l, CK-l, and VG-l belong to the same order Gadida, with the hosts showing a closer relationship to each other than to the host of CA-l, but these bacteria have no immunochemical relationship to each other in terms of LPS in spite of belonging to the same species, P. phosphoreum. These results were unexpected. Thus the immunochemically comparative studies among LPS from various symbionts seems to be a useful means for the study of the relationship between symbiotic luminous bacteria and their hosts, since the usual LPS possess three regions (a-antigenic region, R-core region and lipid A region) and bear the specificity of the strain or species of bacterium in the polysaccharide region which is used for diagnosis in the bacterial taxonomic system. Moreover, further investigations on the LPS from these symbiotic luminous marine bacteria may give us new information in the field of bacterial LPS, since LPS from these symbiotic luminous marine bacteria are characterized by low content (0.2%, wjw) or absence of KDO, which is a basic component of the constitution of the usual LPS and which plays a role in linking the polysaccharide region and the lipid A region, as well as by possessing diverse biological effects (manuscript in preparation). REFERENCES I) Buchanan, R.E., and Gibbons, N.E. (eds), Bergey's manual of determinative bacteriology, 8th ed, Williams and Wilkins Co., Baltimore. 2) Chase, M.W., and Williams, C.A. (eds), Chemical analyses, p In Methods in immunology and immunochemistry, Vol. II, Academic Press, Inc., New York and London. 3) Fitzgerald, J.M Classification of luminous bacteria from the light organ of the Australian pinecone fish, Cleidopus gloriamaris. Arch. Microbiol. 112: ) Fukasawa, S Studies of symbiotic luminous bacteria. Gekkan Kaiyohkagaku 10: (in Japanese). 5) Haneda, Y Luminescent fish with open-type glands containing luminous bacteria or luminous substances of ingested crustaceans. Sci. Rep. Yokosuka City Mus. 27: ) Imai, H Studies of symbiotic luminous bacteria. Seiikai Zasshi 61: (in Japanese). 7) Kuwae, T., Sasaki, T., and Kurata, M Chemical and biological properties of lipopolysaccharide from a marine bacterium, Photobacterium phosphoreum PJ-1. Microbiol. Immunol. 26: ) Lowry, O.H., Rosebrough, N.J., Farr, A.L., and Randall, R.J Protein measurement with the Folin phenol reagent. J. BioI. Chern. 193: ) Ouchterlony, O Diffusion-in-gel methods for immunological analysis. II, p In Kallas, P., and Waksman, RH. (eds), Progress in allergy, Vol. VI, Karger, Basel and New York. 10) Reichelt, J.L., Nealson, K., and Hastings, J.W The specificity of symbiosis: pony fish and luminescent bacteria. Arch. Microbiol. 112: II) Westphal, 0., and Jann, K Bacteriallipopolysaccharides: extraction with phenol water and further applications of the procedure, p In Methods in carbohydrate chemistry, Vol. V, Academic Press, Inc., New York and London. (Received for publication,june 17, 1982)