Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Journal of General Microbiology (1g77), roo, 355-361 Printed in Great Britain 355 Genetic Determinants of the Synthesis of the Polysaccharide Capsular Antigen K27(A) of Escherichia coli By G. SCHMIDT, BARBARA J A N N A N D K. J A N N Max-Planck-lnstitutfur Immunbiologie, D 7800, Freiburg, Germany I D A 0 R S K O V A N D F. 0 R S K O V Statens Seruminstitut, WHO Collaborative Centrefor Reference and Research on Escherichia, DK 2300, Copenhagen, Denmark (Received 22 November 1976; revised 8 February 1977) SUMMARY Most of the his+ hybrids from crosses between the Escherichia coli donor Hfr45(08 :K27) and different E. coli Og recipients expressed the donor 0 8 antigen specificity and produced the capsular antigen K27. Therefore these hybrids must have inherited the his-linked donor rfb region determining the synthesis of 08specific polysaccharides as well as his-linked genes involved in K27 antigen synthesis. In the living state these hybrids were inagglutinable in 0 8 antiserum like the donor cells. However, when E. coli K I and ~ 0 8 : Q 2 - were used as recipients most of the his+ hybrids were agglutinable in 0 8 and K27 antisera. The amounts of K27 antigen present in these hybrids, designated as K27i (intermediate) forms, were sufficient to evoke the production of K27 antibodies in rabbits, but insufficient to inhibit 0-agglutination of the respective cells. The additional transfer of the trp region of E. coli 0 8 :K27 into such K27i forms frequently resulted in 0-inagglutinable K27f hybrids. This is attributed to the introduction of trp-linked genes which apparently play a role in the synthesis of K27 capsular antigen. Thus it is concluded that at least two gene loci, one close to his and the other close to trp, are required for the synthesis of the complete capsular antigen K27. INTRODUCTION Many Escherichia coli serotypes produce capsular antigens (K antigens) which usually consist of acidic polysaccharides (Jann & Westphal, 1975). A strong indication of the presence of K antigens is the inagglutinability of living bacteria in homologous 0 antiserum (Kauffman, I 966). Differences in the temperatures required to overcome this 0-inagglutinability and in the thermoresistance of K antibody-binding and agglutinating capacities have been decisive for the division of K antigens into three types, the A, B and L antigens {Kauffmann & Vahlne, 1945; Kauffmann, 1966). The A antigens are relatively thermostable capsules which, in contrast to B and L antigens, still inhibit agglutination of the bacteria in homologous 0 antiserum after heating at IOO "C for 2-5 h. Strains with K(A) antigens have to be autoclaved to make them agglutinable in 0 antiserum. Although in recent years there has been considerable progress in understanding the nature and role of K antigens (0rskov et al., 1971; Kaijser, 1973; McCabe et al., 1975; Nicholson & Glynn, 1g75), knowledge of their genetics is scanty. Recently, a gene locus participating in the synthesis of the K(L) antigens KI, K4, KIOand K54 has been found closely linked to serA and was termed kpsA (0rskov & Nyman, 1974; 0rskov, Sharma & 0rskov, 1976). Another gene apparently needed for the expression of these K(L) antigens has not yet been Downloaded from www.microbiologyresearch.org by IP: 78.47.27.170 On: Thu, 20 Oct 2016 11:30:53 356 G. S C H M I D T A N D OTHERS Table Strain I. Characteristics of E. coli strains Derived from Serotype Relevant genotypet Donor Hfr4o(cw)* ~56b 0 8 :K27 Hfr45(ccw) ~56b 0 8 :K27 XYl KL25(CW) K12 Rough thi Recipient F7 1 ~56b 0 8 :K27 his trp ara lac str F464 ~56b 0 8 :K27his met pro ara mtl str F713 Bil61142 0 9 :K29his trp str ~720 SU397314I 09:K3 I his str F730 A295b 0 8 :K42his trp ara str 2402 E69 O9:K30his str 2578 K12 Rough his trp str * cwlccw, Clockwise/counterclockwisedirection of transfer. t Genes for his, histidine; met, methionine; pro, proline; thi, thiamin biosynthesis: ara, arabinose; lac, lactose; mtl, mannitol; xyl, xylose utilization: str, streptomycin resistance. mapped. Genes participating in the synthesis of the K(A) antigen of E. coli Og :K26 (Orskov & Orskov, 1962) are located close to the his-linked rfb region which is involved in the synthesis of 0 antigens (Stocker & Makela, 1971). This paper describes experiments demonstrating that, in addition to this his-linked gene, a second trp-linked gene is needed for the synthesis of the complete K27(A) antigen of E. coli 0 8 :K27. METHODS Bacterial strains. These are listed in Table I. The donor strains Hfr4o and Hfr45 have their points of origin for the chromosome transfer near xyl and his, respectively (Schmidt, Jann & Jann, 1970). The E. coli K 1 2 donor ~ ~ transfers 2 5 its chromosome with ilv as a leading marker (Low, 1968). Media. D1,,-agar (Schlecht & Westphal, 1967) and Merck Standard I broth were used as complete media. Recombinants were selected on minimal agar (Lederberg, 1950) containing streptomycin (100,ug ml-l) to inhibit the growth of the donors. When required, amino acids were added to a final concentration of 20 ,ug ml-l. Mating experiments. These were performed as described previously (Schmidt et al., I 970). Recombinant clones were first transferred on to the same medium used for primary selection and then streaked on to complete agar from which single colonies were isolated. Serologicalmethods. The purified recombinants were scored for their antigenic properties by both slide and tube agglutination tests (Kauffmann, 1966). Absorption of antisera with whole bacteria was performed as described previously (Mayer & Schmidt, 1973). Some recombinants were also examined by immunoelectrophoresis. Antisera. Rabbit 0 antisera were prepared as described previously (Mayer & Schmidt, 1973) using boiled suspensions of non-capsulated E. coli strains (08 :K27- and Og :K29-) as antigens. K27 antiserum was obtained by immunization with a formalin-treated suspension of E. coli ~ 7 8 2which , is a rough form (08-) encapsulated with K27 antigen (Whang et al., 1972). Imrnunoelectrophoresis. This was done as described by Scheidegger (I 955) using extracts prepared as described by Orskov et al. (1971). Downloaded from www.microbiologyresearch.org by IP: 78.47.27.170 On: Thu, 20 Oct 2016 11:30:53 Genetic determinants of K antigens 357 Table 2. Serological analysis of his+ recombinants selected from crosses between the donor Hfr45(08 :K27) and various acapsular recipients Recipient Serotype of recipient No. of recombinants analysed Serological types identified* \ A O8:K- 0g:K- K27 Rough K27i I12 8 O8:K270 104 0 0 F464 F7I 3 Og :K2g7 86 0 0 I00 7 ~720 Og:K318 87 0 0 I00 5 2402 0 9 :K3o9 91 0 0 I00 0 F730 0 8 :K42257 I0 0 7 0 240 Rough (K12) 0 3 8 109 2578 I22 2 * O8:K-, agglutinable in 0 8 antiserum, non-reactive in K27 antiserum. O9:K-, agglutinable in Og antiserum, non-reactive in K27 antiserum. K27, agglutinable in K27 antiserum, non-reactive in 0 antisera. Rough, unspecific agglutination in 3.5 % saline. K27i, agglutinable in 0 8 and K27 antisera. Table 3. Inheritance of trp and serologicalpatterns of 257 his+ hybridsfrom the cross between the donor Hfr45(08 ;K27) and F730(08 :K42-) his+ hybrids with K forms* No. r K27- A K27i \ K27+ 248 9 238 I Recipient trp 9 I 2 6 Donor trp+ Total 257 I0 240 7 * K27-, agglutinable in 0 8 antiserum, non-reactivein K27 antiserum. K27i, agglutinable in K27 and 0 8 antisera. K27+, agglutinable in K27 antiserum, non-reactivein 0 8 antiserum. RESULTS Genetic transfer of the K27 antigen Earlier studies demonstrated a close linkage between the his operon and genetic determinants of the K26(A) antigen (Orskov & 0rskov, 1962) in genetic crosses. Recent experiments have suggested that a similar situation exists with regard to the K27(A) antigen (Olson, Schmidt & Jann, 1969). To confirm this we crossed the E. coli donor Hfr45(08 :K27) with several acapsular (K27-, K29-, K ~ o - ,K31- or K42-) his recipients of 0 8 or 0 9 serotypes and the rough K 1 2 strain and selected recombinants which had received the his+ allele of the donor. The his+ hybrids were divided on the basis of slide agglutination tests into several serological types (Table 2). Most of the his+ hybrids derived from crosses with the K- mutant ~464(08:K27-)or the three Og :K- strains as recipients were inagglutinable in the 0 antisera, but were strongly agglutinated in K27 antiserum, Thus these hybrids behave like K27+ forms and consequently must have inherited the genetic information for K27 antigen synthesis together with the donor his+ allele. In crosses with Og recipients the K27+ hybrids synthesized the 0 8 antigen instead of the recipient’s Og antigen, as detected by the sensitivity of the hybrids to the 0 8 antigen-specific phage Q8 (Jann et al., 1971). Accordingly, the his-linked donor rfb region responsible for 08-specific polysaccharide synthesis had been introduced together with the ‘K27’ genes into the hybrids. Using strains ~ 7 3 0(K- of 0 8 : Q 2 ) and 2578 (KI2) as recipients an additional hybrid class was found which comprised the majority of the his+ hybrids. These hybrids were agglutinated in 0 8 as well as in K27 antisera (Table 2). The reactivity in K27 antiserum indicates the presence of K27 antigen. The 0-agglutinability, however, suggests that the Downloaded from www.microbiologyresearch.org by IP: 78.47.27.170 On: Thu, 20 Oct 2016 11:30:53 358 G. S C H M I D T A N D OTHERS Table 4. Serological analysis of trp recombinants selected from crosses bet ween the donor Hfr4o and K27i recipients f K forms* K27i recipient Dzrived from No. of trpf recombinants analysed K27 K27i K- F733 2592 F730 2578 I27 62 91 52 34 8 2 2 r \ A * K27, agglutinable in K27 antiserum, non-reactive in 0 8 antiserum. K27i, agglutinable in K27 and 0 8 antissa. K-, agglutinable in 0 8 antiserum, non-reactive in K27 antiserum. Table 5. Reciprocal agglutination titres of K27, K27i and K27- bacteria in various antisera ~777(08:K27i) antiserum Bacteria for agglutination & I Unabsorbed ~ 7 8 2 6 2 7 antiserum ) \ Absorbed with E. coli 0 8 :K27- Unabsorbed Live or formalin-treated cultures 80 80 320 F7IW27) F492627-1 25120 < 20 < 20 2592cK27i) 2 2560 (I 280)* ( 22560) F7336 2 7 9 2 2560 (1280) ( 22560) F955-79 2 2560 ( 22560) (B 2560) Heat-killed cultures? I 60 80 640 F7I W27) 20 < 20 25120 F492627-) < 20 < 20 25120 25 92(K270 < 20 < 20 25120 F733cK2 70 < 20 < 20 25120 F955cK27i) * Titres in parentheses indicate finely grained agglutinates. t Bacterial susp2nsions were heated for I h at 100 "Cand washed once with saline. hybrids are intermediate between normal K27+ and K27- forms. Therefore these hybrids were designated as K27 intermediate forms, abbreviated as K27i (Table 2). Only 7 out of 257 hybrids from a cross with the recipient ~ 7 3 0were inagglutinable in 0 8 antiserum like K27+forms. A detailed analysis of these hybrids (Table 3) showed that nine had also received the non-selected trpf from the donor; among those were six of the seven K27f hybrids. A similar result was obtained with the E. coli K I 2 recipient 2578 (Table 2). The majority of the his+ hybrids were K27i forms agglutinable in 0 8 antiserum, indicating that they had also received the donor rfb region. Two of the three K27+ hybrids found among the his+ hybrids were trpf. All the other hybrids had retained the Trp- phenotype of the recipient. These results suggested that synthesis of the complete K27 antigen requires gene(s) in the trp region in addition to the his-linked genes. This was confirmed when the trp defect in K27i hybrids derived from crosses with E. coli ~ 7 3 0and 2578 was cured in crosses with the E. coli donor Hfr4o : 70 to 80 % of the trp+ hybrids behaved like normal K27+forms (Table 4). Transfer of the trp1regionfrom E. coli KI 2 to E. coli 0 8 :K27 If the E. coli KI 2 trp region is introduced into E. coli 0 8 :K27 recipients, one would expect some of these hybrids to be K27i forms as a result of the substitution of trp-linked 'K27' genes by an allelic KI 2 region lacking these determinants. Therefore we crossed the E. coli K I donor ~ ~ ~ with 2 the 5 recipient F ~ (08 I :K27, his, trp, ara, lac) and selected trp+ hybrids. Downloaded from www.microbiologyresearch.org by IP: 78.47.27.170 On: Thu, 20 Oct 2016 11:30:53 Genetic determinants of K antigens 359 All the 128 trpf recombinants had retained the His- phenotype of the recipient and about 10% of them had incorporated donor genes for arabinose and lactose utilization. Of the 128 trp+ recombinants, 94 (73%) behaved like K27i forms, while 32 were K27f like the recipient, and two were K27- (still 08.). Serologicalproperties of Ki hybrids To further characterize the K27i hybrids we performed tube agglutination tests with heat-killed (roo "C,I h) and formalin-treated cultures. The heat-killed bacterial antigen suspensions were washed once with saline. We used ~782(K27)antiserum and an OK antiserum produced with a formalin-treated suspension of a K27i hybrid (~777,0 8 :K27i). From the latter serum a pure K antiserum was obtained by absorption with a heated culture of E. coli 0 8 :K27-. The reciprocal agglutination titres of some K27i and K27 hybrids from different crosses and of normal K27 forms in various sera are given in Table 5. The titres of I :80 with formalin-treated and of I :80 to I : 160 with heated K27 bacterial suspensions in F777 antiserum showed that the K27i form can evoke the production of K27 antibodies. It was striking that in K27 antiserum the formalin-treated K27i hybrids formed very finegrained agglutinates compared with the heavy conglomerates of K27 cells. Similar observations were made in slide agglutination tests. The K27i forms gave much higher K titres than K27 cells. The titres obtained with formalin-treated K27i cells in F777(0K) antiserum were certainly due to mixed K and 0 agglutination. Another difference between K27i forms and K27 forms became evident when heated, washed cell suspensions were used for agglutination tests. In contrast to K27 cells the heated K27i forms had lost their agglutinability in pure K27 serum. This may be due to a loss of K antigen from the pretreated cells. Thus the reactivity of heated K27i hybrids in unabsorbed F777(oK) antiserum must be an 0 antigen agglutination. It is known that the K27 antigen retains its antibody-binding capacity after heat treatment (Kauffmann, 1966). To see whether this was also true for K27i hybrids, we absorbed K27 antiserum with heated and living cells of K27i hybrids and tested for any remaining K27 antibody in the absorbed sera. In no case did we observe agglutination with formalin-treated cultures of K27f cells. This indicated that the K antibody-binding capacity of K27i forms was not abolished by heat treatment. Two antisera prepared with formalin-treated suspensions of two K27i hybrids which had received the his but not the trp marker from the donor were tested by immunoelectrophoresis. In both antisera both 0 and K precipitation lines (0rskov et al., 1971) were formed with extracts of E. coli 08:K27 and of two K27 recombinants to which the his and the trp loci had been transferred. In contrast, extracts from the K27i cultures used for immunization showed only an 0 antigen precipitation line and no K precipitation line. DISCUSSION The K27 antigen is an acidic polysaccharide, the repeating units of which consist of glucose, galactose, fucose and glucuronic acid (Jann, Jann & Schneider, 1968). As shown above, two gene loci appear to participate in the synthesis of the K27 antigen of E. coN 0 8 :K27, one closely linked to his, the other to trp. The serological results suggest that lack of the trp-linked K27 gene(s) led to a reduced expression of the K27 antigen, as judged from the 0-agglutinability of such hybrids and the fact that the Kq-specific precipitation line was absent in their immune electrophoretic pattern. Because of the intermediate position of such hybrids between K27+ and K27- forms, they were designated as intermediate forms (K27i). Downloaded from www.microbiologyresearch.org by IP: 78.47.27.170 On: Thu, 20 Oct 2016 11:30:53 360 G. S C H M I D T A N D OTHERS In matings with K- recipients derived from E. coli Og :K3o and Og :K3 I it was sufficient to introduce the his-linked genes to obtain hybrids with complete K27 antigen. This means that these recipients must have the trp-linked genes which are required for the synthesis of the complete K27 antigen. Like E. coli 0 8 : K27, the E. coli Og strains mentioned above have glucuronic acid as a constituent of their K antigens (Jann & Westphal, 1975). Escherichia coli 0 8 :K42, which lacks the trp-linked gene needed for complete K27 antigen synthesis, contains galacturonic acid in its K antigen (Jann & Westphal, 1975). Thus it appears possible that the trp-linked locus is somehow connected with glucuronic acid, the acidic constituent of the K27 and other K antigens. It is not likely that the product of the trp-linked 'K27' gene acts on the level of glucuronic acid precursors because UDPglucuronic acid is a precursor of UDPgalacturonic acid, which in turn is needed for K42 antigen synthesis. However, such a gene product could possibly be a polymerase (or a subunit of such an enzyme) which recognizes glucuronic acid as a reactive site. The trp-linked 'K27' gene would then correspond to the Salmonella rfc locus which controls the polymerization of 0-specific repeating units in certain Salmonella strains (Stocker & Makela, 1971). If this were true, the K27-specificity of K27i cells would not be due to a polysaccharide, but only to single oligosaccharide repeating units linked to an acceptor (possibly core-lipid A). It is conceivable that the transfer of one repeating unit in K27i cells is less effective than that of the wild-type polysaccharide in K27 cells. Thus the K27 antigen of K27i forms would differ from that of the wild-type K27 strain quantitatively and qualitatively. This reduction of the K27 antigen in size and amount might explain the 0-agglutinability of K27i cells and the lack of the K27i precipitation line in immune electrophoresis. REFERENCES JANN,K. & WESTPHAL,0. (1975). Microbial polysaccharides. In The Antigens, vol. 111, pp. 1-125. Edited by M. Sela. New York, San Francisco, London: Academic Press. JANN,K., JANN,B. & SCHNEIDER, K. F. (1968). Immunochemistry of K antigens of Escherichia coli. 5. The K antigen of E. coli 0 8 :K27 :H-. European Journal of Biochemistry 5,456-465. E. (1971). Isolation and characterization of JANN,K., SCHMIDT, G., WALLENFELS, B. & FREUND-MOLBERT, Escherichia coli bacteriophage R8 specific for E. coli strains belonging to sero-group 08. Journal of General Microbiology 67,289-297. KAIJSER, B. (1973). Immunology of Escherichia coli: K antigen and its relation to urinary-tract infection. Journal of Infectious Diseases 127,670-677. KAUFFMANN, F. (I 966). The Bacteriology of Enterobacteriaceae. Copenhagen: Munksgaard. KAUFFMANN, F. & VAHLNE,G. (1945). Uber neue thermolabile Korperantigene der Coli-Bakterien. Acta pathologica et microbiologica scandinavica 20, 2 1-44. LEDERBERG, J. (I 950). Isolation and characterization of biochemical mutants of bacteria. Methods in Medical Research 3, 5-22. Low, B. (1968). Formation of merodiploids in matings with a class of rec- recipient strains of Escherichia coli K-12. Proceedings of the National Academy of Sciences of the United States of America 60,160-167. MCCABE, W. R., CARLING, P. C., BRUINS, S. & GREELY, A. (1975). The relation of K antigen to virulence of Escherichia coli. Journal of Infectious Diseases 131,6-10. MAYER, H. & SCHMIDT,G. (1973). The occurrence of three different lipopolysaccharide cores in Shigella and their relationship to known enterobacterial core types. Zentralblatt fur Bakteriologie, Parasitenkunde, Infektionskrankheiten und Hygiene (Originale A ) 224,345-354. NICHOLSON, A. M. & GLYNN,A. A. (1975). Investigation of the effect of K antigen in Escherichia coli urinary tract infections by use of a mouse model. British Journal of Experimental Pathology 56, 549-553. K. (1969). Biochemistry of the K antigens in Escherichia coli. European OLSON,A. C., SCHMIDT,G. & JANN, Journal of Biochemistry 11,376-385. ~ R S K O V ,I. & NYMAN, K. (r974). Genetic mapping of the antigenic determinants of two polysaccharide K antigens, KIOand K54, in Escherichia coli. Journal of Bacteriology 120,43-51. ~ R S K O V ,F. & ~ R S K O V ,I. (1962). Behaviour of E. coli antigens in sexual recombination. Actapathologica et microbiologica scandinavica 55,99-109. Downloaded from www.microbiologyresearch.org by IP: 78.47.27.170 On: Thu, 20 Oct 2016 11:30:53 Genetic determinants of K antigens ~ R S K O V , F., ~ R S K O V ,I., JANN,B. 361 & JANN,K. (1971). Immunoelectrophoretic patterns of extracts from all Escherichia coli 0 and K antigen test strains. Correlation with pathogenicity. Acta pathologica et microbiologica scandinavica, B79, 142-1 5 2 . ~ R S K O V ,I., SHARMA, V. & ~ R S K O V ,F. (1976). Genetic mapping of the K I and Kq antigen (L) of Escherichia coli. Acta pathologica et microbiologica scandinavica BS4,125-13 I . SCHEIDEGGER, J.-J. (I 955). Une micro-mCthode de I’immunoelectrophorCse.International Archives of Allergy and Applied Immunology 7,103-1 10. SCHLECHT, S. & WESTPHAL, 0. (I967). Wachstum und Lipopolysaccharidgehalt von Salmonellen bei Zuchtung auf Agarnahrboden. Zentralblatt fiir Bakteriologie, Parasitenkunde, Infektionskrankheiten und Hygiene (Originale) 200, 241-259. SCHMIDT, G., JANN,B. & JANN,K. (1970). Studies on R mutants with an incomplete core, derived from Escherichia coli 0 8 :K27. European Journal of Biochemistry 16,382-392. STOCKER, B. A. D. & MAKELA,P. H. (rg71). Genetic aspects of biosynthesis and structure of Salmonella lipopolysaccharide. In Microbial Toxins, vol. 4, pp. 369-433. Edited by G. Weinbaum, S. Kadis and S. J. Ajl. New York: Academic Press. WHANG,H. Y., MAYER,H., SCHMIDT, G. & NETER, E. (1972). Immunogenicity of the common enterobacterial antigen produced by smooth and rough strains. Infection and Immunity 6,533-539. Downloaded from www.microbiologyresearch.org by IP: 78.47.27.170 On: Thu, 20 Oct 2016 11:30:53