Separatum ACTA PATHOLOGICA ET MICROBIOLOGICA SCANDINAVICA Vol. 51, pp. 280-290, 1961 FROM DEPARTMENT OF GENETICS, UNIVERSITY OF WISCONSIN, MADISON, WIS. AND THE INTERNATIONAL SALMONELLA & ESCHERICHIA CENTRE, STATENS SERUMINSTITUT, COPENHAGEN, DENMARK THE FERTILITY OF ESCHERICHIA COLI ANTIGEN TEST STRAINS IN CROSSES WITH K 12 By Frits Orskov and Ipa ORSKoV Received 5.vii.60 Sexual recombination in Escherichia coli has been studied almost exclusively with strain “K 12”. The initial choice of this strain was entirely fortuitous (Lederberg & Lederberg 1956, Gray & Tatum 1944). While many aspects of the life cycle have been successfully analyzed with this particular strain, at least two considerations have motivated a search for the distribution of sexual interfertility among a wider group of strains: (1) the interest in the status of the Escherichia coli group, as a gene pool and (2) the possibility that some genctic differ- ences not evident from mutation in the laboratory might be found among diverse strains from natural habitats. In the latter category, the most interesting features might well involve the mechanism of sexual compatibility among various serotypes. A preliminary sereening of wild type strains seemed to justify these expectations (Lederberg 1951). This study was however conducted before the Hfr, Ft, F- system of sexual compatibility was adequately understood (Cavalli 1950, Lederberg, Cavalli & Lederberg 1953, Hayes 1953). In view of the promise of immunogenctic factors in future work it was, therefore, decided to take advantage of the collection of estab- lished serotypes of Escherichia coli for a review of genetic interactions in the group. Strain “K 12” itself poses formidable difficulties for immunological studies in account of its virtual loss of O and K antigens, the strain having been cultivated for other purposes and without special scrutiny for its serological properties since 1922. The establishment of a serological scheme for E. coli (Kauffmann 1954; Ewing 1956) indeed represents a substantial investment that can he exploited in further genotypic analysis of the species. These studies were conducted, for the most part, at the Department of Genetics, University of Wisconsin, Madison, Wis. The work was supported by research grants (to Professor Lederberg) from the National Science Foundation and from the National Cancer Institute (C-2157), U. S. Public Health Service. I, Orskov was a fellow of the American Association for University Women. 280 281 The E. coli Antigenic Scheme. Three main types of antigens are depicted in the E. coli group: (1) The O antigens which are thermostable lipopolysaccharide-complexes con- stituting part of the cell wall, (2) the K antigens which are envelope or capsule antigens of polysaccharide nature; historically three types of K antigens A, B and L have been described. They are differentiated by the varying thermostability of their agglutinability, their capacity to evoke the formation of agglutinins and their agglutinin-binding capacity. (3) The H- or flagellar antigens; by contrast with Salmonella, which frequently display alternative flagellar forms only monophasic strains have been found so far in the coli group. The first antigenic scheme for the E. coli group was published by Kauffmann in 1944. This scheme contained 20 different O groups, 17 K antigens and 3 H antigens. Since then the scheme has been steadily extended and it com- prised when the present examinations were carried out 137 O antigens 80 K antigens 43 H antigens corresponding to a somewhat smaller number of type strains, since some of the K and H type strains are also represented in the O series. The serotype of a coli strain is given as follows: 0 111: K 58:H 21. The O, K and H antigens exist in a great number of different combinations; if they could recombine freely the absolute number of coli serotypes containing the known antigens would be close to half a million. In practice the number is reduced, as it has been observed that a number of O and K antigens (especially OA and OB) are closely connected. For example it could be mentioned that the combinations O 55:K 59 and 0 111:K 58 are known only in the indicated pairs. On the other hand that the H antigens can be combined frecly with diverse OK groups is illustrated by the experience from the 0 55: K 59 and O111: K 58 groups. Among the rather limited number of these strains whose H antigens have been recorded at least 8 different H antigen combina- tions for O 111: K 58 and 10 for O 55: K 59 have been detected (Ewing 1956, Arskov 1956). Sexual Compatibility in E. coli. Sexual recombination in E. coli is mediated by contact between cells of different mating type. In order to give a productive cross one type of cells (F+) acts as genetic donor or male and the other as recipient or female (F-). However, F* are ambivalent and F* x F* crosses give a few recombinants. Most of the “K 12” strains are F+: upon mixing with “K 12” F- culture they exhibit a low frequency of recombination 1 According to earlier convention this would have been designated O111: B4: H2; the notations here follows Kauffmann, Orskov & Eving (1956). 282 (10-5 or less expressed as the observed ratio of recombinant to input parental cells). Strains termed Hfr (Cavalli 1950, Hayes 1953) show a high frequency of recombination in crosses with F- strains (10-1 to 10-3). In most sexual crossing experiments carried out so far both parent strains have been auxotrophic mutants which were unable to synthesise one or more compounds necessary for growth, most often amino acids. This procedure requires considerable handling of each parent strain to produce the necessary mutants. To simplify the screen- ing of a large number of strains the so-called SRP-technique (strepto- mycin-resistance-prototroph) was developed (Lederberg 1951). Re- combinants are selected from test crosses between auxotrophic, strepto- mycin-resistant tester strains and various prototrophic, streptomycin- sensitive te. wild type strains. Thus a large number of coli wild type strains could be screened both for fertility and for mating type. METHODS The parent strains were grown in penassay broth Difco for 20 hours and 0.5 ml from each parent culture was inoculated into a fresh broth. The mixed culture was incubated for another 20 hours, centrifuged and the pellet resuspended in 0.5 ml distilled water. One drop of this suspension was spread onto minimal medium (see below). Controls were provided by plating the washed parent cultures onto the same medium as used for crosses. The plates were read after 48 hours incubation at 37° C. No further analysis of the recombinant colonies was carried out. (Leder- berg 1951). Media, For fluid medium penassay broth Difco was most often used. In some cases autoclaved ox heart infusion broth, produced in Statens Seruminstitut, Copenhagen, was employed, No difference as to the usefulness of these two media could be found. EMS agar plates (Lederberg 1950) supplemented with dihydrostreptomycin 100 «g/ml and galactose 1 per cent was used as selective medium for the recombinants. Strains, All type strains for coli antigens which were established up till 1957 were examined: O antigen test strains: 137 K antigen test strains: 80 H antigen test strains: 43 Two of the O type strains were found to be streptomycin resistant (0126 and 0127), four O type strains have previously been found not to belong to E, coli but to the Citrobacter group (Orskov 1956) and finally 047 has been lost many years ago. Of the K type strains, 29 were identical with different O type strains. Among the 43 H antigen type strains 23 were already represented among the O or K type strains; two H type strains have been found to belong the Citrobacter group. The reduced number of different type strains examined was therefore: O antigen test strains: 130 K antigen test strains: 51 H antigen test sttrains: 18 Total: 199 76 E, coli strains belonging to the OK group 026; K60, 055: K59 and O111: K58, in many different H antigen combinations, were further examined. (Table 1). 72 Klebsiella type strains representing the same number af different capsule antigens were also included. The auxotrophic coli strains with known fertility used in the crosses can be found in Table 2. These last mentioned strains were all developed in Dept. of Genetics. University of Wisconsin, Madison, Wisconsin, U.S.A. 283 TABLE 1 E. coli Strains Belonging to the OK-Types 026: K60, 055: K59 and O111: K58 Tested. Serotype antigens Nuntber of strains Oo K I 26: 60: - 7 26: 60: 8 1 26: 60: 11 5 26: 60: 32 3 55: 59: — 8 55:59: 2 3 55:59: 4 2 55: 59: 6 9 95:59: 7 4 55:59: 8 3 95: 59:10 2 55: 59: 11 2 55: 59: 16 1 55:59: 21 1 55: 59: 25 1 55:59: 27 3 95:59: 32 3 35:59: 34 1 111: 58: - 6 111: 58:2 4 111: 58:4 2 111: 58:11 1 111: 58: 12 1 111: 58: 16 1 111: 58: 21 1 111: 58: 27 1 Totally 76 TABLE 2 Aunxotrophic Strains Used, Source desigantion Markers Scrolype K12 W1607 F- M-Sr O-: K?; H48 - W3287 Hfrig M-Sr i Wa4 W3703 Hfr L°Tryp-Sr 025; K-: nm H509a W3479 F- H-Isol-Sr 0100: K+: H2 = W3482 Fr H-Isol-Sr i Markers without relation to the study have been omitted. M — methionine. L. == leucine, Tryp = tryptophane, H — histidine, Isol — isoleucine and St —= streptomycin resistant, W 3287 is an Hfr strain isolated from K12 F*M-Lact by J. Lederberg using indirect selection (replica plating UV-survivors against a Lac’ indicator). Streptomycin resistance was introduced by selection of a spontaneous mutant on streptomycin medium. The K12 strains are characterized as rough, no O antigen has been found till now. The presence of an ordinary K antigen is doubtful. Th H antigen is num- bered H48 (Orskov & @rskov 1960b). WGA4 was isolated in 1950 from a urine culture submitted to the Wisconsin Public Health Laboratory, The markers L*, Tryp- and Sr were introduced by conventional 284 methods (Lederberg 1950). After having been converted to the F* state an Hfr mutant W3703 was isolated by the authors using indirect selection against a histidine-less indicator. W3703 has O antigen 25, no detectable K antigen and is non-motile (non- flagellate, therefore missing H antigen). H509a is the type strain of E. coli O antigen 100. The markers H-, Isol- and Sr were introduced by conventional methods. W3482 received its F factor from the K12 strain W1876 F* (@rskov & Orskov 1960a). W3479 and W3482 have O antigen 0100, a K antigen not yet numbered, and H antigen 2. TABLE 3 Crosses between Auxotrophic Hfr, F+ and F- Strains and Prototrophic E. coli Antigen Type Strains. Strain Serotype W3287 W3703 W3482 w3i79 | W607 ae antigens ' K12 WG4 0100 0100 | R12 no, O K E Hfr Hfr pe ! FE | Ir O-type strains. U4/41 4a: 3(L):5 ++ _ - _ Bi623/42 17: 10(L): 10 ++ +a+ - - - F10018/41 18: 76B:14 tH++ fot ++ Ki2a 17:16L:18 +4 ae _ _ F8188/41 19:-:7 - +4 - _ - E47a 25: 19L: 12 ++++ fat + - _ P6a 32: 2:19 t+4++ +++ - - H304 34: 2:10 ++ - - ~ _ E77a 35: 2:10 +444 +a + _ H510ce 37: 2.210 - 7 +4 - - F11621/41 382.226 4+4+4++4 +oo+He +++ _ _ H7 392 0 +++ - - - H316 40: .:4 ++4+4+ t+at+ t+ _ H710¢ Af: .: 40 ++4++ ++ t+ - Plia 42: .:37 ++4++ +++ + - U19/41 Sf: 2224 4 4 + ~ 20/41 52: .:10 +++ ++++ t+++ - - Bi7327/41 533.53 4+4+-+4++ ++ +++ ++ $u3684/41 562.3 - ++++ t+ - - F8198/41 573 63 - ++ ++ +++ - _ F8962/41 58: 2:27 tat + - - 4h F10524/41 62: 2:30 - +++ - - K6b 64: .3- +++ ++ ++ - - Killa G52. 2- +++ ao _ _ _ Pla 66: .225 +++ +++ - - - P9b 69: .: 38 toy ++ _ Pida 71: 2:12 t+4++ +++ - - - Pl2a 733.7331 +44 +++ - E3a 74: . 239 +++ +4+++ ++ + Edd 762.38 +444 ++ - 7 - E71 80: . 326 +4-+ +e+ + + H19 84: 2:21 ++++ f+ +Ho+ _ H35 86: 2326 ++ +4 ++ + - H40 87: 2:12 ++4+4+-+ toe - - H68 89: 2:16 +444 +++ +4 _ _ H77 90: .4- 4 -- 4- — _ _ H307b 91: .2- ++ - - - _ H308a 92: 2.333 - +++ - _ H319 96: .:19 ++ tH _ _ H504e 99: . 233 +++ - - - - 285 TABLE 3 (cont.) | . “De 3987 13705 "3482 | wsi79 | "1607 Strain ania ‘Se Ya Ni O100 Nay a Oo K H Hir Ulfr rE oa be O-—type strains, H509a 100: .:2 ++++ +4-+ ++ - _ H511 102: .:8 ++++ +++ ++ _ H519 104: .:12 +-+++ +++ ++ - H521la 106; .:33 +++ f+ 4 - _ _ H705 107: .:27 +++ 4-4 . _ H708b 108: .:10 +++ +++ - _ 26w 114: 2332 ++++ _ tH - 28w 116: .:10 +44 + t+ _ - 30w 117: 2:4 +44 + - _ _ 3lw 118: .:- +44 t4 - - 43w 123: .:16 - +++ +++ 7 - Canioni 125: 70B: 19 +++ - ++ - 178/54 129: 2:11 b+++ t+ ~t+ _ _ 4866/33 130: .:9 +++ +++ toe - _ $239 131: . : 26 +++ - + — _ N87 132: . 128 ++ + +4 _ 4370/53 134: ..:35 +++ +4 a _ _ coli Pecs 135: .3- +44 ++ +--+ - _ K-type strains not listed above, Pus 3422/41 7:7L:4 ++ +44 toe ~ ~ H67 23: 22L: — -ot tet + 7 _ Bi449/42 9a: 264: —- + +oH+ + _ _ E56b 8: 27A:— + +t - _ H36 9:32A:19 - +4 7 7 _ A198a 9:36A:19 — ++ f+ _ _ A262a 9: 38A:— ~ ++ _ i _ Al12b 6: 54L: — +++ +++ - - _ N24c 9: 554A: — ~ +4 _ _ _ 5017/53 86ab: 64B: 36 +4 +4 _ _ 2160/53 127ab: 65B: 4 +4 +3 _ _ _ H-type strains not listed above. Bi7575/41 8: 25B: 9 +4 ++ - _ _ H330b 8: . 1:20 44+ tH _ _ _ K42 45: ..:23 - + +++ - _ K72 Bl: 2:24 - +H+ _ _ K181 11: .:33 ++ + ~ - _ & = 1-2 colonies, + = 10 - ++ 10-50 - +++ = 50-200 - +$++ = >2000 - In the serotype formulas - means, not examined, — means, antigen not present. Only productive crosses have been recorded. The antigen for which the corresponding strain is the official type strain has been written in italics in the serotype column. 286 EXPERIMENTAL 199 different E.coli antigen type strains were crossed with the following strains of known fertility: Two Hfr strains W 3287 and W 3703 coming from the “K 12” and the WG 4 line respectively, further one F* strain W 3482 derived from the coli O 100 type strain. Finally crosses with two F- strains, W 1607 from the ”K 12” line, and W 3479 derived from coli 0100 were carried out. The two last mentioned strains were included to see if any F* strains were represented among the type strains. All positive crosses were carried out at least twice. The control plates on which the two parent strains were inoculated separatively showed no growth in the recorded cases. TABLE 4 Crosses between Auxotrophic Hfr, F* and F- Strains and Prototrophic E. coli Strains Isolated from Infantile Diarrhoea. Strain Serotype | a | Nope ihe : oid Country no. oO gk Her | et Iho: where isolated | ' | C24-55 26: 60: 32 $44 +4 - Switzerland C116-55 26: 60: 32 +++ +++ - Mexico C56-56 26: 60: 32 4-4 = - - Wales F53-50 55:59: 2 + + + + Sweden C3572-54 55:59: 2 ++4+-+ - - Germany C218-54 55:59: 10 4+44++ + a - C50-56 55:59:11 +t+ttst - Wales €222-53 55:59:11 ++4 + - France C223-53 55: 59: 21 +++ ++ - 7 C238-56 55:59: 25 t+4+ oF - - C87--53 111: 58: -- t+4 4. + = 1-2 colonies, + == 10 - ++ = 10-50 - +++ == 50-200 +b t+ = > 2000 - In the serotype formulas - means, not examined, — means, antigen not present. The outcome of the crosses is shown in Table 3. It appears that 74 different strains were fertile in crosses with W 3287 or W 3703, ie. 37 per cent. Sixtythree strains were fertile with W 3287 and sixtyfour with W 3703. Thirtyseven strains were fertile in crosses with the F- strain W 3482. Generally there was good agreement between the fertile strains found in the three series of crosses; only 10 strains were fertile with W 3287 and not with W 3703, and 11 strains fertile with W 3703 and not with W 3287. All strains fertile with W 3482 gave productive crosses with either W 3287 or W 3703. Generally the number of re- combinants was largest in the W 3287 crosses. Only a few examples of F* strains i.e. strains giving recombinants in crosses with F- strains, were detected, and the number of reeombinants was low in these cases. 287 All crosses were carried out at least twice and as could be expected the number of recombinants produced in a given cross could vary to a large extent from time to time. In addition to the coli type strains further 76 coli strains belonging to the OK groups O 26:K 60, 0 55:K 59 and O 111:K 58 in a variety of H antigen combinations were tested (Table 4). These strains were isolated from outbreaks and single cases of infantile diarrhoea and in some cases from healthy or diseased animals. Eleven strains gave productive crosses with W 3287, of these nine could also be crossed with W 3482, but these crosses were less productive. None were fertile with the F- strains W 3479 and W 1607. It should be pointed out that two strains out of three belonging to O 55: K 59:H 2, and that two out of two belonging to O 55: K 59:H 11 and furthermore three out of three belonging to O 26:K 60:H 32 were fertile in crosses with W 3287. In order to show that there were no direct epidemiological connection between these fertile strains of the same serotype, the place where the strains were originally isolated have been recorded in the table. In order to test if some Klebsiella strains were fertile in crosses with W 3287, the 72 Alebsiella capsule type strains were examined. Eight were found to be streptomycin resistant and could not be tested with the SRP technique. In none of the remaining cases recombinants could be detected. DISCUSSION It appears from the recorded experiments that about one third of two hundred serologically different EF. coli strains are fertile in crosses with cither or both of two Hfr testers having the K 12 fertility factor. This figure is considerably higher than earlier reported figures, (Lederberg et al. 1952). As already mentioned Lederbergs screening of the fertility of wild type strains was carried out before the Hfr, F* and F- mating system was adequately understood and an F* strain was used as donor instead of the Hfr strain used here. An F* strain from the “K 12” line was not included among the donor strains in this study, but even the O 100 F* strain, W 3482, seems to detect more fertile coli wild type strains (18 per cent) than the “K 12” F+ strain used in Leder- bergs investigation. One explanation for this discrepancy could be that the criteria for naming a strain E. coli perhaps were less strict among the strains in the earlier reported series, than they were among the type strains which never deviated much from the recognized biochemical pattern of E.coli; the “K 12” wild type strain will show that pattern without deviations. Another explanation might be the different origin and the different age as laboratory strains of the two series of strains. The strains used by Lederberg et al. were mostly freshly isolated strains from pathological conditions, while most of the coli type strains have been kept in the laboratory for many years and only part of them 288 are from pathological conditions. It is well known that the O and K antigens of freshly isolated strains are bettter developed than in old laboratory strains and further several investigators (Kauffmann 1944, VahlIne 1945) have shown that the frequency of O-inagglutinable strains, indicating strains with well developed K antigens, is greater among strains from pathological conditions than among _ strains isolated from normal faeces. It is therefore probable that the strains examined by Lederberg had better developed K and O antigens, and further that the readiness with which they mutated to R forms was less than that of the old laboratory strains used in this study. No- body has yet reported if fertility in E.coli is influenced by qualitative and quantitative differences of the O and K antigens, but it may be that strains with poor development of these antigens are more fertile. In this connection it could also be pointed out that none of the 72 Klebsiella strains, which all have very large capsules were found to give productive crosses with the Hfr and F* strains used here. In another investigation (Orskov, Orskov & Kauffmann 1961) it was found that more than 50 per cent of randomly selected Salmonella strains representing all Salmonella O groups were fertile with the “K 12” Hfr strain used in this study; it is known that K antigens if present, are only poorly developed in Salmonella strains. Many of the O antigens of the coli type strains are interrelated, but only in very few cases have these relationships been definitely elu- cidated i.e. the coli antigenic scheme is not so explicitly developed as to give the highly differentiated tabulation of the different O antigen fractions which is characteristic of the Kauffmann-White scheme for the Salmonella group. What does exist is a listing of the O antigen agglutination titres from mutual agglutinations of O test strains in O antigen typing sera (Ewing 1956). When such a list is compared with the result of the compatibility examination recorded in this paper, it is difficult to find any relationship between special O antigens or O an- tigen fractions and fertility. The K antigens of the examined strains are to a great extent not numbered yet, but from unpublished findings it is known that a large fraction of the unnumbered K antigens from the O series represent new and different K antigens most of them probably B antigens. It is therefore also difficult to find any relationship between compatibility with the testers used and the specificity of the K antigens. When we finally turn to the H antigens it is not possible to find any correlation between compatibility and the different H antigens. One exception is H 10 which is found in 9 out of 130 strains in the O series; 7 of these are fertile with one or more of the male testers. The outcome of the crosses involving more identical strains of the same serotype seem to tell more of the role of antigenic components in compatibility studies. It appears that neither O nor K antigens alone can determine the compatibility, because only some of the strains with 289 the O 26:K 60 or O 55:K 59 complex are fertile with the testers used. The results seem to imply that the serotype is of some importance, because more cases were recorded where strains of the same serotype of independant origin were compatible. The simplest explanation would be that such fertile strains of the same serotype were derived from the same ancestral strain. With this explanation in consideration we cannot draw any con- clusions from the recorded results as to a possible genetical or physio- logical connection between the antigenic composition of a certain sero- type and its fertility. No large scale examination of the interfertility of the type strains that were found to be fertile in these studies have been carried out. In a limited number of cases such fertile strains, after conversion to the F* state could also be crossed with one another (@rskov & Orskov). Two thirds of the examined strains were found to be sterile. Future research will show if those strains are completely sterile or if they belong to other independant breeding groups. Finally it should be kept in mind that the sterility detected might disappear when a different crossing technique was employed. SUMMARY A sercening for fertility of 199 E. coli antigenic type strains (O, K and H-antigens), using a number of testers of known fertility showed that about 35 per cent were fertile. With few exceptions the strains were found in the F- state. No definite correlation between fertility and single antigens could be found. A similar screening of a number of coli strains having relations to types found in infantile diarrhoea was performed. In a number of eases there seemed to be some correlation between the serotype and the fertility. This fact is considered attributable, not to a connection of antigenic structure with fertility but to a common origin of the strains. Finally all Klebsiella capsule type strains were examined: none were found to be fertile. REFERENCES Cavalli Sforza, L. L.: La sessualita nei batteri, Boll, Ist. Sieroter. Milan. 29: 281, 1950. Ewing, W.H., Tatum, H. W., Davis, B. R. & Reavis, R. W.: Studies on the Serology of the Escherichia coli Group. Public Health Service, Communicable Disease Center, Atlanta, Georgia, 1956. Gray, C.H. & Tatum, E.L.; X-ray induced growth factor requirements in bacteria. Proc, Natl, Acad, Sci. U. S. 30: 404, 1944. Hayes, W.: The Mechanism of Genetic Recombination in Escherichia coli, In Cold Spr. Harb. Symp. 18: 75, 1953. Kauffmann, F.: Zur Serologic der Coli-Gruppe. Acta path. et microbiol. seandinav. 21: 20, 1944, Kauffmann, F.; Enterobacteriaceae, 2. edition, Munksgaard, Copenhagen 1954. 19 acTa PATH. 51,3 290 Kauffmann, F., Orskov, F, & Ewing,W.H.: Designations for the K antigens of Escherichia coli serotypes. Int. Bull. Nomen. and Taxon. 6: 63, 1956. Lederberg, J.: Isolation and characterization of biochemical mutants of bacteria. Methods in Medical Research, 3: 5, 1950. Lederberg, J.: Prevalence of Escherichia coli strains exhibiting genetic reeombina- tion. Science. 114: 68, 1951. Lederberg. J., Cavalli, L. L, & Lederberg, Esther M.: Sex compatibility in Escherichia coli, Genetics. 37: 720, 1952. Lederberg, J. & Lederberg, Esther M.: “Infection and Heredity” in ‘Cellular Mechan- isms in Differentiation and Growth”, 14th Symposium of the Society for the Study of Development and Growth. Princetown University Press, Princetown, New Jersey 1956. Vahine,G.: Serological Typing of Colon Bacteria. Gleerupska Univ.-Bokhandels, Lund 1945. Orskov, F.: Escherichia coli, Om type bestemmelse specielt med henblik pa stammer fra infantil diarrhoea, Nyt Nordisk Forlag, Copenhagen 1956, Orskov, F., Orskov,Ida & Kauffmann, F.: The fertility of Salmonella strains deter- mined in mating experiments with Escherichia strains, Acta path. et microbiol. scandinay. 57: 291, 1961. Orskov, Ida & Orskov, F.: An antigen termed f+ occuring in F+ E, coli strains, Acta path. et microbiol, scandinav, 48:37, 1960a. Orskov,Ida & Orskov, F.: The H-antigen of the “K12” strain. A new E. coli, H anti- gen: H48. Acta path, et microbiol. scandinav, 48:47, 1960b. Orskov, [da & Orskov, F.: Unpublished observations,