Zbl. Bakt. Hyg. A 266, 323-329 (1987)

Studies on Clonal Assignment of Urinary Escherichia coli01:Kl Strains"

WOLFGANG NIMMICH I, GERHARD ZINGLER I,

GUNTRAM SELTMANN 2, ERHARD TIETZE 2

, and GERD HORN 2

1 Institut fur Mikrobiologie (Direktor: OMR Prof. Dr. sc. med. G. Naumann) des BereichesMedizin der Wilhelm-Pieck-Universitat Rostock, DDR, und

2 Institut fur Experimentelle Epidemiologie (Direktor: OMR Prof. Dr. med. habil. H.Rische) Wernigerode, DDR

With 1 Figure· Received September 3, 1986 . Accepted April 30, 1987

Abstract

Fifty-nine Escherichia coli strains belonging to two clonal groupings were investigated formajor outer membrane proteins, colicin production, and partly for plasmid DNA content.The membrane protein patterns of the 01:Kl:H7(H-):Fll and 01:Kl:H-:F9 strains ob­tained by sodium dodecyl sulfate-polyacrylamide gel electrophoresis were distinctly diffe­rent from each other and, therefore, are useful for clonal assignment. All of the Fll isolateshad one plasmid of about 85 Md in common which is suggested to be characteristic for theclone. Their content of smaller plasmid DNA was heterogeneous and showed geographicaldifferences. The F9 strains showed a plasmid pattern different from the Fll strains. Aplasmid of 2.6 Md was found in all of these isolates. Colicin production was found to beuseful for clonal assignment only in two thirds of the F9 and not in the Fll strains. Somestrains with identical properties seem to be of epidemic importance.

Zusammenfassung

Zwei klonalen Gruppen angehorende 59 E. coli O'lK'l-Starnrne wurden hinsichtlichihrer augeren Membranproteine, ihrer Colicin-Bildung und zum Teil auf ihren Plasmid­Gehalt untersucht. Die mittels SDS-PAGE ermittelten Mernbranprotein-Muster zeigten ein­deutige klonspezifische Unterschiede zwischen 01:Kl:H7(H-):Fll- und 01:Kl:H-:F9­Isolaten. Bei allen Fl l-Stammen konnte ein offensichtlich Klon-charakteristisches grolsesPlasmid von 85 Md nachgewiesen werden, wahrend bei den kleineren Plasmiden geo­graphische Unterschiede registriert wurden. Die untersuchten F9-Isolate besafen eingemeinsames Plasmid von 2,6 Md, zeigten jedoch generell ein anderes Plasmid-Muster alsdie Fl l-Stamme. Die Colicin-Bildung konnte nur bei F9-Stammen, und auch nur bei 2/3 derIsolate, fur eine Klon-Zuordnung herangezogen werden. Einige Starnme mit vollig uberein­stimmenden Eigenschaften scheinen epidemischen Charakter zu besitzen.

* Dedicated to the 60th birthday of OMR Prof. Dr. sc. med. G. Naumann.

22 Zbl. Bakt. Hyg. A 266/1--4

324 W. Nimmich, G. Zingler, G. Seltmann, E.Tietze, and G. Horn

Introduction

Escherichia coli 01:K1 strains are among the most common serotypes causingurinary tract infections (UTI) (13, 14, 18, 22). Such strains are known to carry P­fimbriae which enable them to specifically adhere to uroepithelial cells and thus givingrise to colonization and infection of the urinary tract. Recently we have demonstratedtwo different clones in 01 :K1 isolates by serological typing of the fimbrial antigens.One clone was represented by 01:K1:H7:Fll strains and a second could be charac­terized by serotype 01:K1:H-:F9 (15). The clonal distinction was supported by biotyp­ing, nutritional requirements, to a certain extent also by phage typing(13), and by 01antigen factor analysis (13, 24). The electrophoretic mobility of major outer membraneproteins by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE),plasmid content, colicin production (2, 11, 22), alloenzyme analysis (16), and SDS­PAGE of lipopolysaccharide patterns (4, 11) have also been used for clonal assignmentof 01:K1 strains of diverse origin.

In the present paper we report on outer membrane protein patterns, plasmid profile,and colicin types in relation to clonal grouping of Escherichia coli 01 :K1 strains fromunne.

Material and Methods

Bacterial strains. Fifty-nine Escherichia coli 01:Kl strains from female patients withurinary tract infections, mainly pyelonephritis, were examined. Detailed strain informationis given in reference (13). The test strains for Escherichia coli 01 (U5/41) and fimbrialantigen Fll (C1976-79) from the International Escherichia and Klebsiella Centre, StatensSeruminstitute, Copenhagen, were included in the study.

o and H serotyping was performed by standard laboratory methods using bacterialagglutination techniques (13).

Kl antigen was determined by specific phages (13).Fimbrial antigens were identified by crossed-line immuno-electrophoresis (15).Major outer membrane proteins were prepared essentially as described by Achtman et

al.(2). Briefly, the bacterial cells were broken by sonication and centrifuged off at 50000 xg for 60 min. The pellet was suspended in Tris/Sarkosyl buffer (1.34 g Tris and 16.7 gSarkosyl per 1 I) of pH 7.6 for solubilization of the cytoplasmic membranes. The outermembranes were obtained by centrifugation at 50000 x g for 90 min. Membrane patternswere determined by SDS-PAGE (5) using 11% acrylamide and 0.2% methylene bisacryl­amide in the running gel.

Plasmid analysis was carried out by two alternative methods as described by Meyers et al.(9) and Tietze and Tschape (21). Samples were run in 0.7% agarose (Seakem, MarineColloids, Inc.) dissolved in Tris borate buffer (89 mM Tris, 89 mM boric acid, 2.5 mMNa2EDTA, pH 8.0) for about 3 h at 40 rnA. Gels were stained in a solution of ethidiumbromide (0.4 ug/rnl) in water and photographed in UV light. The molecular weights werecalculated from the mobility of various reference plasmids; pMB8 (1.87 Md), RSF 1010 (5.0Md), R6K (26 Md), RI00 (60 Md), R27 (112 Md), and R478 (160 Md) (21).

Colicin production of the strains was tested after chloroform treatment of the bacterialgrowth on agar plates using the overlay method and growth inhibition of Escherichia coli K­12 (P400) as indicator (2). Colicin typing was carried out according to Abbott and Shannonusing the indicator strains for colicin types lA, IB, 2, 3, 3A, 4-14 (1).

E.coli 01 :Kl Clones 325

Results

The migration patterns of outer membrane proteins of Escherichia coli 01:Klstrains were classified by SDS-PAGE. A correlation was observed between membranepatterns (MP) and serotype, especially when the fimbrial antigens were considered.After submitting our paper to press we learned that MP data of about half of ourstrains placed at disposal have been published (11). The results of the present studyinclude additional thirty strains. They are shown in Figure 1 as characteristic patternsonly.

The 01 :Kl :H7:Fll strains showed five outer membrane proteins which correspondto MP 9 of Achtman et al.(2). The plasmid coded protein (PCP), however, was notdetected in each of the strains and was mainly a minor constituent. The reference strainfor F11 (C 1976) had no PCP and should be assigned to MP 6 (2) and not MP 9 asfound by Mollet al. (11). This strain has already previously been found to be biochemi­cally different from other Fll strains (13). A slightly different MP was also observed inthe immotile variants of this serotype, i.e. H-:Fll strains, containing only one line inthe porin region. A distinctly different MP was found in isolates with fimbrial antigenF9. This is obviously identical with MP 5 (2) and also present in H-:r strains.

All of the 01:Kl isolates examined contained plasmid DNA. The number of plas­mids per strain ranged from one to five with an average of 2.9. An association wasobserved between O:K:H:F serotype and plasmid content. Isolates of serotype01:Kl:H7:Fl1, including the immotile variants with Fl1, had one large plasmid ofabout 85 Md in common. However, a heterogeneous content of smaller plasmids was

Porins :::::: .K­

OmpA­PCP -

H7 :F11 H- : F 11 H - :F9 H- : F- H7 : F11( U 5 / 41)

H7 :F11( C 1976 )

Fig. 1. Migration patterns of major outer membrane proteins of E. coli 01 :Kl :H:F strains.

Table 1. Plasmid DNA content of E. coli 01:K1 strains with fimbrial antigen Fll

Strain Serotype Plasmid DNA species (Md)

C 1976-79 01:K1:H7:Fll 85 42 1,6U5/41 01:Kl:H7:F1l 85R33 0l:Kl:H7:F11 85 2,8 1,35R52 0l:Kl:H7:F1I 85 1,35RI04 01:Kl:H7:Fll 85 1,35M8137 01:K1:H7:FIl 85 48 4,0 2,4M8278 01:K1:H7:FIl 85 3,6 2,8

R12 01:Kl:W:F11 85 3,8 1,0R39 01 :K1:H-:F11 85 70 3,8 1,35R48 01 :K1:W:F1I 85 3,8 1,35 1,0

326 W. Nimrnich, G. Zingler, G. Seltmann, E.Tietze, and G. Horn

found as shown in Table 1. In this respect two strains isolated in Denmark (CI976-79,US-41) were different from each other as well as from the two strains from the USSR(M8137, M8278). No resemblance in patterns of small plasmid species exists betweenthe strains from the different countries. The strains from Rostock (R33, RS2, RI04,R12, R39, R48) seem to be more closely related as evidenced by some small plasmidsof the same size. Different plasmid profiles were found in 01:Kl:H-:F9 strains. Table2 shows the plasmid DNA species observed in six of these strains isolated in Rostock.Three of them may be considered identical. One plasmid DNA of 2.6 Md was found tobe common for all of the F9 strains. Phage patterns are also shown in Table 2 givingfurther evidence for relationship in these strains.

The colicin production of the 01: K1 strains was examined according to Abbott andShannon (1). This system originally introduced for Shigella sonnei was also found to beuseful for epidemiological studies in Escherichia coli (7, 12). As shown in Table 3members of the two 01:Kl clones produced different colicin types. Eight differentcolicin types were identified among theFll strains. Only some of them were commonto both H7 and H- isolates. Strains belonging to the F9 clone showed only one colicintype. Twenty out of thirty 01 :Kl :H-:F9 isolates produced colicin type 3. The other tenstrains were non-colicinogenic.

Table 2. Plasmid DNA content and phage patterns of E. coli 01:Kl strains with fimbrialantigen F9

Strain Serotype Plasmid DNA species (Md) Phage patterns (13)Kl others

Rll 01:K1:H-:F9 55 32 2,8 2,6 1,25 ABCE NTR61 01:K1:H-:F9 55 32 2,8 2,6 1,25 ABCE NTR88 Ol:Kl:H-:F9 55 32 2,8 2,6 1,25 ABCE NT

R36 Ol:Kl:H-:F9 2,6 NT AMR34 Ol:Kl :H-:F9 2,6 NT NTR15 Ol:K I :H-:F9 55 2,6 NT NT

NT = not typabJe

Table 3. Colicin types (1) of 59 E. coli 01:Kl strains

Serotype No. of 1B 2 3 3A 4 5 11 14 not iden- no pro-strains tified ducers

01:Kl:H7:Fll 18 4 5 1 2 3 201:Kl :H-:Fl1 11 8 1 001 :Kl :H-:F9 30 20 - 10

Discussion

Escherichia coli Ol:Kl strains have been subject for recent investigations (2,4, 11,13-16, 18, 22, 24). Summarized epidemiological data revealed that 01:Kl bacteriawere common in cases of pyelonephritis but rarely isolated from patients with neonatalmeningitis (19).

E.coli 01:K1 Clones 327

The results of our studies on the membrane protein patterns in urinary 01:K1strains are generally in agreement with findings of the Achtman group (2, 11). Twodistinctly different patterns were found in the isolates each being characteristic for oneclone. No differences were observed in strains from different countries. Thus themigration patterns of major outer membrane proteins are stable markers and useful forrecognizing bacterial clones.

The plasmid content of the strains investigated was found to be relatively variable innumber and size. Some clonal characteristics were apparent. In the 01:K1:H7:F11 andH-:F11 isolates one plasmid of 85 Md size was present. These plasmids were detectedin all strains independent of the geographical origin and seem to be characteristic forthe clone. A random occurrence of different plasmids of the same size seems to beunlikely if recent investigations of Mercer et al. are considered (8). These authors coulddemonstrate by restriction endonuclease digestion of the plasmids of 01 :Kl bacteriathat the large plasmids present in such strains were very similar to each other. TheEscherichia coli reference strain for 01 antigen (U5-41) containing the single plasmidof 82 Md size was included in that study. This is in good agreement with the resultsshown in our study (85 Md, Table 1). The large plasmid was shown to code for colicinproduction and the major outer membrane protein PCP (2, 8). Two of our strains wereno producers of colicins and some without detectable amounts of PCP in spite of thepresence of this particular plasmid, however.

The plasmid content of 01:K1:H-:F9 isolates was different from that of the Fllstrains. They do not harbour the large plasmid present in the latter. The restrictionendonuclease analysis of comparable 01:K1 strains with the same MP 5 did not giveany similarity of digestion patterns neither within this clonal group nor with strains ofMP 9 (8). The strain collection used in that study differed from our material particular­ly in that most of the strains were faecal isolates without P-fimbriae and with negativeindole reaction (2, 4). A similar but not identical plasmid content in 01:K1 strains hasalso been found in other studies (20, 22). Plasmid analysis of our strains revealed oneseemingly common plasmid DNA species of 2.6 Md. Their plasmid profiles generallywere found to be more closely related.

Colicin production has already been studied in Escherichia coli strains of varioussources and from different geographical locations (2,7, 10, 12,20). The production ofcolicins was found to be of variable nature and subject to continued evolution. Somekind of association seems to exist between O:K:H serotype and colicin production assome serotypes, for example 018ac:K1:H7 isolates of MP 6, were non-colicinogenicand others were producers (2). No uniform colicins have been detected among 01:K1strains of MP 9 which is accordant with our own results (2). Colicin E1 was the mostcommon colicin evolved from 01:K1 bacteria of MP 5 (2). A plasmid of 4.2 Md hasbeen described as being responsible for colicin E1 production (3) but no plasmid of thatsize was found in our strains. Colicin type 3 was the only colicin detected in the01:K1:H-:F9 strains. There is no similarity between the colicin of Abbott-Shannontype 3 and colicin E1, the latter being type 6 in the Abbott-Shannon scheme, which hasnever been found in the material presented. The difference in the strain collectionsmentioned above may account also for these findings.

The clone concept implies that bacterial strains descending from one and the sameancestral cell are expected to have preserved all the properties from the progenitor and,therefore, are identical in their phenotypic and genetic traits (17). In the strict sense ofthis term only three of the strains (No. RII, R61 and R88) could be designated asmembers of a clone. These strains are identical in O:K:H serotype, fimbrial antigens

328 W. Nimmich, G. Zingler, G. Seltmann, E.Tietze, and G. Horn

(15), biochemical properties, 01 factor analysis, and phage patterns (13). In the pre­sent study further evidence for identity was found in membrane protein patterns,plasmid content, and colicin production. Two of the strains (No. R61, No. R88) wereisolated in 1981 from women with pyelonephritis in two different wards of the Ros­tock University Hospital for Obstetrics and Gynecology. The third strains (No. Rll)was also from a female patient with pyelonephritis but from a hospital outside ofRostock. These isolates certainly are of the same origin, appear to have spread inRostock and surroundings and should be considered of epidemic importance. A newlystarted study on the epidemiology of 01:K1 strains in Rostock will show whether theharbouring of special plasmids or other properties are of some importance for theepidemic occurrence of such strains.

Strains No. 34 and No. 36 of the same clonal group have been isolated from onewoman with recurrent urinary tract infection in a two months distance. They areidentical in all the traits mentioned above including antibiotic susceptibility except forthe phage patterns. The first isolate (No. R36) was sensitive to phages A and M (6,13)while the second (No. R34) was resistant. Phage susceptibility, therefore, seem to beprone to evolutionary variation already within a short time.

Epidemiological studies have shown that plasmid profiles are relatively stable mar­kers of bacterial strains, even if originating from various locations (21). On the otherhand plasmids and plasmid encoded phenotypes such as colicin production are knownto be the most variable features of bacteria in evolution (2, 8, 17,20,21,23). From ourdata, two clonal lines of Escherichia coli 01:K1, closely related with respect to manyproperties, could be shown to give rise to genetic variants in course of evolution. This isreflected by different plasmid patterns and indicates individual epidemic processesinvolving particular strains.

References

1. Abbott, ]. D. and R. Shannon: A method for typing Shigella sonnei using colicineproduction as a marker. J. Clin. Path. 11 (1958) 71

2. Achtman, M., A. Mercer, B. Kusecek, A. Pohl, M. Heuzenroeder, W. Aaronson, A.Sutton, and R. P. Silver: Six widespread bacterial clones among Escherichia coli K1isolates. Infect. Immun. 39 (1983) 315-335

3. Bazaral, M. and D. R. Helsinki: Characterization of multiple circular DNA forms ofcolicinogenic factor £1 from Proteus mirabilis. Biochemistry 10 (1968) 3513-3519

4. Kusecek, B., H. Wloch, A. Mercer, V. Yiiisanen, G. Pluschke, T. Korhonen, and M.Achtman: Lipopolysaccharide, capsule, and fimbriae as virulence factors among 01,07,016,018 or 075 and K1, K5, or K100 Escherichia coli. Infect. Immun. 43 (1984)368-379

5. Lugtenberg, B.,]. Meijers, R. Peters, P. van der Hoek, and L. van Alphen: Electrophore­tic resolution of the "major outer membrane protein" of Escherichia coli K12 into fourbands. FEBS Lett. 58 (1975) 254-258

6. Marsik, F. ]. and]. T. Parisi: Bacteriophage types and 0 antigen groups of Escherichiacoli from urine. Appl. Microbiol. 22 (1971) 26-31

7. McGeachie, J.: Bacteriocin typing in urinary infection. Zbl. Bakt. Hyg., I. Abt, Orig.196 (1965) 377-384

8. Mercer, A., G. Morelli, M. Heuzenroeder, M. Kamke, and M. Achtman: Conservationof plasmids among Escherichia coli K1 isolates of diverse origins. Infect. Immun. 46(1984) 649-657

E. coli 01 :Kl Clones 329

9. Meyers, J. A., D. Sanchez, L. P. Elwell, and S. Falkow: Simple agarose gel electrophore­tic method for the identification and characterization of plasmid DNA. J. Bact. 127(1976) 1529-1537

10. Minshew, B. H.,]. Jorgensen, G. W. Counts, and S. Falkow: Association of hemolysinproduction, hemagglutination of human erythrocytes, and virulence for chicken em­bryos of extraintestinal Escherichia coli isolates. Infect. Immun. 20 (1978) 50-54

11. Moll, A., B. Kusecek, G. Pluschke, G. Morelli, M. Kamke, B. [ann, K. [ann, and M.Achtman: A reexamination of the 01 lipopolysaccharide antigen group of Escherichiacoli. Infect. Immun. 53 (1986) 257-263

12. Mustafa, M. A. and]. M. Dunbar: The use of colicine typing in a study of the relation­ship of infecting urinary organism to the fecal flora in pregnant patients with significantbacteriuria. J. Obstet. Gyn. 77 (1970) 544-547

13. Nimmich, W. and G. Zingler: Biochemical characteristics, phage patterns, and 01factor analysis of Escherichia coli 01:Kl:H7:F11 and 01:Kl:H-:F9 strains isolatedfrom patients with urinary tract infections. Med. Microbiol. Immunol. 173 (1984)75-85

14. Nimmich, W., G. Zingler, U. Falkenhagen, E. Straube und Yu. A. Ratiner: Haufigkeitverschiedener Escherichia coli O:Kl:H-Serotypen bei Patienten mit Harnweginfek­tionen. Z. Klin. Med. 41 (1986) 1759-1761

15. Nimmich, W., G. Zingler. and I. 0rskov: Fimbrial antigens of Escherichia coli01:Kl:H7 and 01:Kl:H- strains isolated from patients with urinary tract infections.Zbl. Bakt. Hyg. A 258 (1984) 104-111

16. Ochman, H. and R. K. Selander: Evidence for clonal population structure in Escherichiacoli. Proc. Nat!. Acad. Sci. (USA) 81 (1984) 198-201

17. 0rskov, F. and I. 0rskov: Summary of a workshop on the clone concept in theepidemiology, taxonomy, and evolution of the Enterobacteriaceae and other bacteria. J.Infect. Dis. 148 (1983) 346-357

18. 0rskov, I., F. 0rskov. A. Birch-Andersen, M. Kanamori, and C. Suanborg-Eden: 0, K,H, and fimbrial antigens in Escherichia coli serotypes associated with pyelonephritis andcystitis. J. Infect. Dis. Suppl. 33 (1982) 18-25

19. Pluschke, G., A. Mercer, B. Kusecek, A. Pohl, and M. Achtman: Induction of bac­teremia in newborn rats by Escherichia coli Kl is correlated with only certain 0(lipopolysaccharide) antigen types. Infect. Immun. 39 (1983) 599-608

20. Silver, R. P., W. Aaronson, A. Sutton, and R. Schneerson: Comparative analysis ofplasmids and some metabolic characteristics of Echerichia coli Kl from diseased andhealthy individuals. Infect. Immun. 29 (1980) 200-206

21. Tietze, E. and H. Tschdpe: Plasmid pattern analysis of natural bacterial isolates and itsepidemiological implication. ]. Hyg. (Camb.) 90 (1983) 475-488

22. Yaisdnen-Rben, V.,]. Elo, E. Viiisiinen, A. Siitonen, I. 0rskov, F. 0rskov, S. B. Sven­son, P. H. Miike/ii, and T. K. Korhonen: P-fimbriated clones among uropathogenicEscherichia coli strains. Infect. Immun. 43 (1984) 149-155

23. Wachsmuth, K., ]. DeBoy, K. Birkness, D. Sack, and]. Wells: Genetic transfer ofantimicrobial resistance and enterotoxigenicity among Escherichia coli strains. Antimi­crob. Agents Chemother. 23 (1983) 278-283

24. Zingler, G., W. Nimmich, G. Naumann, U. Falkenhagen, E. Straube, I. Ulisko, Yu. A.Ratiner und I. V. Golubeva: O-Antigenfaktoren bei Escherichia coli O'l-Stammen vonPatienten mit Harnweginfektionen. Z. ges. Hyg. 31 (1985) 44-46

Prof. Dr. sc. med. Dr. rer. nat. Wolfgang Nimmich, Institut fur Mikrobiologie des Be­reiches Medizin der Wilhelm-Pieck-Universitat Rostock, Leninallee 70, DDR-2500 Rostock