- Email: [email protected]
Persistence and spread of a chloramphenicol resistance-mediating plasmid in antigenic types of Escherichia coli, pathogenic for piglets
PLASMID
4, 123-129 (1980)
Persistence
and Spread of a Chloramphenicol Resistance-Mediating Plasmid in Antigenic Types of Escherichia co/i, Pathogenic for Piglets
SIGRID TUE JORGENSEN,’ JOHN GRINSTED, Department
of Bacteriology,
University
Received November
AND MARK H. RICHMOND
PETER BENNETT,
of Bristol,
University
Walk, Bristol BS8 ITD, England
26, 1979; revised January 31, 1980
All chloramphenicol-resistant Escherichia coli strains isolated from piglets in the State veterinary Serum Laboratory, Copenhagen, in 1974- 1975 harbored plasmids of IncFII group with largely the same resistance markers. Two strains from 1978 carried plasmids with similar characters. Restriction enzyme analysis of DNA from these plasmids with restriction endonucleases EcoRI, BgIII, and PstI shows that the Cm plasmids are extremely closely related; but the patterns obtained (particularly fromPsr1 digests) enable the classification of the plasmids into groups. These bear a strong relation to time and place of isolation so that plasmids isolated on the same farm belong to the same group even when their host strains are of different antigenic types. It is concluded that these plasmids have evolved from a single plasmid.
Chloramphenicol resistance (Cm) plasmids have been detected at a low frequency among various antigenic types of Escherichia coli isolated from outbreaks of diarrhea among piglets in Denmark (Jorgensen and Poulsen, 1976). Since it is undesirable to find resistance plasmids of any kind in bacteria isolated from animals destined for human consumption, it was decided to investigate the further occurrence of these plasmids by monitoring the resistance patterns of strains of E. coli pathogenic for piglets. These strains were isolated in the State Veterinary Serum Laboratory in Copenhagen over a period of time. Using this approach, a total of 19 strains carrying Cm plasmids was isolated from April 1974 to October 1975. The screening for Cm plasmids was then stopped, but it was resumed in 1978 with the purpose of looking for any similarity between the plasmids isolated then and those found earlier. The broad characteristics of the plasmids isolated in 1974- 1975 (transmissibility, in-
compatibility, and their ability to be transduced by phage P 1kc) have already been reported (Jorgensen, 1978). However, it proved impossible to distinguish between the plasmids on these grounds, and therefore molecular studies, among which was restrictionenzyme analysis, were instituted to see whether differences in the plasmids would emerge by the use of these more discriminating techniques. The application of restriction-enzyme analysis as a means to study the relatedness between plasmids was first described by Thompson et al. (1974). The method was subsequently applied in a survey on SmSu plasmids (Grinter and Barth, 1976) and degradative plasmids (Duggleby ef al., 1977). More recently Causey and Brown (1978) confirmed its usefulness, but also pointed out its limitations. In the present paper we present the results of its application on Cm plasmids from piglets. The data confirm that all Cm plasmids in the study evolved from one single plasmid. MATERIALS
1 Present address: Department of Animal Genetics, Royal Veterinary and Agricultural University, Biilowsvej 13, DK-1870 Copenhagen V, Denmark.
AND METHODS
Strains. The strains used in these studies (Table 1) were naturally occurring chloram123
0147-619X’80/050123-07$02.00/O Copyright 0 1980 by Academic Press, Inc. All rights of reproduction in any form reserved.
124
JORGENSEN
phenicol-resistant Escherichiu coli belonging to three antigenic types commonly associated with diarrhea in piglets: specifically 08:K87, 0141:K85ac, and 0149:K91 (Sojka et ul., 1960; Brskov et al., 1969). The H antigens of these strains were not determined. The strains originated on several farms, all but one of which (Farm A) are located in Jutland (see map, Fig. 1). A number of these farms are known to be “problem farms” where diarrhea among piglets occurs frequently.
Separation of Cm plasmids from other plasmids and isolation of plasmid DNA. The Cm plasmids were transferred from the naturally occurring strains to a Nal*luc~F~ strain of E. co/i K-12 (obtained from H. Williams Smith, 1970)by conjugation as described previously (Jorgensen, 1978). Plasmid DNA was isolated from chloramphenicolresistant transconjugants in CsCl-ethidium bromide dye-buoyant density gradients as described by Cornelis et al. (1976) for the iso-
TABLE CHARACTERISTICS
ET AL.
1 E. coli STRAINS
OF CHLORAMPHENICOL-RESISTANT
AND THEIR
R PLASMIDS”
Plasmid
strain
Origin
Date of isolation
0:K
type
tra characters of transductants
Plasmid designation
group
Size ( Mdal : EM measuremerits)
+
pHG23
FII
53
+ + -
i i +
pHG24 pHG27 pHG30 pHG25 pHG28 pHG31
FII n.d.” FII FII n.d. FII
52 81 64 52 80 63
1
pHG2 1
FII
60
pHG1 pHG2 pHG3 pHG4 pHG5 pHG6 pHG7 pHG8 pHG9 pHGl0 pHGl1
FII FII FII FII FII FII FII FII FII FII FII
52 52 52 52 52 52 52 52 52 52 53
pHG18 pHG29
FII FII
52 52
pHG 16
FII
53
pHG19
FII
58
pHG33
FII
53
pHG35
FII
69
All R markers cotransduced
Resistance marker
Farm D
1974 April
0141:K85ac
CmSmSu
+
Farm B
1974 April
08:K87
CmSmSu SmTc CmSmSuTc CmSmSu SmTc CmSmSuTc
+
Farm C
1974 June
0141:K85ac
CmSmSu
1
Farm A
1975 January
0149:K91
CmSmSu
+ + + + + + + + + + I
Farm C
1975 July 1975 July
0149:K91 014l:K85ac
CmSmSu CmSmSu
Farm E
1975 September
0149zK91
CmSmSu
Farm F
1975 October
0149:K91
ApCmSmSu
Farm G
1978 January
08:K87
CmSmSu
Farm H
1978 July
08:K87
ApCmKmSmSu
n Part of this table has been published previously b n.d.. not determined.
+ + + All R markers cotransformed
(Jergensen,
tra characters of transformation +
1978). tra, transmissibility;
IIK
Inc, incompatibility.
Cm PLASMID
IN PORCINE
lation of unlabeled plasmid DNA, and recovered, after removal of ethidium bromide, by precipitation from the CsCl solution with sodium acetate and ethanol at -20°C for 18 h. The DNA precipitate was pelleted by centrifugation and then resuspended in 10 mM Tris, pH 8, containing 0.1 mM EDTA.2 On occasions where other plasmids were conjugally cotransferred, plasmid DNA was isolated from a transductant or a transformant which carried the Cm plasmid only. Transduction and transformation of Cm plasmids. Transduction from E. coli K-12 transconjugants was performed with phage Plkc (Lennox, 1955), using E. coli C600 as the recipient organism. Selection of transductants was on nutrient agar supplemented with chloramphenicol (15 &ml). Transformation, using plasmid DNA isolated from chloramphenicol-resistant transconjugants, was performed essentially as described by Cohen et al. (1972). Transformants were selected on nutrient agar supplemented with chloramphenicol (15 pg/ml). Restriction enzyme analysis was carried out as described by Grinsted et al. (1977). The restriction endonucleases used had been prepared in the laboratory. DNA fragments generated by restriction-endonuclease cleavage were separated on vertical slab gels (0.7% or 1% agarose) or on horizontal slab gels (0.5% agarose). Horizontal gels were run overnight at low voltage to achieve separation of large fragments. Electronmicroscopy of plasmid DNA was performed as described by Robinson et al. (1977). RESULTS
Table 1 summarizes previously published properties of the plasmids in this survey plus some new information. Despite the fact that they originated in E. coli strains of different serotypes isolated from a variety of geographical locations (Fig. 1) most of the plasmids appear to be very similar if not identical. Thus all 23 Cm plasmids are conjugative, can be transduced by phage Pl , and are IncFII. 2 Abbreviations
used: Mdal, megadaltons.
E. coli PATHOGENS
FIG. 1. Geographical distribution of farms in Denmark with Cm-resistant Escherichia co/i in piglets.
In addition to Cm resistance the majority of the plasmids carry resistance to SmSu and are 52-53 Mdal of size. Four plasmids, however, confer additional resistance to Tc, Ap, or ApKm and consequently they are somewhat larger. The two SmTc plasmids pHG27 and pHG28 are not related to the Cm plasmids and appear in Table 1 only because Tc, and possibly Sm, proved to translocate from pHG27 and pHG28 on to pHG24 and pHG25, respectively, thus creating “new” Cm plasmids (Jargensen, 1978;Jorgensen, Oliva, Grinsted & Bennett, in preparation). Restriction Enzyme Analysis Plasmid DNA isolated from 19 CmSmSu, 2 CmSmSuTc, 1 ApCmSmSu, and 1 ApCmKmSmSu resistant transconjugants was digested with EcoRI, BgfII, or PstI restriction endonucleases, and the fragment patterns obtained were compared. EcoRI Endonuclease Analysis Eighteen of the CmSmSu plasmids and one ApCmSmSu plasmid gave apparently iden-
126
JORGENSEN
ticalEcoRI band patterns comprising 10bands (Fig. 2a). This band pattern is regarded as the standard EcoRI pattern for the purposes of comparison. EcoRI digestion of pHG21 (Table 1) generated a slightly different pattern from the standard. EcoRI band 2 was missing while band 1 appeared as a doublet (Fig. 2a), consistent with an increase in plasmid size (Table 1). The single ApCmKmSmSu plasmid examined in this survey, pHG35, yielded a
ET AL.
band pattern which differed markedly from the standard pattern (Fig. 2a), although some similarity remained, i.e., bands 2, 3, and 6-10 of the standard pattern appear to be present in EcoRI digests of pHG35 DNA. Rglm Endonuclease Analysis BglII digestion also failed to distinguish between the 18 CmSmSu plasmids which gave identical EcoRI band patterns. In these
a kpR1
AlECOR -
standard type
pHG21
pH635
-;---
4s6--
b pattern pHG19
pHGZ1 -
l*-..‘.-z
-- -E
3--
&I II standard type
pattern
3-
-
pHG35 --__
-
-
,-a-‘,--
_Y _-__ -
lU--
4-m--
____
c
st:;;:rd (farm
----
pH616 B)
&tl pattern pHG21 pH619 (tad)
band not present double band
In the
standard
pattern
FIG. 2. Schematic illustration of restriction patterns obtained after digestion of plasmid DNA with restriction endonucleases EcoRI (Fig. 2a), BglII (Fig. 2b), and PsrI (Fig. 2~).
Cm PLASMID
IN PORCINE
cases, four bands were seen (BgZII standard pattern, Fig. 2b), the second of which was a doublet. pHG19, the ApCmSmSu plasmid which showed an apparently standard EcoRI band pattern, gave a pattern which differed slightly from the standard when digested with BgZII. One of the fragments forming the doublet band (band 2, Fig. 2b) was missing and in its place a slightly larger fragment appeared. This is consistent with pHG19 being somewhat larger than the CmSmSu plasmids, 58 Mdal as against 53 Mdal (Table l), and suggests that the additional material is incorporated into a nucleotide sequence which comprises one of the BgZII fragments which form the doublet band (band 2, Fig. 2b). Plasmid pHG21 also gave an almost standard band pattern with BgZII, but one of the fragments which form the doublet in the standard pattern was missing and a larger band was present (Fig. 2b). Plasmid pHG35 showed a band pattern in which some bands appeared to be the same as in the standard but one or two changes were also apparent (Fig. 2b).
127
E. coli PATHOGENS
from both Farm A and Farm B. These plasmids had lost bands 5 and 7 of the standard pattern and had acquired two new bands of a slightly larger size, 3.6 and 4 Mdal as compared with 2.4 and 2.05 Mdal. Moreover, this group displayed internal variation in that the PstI band pattern of pHG29 (not shown) also lacked a fragment of size 1.2 Mdal (band 12, Fig. 2c) which was present in the other two plasmids. Predictably, plasmids pHG 19 and pHG35 each showed some changes from the standard pattern (Fig. 2c), and in addition pHG16 could now be distinguished from the other CmSmSu plasmids in that band 12 of the standard had disappeared while a new one appearedbetween bands 5 and 6 (Fig. 2~). Restriction Enzyme Analysis and pHG31
of pHG30
The restriction patterns of pHG30 and pHG3 1 with EcoRI, BglII, and PstI are very similar, but not identical to the three standard patterns. For example, after EcoRI digestion, agarose-gel analysis shows that both plasmids generate 11fragments rather than 10. In PstZ Endonuclease Analysis the caseof pHG3 1, band 6 (of standard EcoRI Whereas the EcoRI band patterns of the pattern, Fig. 2a) was missing while two new majority of the CmSmSu plasmids were bands appeared between bands 2 and 3 of indistinguishable, as were those generated the standard pattern. In the case of pHG30, with BgZII, digestion with PstI gave patterns the standard lo-band pattern appeared to which allowed distinctions to be made. In be intact with the addition of an extra band most casesPstI generated a pattern made up between bands 2 and 3. Digestion of these plasmids with BgZII or with PstI also genof 21 bands (see PstI standard pattern, Fig. erated patterns with minor variations from 2c), and although all the patterns were extremely similar, some slight variations were the standard band patterns. Both pHG30 and pHG3 1 are approximately 12 Mdal larger than found. Thus the plasmids carried by strains isolated on Farm A (data not given) differed the standard CmSmSu plasmids and further from those carried by strains isolated on Farm data (to be published) indicate that both these CmSmSuTc plasmids have arisen as the reB in that a small fragment (band 15, PstI standard pattern, Fig. 2c) from the former sult of transposition of a Tc resistance determinant from a SmTc resistance plasmid, also plasmids was slightly larger than the equivcarried in the farm isolate, to a CmSmSu alent fragment from the latter plasmids (0.88 plasmid. Mdal as against 0.85 Mdal). All plasmids from Farm A yielded indistinguishable band DISCUSSION patterns, while the same was true of plasmids from Farm B. The results presented in this paper confirm All three plasmids from Farm C (pHG18, that all the Cm plasmids studied in the survey, pHG21, pHG29, Table 1) differed from those with the exception of pHG35, are very closely
128
JORGENSEN ET AL.
related despite the fact that they were found in naturally occurring E. co/i strains of different serotype, and despite the fact that the strains were isolated on several different farms in Jutland and Zealand. It is not unreasonable, therefore, to conclude that all of these Cm plasmids originated from a common ancestral plasmid, but it is impossible to say where and when that plasmid emerged. Certainly, since its emergence the plasmid appears to have continued to evolve as witnessed by the accretion of further resistance markers, e.g., Ap in pHG19; Tc in pHG30 and pHG31; and ApKm in pHG35. Although, even plasmids with the same marker pattern may show slight variation, e.g., pHG16, pHG21, and pHG24 (Table 1, Fig. 2). The minor variations observed among these otherwise similar plasmids bear a strong relationship to their time and place of isolation. Hence plasmids isolated from the same outbreak of disease on a single farm, e.g., Farm C, pHG18 and pHG29, appear to be more closely related one to the other than to plasmids isolated in other places, even when the plasmids are carried by strains of different serotypes (Table 1 and Results). Indeed, plasmids isolated at different times, but from the same place (e.g., Farm C) appear to be more closely related one to the other than to plasmids isolated in other farms (see data concerning pHG18, pHG21, and pHG29 from Farm C; cf. plasmids from Farm B and Farm A, Table 1, Fig. 2). However, the fact that plasmids were isolated on one farm does not imply that they differ from those isolated elsewhere. Thus pHG23 and pHG33 appear to be identical to the CmSmSu plasmids isolated on Farm B (Table 1, Fig. 2). The occurrence of identical conjugative CmSmSu plasmids in E. coli strains of different serotype demonstrates the ability of a plasmid to transfer successfully from one organism to another under natural conditions, so facilitating its spread. It is likely that these plasmids and the strains harboring them have been disseminated among swine herds via purchase of piglets. In spite of its ability to persist for several
years in the E. coli population and its successful spread among various antigen types, the emergence of the CmSmSu resistance plasmid has not led to a dramatic increase in the frequency of isolation of Cm-resistant E. co/i as determined from the records of the State Veterinary Serum Laboratory. Presumably, therefore, it is maintained at a low level in the E. coli population. It is surprising that these Cm-resistant pathogenic strains of E. coli were isolated almost exclusively on farms in Jutland even though the transfer of piglets between the various parts of Denmark is common. Our data confirm that in nature, even a stable well-established plasmid may undergo spontaneous DNA sequence alterations leading to a shifting of restriction-endonuclease sites as well as other major changes such as the pickup of new resistance markers. This is in agreement with Timmis et al. (1978) although they observed a somewhat higher instability in the DNA sequences of their plasmid, R6-5. In Denmark, the use of chloramphenicol in animals destined for human consumption has now been discontinued. It will be interesting to see if the CmSmSu plasmid persists in the E. cofi population. Certainly the carriage of the resistance determinants Sm and Su should aid its survival since these drugs are still widely used in veterinary practice. ACKNOWLEDGMENTS We are grateful to Evelyn Lewis, University of Bristol for the EM work and to A. Dam and K. B. Pedersen, State Veterinary Serum Laboratory, Copenhagen, for strains. S. Tue Jorgensen was supported by a traveling research fellowship from The Carlsberg Foundation, Copenhagen. This is gratefully acknowledged.
REFERENCES BENNETT,P. M., AND RICHMOND,M. H. (1976).Translocation of a discrete piece of deoxy-ribonucieic acid carrying an amp gene between replicons in Escherichia co/i. J. Bacterial.
126, 1-6.
CAUSEY,S. C., AND BROWN,L. R. (1978). Transconju-
Cm PLASMID IN PORCINE E. co/i PATHOGENS gant analysis: Limitations on the use of sequencespecific endonucleases for plasmid identification. J.
129
JORGENSEN,S. T., AND POULSEN,A.-L. (1976). Antibiotic resistance and Hly plasmids in serotypes of Bacreriol. 135, 1070- 1079. Escherichia coli associated with porcine enteric COHEN, S. N., CHANG, A. C. Y., AND Hsu, L. (1972). disease. Antimicrob. Agents Chemother. 9, 6- 10. Nonchromosomal antibiotic resistance in bacteria: LENNOX, E. S. (1955). Transduction of linked genetic genetic transformation of Eschen’chia co/i by Rcharacters of the host by bacteriophage Pl. Virology 1, factor DNA. Proc. Nat. Acad. Sci. USA 69, 190-206. 2110-2114. ~RSKOV,I., ~RSKOV,F., WITTIG, W., AND SWEENEY, CORNELIS, G., BENNETT, P. M., AND GRINSTED, E.. J. (1%9). A new E. coli serotype 0149:K91(B): J. (1976). Properties of pGC1 a lac plasmid originating K88ac(L):HlO isolated from diseased swine. Acta in Yersinia entuocolitica 842. J. Bacterial. 127, Pathol. Microbial. &and. 75, 491-498. 1058- 1062. ROBINSON,M. K., BENNETT, P. M., AND RICHMOND, DUGGLEBY, C. J., BAYLEY, S. A., WORSEY,M. J., M. H. (1977). Inhibition of TnA translocation by WILLIAMS, P. A., AND BRODA, P. (1977). Molecular TnA. J. Bacieriol. 129, 407-414. sizes and relationships of TOL plasmids in Pseudo- SOJKA, W. J., LLOYD, M. K., AND SWEENEY, E. J. monas. J. Bacterial. 130, 1274-1280. (1960). Escherichia coli serotypes associated with GRINSTED,J., BENNETT,P. M., AND RICHMOND,M. H. certain pig diseases. Res. Vet. Sci. 1, 17-27. (1977). Restriction map of RPl. Plasmid 1, 34-37. TIMMIS, K. N., CABELLO,F., ANDRBS,I., NORDHEIM, GRINTER, N. J., AND BARTH, P. T. (1976). CharacteriA., BURKHARDT, H. J., AND COHEN, S. N. (1978). zation of SmSu plasmids by restriction endonuclease Instability of plasmid DNA sequences: Macro and micro evolution of the antibiotic resistance plasmid cleavage and compatibility testing. J. Bacreriol. 128, 394-400. R6-5. Mol. Gen. Gene!. 167, 11-19. JORGENSEN,S. T. (1978). Chloramphenicol resistance THOMPSON,R., HUGHES,S. G., AND BRODA,P. (1974). plasmids in Escherichia co/i isolated from diseased Plasmid identification using endonucleases. Mol. piglets. Antimicrob. Agents Chemother. 13,710-715. Gen. Genet. 133, 141-149.
4, 123-129 (1980)
Persistence
and Spread of a Chloramphenicol Resistance-Mediating Plasmid in Antigenic Types of Escherichia co/i, Pathogenic for Piglets
SIGRID TUE JORGENSEN,’ JOHN GRINSTED, Department
of Bacteriology,
University
Received November
AND MARK H. RICHMOND
PETER BENNETT,
of Bristol,
University
Walk, Bristol BS8 ITD, England
26, 1979; revised January 31, 1980
All chloramphenicol-resistant Escherichia coli strains isolated from piglets in the State veterinary Serum Laboratory, Copenhagen, in 1974- 1975 harbored plasmids of IncFII group with largely the same resistance markers. Two strains from 1978 carried plasmids with similar characters. Restriction enzyme analysis of DNA from these plasmids with restriction endonucleases EcoRI, BgIII, and PstI shows that the Cm plasmids are extremely closely related; but the patterns obtained (particularly fromPsr1 digests) enable the classification of the plasmids into groups. These bear a strong relation to time and place of isolation so that plasmids isolated on the same farm belong to the same group even when their host strains are of different antigenic types. It is concluded that these plasmids have evolved from a single plasmid.
Chloramphenicol resistance (Cm) plasmids have been detected at a low frequency among various antigenic types of Escherichia coli isolated from outbreaks of diarrhea among piglets in Denmark (Jorgensen and Poulsen, 1976). Since it is undesirable to find resistance plasmids of any kind in bacteria isolated from animals destined for human consumption, it was decided to investigate the further occurrence of these plasmids by monitoring the resistance patterns of strains of E. coli pathogenic for piglets. These strains were isolated in the State Veterinary Serum Laboratory in Copenhagen over a period of time. Using this approach, a total of 19 strains carrying Cm plasmids was isolated from April 1974 to October 1975. The screening for Cm plasmids was then stopped, but it was resumed in 1978 with the purpose of looking for any similarity between the plasmids isolated then and those found earlier. The broad characteristics of the plasmids isolated in 1974- 1975 (transmissibility, in-
compatibility, and their ability to be transduced by phage P 1kc) have already been reported (Jorgensen, 1978). However, it proved impossible to distinguish between the plasmids on these grounds, and therefore molecular studies, among which was restrictionenzyme analysis, were instituted to see whether differences in the plasmids would emerge by the use of these more discriminating techniques. The application of restriction-enzyme analysis as a means to study the relatedness between plasmids was first described by Thompson et al. (1974). The method was subsequently applied in a survey on SmSu plasmids (Grinter and Barth, 1976) and degradative plasmids (Duggleby ef al., 1977). More recently Causey and Brown (1978) confirmed its usefulness, but also pointed out its limitations. In the present paper we present the results of its application on Cm plasmids from piglets. The data confirm that all Cm plasmids in the study evolved from one single plasmid. MATERIALS
1 Present address: Department of Animal Genetics, Royal Veterinary and Agricultural University, Biilowsvej 13, DK-1870 Copenhagen V, Denmark.
AND METHODS
Strains. The strains used in these studies (Table 1) were naturally occurring chloram123
0147-619X’80/050123-07$02.00/O Copyright 0 1980 by Academic Press, Inc. All rights of reproduction in any form reserved.
124
JORGENSEN
phenicol-resistant Escherichiu coli belonging to three antigenic types commonly associated with diarrhea in piglets: specifically 08:K87, 0141:K85ac, and 0149:K91 (Sojka et ul., 1960; Brskov et al., 1969). The H antigens of these strains were not determined. The strains originated on several farms, all but one of which (Farm A) are located in Jutland (see map, Fig. 1). A number of these farms are known to be “problem farms” where diarrhea among piglets occurs frequently.
Separation of Cm plasmids from other plasmids and isolation of plasmid DNA. The Cm plasmids were transferred from the naturally occurring strains to a Nal*luc~F~ strain of E. co/i K-12 (obtained from H. Williams Smith, 1970)by conjugation as described previously (Jorgensen, 1978). Plasmid DNA was isolated from chloramphenicolresistant transconjugants in CsCl-ethidium bromide dye-buoyant density gradients as described by Cornelis et al. (1976) for the iso-
TABLE CHARACTERISTICS
ET AL.
1 E. coli STRAINS
OF CHLORAMPHENICOL-RESISTANT
AND THEIR
R PLASMIDS”
Plasmid
strain
Origin
Date of isolation
0:K
type
tra characters of transductants
Plasmid designation
group
Size ( Mdal : EM measuremerits)
+
pHG23
FII
53
+ + -
i i +
pHG24 pHG27 pHG30 pHG25 pHG28 pHG31
FII n.d.” FII FII n.d. FII
52 81 64 52 80 63
1
pHG2 1
FII
60
pHG1 pHG2 pHG3 pHG4 pHG5 pHG6 pHG7 pHG8 pHG9 pHGl0 pHGl1
FII FII FII FII FII FII FII FII FII FII FII
52 52 52 52 52 52 52 52 52 52 53
pHG18 pHG29
FII FII
52 52
pHG 16
FII
53
pHG19
FII
58
pHG33
FII
53
pHG35
FII
69
All R markers cotransduced
Resistance marker
Farm D
1974 April
0141:K85ac
CmSmSu
+
Farm B
1974 April
08:K87
CmSmSu SmTc CmSmSuTc CmSmSu SmTc CmSmSuTc
+
Farm C
1974 June
0141:K85ac
CmSmSu
1
Farm A
1975 January
0149:K91
CmSmSu
+ + + + + + + + + + I
Farm C
1975 July 1975 July
0149:K91 014l:K85ac
CmSmSu CmSmSu
Farm E
1975 September
0149zK91
CmSmSu
Farm F
1975 October
0149:K91
ApCmSmSu
Farm G
1978 January
08:K87
CmSmSu
Farm H
1978 July
08:K87
ApCmKmSmSu
n Part of this table has been published previously b n.d.. not determined.
+ + + All R markers cotransformed
(Jergensen,
tra characters of transformation +
1978). tra, transmissibility;
IIK
Inc, incompatibility.
Cm PLASMID
IN PORCINE
lation of unlabeled plasmid DNA, and recovered, after removal of ethidium bromide, by precipitation from the CsCl solution with sodium acetate and ethanol at -20°C for 18 h. The DNA precipitate was pelleted by centrifugation and then resuspended in 10 mM Tris, pH 8, containing 0.1 mM EDTA.2 On occasions where other plasmids were conjugally cotransferred, plasmid DNA was isolated from a transductant or a transformant which carried the Cm plasmid only. Transduction and transformation of Cm plasmids. Transduction from E. coli K-12 transconjugants was performed with phage Plkc (Lennox, 1955), using E. coli C600 as the recipient organism. Selection of transductants was on nutrient agar supplemented with chloramphenicol (15 &ml). Transformation, using plasmid DNA isolated from chloramphenicol-resistant transconjugants, was performed essentially as described by Cohen et al. (1972). Transformants were selected on nutrient agar supplemented with chloramphenicol (15 pg/ml). Restriction enzyme analysis was carried out as described by Grinsted et al. (1977). The restriction endonucleases used had been prepared in the laboratory. DNA fragments generated by restriction-endonuclease cleavage were separated on vertical slab gels (0.7% or 1% agarose) or on horizontal slab gels (0.5% agarose). Horizontal gels were run overnight at low voltage to achieve separation of large fragments. Electronmicroscopy of plasmid DNA was performed as described by Robinson et al. (1977). RESULTS
Table 1 summarizes previously published properties of the plasmids in this survey plus some new information. Despite the fact that they originated in E. coli strains of different serotypes isolated from a variety of geographical locations (Fig. 1) most of the plasmids appear to be very similar if not identical. Thus all 23 Cm plasmids are conjugative, can be transduced by phage Pl , and are IncFII. 2 Abbreviations
used: Mdal, megadaltons.
E. coli PATHOGENS
FIG. 1. Geographical distribution of farms in Denmark with Cm-resistant Escherichia co/i in piglets.
In addition to Cm resistance the majority of the plasmids carry resistance to SmSu and are 52-53 Mdal of size. Four plasmids, however, confer additional resistance to Tc, Ap, or ApKm and consequently they are somewhat larger. The two SmTc plasmids pHG27 and pHG28 are not related to the Cm plasmids and appear in Table 1 only because Tc, and possibly Sm, proved to translocate from pHG27 and pHG28 on to pHG24 and pHG25, respectively, thus creating “new” Cm plasmids (Jargensen, 1978;Jorgensen, Oliva, Grinsted & Bennett, in preparation). Restriction Enzyme Analysis Plasmid DNA isolated from 19 CmSmSu, 2 CmSmSuTc, 1 ApCmSmSu, and 1 ApCmKmSmSu resistant transconjugants was digested with EcoRI, BgfII, or PstI restriction endonucleases, and the fragment patterns obtained were compared. EcoRI Endonuclease Analysis Eighteen of the CmSmSu plasmids and one ApCmSmSu plasmid gave apparently iden-
126
JORGENSEN
ticalEcoRI band patterns comprising 10bands (Fig. 2a). This band pattern is regarded as the standard EcoRI pattern for the purposes of comparison. EcoRI digestion of pHG21 (Table 1) generated a slightly different pattern from the standard. EcoRI band 2 was missing while band 1 appeared as a doublet (Fig. 2a), consistent with an increase in plasmid size (Table 1). The single ApCmKmSmSu plasmid examined in this survey, pHG35, yielded a
ET AL.
band pattern which differed markedly from the standard pattern (Fig. 2a), although some similarity remained, i.e., bands 2, 3, and 6-10 of the standard pattern appear to be present in EcoRI digests of pHG35 DNA. Rglm Endonuclease Analysis BglII digestion also failed to distinguish between the 18 CmSmSu plasmids which gave identical EcoRI band patterns. In these
a kpR1
AlECOR -
standard type
pHG21
pH635
-;---
4s6--
b pattern pHG19
pHGZ1 -
l*-..‘.-z
-- -E
3--
&I II standard type
pattern
3-
-
pHG35 --__
-
-
,-a-‘,--
_Y _-__ -
lU--
4-m--
____
c
st:;;:rd (farm
----
pH616 B)
&tl pattern pHG21 pH619 (tad)
band not present double band
In the
standard
pattern
FIG. 2. Schematic illustration of restriction patterns obtained after digestion of plasmid DNA with restriction endonucleases EcoRI (Fig. 2a), BglII (Fig. 2b), and PsrI (Fig. 2~).
Cm PLASMID
IN PORCINE
cases, four bands were seen (BgZII standard pattern, Fig. 2b), the second of which was a doublet. pHG19, the ApCmSmSu plasmid which showed an apparently standard EcoRI band pattern, gave a pattern which differed slightly from the standard when digested with BgZII. One of the fragments forming the doublet band (band 2, Fig. 2b) was missing and in its place a slightly larger fragment appeared. This is consistent with pHG19 being somewhat larger than the CmSmSu plasmids, 58 Mdal as against 53 Mdal (Table l), and suggests that the additional material is incorporated into a nucleotide sequence which comprises one of the BgZII fragments which form the doublet band (band 2, Fig. 2b). Plasmid pHG21 also gave an almost standard band pattern with BgZII, but one of the fragments which form the doublet in the standard pattern was missing and a larger band was present (Fig. 2b). Plasmid pHG35 showed a band pattern in which some bands appeared to be the same as in the standard but one or two changes were also apparent (Fig. 2b).
127
E. coli PATHOGENS
from both Farm A and Farm B. These plasmids had lost bands 5 and 7 of the standard pattern and had acquired two new bands of a slightly larger size, 3.6 and 4 Mdal as compared with 2.4 and 2.05 Mdal. Moreover, this group displayed internal variation in that the PstI band pattern of pHG29 (not shown) also lacked a fragment of size 1.2 Mdal (band 12, Fig. 2c) which was present in the other two plasmids. Predictably, plasmids pHG 19 and pHG35 each showed some changes from the standard pattern (Fig. 2c), and in addition pHG16 could now be distinguished from the other CmSmSu plasmids in that band 12 of the standard had disappeared while a new one appearedbetween bands 5 and 6 (Fig. 2~). Restriction Enzyme Analysis and pHG31
of pHG30
The restriction patterns of pHG30 and pHG3 1 with EcoRI, BglII, and PstI are very similar, but not identical to the three standard patterns. For example, after EcoRI digestion, agarose-gel analysis shows that both plasmids generate 11fragments rather than 10. In PstZ Endonuclease Analysis the caseof pHG3 1, band 6 (of standard EcoRI Whereas the EcoRI band patterns of the pattern, Fig. 2a) was missing while two new majority of the CmSmSu plasmids were bands appeared between bands 2 and 3 of indistinguishable, as were those generated the standard pattern. In the case of pHG30, with BgZII, digestion with PstI gave patterns the standard lo-band pattern appeared to which allowed distinctions to be made. In be intact with the addition of an extra band most casesPstI generated a pattern made up between bands 2 and 3. Digestion of these plasmids with BgZII or with PstI also genof 21 bands (see PstI standard pattern, Fig. erated patterns with minor variations from 2c), and although all the patterns were extremely similar, some slight variations were the standard band patterns. Both pHG30 and pHG3 1 are approximately 12 Mdal larger than found. Thus the plasmids carried by strains isolated on Farm A (data not given) differed the standard CmSmSu plasmids and further from those carried by strains isolated on Farm data (to be published) indicate that both these CmSmSuTc plasmids have arisen as the reB in that a small fragment (band 15, PstI standard pattern, Fig. 2c) from the former sult of transposition of a Tc resistance determinant from a SmTc resistance plasmid, also plasmids was slightly larger than the equivcarried in the farm isolate, to a CmSmSu alent fragment from the latter plasmids (0.88 plasmid. Mdal as against 0.85 Mdal). All plasmids from Farm A yielded indistinguishable band DISCUSSION patterns, while the same was true of plasmids from Farm B. The results presented in this paper confirm All three plasmids from Farm C (pHG18, that all the Cm plasmids studied in the survey, pHG21, pHG29, Table 1) differed from those with the exception of pHG35, are very closely
128
JORGENSEN ET AL.
related despite the fact that they were found in naturally occurring E. co/i strains of different serotype, and despite the fact that the strains were isolated on several different farms in Jutland and Zealand. It is not unreasonable, therefore, to conclude that all of these Cm plasmids originated from a common ancestral plasmid, but it is impossible to say where and when that plasmid emerged. Certainly, since its emergence the plasmid appears to have continued to evolve as witnessed by the accretion of further resistance markers, e.g., Ap in pHG19; Tc in pHG30 and pHG31; and ApKm in pHG35. Although, even plasmids with the same marker pattern may show slight variation, e.g., pHG16, pHG21, and pHG24 (Table 1, Fig. 2). The minor variations observed among these otherwise similar plasmids bear a strong relationship to their time and place of isolation. Hence plasmids isolated from the same outbreak of disease on a single farm, e.g., Farm C, pHG18 and pHG29, appear to be more closely related one to the other than to plasmids isolated in other places, even when the plasmids are carried by strains of different serotypes (Table 1 and Results). Indeed, plasmids isolated at different times, but from the same place (e.g., Farm C) appear to be more closely related one to the other than to plasmids isolated in other farms (see data concerning pHG18, pHG21, and pHG29 from Farm C; cf. plasmids from Farm B and Farm A, Table 1, Fig. 2). However, the fact that plasmids were isolated on one farm does not imply that they differ from those isolated elsewhere. Thus pHG23 and pHG33 appear to be identical to the CmSmSu plasmids isolated on Farm B (Table 1, Fig. 2). The occurrence of identical conjugative CmSmSu plasmids in E. coli strains of different serotype demonstrates the ability of a plasmid to transfer successfully from one organism to another under natural conditions, so facilitating its spread. It is likely that these plasmids and the strains harboring them have been disseminated among swine herds via purchase of piglets. In spite of its ability to persist for several
years in the E. coli population and its successful spread among various antigen types, the emergence of the CmSmSu resistance plasmid has not led to a dramatic increase in the frequency of isolation of Cm-resistant E. co/i as determined from the records of the State Veterinary Serum Laboratory. Presumably, therefore, it is maintained at a low level in the E. coli population. It is surprising that these Cm-resistant pathogenic strains of E. coli were isolated almost exclusively on farms in Jutland even though the transfer of piglets between the various parts of Denmark is common. Our data confirm that in nature, even a stable well-established plasmid may undergo spontaneous DNA sequence alterations leading to a shifting of restriction-endonuclease sites as well as other major changes such as the pickup of new resistance markers. This is in agreement with Timmis et al. (1978) although they observed a somewhat higher instability in the DNA sequences of their plasmid, R6-5. In Denmark, the use of chloramphenicol in animals destined for human consumption has now been discontinued. It will be interesting to see if the CmSmSu plasmid persists in the E. cofi population. Certainly the carriage of the resistance determinants Sm and Su should aid its survival since these drugs are still widely used in veterinary practice. ACKNOWLEDGMENTS We are grateful to Evelyn Lewis, University of Bristol for the EM work and to A. Dam and K. B. Pedersen, State Veterinary Serum Laboratory, Copenhagen, for strains. S. Tue Jorgensen was supported by a traveling research fellowship from The Carlsberg Foundation, Copenhagen. This is gratefully acknowledged.
REFERENCES BENNETT,P. M., AND RICHMOND,M. H. (1976).Translocation of a discrete piece of deoxy-ribonucieic acid carrying an amp gene between replicons in Escherichia co/i. J. Bacterial.
126, 1-6.
CAUSEY,S. C., AND BROWN,L. R. (1978). Transconju-
Cm PLASMID IN PORCINE E. co/i PATHOGENS gant analysis: Limitations on the use of sequencespecific endonucleases for plasmid identification. J.
129
JORGENSEN,S. T., AND POULSEN,A.-L. (1976). Antibiotic resistance and Hly plasmids in serotypes of Bacreriol. 135, 1070- 1079. Escherichia coli associated with porcine enteric COHEN, S. N., CHANG, A. C. Y., AND Hsu, L. (1972). disease. Antimicrob. Agents Chemother. 9, 6- 10. Nonchromosomal antibiotic resistance in bacteria: LENNOX, E. S. (1955). Transduction of linked genetic genetic transformation of Eschen’chia co/i by Rcharacters of the host by bacteriophage Pl. Virology 1, factor DNA. Proc. Nat. Acad. Sci. USA 69, 190-206. 2110-2114. ~RSKOV,I., ~RSKOV,F., WITTIG, W., AND SWEENEY, CORNELIS, G., BENNETT, P. M., AND GRINSTED, E.. J. (1%9). A new E. coli serotype 0149:K91(B): J. (1976). Properties of pGC1 a lac plasmid originating K88ac(L):HlO isolated from diseased swine. Acta in Yersinia entuocolitica 842. J. Bacterial. 127, Pathol. Microbial. &and. 75, 491-498. 1058- 1062. ROBINSON,M. K., BENNETT, P. M., AND RICHMOND, DUGGLEBY, C. J., BAYLEY, S. A., WORSEY,M. J., M. H. (1977). Inhibition of TnA translocation by WILLIAMS, P. A., AND BRODA, P. (1977). Molecular TnA. J. Bacieriol. 129, 407-414. sizes and relationships of TOL plasmids in Pseudo- SOJKA, W. J., LLOYD, M. K., AND SWEENEY, E. J. monas. J. Bacterial. 130, 1274-1280. (1960). Escherichia coli serotypes associated with GRINSTED,J., BENNETT,P. M., AND RICHMOND,M. H. certain pig diseases. Res. Vet. Sci. 1, 17-27. (1977). Restriction map of RPl. Plasmid 1, 34-37. TIMMIS, K. N., CABELLO,F., ANDRBS,I., NORDHEIM, GRINTER, N. J., AND BARTH, P. T. (1976). CharacteriA., BURKHARDT, H. J., AND COHEN, S. N. (1978). zation of SmSu plasmids by restriction endonuclease Instability of plasmid DNA sequences: Macro and micro evolution of the antibiotic resistance plasmid cleavage and compatibility testing. J. Bacreriol. 128, 394-400. R6-5. Mol. Gen. Gene!. 167, 11-19. JORGENSEN,S. T. (1978). Chloramphenicol resistance THOMPSON,R., HUGHES,S. G., AND BRODA,P. (1974). plasmids in Escherichia co/i isolated from diseased Plasmid identification using endonucleases. Mol. piglets. Antimicrob. Agents Chemother. 13,710-715. Gen. Genet. 133, 141-149.