Veterinary Microbiology, 31 ( 1 9 9 2 ) 2 7 3 - 2 8 5 Elsevier Science Publishers B.V., A m s t e r d a m
273
Typing of Australian isolates of Treponema hyodysenteriae by serology and by DNA restriction endonuclease analysis B.G. Combs, D.J. Hampson and S.J. Harders School of Veterinary Studies, Murdoch University, Murdoch, WA, Australia (Accepted 23 October 1991 )
ABSTRACT Combs, B.G., Hampson, D.J. and Harders, S.J., 1992. Typing of Australian isolates of Treponema hyodysenteriae by serology and by DNA restriction endonuclease analysis. Vet. Microbiol., 31:273285. A total of 91 isolates of Treponema hyodysenteriae which were obtained from 62 piggeries located around Australia were typed by serology and by DNA restriction endonuclease analysis (REA). The isolates fell into eight serogroups, of which groups B and D were the most common. Isolates with different REA patterns were recognised within serogroups, whilst a few isolates with the same REA pattern were placed into different serogroups. Some of the latter isolates were either from the same piggery or from farms with epidemiological links, thus indicating the bacteria may have altered their antigenic properties. A total of 31 different REA patterns were recognised amongst the Australian isolates. These comprised eight major patterns, with four of these being subdivided on the basis of minor differences in banding. Where a number of isolates were obtained from individual piggeries these all had the same REA pattern, and in one piggery isolates with the same pattern were recovered over a five year period. Plasmid bands were observed in 70 of the Australian isolates (77%), and in six of the seven overseas isolates included in the study for comparison. These plasmids did not affect the REA pattern. Of the States from which substantial numbers of isolates were examined, the greatest number of different strains ( 12 amongst 19 piggeries) were found from Victoria, and there were 10 REA patterns in strains from 24 piggeries in Queensland. Despite the large total number of different strains of T. hyodysenteriae in Australia, only three were found in more than one State.
INTRODUCTION
The spirochaete Treponerna hyodysenteriae is the primary aetiological agent of swine dysentery (SD), a major disease of pigs throughout the world (Roncalli and Leaning, 1976; Harris and Glock, 1986). A recent serological survey in Western Australia has shown that the infection is present in around one C o r r e s p o n d e n c e to: B.G. Combs, School of Veterinary Studies, M u r d o c h University, M u r d o c h , WA 6150, Australia.
0 3 7 8 - 1 1 3 5 / 9 2 / $ 0 5 . 0 0 © 1992 Elsevier Science Publishers B.V. All rights reserved.
274
B.G, COMBS ET AL.
third of herds ( M h o m a et al., 1992), and a similar prevalence has been reported in Victoria (Cutler and Gardner, 1988). Although SD is considered to be the single most economically important pig disease in Australia (Cutler and Gardner, 1988), little is known about many aspects of the epidemiology of the infection. Swine dysentery has not been easy to investigate: T. hyodysenteriae is relatively difficult both to isolate and to identify, and until recently there have been few methods available for differentiation within the bacterial species. Without the ability to identify individual strains, it has been difficult to trace outbreaks of infection, and to appreciate the diversity of strains within a geographical location or country. In the present study a large collection of Australian isolates of T. hyodysenteriae was assembled and typed using serological methods (Hampson et al., 1989) and DNA restriction endonuclease analysis (REA) (Combs et al., 1989). The purpose of this work was to gain essential background information about strain diversity and dissemination of T. hyodysenteriae for use in future work on the local epidemiology and control of the infection in Australia.
MATERIALS AND METHODS
Treponemes A total of 91 isolates of T. hyodysenteriae from 62 different piggeries located in five States of Australia were examined (Table 1 ). Western Australian isolates were either isolated in our laboratory or were received through Ms N. Buller of the Western Australian Department of Agriculture. Isolates from South Australia were obtained through Dr A. Pointon, South Australian Department of Agriculture; those from Victoria and New South Wales were obtained either through Dr R.T. Jones, Regional Veterinary Laboratory, Bendigo or through Dr P.J. Coloe, Royal Melbourne Institute of Technology: Queensland isolates were provided through Ms C. Stephens, Regional Veterinary Laboratory, Toowoomba or through Mr R. Thomas, Animal Research Institute, Yeerongpilly. All isolates which were studied originated from piggeries between 1986 and 1990. Strains B78, B234, B204, B169, A1, KF9 and M C 5 2 / 8 0 from the USA, Canada and the U K were included in the study for comparison. These were provided through Dr R.J. Lysons, Institute for Animal Health, Compton, England, and have previously been typed by us (Combs et al., 1989; Hampson et al., 1989). All isolates used in the study were confirmed by us as being T. hyodysenteriae on the basis of their strong beta-haemolysis, production of indole, and a characteristic enzyme profile using the API-ZYM system (Lemcke and Burrows, 1981 ; Hunter and Wood, 1979).
T Y P I N G BY SEROLOGY AND BY DNA RESTRICTION ENDONUCLEASE ANALYSIS
275
Serogrouping of isolates All isolates were grown and then serogrouped in agarose gel double immunodiffusion tests ( A G D P ) using extracted lipopolysaccharide (LPS), as previously described (Hampson et al., 1989; Hampson et al., 1990; Hampson, 1990). Restriction endonuclease analysis The bacteria were grown, their DNA extracted and restriction endonuclease analysis conducted as previously described (Combs et al., 1989). The DNA of the isolates was initially digested with Hindl 11 (Toyobo, Japan), to obtain the REA pattern. Where isolates with identical REA patterns were found the enzyme Hael 11 (Toyobo, Japan) was then used to see whether the isolates could be further differentiated. The enzymes EcoRV, Sau3A 1, Xba 1 (Toyobo, Japan), Rsa 1, Pvu 11 and A i u 1 (Boehringer Mannheim, Australia) were also found to be capable of digesting DNA from T. hyodysenteriae, and some of these enzymes were occasionally used for confirming the REA pattern designation of very similar isolates. Restriction endonucleases that would either not digest or only partially digest DNA from T. hyodysenteriae isolates were Dra 1 (Boehringer Mannheim, Australia), EcoR 1, BamH 1, Xho 1, Pst 1 and Notl (Toyobo, Japan). The REA pattern of each isolate was only designated after it had been compared with all patterns, and its DNA had been run on the same gel as DNA from isolates with apparently similar patterns. Distinct patterns were given letters of the alphabet. Some isolates had very minor differences in one or two DNA bands from those of other isolates. Where this occurred and was shown to be consistent on repeated analysis, the isolates were identified by the same letter but were given different numerals, eg L1 and L2. Each of these patterns was treated as indicating distinct but related strains of T. hyodysenteriae. Presence of extrachromosornal DNA and influence on REA pattern For each isolate 20/tl of undigested DNA solution was mixed with 5 #1 of loading buffer, and electrophoresed in a 10-cm long horizontal 0.7% agarose gel for 30 min at 75 v, with a 0.89 M Tris-borate (pH 8.0), 0.04 M EDTA electrophoresis buffer. The gel was stained with ethidium bromide and the presence of plasmid DNA bands which were distinct from chromosomal DNA and RNA were recorded. In order to determine whether plasmid DNA contributed significantly to the REA patterns observed, patterns were compared between isolates for the same piggery which either possessed plasmid DNA or did not (e.g. Vic 4 and Vic 5, and Vic 15 and Vic 16 ). Patterns obtained for isolates such as A 1 and WA2, both of which lost their plasmid band after in vitro culture were also compared before and after loss of the extrachromosomal DNA.
276
B.G. COMBS ET AL.
TABLEI Origin, serogroup, R E A patterns a n d presence o f p l a s m i d b a n d s in isolates of
Treponema hyodysenteriae
Isolate ~
Origin 2
Serogroup 3
R E A pattern
B234 B78 B204 B 169 AI MC52/80 KF9 WA 14 WA15 ~1 WA16 J WA27 WA28 WA 1 WA 2 WA4 WA 7 WA8 to WA 13 ~ WA26 WA3 WA 6 WA 5 SA3 SA4 SAI a n d SA2 ~ NSW 1 NSW2 NSW3 Vic39 Vic27 a n d Vic28 ~ V i c l 0 to Vic21 ~
USA USA USA Canada UK UK UK WA
A A
A B
B
C
+
C D E E
D E F G
+ + + +
A
J
+
WA
A N/T A A
J J J H4
+ + + +
B B B B B B
HI H1 HI HI H1 H3
+ + + + + +
E E E A A D
11 I1 I2 N5 N2 L8
+ + + +
I D B
H7
-
L9
+
H2
+
A A
NI N3
+ +
B
H5
+
Vic30 Vic31 Vic32 Vic33 Vic34 Vic35 Vic36 Vic37 Vic38 Vic4, }~ Vic5 & Vic6 Vic22 to V i c 2 C
Victoria Victorm Victoria Victoria Victoria Victoria Victoria Victoria Victoria Victoria
B B B
H8 H9 H8
+ + +
N/T
L5
+
B B B B B B D
L5 H1 H5 H5 M L6
+ + + + + -
Victoria
D
L6 L5
+ + (Vic 22, Vic 2 3 - )
WA WA WA WA WA WA WA WA WA WA WA WA SA WA SA NSW NSW NSW Victoria Victoria Victoria
Plasmid band(s) +
(Vicl6-)
TYPING BY SEROLOGY AND BY DNA RESTRICTION ENDONUCLEASEANALYSIS
277
Isolate ~
Origin 2
Serogroup 3
REA pattern
Plasmid band(s)
Vic3 Vic7, Vic8 and Vic9 ~ Vic29 Vicl Vic2 Q20 Q 10 Q 11 Q 17 Q22 Q21 Q25 Q26 Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q 12 and Q 13 ~ Q 15 Q 16 Q 19 Q23 Q24 QI8 Q 14
Victoria Victoria
D D
L6 L6
Victoria Victoria Victoria Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland
D F H A A B B G B B B D D D D D D D D D D D D D D E N/T G
L7 K H6 N4 H9 H9 H 10 H 11 L2 H 11 H8 L2 L2 L2 L2 L2 L2 L2 L3 L3 L4 L4 L1 L4 L2 O H10 H9
+ + (Vic 9 - ) + + + + + + + + + + + + + + +
~ J
'multiple samples from the same piggery. 2WA, Western Australia, SA = South Australia, NSW = New South Wales. 3N/T, Not typed.
RESULTS
The results of typing the T. hyodysenteriae isolates are presented in Table 1. Eight different serogroups were identified in Australia, with four of these not currently recorded elsewhere in the world. The most common group was B, present on 22 of the 62 properties (35.5%), followed by D (21 or 33.9%), A (10 or 16.1%), E (4 or 6.5%), G (2 or 3.2%), and F, H and I (1 each, or 1.6%). Three isolates were not typed. The Australian isolates were further divided into 31 different REA patterns, comprising eight major patterns (H to O ) with up to eleven minor vari-
278
B.G. COMBS ETAL.
1 2 3
4
56
il
78
9
i¸
(A) Fig. 1A. REA patterns of strains of Treponema hyodysenteriae which are representatives of the eight major REA patterns found in Australia. Lanes 1-9 respectively, Q24 (pattern O), Vic 39 (pattern N1 ), Vic 38 (pattern M), QI2 (pattern L4), Vicl (pattern K), WA14 (pattern J), WA3 (pattern I1 ), Vic2 (pattern H6), phage lambda (digested with Hindl 11 ).
ations on these, each consistent enough to identify a strain (Fig. 1A). None of these patterns were the same as those obtained with the seven strains which were not from Australia. The predominant major Australian REA patterns were H and L (Fig. 1B), each found on 24 different piggeries (77% of the total n u m b e r examined). These were both found in piggeries in Queensland, New South Wales and Victoria, whilst L was also present in South Australia and H was found in Western Australia. Pattern N was also found in Queensland, Victoria and South Australia. Between two and twelve isolates were obtained from 10 piggeries and in each case all isolates had the same REA pattern. Isolates Vic 10 to Vic 21 were obtained from one piggery over a five year period, and all had REA pattern
TYPING BY SEROLOGY AND BY DNA RESTRICTION ENDONUCLEASE ANALYSIS
279
Fig. 1B. REA patterns of selected strains of Treponema hyodysenteriae which are minor variants on two of the major Australian REA patterns H and L. Lanes 1-10 respectively, Vic31 (pattern H9), Vie30 (pattern H8), NSWI (pattern H7), Vicl9 (patterns H5), NSW2 (pattern L9), Vic4 (pattern L6 ), Q 12 ( pattern L4 ), Q 16 (pattern I~1 ), phage lambda (digested with Hind 111 ).
H5. Throughout Australia 13 of the 31 different REA patterns were found in more than one piggery, although only three patterns (H 1, H8 and H9) were found in more than one State. REA pattern L2 was found in nine different piggeries in Queensland, and H 1 in five piggeries in Western Australia. Overall, Victoria had the greatest number of different patterns ( 12 on 19 piggeries), whilst there were 10 REA patterns on 24 piggeries in Queensland and five on 13 piggeries in Western Australia. Isolates with the same serogroup could be divided into a number of different REA patterns. Generally, all isolates with the same REA pattern shared the same serogroup. However of three isolates with REA pattern L6 on one Victorian piggery, two (Vic 5 and Vic 6 ) were types as serogroup D and one (Vic 4) as serogroup B. Furthermore two isolates from a Queensland piggery
280
B.G. COMBSET AL.
1
2
3
4
5
6
7
8
9
10
Fig. 2. Agarose gel electrophoresis of DNA extracted from selected Australian isolates of Treponema hyodysenteriae with different REA patterns. Lanes 1-10 respectively, Vic28, SA I, Vic29, VicT, Vic26, Q10, Vic32, WA7, WA27, phage lambda (digested with Hindl 11 ). The arrow indicates the extrachromosomal (plasmid) DNA band.
(Q10 and Q11 ) both had REA pattern H9, but one was of serogroup B and the other of serogroup A. Two other epidemiologically linked isolates with REA pattern H9 also had different serogroups. Vic 31 (serogroup B) was from a piggery in Victoria and Q14 (group G) was from a piggery in Queensland which had supplied pigs to the other property. Furthermore, not all isolates within REA patterns L2, L5, L6 and H 11 on different piggeries had the same serogroup. Plasmid bands were seen in 70 of the 91 Australian isolates (77%), and in six of the seven non-Australian strains (86%). In most cases a single high molecular weight band (approximately corresponding to a 6.5 kb linear DNA fragment) was seen running in front of the chromosomal DNA in the agarose gels (Fig. 2), although three lower molecular weight bands were observed in isolate NSW 2. Eight isolates (A1, WA2, WA3, WA5, WA6, QI, Vicl and Vicl0) appeared to lose their plasmids after repeated in vitro subculture. Plasmid DNA did not contribute noticeably to the REA patterns of the isolates. DISCUSSION
Differentiation of bacterial isolates within species is an essential tool in understanding the epidemiology of a large number of bacterial diseases (Wachsmuth, 1986 ). In previous studies on T. hyodysenteriae, isolates were divided into seven serotypes based on the variation of antigenicity of lipopolysaccharides extracted from the cell surface by a hot-water-phenol method (Baum and Joens, 1979; Mapother and Joens, 1985 ). This system was later modified to divide the organisms into five serogroups, each containing distinct serovars or serotypes (Hampson et al., 1989). This system has recently been further extended to include nine serogroups, A to I, and it is likely that others exist (Hampson et al., 1990; Hampson, 1990). We were particularly
T Y P I N G BY SEROLOGY AND BY DNA RESTRICTION ENDONUCLEASE ANALYSIS
281
interested to determine the extent of antigenic diversity amongst Australian isolates of T. hyodysenteriae, since immunity against infection in a colonicloop model has been shown to be LPS-serotype specific (Joens et al., 1983), and experimental vaccines appear to confer the best protection when subsequent infection is with an homologous LPS-serotype of the bacteria (Parizek et al., 1985). The occurrence of eight serogroups of T. hyodysenteriae in Australia emphasises the extent of antigenic diversity amongst the LPS of the bacteria, and suggests that the prospects for developing universally applicable bacterin vaccines may not be good. Fortunately, isolates on 42 (67.7%) of the 62 piggeries were either serogroup B or D, and those on a further 10 (16.1%) piggeries were serogroup A. Should a multivalent bacterin vaccine be developed for use in Australia, these would be the three most appropriate serogroups for inclusion. In Western Australia serogroup E isolates were found on three of 13 piggeries (23%), and in one of 24 in Queensland (4.2%), so consideration should be given to including this serogroup in vaccines for use in those States. In a recent study Smith et al. ( 1990 ) used sodium dodecyl sulfate polyacrylamide gel electrophoresis and immunoblotting to examine outer membrane extracts from 18 Australian isolates of T. hyodysenteriae. Five isolates were classified as being of LPS "serotype" 1 (serogroup A of Hampson et al., 1989 ), six were "serotype" 2 (serogroup B) and the rest reacted weakly with serum raised against strain A1 (serogroup D), but more strongly with sera raised against Australian isolates 5380 and 70A. These latter group of isolates were suggested as representing a new "serotype" of T. hyodysenteriae, but are almost certainly the same as those classified as being of serogroup D in the present work. We have previously reported that Australian isolates of serogroup D are not typical in that they may share weak cross-reactivity with serogroup B (Hampson et al., 1990 ). Generally the results of Smith et al. (1990) are consistent with those presented here, except that in the present study examination of large numbers of isolates has revealed the presence of a greater diversity of serological types of T. hyodysenteriae in Australia. In the future it is likely that other genetic and serological types of the bacterial will be uncovered. Although serological typing provides important information about an isolate, antigenic groups do not necessarily reflect the scope of genetic diversity amongst bacterial species (Cleary et al., 1988; Kristiansen et al., 1984). For this reason we also used REA analysis, a DNA-based typing technique on our isolates to complement the serological typing. In a previous study we showed that REA analysis of T. hyodysenteriae was capable of distinguishing phenotypically and antigenically similar isolates from the USA, the UK, Canada and from Western Australia (Combs et al., 1989). In the present study we noted that the presence of plasmid DNA in many isolates, but fortunately the presence of this DNA was shown not to noticably alter or contribute to REA
282
B.G. COMBS ET AL.
patterns. Although it was known that T. hyodysenteriae may contain plasmids (Joens and Marquez, 1988 ), the widespread distribution of this DNA amongst isolates of T. hyodysenteriae observed here has apparently not previously been reported. The nature and effects of this DNA requires further investigation. The 91 isolates of T. hyodysenteriae obtained from Australia were divided into 31 different REA patterns, each distinct from the seven patterns observed amongst the non-Australian isolates. The REA patterns of some isolates initially appeared to be identical, but as resolution of banding patterns was improved, minor but consistent differences were observed. We interpret each of these distinct and repeatable patterns as representing a "strain" of the bacterium, and we have used them to trace the dissemination of strains between piggeries (Combs, B.G. and Hampson, D.J., unpublished studies). However those isolates sharing the same basic REA pattern, such as the L series (L1-L9) seem likely to represent a clone, all having originated from a single bacterium. This is supported, for example, by the close geographic clustering of isolates which are variants of the L pattern ( L I - L 4 ) in the Darling Down region of south east Queensland. Minor variants of the basic patterns seem likely to have arisen from mutational changes, although the biological significance of these are not clear. We therefore have apparently identified 8 clones and 31 individual strains of T. hyodysenteriae in Australia. Variants of the three basic REA patterns H, L and N constituted 85% of all the patterns found, and representatives of each were present in three or more States. However in only three cases were the identical strain found in more than one State, these being H8 and H9 found in Queensland and Victoria (two properties with H9 isolates being epidemiologically linked), and H1 found in Western Australia and also on a property in Victoria. Although there appears to be only limited dissemination of specific strains between States, the majority of strains in Australia appear to have been derived from only a few of the ancestral types. The latter were presumably introduced and widely dispersed at some time following European settlement of Australia. In each Australian State it was fairly common to find isolates from different properties sharing the same REA pattern ( 13 of the 31 patterns were found on more than one piggery). This was especially noticeable in Queensland and in Western Australia where nine and five piggeries respectively all had isolates with the same REA pattern. Dissemination of strains between piggeries in a local geographical area therefore appears to be relatively common. The REA typing technique was shown to be a stable marker of a strain, with for example a single pattern found amongst six isolates from different pigs in one outbreak, and another pattern found for 12 isolates obtained from one piggery over a five year period. The persistence of a single strain on this large property despite vigorous attempts to control and later eradicate it over this period is also of interest. It was surprising that where multiple isolates were taken from piggeries they
TYPING BY SEROLOGY AND BY DNA RESTRICTION ENDONUCLEASE ANALYSIS
283
all had the same REA type. Studies of other pathogenic bacteria infecting pigs have often shown the presence of multiple REA types on a particular piggery. For example Kakoyiannis et al. (1984) found 18 different REA types amongst 41 isolates of Campylobacter coli in one pig-rearing unit, and Smart et al. (1988 ) found several strains of Haemophilus parasuis in most piggeries that they examined, The present findings may result from only examining a relatively small number of repeat samples on a few piggeries, but could also reflect an ability of certain strains of T. hyodysenteriae to outgrow others and to dominate numerically in pigs which are potentially infected with more than one strain. REA was a more discriminatory typing technique than serogrouping, with, for example, 13 different REA patterns recognised amongst 39 isolates of serogroup B. Generally isolates with the same REA pattern belonged to the same serogroup. This suggests that the serogroup is a stable phenotype, despite possible environmental influences by the i m m u n e system in selecting antigenic variants in a population of T. hyodysenteriae cells infecting a pig. Exceptions were REA patterns L2, L5, L6, H9 and H 11 where different isolates of each had two associated serogroups. In the case of L2, L5 and L6 the serogroups were D and B, and it has already been reported that certain Australian isolates may show cross-reactivity between these two serogroups (Hampson et al., 1990). Two other isolates with the H9 REA pattern but with serogroups A and B were recovered from the same property, and were therefore likely to be of the same strain. Another pair of isolates with the H9 REA pattern but with different serogroups were from different properties, but pigs had been transfered between them. In this pair, although the property transfering pigs was infected with an isolate of the unusual serogroup G, clinical signs were not apparent. The piggery receiving the animals had an isolate of serogroup B recovered and did develop an outbreak of swine dysentery. In view of the epidemiological links in this case it appears that the REA pattern was identifying a single strain. Taken together these results suggest that individual strains of T. hyodysenteriae are capable of altering antigenic properties of their LPS components. If this can be confirmed it will be important in studies of both natural and vaccine-induced immunity to infection. Furthermore the correlation between apparent differences in antigenicity of LPS components and expression of virulence of the bacteria on the properties of origin deserves further investigation. The effects of T. hyodysenteriae LPS have previously been correlated with severity of disease in mouse models (Nuessen et al., 1983 ), and subcultured avirulent strains have been shown to have differences in LPS components from the original virulent strains (Halter and Joens, 1988). In conclusion, this study has demonstrated the presence of a great diversity of strains and antigenic types of T. hyodysenteriae in Australia. There is relatively little sharing of strains between piggeries widely separated over large
284
B.G. COMBS ET AL.
geographical areas, but, particularly in south east Queensland and in the south west of Western Australia, local clustering of strains occurs. Strains generally appear to be antigenically stable within an infected pig population, but possible exceptions have been observed. ACKNOWLEDGMENTS
This work was supported by Grants from the Australian Pig Research and Development Corporation and from Murdoch University. B.G.C, is in receipt of a Pig Research and Development Corporation Fellowship. Thanks are due to all those individuals and laboratories supplying isolates for use in the study.
REFERENCES Baum, D.H. and Joens, L.A., 1979. Serotypes of beta-haemolytic Treponema hyodysenteriae. Infect. Immun., 25: 792-796. Cleary, P.P., Kaplan, E.L. Livdahl, C. and Skyold, S., 1988. DNA fingerprints of Streptococcus pyogenes are M type specific. J. Infect. Dis., 158:1317-1323. Combs, B.G., Hampson, D.J., Mhoma, J.R.L. and Buddle, J.R., 1989. Typing of Treponema hyodysenteriae by restriction endonuclease analysis. Vet. Microbiol., 19:351-359. Cutler, R. and Gardner, I., 1988. A blue print for pig health research. Australian Pig Research Council, Canberra, Australia, pp. 48-50. Halter, M.R. and Joens, L.A., 1988. Lipooligosaccharides from Treponema hyodysenteriae and Treponema innocens. Infect. Immun., 56:3152-3156. Hampson, D.J., 1990. New serogroups of Treponema hyodysenteriae (G, H and 1). Vet. Rec., 127: 524. Hampson, D.J., Mhoma, J.R.L., Combs, B.G. and Buddle, J.R., 1989. Proposed revisions to the serological typing system for Treponema hyodysenteriae. Epidemiol. Infect., 102: 75-84. Hampson, D.J., Mhoma, J.R.L., Combs, B.G. and Lee, J.I., 1990. Serological grouping of Treponema hyodysenteriae. Epidemiol. Infect., 105: 79-86. Harris, D.L. and Glock, R.D., 1986. Swine dysentery and spirochaetal diarrhoea. In: A.D. Leman, R.D. Glock, W.L. Mengeling, R.H.C. Penny, E. Scholl and B. Straw (Editors), Diseases of Swine, 6th Edn. Iowa State University Press. Ames, Iowa., pp 494-507. Hunter, D. and Wood, T., 1979. An evaluation of the AP 1 ZYM system as a means of classifying spirochaetes associated with swine dysentery. Vet. Rec., 104: 383-384. Joens, L.A. and Marquez, R., 1988. The diagnosis of swine dysentery using a labelled nucleic acid probe. Proc. Int. Pig Vet. Soc. Congr., Rio de Janeiro, Brazil., p120. Joens, L.A., Whipp, S.C., Glock, R.D. and Nuessen, M.E., 1983. Serotype-specific protection against Treponema hyodysenteriae infection in ligated colonic loops of pigs recovered from swine dysentery. Infect. Immun., 39: 460-462. Kakoyiannis, C.K., Winter, P.J. and Marshall, R.B., 1984. Identification of Campylobacter coli isolates from animals and humans by bacterial restriction endonuclease DNA analysis. Appl. Environ. Microbiol., 48: 545-549. Kristiansen, B.E., Bjorvatn, B., Lund, V., Linquist, B. and Holten, E., 1984. Differentiation of B 15 strains ofNeisseria meningitidis by DNA restriction endonuclease fingerprinting. J. Infect. Dis., 150: 672-677.
TYPING BY SEROLOGY AND BY DNA RESTRICTION ENDONUCLEASE ANALYSIS
285
Lemcke, R.M. and Burrows, M.R., 1981. A comparative study of spirochaetes from the porcine alimentary tract. J. Hyg., Camb., 86: 173-182. Mapother, M.E. and Joens, L.A., 1985. New serotypes of Treponema hyodysenteriae. J. Clin. Microbiol., 22: 161-164. Mhoma, J.R.L., Hampson, D.J., and Robertson, I.D., 1992. A serological survey to establish the prevalence of infection with Treponema hyodysenteriae in Western Australia. Aust. Vet. J., (In press). Nuessen, M.E., Joens, L.A. and Glock, R.d., 1983. Involvement of lipopolysaccharides in the pathogenicity of Treponema hyodysenteriae. J. Immun., 131: 997-999. Parizek, R., Stewart, R., Brown, K., and Blevins, D., 1985. Protection against swine dysentery with an inactivated Treponema hyodysenteriae bacterin. Vet. Med., 80: 80-86. Roncalli, R.A. and Leaning, W.H.D., 1976. Geographical distribution of swine dysentery. Proc. Int. Pig Vet. Soc. Congr., Ames, IA, p. 417. Smart, N.L., Miniats, P. and MacInnes, J.I., 1988. Analysis of Haemophilus parasuis isolates from Southern Ontario swine by restriction endonclease fingerprinting. Can. J. Vet. Res, 52: 319-324. Smith, S.C., Roddick, F., Ling, Gerraty, N.L. and Coloe, P.J., 1990. Biochemical and immunochemical analysis of strains of Treponema hyodysenteriae. Vet Microbiol., 24:29-41. Wachsmuth, K., 1986. Molecular epidemiology of bacterial infections: examples of methodology and of investigations of outbreaks. Rev. Infect. Dis, 8: 682-692.
273
Typing of Australian isolates of Treponema hyodysenteriae by serology and by DNA restriction endonuclease analysis B.G. Combs, D.J. Hampson and S.J. Harders School of Veterinary Studies, Murdoch University, Murdoch, WA, Australia (Accepted 23 October 1991 )
ABSTRACT Combs, B.G., Hampson, D.J. and Harders, S.J., 1992. Typing of Australian isolates of Treponema hyodysenteriae by serology and by DNA restriction endonuclease analysis. Vet. Microbiol., 31:273285. A total of 91 isolates of Treponema hyodysenteriae which were obtained from 62 piggeries located around Australia were typed by serology and by DNA restriction endonuclease analysis (REA). The isolates fell into eight serogroups, of which groups B and D were the most common. Isolates with different REA patterns were recognised within serogroups, whilst a few isolates with the same REA pattern were placed into different serogroups. Some of the latter isolates were either from the same piggery or from farms with epidemiological links, thus indicating the bacteria may have altered their antigenic properties. A total of 31 different REA patterns were recognised amongst the Australian isolates. These comprised eight major patterns, with four of these being subdivided on the basis of minor differences in banding. Where a number of isolates were obtained from individual piggeries these all had the same REA pattern, and in one piggery isolates with the same pattern were recovered over a five year period. Plasmid bands were observed in 70 of the Australian isolates (77%), and in six of the seven overseas isolates included in the study for comparison. These plasmids did not affect the REA pattern. Of the States from which substantial numbers of isolates were examined, the greatest number of different strains ( 12 amongst 19 piggeries) were found from Victoria, and there were 10 REA patterns in strains from 24 piggeries in Queensland. Despite the large total number of different strains of T. hyodysenteriae in Australia, only three were found in more than one State.
INTRODUCTION
The spirochaete Treponerna hyodysenteriae is the primary aetiological agent of swine dysentery (SD), a major disease of pigs throughout the world (Roncalli and Leaning, 1976; Harris and Glock, 1986). A recent serological survey in Western Australia has shown that the infection is present in around one C o r r e s p o n d e n c e to: B.G. Combs, School of Veterinary Studies, M u r d o c h University, M u r d o c h , WA 6150, Australia.
0 3 7 8 - 1 1 3 5 / 9 2 / $ 0 5 . 0 0 © 1992 Elsevier Science Publishers B.V. All rights reserved.
274
B.G, COMBS ET AL.
third of herds ( M h o m a et al., 1992), and a similar prevalence has been reported in Victoria (Cutler and Gardner, 1988). Although SD is considered to be the single most economically important pig disease in Australia (Cutler and Gardner, 1988), little is known about many aspects of the epidemiology of the infection. Swine dysentery has not been easy to investigate: T. hyodysenteriae is relatively difficult both to isolate and to identify, and until recently there have been few methods available for differentiation within the bacterial species. Without the ability to identify individual strains, it has been difficult to trace outbreaks of infection, and to appreciate the diversity of strains within a geographical location or country. In the present study a large collection of Australian isolates of T. hyodysenteriae was assembled and typed using serological methods (Hampson et al., 1989) and DNA restriction endonuclease analysis (REA) (Combs et al., 1989). The purpose of this work was to gain essential background information about strain diversity and dissemination of T. hyodysenteriae for use in future work on the local epidemiology and control of the infection in Australia.
MATERIALS AND METHODS
Treponemes A total of 91 isolates of T. hyodysenteriae from 62 different piggeries located in five States of Australia were examined (Table 1 ). Western Australian isolates were either isolated in our laboratory or were received through Ms N. Buller of the Western Australian Department of Agriculture. Isolates from South Australia were obtained through Dr A. Pointon, South Australian Department of Agriculture; those from Victoria and New South Wales were obtained either through Dr R.T. Jones, Regional Veterinary Laboratory, Bendigo or through Dr P.J. Coloe, Royal Melbourne Institute of Technology: Queensland isolates were provided through Ms C. Stephens, Regional Veterinary Laboratory, Toowoomba or through Mr R. Thomas, Animal Research Institute, Yeerongpilly. All isolates which were studied originated from piggeries between 1986 and 1990. Strains B78, B234, B204, B169, A1, KF9 and M C 5 2 / 8 0 from the USA, Canada and the U K were included in the study for comparison. These were provided through Dr R.J. Lysons, Institute for Animal Health, Compton, England, and have previously been typed by us (Combs et al., 1989; Hampson et al., 1989). All isolates used in the study were confirmed by us as being T. hyodysenteriae on the basis of their strong beta-haemolysis, production of indole, and a characteristic enzyme profile using the API-ZYM system (Lemcke and Burrows, 1981 ; Hunter and Wood, 1979).
T Y P I N G BY SEROLOGY AND BY DNA RESTRICTION ENDONUCLEASE ANALYSIS
275
Serogrouping of isolates All isolates were grown and then serogrouped in agarose gel double immunodiffusion tests ( A G D P ) using extracted lipopolysaccharide (LPS), as previously described (Hampson et al., 1989; Hampson et al., 1990; Hampson, 1990). Restriction endonuclease analysis The bacteria were grown, their DNA extracted and restriction endonuclease analysis conducted as previously described (Combs et al., 1989). The DNA of the isolates was initially digested with Hindl 11 (Toyobo, Japan), to obtain the REA pattern. Where isolates with identical REA patterns were found the enzyme Hael 11 (Toyobo, Japan) was then used to see whether the isolates could be further differentiated. The enzymes EcoRV, Sau3A 1, Xba 1 (Toyobo, Japan), Rsa 1, Pvu 11 and A i u 1 (Boehringer Mannheim, Australia) were also found to be capable of digesting DNA from T. hyodysenteriae, and some of these enzymes were occasionally used for confirming the REA pattern designation of very similar isolates. Restriction endonucleases that would either not digest or only partially digest DNA from T. hyodysenteriae isolates were Dra 1 (Boehringer Mannheim, Australia), EcoR 1, BamH 1, Xho 1, Pst 1 and Notl (Toyobo, Japan). The REA pattern of each isolate was only designated after it had been compared with all patterns, and its DNA had been run on the same gel as DNA from isolates with apparently similar patterns. Distinct patterns were given letters of the alphabet. Some isolates had very minor differences in one or two DNA bands from those of other isolates. Where this occurred and was shown to be consistent on repeated analysis, the isolates were identified by the same letter but were given different numerals, eg L1 and L2. Each of these patterns was treated as indicating distinct but related strains of T. hyodysenteriae. Presence of extrachromosornal DNA and influence on REA pattern For each isolate 20/tl of undigested DNA solution was mixed with 5 #1 of loading buffer, and electrophoresed in a 10-cm long horizontal 0.7% agarose gel for 30 min at 75 v, with a 0.89 M Tris-borate (pH 8.0), 0.04 M EDTA electrophoresis buffer. The gel was stained with ethidium bromide and the presence of plasmid DNA bands which were distinct from chromosomal DNA and RNA were recorded. In order to determine whether plasmid DNA contributed significantly to the REA patterns observed, patterns were compared between isolates for the same piggery which either possessed plasmid DNA or did not (e.g. Vic 4 and Vic 5, and Vic 15 and Vic 16 ). Patterns obtained for isolates such as A 1 and WA2, both of which lost their plasmid band after in vitro culture were also compared before and after loss of the extrachromosomal DNA.
276
B.G. COMBS ET AL.
TABLEI Origin, serogroup, R E A patterns a n d presence o f p l a s m i d b a n d s in isolates of
Treponema hyodysenteriae
Isolate ~
Origin 2
Serogroup 3
R E A pattern
B234 B78 B204 B 169 AI MC52/80 KF9 WA 14 WA15 ~1 WA16 J WA27 WA28 WA 1 WA 2 WA4 WA 7 WA8 to WA 13 ~ WA26 WA3 WA 6 WA 5 SA3 SA4 SAI a n d SA2 ~ NSW 1 NSW2 NSW3 Vic39 Vic27 a n d Vic28 ~ V i c l 0 to Vic21 ~
USA USA USA Canada UK UK UK WA
A A
A B
B
C
+
C D E E
D E F G
+ + + +
A
J
+
WA
A N/T A A
J J J H4
+ + + +
B B B B B B
HI H1 HI HI H1 H3
+ + + + + +
E E E A A D
11 I1 I2 N5 N2 L8
+ + + +
I D B
H7
-
L9
+
H2
+
A A
NI N3
+ +
B
H5
+
Vic30 Vic31 Vic32 Vic33 Vic34 Vic35 Vic36 Vic37 Vic38 Vic4, }~ Vic5 & Vic6 Vic22 to V i c 2 C
Victoria Victorm Victoria Victoria Victoria Victoria Victoria Victoria Victoria Victoria
B B B
H8 H9 H8
+ + +
N/T
L5
+
B B B B B B D
L5 H1 H5 H5 M L6
+ + + + + -
Victoria
D
L6 L5
+ + (Vic 22, Vic 2 3 - )
WA WA WA WA WA WA WA WA WA WA WA WA SA WA SA NSW NSW NSW Victoria Victoria Victoria
Plasmid band(s) +
(Vicl6-)
TYPING BY SEROLOGY AND BY DNA RESTRICTION ENDONUCLEASEANALYSIS
277
Isolate ~
Origin 2
Serogroup 3
REA pattern
Plasmid band(s)
Vic3 Vic7, Vic8 and Vic9 ~ Vic29 Vicl Vic2 Q20 Q 10 Q 11 Q 17 Q22 Q21 Q25 Q26 Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q 12 and Q 13 ~ Q 15 Q 16 Q 19 Q23 Q24 QI8 Q 14
Victoria Victoria
D D
L6 L6
Victoria Victoria Victoria Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland Queensland
D F H A A B B G B B B D D D D D D D D D D D D D D E N/T G
L7 K H6 N4 H9 H9 H 10 H 11 L2 H 11 H8 L2 L2 L2 L2 L2 L2 L2 L3 L3 L4 L4 L1 L4 L2 O H10 H9
+ + (Vic 9 - ) + + + + + + + + + + + + + + +
~ J
'multiple samples from the same piggery. 2WA, Western Australia, SA = South Australia, NSW = New South Wales. 3N/T, Not typed.
RESULTS
The results of typing the T. hyodysenteriae isolates are presented in Table 1. Eight different serogroups were identified in Australia, with four of these not currently recorded elsewhere in the world. The most common group was B, present on 22 of the 62 properties (35.5%), followed by D (21 or 33.9%), A (10 or 16.1%), E (4 or 6.5%), G (2 or 3.2%), and F, H and I (1 each, or 1.6%). Three isolates were not typed. The Australian isolates were further divided into 31 different REA patterns, comprising eight major patterns (H to O ) with up to eleven minor vari-
278
B.G. COMBS ETAL.
1 2 3
4
56
il
78
9
i¸
(A) Fig. 1A. REA patterns of strains of Treponema hyodysenteriae which are representatives of the eight major REA patterns found in Australia. Lanes 1-9 respectively, Q24 (pattern O), Vic 39 (pattern N1 ), Vic 38 (pattern M), QI2 (pattern L4), Vicl (pattern K), WA14 (pattern J), WA3 (pattern I1 ), Vic2 (pattern H6), phage lambda (digested with Hindl 11 ).
ations on these, each consistent enough to identify a strain (Fig. 1A). None of these patterns were the same as those obtained with the seven strains which were not from Australia. The predominant major Australian REA patterns were H and L (Fig. 1B), each found on 24 different piggeries (77% of the total n u m b e r examined). These were both found in piggeries in Queensland, New South Wales and Victoria, whilst L was also present in South Australia and H was found in Western Australia. Pattern N was also found in Queensland, Victoria and South Australia. Between two and twelve isolates were obtained from 10 piggeries and in each case all isolates had the same REA pattern. Isolates Vic 10 to Vic 21 were obtained from one piggery over a five year period, and all had REA pattern
TYPING BY SEROLOGY AND BY DNA RESTRICTION ENDONUCLEASE ANALYSIS
279
Fig. 1B. REA patterns of selected strains of Treponema hyodysenteriae which are minor variants on two of the major Australian REA patterns H and L. Lanes 1-10 respectively, Vic31 (pattern H9), Vie30 (pattern H8), NSWI (pattern H7), Vicl9 (patterns H5), NSW2 (pattern L9), Vic4 (pattern L6 ), Q 12 ( pattern L4 ), Q 16 (pattern I~1 ), phage lambda (digested with Hind 111 ).
H5. Throughout Australia 13 of the 31 different REA patterns were found in more than one piggery, although only three patterns (H 1, H8 and H9) were found in more than one State. REA pattern L2 was found in nine different piggeries in Queensland, and H 1 in five piggeries in Western Australia. Overall, Victoria had the greatest number of different patterns ( 12 on 19 piggeries), whilst there were 10 REA patterns on 24 piggeries in Queensland and five on 13 piggeries in Western Australia. Isolates with the same serogroup could be divided into a number of different REA patterns. Generally, all isolates with the same REA pattern shared the same serogroup. However of three isolates with REA pattern L6 on one Victorian piggery, two (Vic 5 and Vic 6 ) were types as serogroup D and one (Vic 4) as serogroup B. Furthermore two isolates from a Queensland piggery
280
B.G. COMBSET AL.
1
2
3
4
5
6
7
8
9
10
Fig. 2. Agarose gel electrophoresis of DNA extracted from selected Australian isolates of Treponema hyodysenteriae with different REA patterns. Lanes 1-10 respectively, Vic28, SA I, Vic29, VicT, Vic26, Q10, Vic32, WA7, WA27, phage lambda (digested with Hindl 11 ). The arrow indicates the extrachromosomal (plasmid) DNA band.
(Q10 and Q11 ) both had REA pattern H9, but one was of serogroup B and the other of serogroup A. Two other epidemiologically linked isolates with REA pattern H9 also had different serogroups. Vic 31 (serogroup B) was from a piggery in Victoria and Q14 (group G) was from a piggery in Queensland which had supplied pigs to the other property. Furthermore, not all isolates within REA patterns L2, L5, L6 and H 11 on different piggeries had the same serogroup. Plasmid bands were seen in 70 of the 91 Australian isolates (77%), and in six of the seven non-Australian strains (86%). In most cases a single high molecular weight band (approximately corresponding to a 6.5 kb linear DNA fragment) was seen running in front of the chromosomal DNA in the agarose gels (Fig. 2), although three lower molecular weight bands were observed in isolate NSW 2. Eight isolates (A1, WA2, WA3, WA5, WA6, QI, Vicl and Vicl0) appeared to lose their plasmids after repeated in vitro subculture. Plasmid DNA did not contribute noticeably to the REA patterns of the isolates. DISCUSSION
Differentiation of bacterial isolates within species is an essential tool in understanding the epidemiology of a large number of bacterial diseases (Wachsmuth, 1986 ). In previous studies on T. hyodysenteriae, isolates were divided into seven serotypes based on the variation of antigenicity of lipopolysaccharides extracted from the cell surface by a hot-water-phenol method (Baum and Joens, 1979; Mapother and Joens, 1985 ). This system was later modified to divide the organisms into five serogroups, each containing distinct serovars or serotypes (Hampson et al., 1989). This system has recently been further extended to include nine serogroups, A to I, and it is likely that others exist (Hampson et al., 1990; Hampson, 1990). We were particularly
T Y P I N G BY SEROLOGY AND BY DNA RESTRICTION ENDONUCLEASE ANALYSIS
281
interested to determine the extent of antigenic diversity amongst Australian isolates of T. hyodysenteriae, since immunity against infection in a colonicloop model has been shown to be LPS-serotype specific (Joens et al., 1983), and experimental vaccines appear to confer the best protection when subsequent infection is with an homologous LPS-serotype of the bacteria (Parizek et al., 1985). The occurrence of eight serogroups of T. hyodysenteriae in Australia emphasises the extent of antigenic diversity amongst the LPS of the bacteria, and suggests that the prospects for developing universally applicable bacterin vaccines may not be good. Fortunately, isolates on 42 (67.7%) of the 62 piggeries were either serogroup B or D, and those on a further 10 (16.1%) piggeries were serogroup A. Should a multivalent bacterin vaccine be developed for use in Australia, these would be the three most appropriate serogroups for inclusion. In Western Australia serogroup E isolates were found on three of 13 piggeries (23%), and in one of 24 in Queensland (4.2%), so consideration should be given to including this serogroup in vaccines for use in those States. In a recent study Smith et al. ( 1990 ) used sodium dodecyl sulfate polyacrylamide gel electrophoresis and immunoblotting to examine outer membrane extracts from 18 Australian isolates of T. hyodysenteriae. Five isolates were classified as being of LPS "serotype" 1 (serogroup A of Hampson et al., 1989 ), six were "serotype" 2 (serogroup B) and the rest reacted weakly with serum raised against strain A1 (serogroup D), but more strongly with sera raised against Australian isolates 5380 and 70A. These latter group of isolates were suggested as representing a new "serotype" of T. hyodysenteriae, but are almost certainly the same as those classified as being of serogroup D in the present work. We have previously reported that Australian isolates of serogroup D are not typical in that they may share weak cross-reactivity with serogroup B (Hampson et al., 1990 ). Generally the results of Smith et al. (1990) are consistent with those presented here, except that in the present study examination of large numbers of isolates has revealed the presence of a greater diversity of serological types of T. hyodysenteriae in Australia. In the future it is likely that other genetic and serological types of the bacterial will be uncovered. Although serological typing provides important information about an isolate, antigenic groups do not necessarily reflect the scope of genetic diversity amongst bacterial species (Cleary et al., 1988; Kristiansen et al., 1984). For this reason we also used REA analysis, a DNA-based typing technique on our isolates to complement the serological typing. In a previous study we showed that REA analysis of T. hyodysenteriae was capable of distinguishing phenotypically and antigenically similar isolates from the USA, the UK, Canada and from Western Australia (Combs et al., 1989). In the present study we noted that the presence of plasmid DNA in many isolates, but fortunately the presence of this DNA was shown not to noticably alter or contribute to REA
282
B.G. COMBS ET AL.
patterns. Although it was known that T. hyodysenteriae may contain plasmids (Joens and Marquez, 1988 ), the widespread distribution of this DNA amongst isolates of T. hyodysenteriae observed here has apparently not previously been reported. The nature and effects of this DNA requires further investigation. The 91 isolates of T. hyodysenteriae obtained from Australia were divided into 31 different REA patterns, each distinct from the seven patterns observed amongst the non-Australian isolates. The REA patterns of some isolates initially appeared to be identical, but as resolution of banding patterns was improved, minor but consistent differences were observed. We interpret each of these distinct and repeatable patterns as representing a "strain" of the bacterium, and we have used them to trace the dissemination of strains between piggeries (Combs, B.G. and Hampson, D.J., unpublished studies). However those isolates sharing the same basic REA pattern, such as the L series (L1-L9) seem likely to represent a clone, all having originated from a single bacterium. This is supported, for example, by the close geographic clustering of isolates which are variants of the L pattern ( L I - L 4 ) in the Darling Down region of south east Queensland. Minor variants of the basic patterns seem likely to have arisen from mutational changes, although the biological significance of these are not clear. We therefore have apparently identified 8 clones and 31 individual strains of T. hyodysenteriae in Australia. Variants of the three basic REA patterns H, L and N constituted 85% of all the patterns found, and representatives of each were present in three or more States. However in only three cases were the identical strain found in more than one State, these being H8 and H9 found in Queensland and Victoria (two properties with H9 isolates being epidemiologically linked), and H1 found in Western Australia and also on a property in Victoria. Although there appears to be only limited dissemination of specific strains between States, the majority of strains in Australia appear to have been derived from only a few of the ancestral types. The latter were presumably introduced and widely dispersed at some time following European settlement of Australia. In each Australian State it was fairly common to find isolates from different properties sharing the same REA pattern ( 13 of the 31 patterns were found on more than one piggery). This was especially noticeable in Queensland and in Western Australia where nine and five piggeries respectively all had isolates with the same REA pattern. Dissemination of strains between piggeries in a local geographical area therefore appears to be relatively common. The REA typing technique was shown to be a stable marker of a strain, with for example a single pattern found amongst six isolates from different pigs in one outbreak, and another pattern found for 12 isolates obtained from one piggery over a five year period. The persistence of a single strain on this large property despite vigorous attempts to control and later eradicate it over this period is also of interest. It was surprising that where multiple isolates were taken from piggeries they
TYPING BY SEROLOGY AND BY DNA RESTRICTION ENDONUCLEASE ANALYSIS
283
all had the same REA type. Studies of other pathogenic bacteria infecting pigs have often shown the presence of multiple REA types on a particular piggery. For example Kakoyiannis et al. (1984) found 18 different REA types amongst 41 isolates of Campylobacter coli in one pig-rearing unit, and Smart et al. (1988 ) found several strains of Haemophilus parasuis in most piggeries that they examined, The present findings may result from only examining a relatively small number of repeat samples on a few piggeries, but could also reflect an ability of certain strains of T. hyodysenteriae to outgrow others and to dominate numerically in pigs which are potentially infected with more than one strain. REA was a more discriminatory typing technique than serogrouping, with, for example, 13 different REA patterns recognised amongst 39 isolates of serogroup B. Generally isolates with the same REA pattern belonged to the same serogroup. This suggests that the serogroup is a stable phenotype, despite possible environmental influences by the i m m u n e system in selecting antigenic variants in a population of T. hyodysenteriae cells infecting a pig. Exceptions were REA patterns L2, L5, L6, H9 and H 11 where different isolates of each had two associated serogroups. In the case of L2, L5 and L6 the serogroups were D and B, and it has already been reported that certain Australian isolates may show cross-reactivity between these two serogroups (Hampson et al., 1990). Two other isolates with the H9 REA pattern but with serogroups A and B were recovered from the same property, and were therefore likely to be of the same strain. Another pair of isolates with the H9 REA pattern but with different serogroups were from different properties, but pigs had been transfered between them. In this pair, although the property transfering pigs was infected with an isolate of the unusual serogroup G, clinical signs were not apparent. The piggery receiving the animals had an isolate of serogroup B recovered and did develop an outbreak of swine dysentery. In view of the epidemiological links in this case it appears that the REA pattern was identifying a single strain. Taken together these results suggest that individual strains of T. hyodysenteriae are capable of altering antigenic properties of their LPS components. If this can be confirmed it will be important in studies of both natural and vaccine-induced immunity to infection. Furthermore the correlation between apparent differences in antigenicity of LPS components and expression of virulence of the bacteria on the properties of origin deserves further investigation. The effects of T. hyodysenteriae LPS have previously been correlated with severity of disease in mouse models (Nuessen et al., 1983 ), and subcultured avirulent strains have been shown to have differences in LPS components from the original virulent strains (Halter and Joens, 1988). In conclusion, this study has demonstrated the presence of a great diversity of strains and antigenic types of T. hyodysenteriae in Australia. There is relatively little sharing of strains between piggeries widely separated over large
284
B.G. COMBS ET AL.
geographical areas, but, particularly in south east Queensland and in the south west of Western Australia, local clustering of strains occurs. Strains generally appear to be antigenically stable within an infected pig population, but possible exceptions have been observed. ACKNOWLEDGMENTS
This work was supported by Grants from the Australian Pig Research and Development Corporation and from Murdoch University. B.G.C, is in receipt of a Pig Research and Development Corporation Fellowship. Thanks are due to all those individuals and laboratories supplying isolates for use in the study.
REFERENCES Baum, D.H. and Joens, L.A., 1979. Serotypes of beta-haemolytic Treponema hyodysenteriae. Infect. Immun., 25: 792-796. Cleary, P.P., Kaplan, E.L. Livdahl, C. and Skyold, S., 1988. DNA fingerprints of Streptococcus pyogenes are M type specific. J. Infect. Dis., 158:1317-1323. Combs, B.G., Hampson, D.J., Mhoma, J.R.L. and Buddle, J.R., 1989. Typing of Treponema hyodysenteriae by restriction endonuclease analysis. Vet. Microbiol., 19:351-359. Cutler, R. and Gardner, I., 1988. A blue print for pig health research. Australian Pig Research Council, Canberra, Australia, pp. 48-50. Halter, M.R. and Joens, L.A., 1988. Lipooligosaccharides from Treponema hyodysenteriae and Treponema innocens. Infect. Immun., 56:3152-3156. Hampson, D.J., 1990. New serogroups of Treponema hyodysenteriae (G, H and 1). Vet. Rec., 127: 524. Hampson, D.J., Mhoma, J.R.L., Combs, B.G. and Buddle, J.R., 1989. Proposed revisions to the serological typing system for Treponema hyodysenteriae. Epidemiol. Infect., 102: 75-84. Hampson, D.J., Mhoma, J.R.L., Combs, B.G. and Lee, J.I., 1990. Serological grouping of Treponema hyodysenteriae. Epidemiol. Infect., 105: 79-86. Harris, D.L. and Glock, R.D., 1986. Swine dysentery and spirochaetal diarrhoea. In: A.D. Leman, R.D. Glock, W.L. Mengeling, R.H.C. Penny, E. Scholl and B. Straw (Editors), Diseases of Swine, 6th Edn. Iowa State University Press. Ames, Iowa., pp 494-507. Hunter, D. and Wood, T., 1979. An evaluation of the AP 1 ZYM system as a means of classifying spirochaetes associated with swine dysentery. Vet. Rec., 104: 383-384. Joens, L.A. and Marquez, R., 1988. The diagnosis of swine dysentery using a labelled nucleic acid probe. Proc. Int. Pig Vet. Soc. Congr., Rio de Janeiro, Brazil., p120. Joens, L.A., Whipp, S.C., Glock, R.D. and Nuessen, M.E., 1983. Serotype-specific protection against Treponema hyodysenteriae infection in ligated colonic loops of pigs recovered from swine dysentery. Infect. Immun., 39: 460-462. Kakoyiannis, C.K., Winter, P.J. and Marshall, R.B., 1984. Identification of Campylobacter coli isolates from animals and humans by bacterial restriction endonuclease DNA analysis. Appl. Environ. Microbiol., 48: 545-549. Kristiansen, B.E., Bjorvatn, B., Lund, V., Linquist, B. and Holten, E., 1984. Differentiation of B 15 strains ofNeisseria meningitidis by DNA restriction endonuclease fingerprinting. J. Infect. Dis., 150: 672-677.
TYPING BY SEROLOGY AND BY DNA RESTRICTION ENDONUCLEASE ANALYSIS
285
Lemcke, R.M. and Burrows, M.R., 1981. A comparative study of spirochaetes from the porcine alimentary tract. J. Hyg., Camb., 86: 173-182. Mapother, M.E. and Joens, L.A., 1985. New serotypes of Treponema hyodysenteriae. J. Clin. Microbiol., 22: 161-164. Mhoma, J.R.L., Hampson, D.J., and Robertson, I.D., 1992. A serological survey to establish the prevalence of infection with Treponema hyodysenteriae in Western Australia. Aust. Vet. J., (In press). Nuessen, M.E., Joens, L.A. and Glock, R.d., 1983. Involvement of lipopolysaccharides in the pathogenicity of Treponema hyodysenteriae. J. Immun., 131: 997-999. Parizek, R., Stewart, R., Brown, K., and Blevins, D., 1985. Protection against swine dysentery with an inactivated Treponema hyodysenteriae bacterin. Vet. Med., 80: 80-86. Roncalli, R.A. and Leaning, W.H.D., 1976. Geographical distribution of swine dysentery. Proc. Int. Pig Vet. Soc. Congr., Ames, IA, p. 417. Smart, N.L., Miniats, P. and MacInnes, J.I., 1988. Analysis of Haemophilus parasuis isolates from Southern Ontario swine by restriction endonclease fingerprinting. Can. J. Vet. Res, 52: 319-324. Smith, S.C., Roddick, F., Ling, Gerraty, N.L. and Coloe, P.J., 1990. Biochemical and immunochemical analysis of strains of Treponema hyodysenteriae. Vet Microbiol., 24:29-41. Wachsmuth, K., 1986. Molecular epidemiology of bacterial infections: examples of methodology and of investigations of outbreaks. Rev. Infect. Dis, 8: 682-692.
