Eur. J. Epidemiol. 0392-2990
EUROPEAN JOURNAL
January 1992, p. 15-21
Vol. 8, No. I
OF EPIDEMIOLOGY
BIOTYPES AND DNA FINGERPRINTS OF CYTOTOXIGENIC HELICOBACTER PYLORI FROM PATIENTS WITH GASTRITIS AND PEPTIC ULCERATION IN ITALY R.J. O W E N *l, N. F I G U R A * * and M. M O R E N O * *National Collection o f Type Cultures - Central Public Health Laboratory - London N W 9 5 H T - UK. **Institute di Malattie Injettive - University o f Siena 53100 - Italy.
Key words: Biotypes - DNA fingerprints - Helicobacter pylori - Cytotoxins Variations in biotypic and DNA characteristics of 21 strains o f Helicobacterpylori from patients in central Italy with histologically defined gastritis and/or peptic ulcers were studied. The strains had the same preformed enzyme biotype but differed in motility and cytotoxigenic activity. The presence or absence of these two characteristics were closely associated in 73% of strains and were used to define three phenotypic subsets. Strains of subset 1 (Mot + Tox +) were most common (52%) particularly amongst patients with peptic ulceration (64~ H. pylori had unique HaelII and HindlII ribopatterns but no one ribopattern or single band within a ribopattern was characteristic of any particular phenotypic subset. H, pylori from patients with gastritis were genomical]y as heterogeneous as those from patients with ulcers. Plasmid DNA was detected in four strains (25~ and although three of these were of the same biotype (phenotypic subset 1), there was no general association apparent between plasmid presence and cytotoxic activity. It was concluded that motility was a useful additional feature for biotyping of H. pylori.
complex and appears to be strain specific (15). In contrast, there is evidence that ribosomal (r) RNA gene patterns (ribopatterns) provide a reliable and simpler basis for strain discrimination (18). Other techniques applied to characterise strains of H. pylori with varying degrees of success include biotyping (23, 32), cytotoxigenicity (5, 9, 11, 14), haemagglutination (35), coagglutination with rabbit antibodies (7), lectin typing (1), immunoblotting (3, 4), plasmid profiling (15, 27, 34) and electrophoretic protein pattern profiling (21). The aim of the present study was to determine if strains of H. pylori defined by biotype, cytotoxin production and disease association had characteristic molecular features based on ribopatterns and plasmid DNA content. Because cytotoxin production is a putative virulence determinant (17, 30), a specific objective of the study was to investigate associations with that feature.
INTRODUCTION
Helicobacter pylori (Campylobacter pylori) is strongly associated worldwide with chronic active type-B gastritis and higher duodenal ulcer relapse rates, and is possibly one of the causes of non-ulcer dyspepsia (16). A number of epidemiological features of H. pylori, such as its incidence and prevalence in population groups, are well documented (26) but the detailed surveillance of individual strains of H. pylori is hindered by the lack of reliable and accurate typing methods. To date, the most precise and widely used method of identifying strains has been DNA restriction endonuclease digest analysis (2, 13, 15, 18, 19, 22, 32, 33), although the technique has limitations as a basis for typing because each DNA pattern is J Corresponding author.
15
Owen R.J. et al.
Eur. J. Epidemiol.
Enzyme biotyping
MATERIALS AND METHODS
Enzymic biotype was determined from preformed enzyme production in API Zym kits as described previously (23).
The 21 clinical strains of H. pylori were from patients undergoing diagnostic gastroduodenal endoscopy at the Ospedale di Grosseto, Grosseto, Italy. Biopsies were obtained from the gastric antrum and the edges of ulcers (when present). Table 1 lists the categories of patients according to symptoms and the degree of gastritis in each category. The severity of gastritis was assessed by the degree of round and polymorphonuclear cell infiltration and the presence of erosions. Methods of isolation and primary identification of strains were described previously (9). Reference strain NCTC 11637 (= RPH 13487) from Perth (Australia) was obtained from the National Collection of Type Cultures (NCTC). All isolates were grown on Oxoid brain-heart infusion agar (BHI) containing 5~ (v/v) horse blood and supplemented with 1% Isovitalex (BBL Microbiology Systems, Becton Dickinson, Cowley, Oxford, UK). Cultures were incubated for 48h at 37~ under microaerobic conditions (5%, 02; 5%, CO2; 2%, H2; 88%, Nz) in a Variable Atmosphere Incubator (Don Whitley Scientific Ltd., Shipley, Yorks). For transportation, strains were maintained in blood agar (25), and stocks of all cultures were preserved in 10% v/v glycerol in Oxoid nutrient broth over liquid nitrogen.
Cytotoxin testing Cytotoxin production tests against mammalian cells (CHO and Vero cells) in vitro were performed as described previously (9). Cells were examined after 24 and 48 h of incubation for the presence of the typical toxic effect (intracytoplasmic vacuolization) at one or more dilutions. DNA isolation, restriction digestion, and ribosomal RNA gene probe hybridization Chromosomal DNA was isolated and purified using the guanidium thiocyanate reagent method (29). All DNA samples (5 lag) were digested for 4 h at 37~C with HaelIl and HindlII (2-3U/lag of DNA) according to the manufacturer's instructions (Northumbria Biochemicals Ltd., Cramlington, Yorkshire). The digested DNA was electrophoresed at 25 V for 16 h in a horizontal 0.70/0 (w/v) agarose gel in a buffer containing 89 mM Tris hydrochloride, 89 mM boric acid, and 2 mM disodium ethylenediaminetetra-acetic acid (EDTA) (pH 8.3). After electrophoresis, the DNA fragments were transferred to Hybond-N nylon membranes (Amersham International) by vacuum transfer blotting. The membranes were then hybridized by standard procedures for 18 h at 42~ C with a biotinylated cDNA probe prepared from a 16S and 23S rRNA mixture ofH.pylori NCTC 11638 using Moloney mouse leukaemia virus reverse transcriptase (Gibco-BRL) (28). Biotinylation was achieved by the incorporation of biotin-16-dUTP. The membranes were washed after hybridization and the hybridized probe detected colorimetrically using a nonradioactive detection kit-BIuGENE (Gibco-BRL). The pattern of bands obtained was designated the ribopattern. Full details of the methods of hybridization and band size estimation are given elsewhere (20).
TABLE 1. - Italian strains of H. pylofi used grouped according to patient symptoms. Patient group
No. of isolates (n = 21)
Gastritis (G) moderate (Gm) severe (Gs) erosive (Ge)
2 3 5
Duodenal ulcer (DU) moderate gastritis severe gastritis
2 4
Gastric ulcer (GU) erosive gastritis
4
Duodenal and gastric ulcer (DU and GU) severe gastritis
1
Plasmid DNA extraction and electrophoresis For plasmid DNA extraction, organisms were cultured as described above. Single colonies were picked and plasmid DNA was prepared according to the method of Kado and Liu (12) as modified previously (24). Samples were electrophoresed in a horizontal 0.7% agarose gel for 2 h at 100 V. After staining With ethidium bromide, the gels were viewed on a UV transilluminator and photographed.
Motility testing Isolates were grown on slopes of BHI blood agar with added Isovitalex as described above. After at least two subcultures at 3-4d intervals, the liquid phase culture at the base of the slope of a 2d-old culture was removed and bacterial motility was determined on a microscope slide by standard methods (6). Cultures that were weakly (sluggishly) motile or non-motile were recultured and retested for motility.
RESULTS
16
Enzyme biotypes All the strains ofH. pylofi were enzyme biotype II. They produced leucine arylamidase and
Biotypes and DNA fingerprints of H. pylori
Voh 8, 1992
p h o s p h o h y d r o l a s e b u t n o t esterases C4 and C8 in the API Z y m galleries.
strain was originally isolated is also indicated, b u t n o further distinctions based o n the degree o f gastritis were made. Subset 1 (Mot ~, Tox*) comprised eleven (52%) Italian strains a n d the reference strain ( N C T C 11637) which were actively motile. T h e r e were visible differences b e t w e e n s o m e strains in the degree o f motility, which was expressed in terms o f the p r o p o r t i o n of cells actively moving. Six strains
Motility and cytotoxic activity T h e results o n motility a n d cytotoxic activity are given in T a b l e 2, in which strains are g r o u p e d into three subsets o n the basis o f these two features. T h e s y m p t o m category o f the patient from w h o m the
T A B L E 2. - Biotypic a n d D N A characteristics o f H. pylori from Italy.
Strain no.
Subset 1
Diagnosis group,
Motility
Cytotoxin productionb CHO
Vero
Plasmid present
Genome DNA cut by HaellI
(Mot + Tox §
Gll
DU
(+)
1/2
-
-
-
G27
Ge
+
1/3
-
+
+
G29
Gs
(+)
1/10
I/I0
-
+
G33
Ge
+
1/10
1/10
-
+
G39
DU
+
1/3
1/5
-
+
G89
GU
+
1/3
1/3
-
-
-
-
G103
Gm
(+)
1/5
-
G105
DU+GU
+
1/3
1/5
+
+
G106
DU
(+)
1/2
-
-
+
G126
GU
+
1/3
1/3
-
+
G204G
DU
(+) +
1/5 +
1/5 +
+ -
+ +
NCTC 11637 Subset 2
G (Mot- Tox §
G2
DU
G32
GU
G65
GU
m
G102
Ge
m
G104
Ge
m
G109
Gs
Subset 3
m
D
m
1/2
+
m
1/10
1/10
1/3
1/5
+
+
1/5
+
1/3 1/5
+
1/3
(Mot- Tox-)
G21
Gm
G50
D
m
Ge
D
m
D
G121Y
Gs
m
m
m
G123 c
DU
i
+ +
+
= See Table 1. b = Highest filtrate dilution giving vacuolization on cells. + only, dilution not known. c = Broth culture filtration concentrates contained cytotoxin-asssociated proteins that blotted with serum from subjects infected by cytotoxigenic strains of H. pylori. + = positive result; (+) = weak positive result; - = negative result. 17
Owen R.J. et al.
Eur. J. Epidemiol.
recorded as motile had a higher proportion of definitely motile cells than the five strains with few motile cells that were recorded as weakly motile. The results (Table 2) show that four of the five weakly motile strains were also non-toxigenic for Vero cells. Seven of these eleven strains were from ulcer patients. Subset 2 (Mot, Tox§ contained six strains (29~ which were non-motile and toxigenic for CHO cells, although three strains were non-toxigenic for Vero cells (Table 2). Three of the six strains were from ulcer patients. Subset 3 (Mot-, Tox-) comprised four (19~ nontoxigenic strains that were non-motile after repeated testing. Most (three) of these strains were from patients with gastritis.
HaellI ribopatterns Chromosomal DNA from 15 isolates and NCTC 11637 was completely digested with HaeIII and each gave a unique multiband ribopattern (Fig. 2). No bands were common to every strain ribopattern except for 1.0 -----0.1kb and 1.2 -+ 0.1kb bands occurring singly or as a doublet in patterns of every strain except G102 (subset 2), which contained neither band. DNAs of the following six strains were not cut by HaeIII even
Plasmid content Plasmid DNA (a single band) was detected in three strains of subset 1 and in one strain of subset 2. All four strains were toxigenic for CHO cells and two strains were also toxigenic for Vero cells. No plasmid DNA was detected in any of the other toxigenic strains. HindlII ribopatterns Chromosomal DNA from all 21 isolates and NCTC 11637 was completely or partially digested with HindIII and most gave a unique multiband ribopattern (Fig. 1) with bands between 1 and 10Kb. Nineteen different ribopatterns were detected although strains G21 and G29 had identical patterns as did strains G39 and G106. No bands were common to every ribopattern but a 1.5 _ 0.1kb band was found in all strain patterns except G27, G32 and G103. Most DNA ribopattems also contained a 3.5 + 0.1kb band.
Figure 2. - The rRNA gene patterns (ribopatterns). HaelIl digests of H. pylori (see legend to Fig. 1 for other details). The Siena strain numbers were as follows. Lanes: 2, G2; 3, G21; 4, G27; 5, G29; 6, G32; 7, G33; 8, G39; I1, G50; 12, G102; 13, G104; 14, G105; 15, G106; 16, G123; 17, G126; 18, G204G. Lanes 1 and 10, bacteriophage ;~-HindllI digests. Lane 9, NCTC 11637 - HaelII digest.
Figure 1. - The rRNA gene patterns (ribopattems). HindlII digests of chromosomal DNA probed with biotinylated cDNA from rRNA of H. pylori NCTC 11638, for the following Siena strains of H. pylori. lanes: 1, G2; 2, Gll; 3, G21; 4, G27; 5, G29; 6, G32; 8, G33; 9, G39; 10, G50; 11, G65; 12, G89; 13, G102; 16, G103; 17, G104; 18, G105; 19, G106; 20, G109; 21, G120; 22, G123; 23, G126; 24, G204G. Lanes 7 and 15, bacteriophage h-HindllI digests to provide molecular size markers. Lane 14, NCTC 11637 - HindllI digest.
18
Biotypes and DNA fingerprints of H. pylori
Vol. 8, 1992
after repeated testing on new DNA samples: G11, G89 and G103 (subset 1); G65 and G109 (subset 2) and G120 (subset 3).
rarely reported for H. pylori but members of the species are generally characterised as being actively motile (10). Our results showed an agreement (positive and negative matches) between motility and toxigenicity for 73% of strains (subset 1 and 3). Only for the strains of subset 2 was there no coincidence between these two features. The distribution of the different subset strains in the ulcer and gastritis (nonulcer) patient groups, illustrated in figure 3, indicates that the subset 1 strains were slightly less prevalent in the gastritis only patients, whereas the subset 3 strains were more prevalent in that patient group. The subset 2 strains were present in equal numbers in each group. Little is known about the precise pathogenic role or the virulence mechanisms of H. pylori but it has been speculated that active motility may facilitate colonization of the gastric mucosa and that the cytotoxin may contribute to gastric inflammation and erosion in vivo (17, 30). Our results support the data of Eaton and colleagues (8) who observed an association between motility and cytotoxicity but only a few strains were examined in that study. They found, also, that non-motile strains of H. pylori were less virulent for gnotobiotic piglets than a motile toxigenic strain, and concluded that virulence correlated well with motility but not as well with cytotoxin production. Our data suggest that the subset 1 (Mot" Tox0 strains could possibly be the most virulent (ulcerogenic) as they were predominantly (640/0) from patients with ulcers whereas subset 3 (Mot- Tox-) strains were predominantly (75%) from gastritis-only
DISCUSSION
In 1989, Figura and colleagues (9) demonstrated that cytotoxin producing strains ofH. pylori were more prevalent in patients with ulcers and chronic gastritis than in patients with gastritis only. The association was considered to be statistically significant and was proposed as evidence that a cytotoxin was involved in the development of peptic ulceration in those patients. The aim of the present study was to investigate the strains further to determine if those from ulcer and non-ulcer patients possessed any distinguishing biotypic or DNA features. Associations with enzyme biotype All strains were biotype 2, which was found previously to be the commonest biotype of H. pylori (23). There was no evidence of an association between biotype and cytotoxin production or ulcer presence. Associations with motility Twelve strains (55~ were observed to be consistently motile but it was unexpected that as many as 10 strains (45%) were non-motile. The feature is
Number of strains l 7
[ - ............................................ l
Subset 1
Subset 3
Subset 2
........
6 5 4 ............................................................................................................................................................
~
......
-:.;.;.;.;.:, . .....
O
,
.........-..
Ulcers
Gastritis only
Patient group Figure 3. - Distribution of H. pylori phenotypic subset strains in different patient groups. 19
Owen R.J. et al.
Eur. J. Epidemiol.
patients. We also observed that three subset 1 (gastritis-only) isolates were atypical either in being weakly motile and/or in producing no cytotoxic activity on Vero ceils. This suggested a gradation in virulence between strains within subset 1. However, on the criterion of motility alone, there was no direct correlation with the diagnostic group of the patient. Our study also revealed intermediate strains (subset 2) that were non-motile and CHO toxin producing, although three of these strains were non-toxigenic for Vero cells. There was no clear correlation between the phenotypic characteristics of these subset 2 strains and diagnosis group o f patients.
Associations with ribopatterns
Aclmowledgement
MM was the recipient of a Junior Training Grant (BAP0568-UK) provided within the framework of the CEC Research Action Programme for Biotechnology. The Procter & Gamble Company (Cincinnati) is thanked for financial support. REFERENCES
1. Ascencio F., Guruge J.L., Lungh A., Megraud F., Wei S. and Wadstrom T. (1990): Lectin typing of Helicobacter pylori, In: Malfertheiner P. and Ditschuneit. H. (eds.), Helicobacter pylori gastritis and peptic ulcer. Springer-Verlag, Heidelberg, pp. 104-109. 2. Beji A., Vincent P., Dachis I., Husson M. 0., Cortol A. and Leclerc H. (1982): Evidence of gastritis with several Helicobacter pylori strains. Lancet ii: 14021403.
There was considerable diversity at the genomic level between the strains of H. pylori with most having different HindlII and HaelII ribopatterns. There were matches in HindlII-ribopatterns in two strain pairs but they differed in HaelH ribopatterns. There was no evidence of correlation between total ribopattern or particular bands within the ribopattern, and phenotypic subsets based on motility and cytotoxin production, or site of isolation, or patient diagnosis group. However, the ribopatterns provided a precise fingerprint for individual strain identification as found previously (18) and were clearly more discriminatory than the phenotypic differences observed between H. pylori. We observed that some strain DNAs were not cut by HaelII, and this was thought most likely to be due to some degree of methylation o f the DNA. Strains exhibiting this feature were not associated with any particular phenotypic subset or clinical group and all had different HindlII ribopatterns.
3. Burnie J., Lee W., McNulty C. and Dent J. (1988): Virulence of Campylobacter strains and degree of gastritis. - Lancet i: 302. 4. Burnie J., Lee W., Dent J.C. and McNulty C.A.M. (1988): Immunoblot fingerprinting of Campylobacter pylori. - J. Med. Microbiol. 27: 153-159. 5. Cover T.L., Dooley C.P. and Blaser M.J. (1990): Characterization of and human serologic response to proteins in Helicobacter pylori broth culture supernatants with vacuolizing cytotoxin activity. Infect. Immun. 58: 603-610. 6. Cowan S.T. (1974): Cowan and Steeles' Manual for the identification of medical bacteria, 2nd ed. Cambridge University Press, London. 7. Danielson D., Blomberg B., Jarnerot G. and Kosunen T.U. (1988): Heterogeneity of Campylobacterpylori as demonstrated by co-agglutination testing with rabbit antibodies - Scand. J. Gastroenterol. 23: (suppl. 142) 58-63.
Association with plasmid DNA content The number of strains containing plasmids was relatively small (180/0) and contrasted with earlier studies in which plasmid occurrence rates o f about 45~ were reported (15, 27, 33). The four strains containing plasmids were cytotoxin producers but as the majority o f strains examined in this study did not contain a plasmid, it appeared highly unlikely the property was plasmid-determined in these particular H. pylori. Our study showed that strains of H. pylori from ulcer patients were genomically as heterogeneous as those from patients with gastritis only (no ulcers). The results suggest that motility is closely associated with cytotoxin production in the majority of strains irrespective of the clinical symptoms of the patients. These results were based on a relatively small patient sample, however, and the associations observed need to be investigated on strains from a larger group of patients. As strains o f H . pylori differed consistently in motility, we conclude, however, that it provides a useful additional character for biotyping of isolates.
8. Eaton K.A., Morgan D.R. and Krakowka S. (1989): Campylobacter pylori virulence factors in gnotobiotic piglets. - Infect. Immun. 57:1119-1125. 9. Figura N., Guglielmetti P., Rossolini A. et al. (1989): Cytotoxin production by Campylobacter pylori strains isolated from patients with peptic ulcers and from patients with chronic gastritis only. J. Clin. Microbiol. 27: 225-226. 10. Goodwin S.C. and Armstrong J.A. (1990): Microbiological aspects of Helicobacter pylori (Campylobacter pylori). - Eur. J. Clin. Microbiol. Infect. Dis. 9: 1-13. 11. Hupertz V. and Czinn S. (1988): Demonstration of a cytotoxin from Campylobacter pylori. - Eur. J. Clin. Microbiol. Infect. Dis. 7: 576-578. 12. Kado C.L and Liu S.T. (1981): Rapid procedure for detection and isolation of large and small plasmids. J. Bacteriol. 145: 1365-1373. 20
Biotypes and DNA fingerprints of H. pylori
Vol. 8, 1992
13. Langenberg W., Rauws E.A.J., Widjojokusumo A., Tytgat G.N.J. and Zanen H.C. (1986): Identification of Campylobacter pylori isolates by restriction endonuclease analysis. - J. Clin. Microbiol. 24: 414417.
24. Owen R.J. and Hernandez J. (1990): Occurence of plasmids in "Campylobacter upsaliensis" (catalase negative or weak group) from geographically diverse patients with gastroenteritis or bacteraemia. - Eur. J. Epidemiol. 6: 111-117.
14. Leunk R.D., Johnson P.T., David B.C. et al. (1988): Cytotoxic activity in broth-culture filtrates of Campylobacter pylori. - J. Med. Microbiol. 26: 93-99.
25. Owen R.J., On S.L.W. and Costas M. (1988): Potential transport medium for Campylobacter pylori - J: Clin. Pathol. 41: 1337-1339. 26. Parsonnet J. (1989): The epidemiology of C. pylori. In: Campylobacter pylori gastritis and peptic ulcer disease. Blaser M.J. (Ed.). - Igaku Shoin Medical Publishers, New York, pp. 51-60.
15. MajewsM S.LH. and Goodwin C.S. (1988): Restriction endonuclease analysis of the genome of Campylobacterpylori with a rapid extraction method: evidence for considerable genomic variation. - J. Infect Dis. 157: 465-471.
27. Penfold S.S., Lastovica A.J. and Elisha B.G. (1988): Demonstration of plasmids in Campylobacterpylori, J. Infect. Dis. 157: 850-851.
16. McNulty C.A.M. (1990): Helicobacter pylori (Campylobacter pylori) - its clinical significance. Rev. Med. Microbiol. 1: 76-82.
28. Pitcher D.G., Owen R.J., Dyal P. and Beck A. (1987): Synthesis of a biotinytated DNA probe to detect ribosomal RNA cistrons in Providencia stuartii. FEMS Microbiol. Lett. 48: 283-287.
17. Morgan D.R. and Leunk R.D. (1989): Pathogenesis of infection by C. pylori. In: Campylobacter pylori in gastritis and peptic ulcer disease. Blaser M.J. (Ed.) Igaku Shoin Medical Publishers, New York, pp. 115133.
29. Pitcher D.G.', Saunders N.A. and Owen R.J. (1989) Rapid extraction of bacterial genomic DNA with guanidium thiocyanate. - Lett. Appl. Microbiol. 8: 151-156.
18. Morgan D.D. and Owen R.J. (1990): Use of DNA restriction endonuclease digest and ribosomal RNA gene probe patterns to fingerprint Helicobacter pylori and Helicobacter mustelae isolated from humans and animal hosts. - Mol. Cell. Probes 4: 321-334.
30. Rathbone B.J. and Heatley R.V. (1989): Possible pathogenic mechanisms in Campylobacter pylori infection. In: Campylobacter pylori and gastroduodenal disease. Rathbone B.J. and Heatley R.V. (Eds.). Blackwell Scientific Publications, Oxford, pp. 203-208.
19. Oudbier J.H., Langenberg W., Rauws E.A.J. and BruinMosch C. (1990): Genotypical variation of Campylobacter pylori from gastric mucosa. - J. Clin. Microbiol. 28: 559-565.
31. Reina J. and Alomar P, (1989): Biotypes of Campylobacter pylori isolated in gastroduodenal biopsies. - Europ. Clin. Microbiol. Infect. Dis. 8: 175177.
20. Owen R.J., Costas M. and Dawson C. (1989): Application of different chromosomal DNA restriction digest fingerprints to specific and subspecific identification of Campylobacter isolates. J. Clin. Microbiol. 27: 2338-2343.
32. Shames B., Krajden S., Babida C., Gurgis R.V. Kurjanc~k L. and Penner J.L.(1989): Investigation of Campylobacter pylori strains isolated from a husband and wife. - In: Megraud F. and Lamouliatte H., (eds.), Gastroduodenal Pathology and Campylobacter pylori. - Elsevier, Amsterdam, pp. 439-442.
21. Owen R.J., Costas M., Morgan D.D., On S.L.W., Hill L.R., Pearson A.D. and Morgan D.R. (1989): Onedimensional electrophoretic protein fingerprints for identifying and typing Campylobacter pylori. In: Megraud F. and Lamouliatte H. (eds.), Gastroduodenal pathology and Campylobacterpylori. Elsevier, Amsterdam, pp. 95-98.
33. Simor A.E., Shames B., Drumm B. et. al. (1990): Typing of Campylobacter pylori by bacterial DNA restriction endonuclease analysis and determination of plasmid profile. - J. Clin. Microbiol. 28: 83-86.
-
34. Tjia T.N., Harper W.E.S., Goodwin c.S. and Grubb W.B. (1987): Plasmids in Campylobacter pyloridis. Microbios Lett. 36: 7-11.
22. Owen R.J., Fraser J., Costas M., Morgan D. and Morgan D.R. (1990): Signature patterns of DNA restriction fragments of Helicobacterpylori before and after treatment. - J. Clin. Microbiol. 43: 646-649.
35. Wadstrom T., Guruge J.L., Wei S., Aleljung P. and Ljungh A. (1990): Helicobacter pylori hemagglutinins possible gut mucosa adhesins. - In: Malfertheiner P. and Ditschneit H. eds. Helicobacter pylori gastritis and peptic ulcer. Springer-Verlag, Heidelberg, pp. 96103.
23. Owen R.J. and Desai M. (1990): Preformed enzyme profiling of Helicobacter pylori and Helicobacter mustelae from human and animal sources. - L e t t . Appl. Microbiol. 11: 103-105.
21
EUROPEAN JOURNAL
January 1992, p. 15-21
Vol. 8, No. I
OF EPIDEMIOLOGY
BIOTYPES AND DNA FINGERPRINTS OF CYTOTOXIGENIC HELICOBACTER PYLORI FROM PATIENTS WITH GASTRITIS AND PEPTIC ULCERATION IN ITALY R.J. O W E N *l, N. F I G U R A * * and M. M O R E N O * *National Collection o f Type Cultures - Central Public Health Laboratory - London N W 9 5 H T - UK. **Institute di Malattie Injettive - University o f Siena 53100 - Italy.
Key words: Biotypes - DNA fingerprints - Helicobacter pylori - Cytotoxins Variations in biotypic and DNA characteristics of 21 strains o f Helicobacterpylori from patients in central Italy with histologically defined gastritis and/or peptic ulcers were studied. The strains had the same preformed enzyme biotype but differed in motility and cytotoxigenic activity. The presence or absence of these two characteristics were closely associated in 73% of strains and were used to define three phenotypic subsets. Strains of subset 1 (Mot + Tox +) were most common (52%) particularly amongst patients with peptic ulceration (64~ H. pylori had unique HaelII and HindlII ribopatterns but no one ribopattern or single band within a ribopattern was characteristic of any particular phenotypic subset. H, pylori from patients with gastritis were genomical]y as heterogeneous as those from patients with ulcers. Plasmid DNA was detected in four strains (25~ and although three of these were of the same biotype (phenotypic subset 1), there was no general association apparent between plasmid presence and cytotoxic activity. It was concluded that motility was a useful additional feature for biotyping of H. pylori.
complex and appears to be strain specific (15). In contrast, there is evidence that ribosomal (r) RNA gene patterns (ribopatterns) provide a reliable and simpler basis for strain discrimination (18). Other techniques applied to characterise strains of H. pylori with varying degrees of success include biotyping (23, 32), cytotoxigenicity (5, 9, 11, 14), haemagglutination (35), coagglutination with rabbit antibodies (7), lectin typing (1), immunoblotting (3, 4), plasmid profiling (15, 27, 34) and electrophoretic protein pattern profiling (21). The aim of the present study was to determine if strains of H. pylori defined by biotype, cytotoxin production and disease association had characteristic molecular features based on ribopatterns and plasmid DNA content. Because cytotoxin production is a putative virulence determinant (17, 30), a specific objective of the study was to investigate associations with that feature.
INTRODUCTION
Helicobacter pylori (Campylobacter pylori) is strongly associated worldwide with chronic active type-B gastritis and higher duodenal ulcer relapse rates, and is possibly one of the causes of non-ulcer dyspepsia (16). A number of epidemiological features of H. pylori, such as its incidence and prevalence in population groups, are well documented (26) but the detailed surveillance of individual strains of H. pylori is hindered by the lack of reliable and accurate typing methods. To date, the most precise and widely used method of identifying strains has been DNA restriction endonuclease digest analysis (2, 13, 15, 18, 19, 22, 32, 33), although the technique has limitations as a basis for typing because each DNA pattern is J Corresponding author.
15
Owen R.J. et al.
Eur. J. Epidemiol.
Enzyme biotyping
MATERIALS AND METHODS
Enzymic biotype was determined from preformed enzyme production in API Zym kits as described previously (23).
The 21 clinical strains of H. pylori were from patients undergoing diagnostic gastroduodenal endoscopy at the Ospedale di Grosseto, Grosseto, Italy. Biopsies were obtained from the gastric antrum and the edges of ulcers (when present). Table 1 lists the categories of patients according to symptoms and the degree of gastritis in each category. The severity of gastritis was assessed by the degree of round and polymorphonuclear cell infiltration and the presence of erosions. Methods of isolation and primary identification of strains were described previously (9). Reference strain NCTC 11637 (= RPH 13487) from Perth (Australia) was obtained from the National Collection of Type Cultures (NCTC). All isolates were grown on Oxoid brain-heart infusion agar (BHI) containing 5~ (v/v) horse blood and supplemented with 1% Isovitalex (BBL Microbiology Systems, Becton Dickinson, Cowley, Oxford, UK). Cultures were incubated for 48h at 37~ under microaerobic conditions (5%, 02; 5%, CO2; 2%, H2; 88%, Nz) in a Variable Atmosphere Incubator (Don Whitley Scientific Ltd., Shipley, Yorks). For transportation, strains were maintained in blood agar (25), and stocks of all cultures were preserved in 10% v/v glycerol in Oxoid nutrient broth over liquid nitrogen.
Cytotoxin testing Cytotoxin production tests against mammalian cells (CHO and Vero cells) in vitro were performed as described previously (9). Cells were examined after 24 and 48 h of incubation for the presence of the typical toxic effect (intracytoplasmic vacuolization) at one or more dilutions. DNA isolation, restriction digestion, and ribosomal RNA gene probe hybridization Chromosomal DNA was isolated and purified using the guanidium thiocyanate reagent method (29). All DNA samples (5 lag) were digested for 4 h at 37~C with HaelIl and HindlII (2-3U/lag of DNA) according to the manufacturer's instructions (Northumbria Biochemicals Ltd., Cramlington, Yorkshire). The digested DNA was electrophoresed at 25 V for 16 h in a horizontal 0.70/0 (w/v) agarose gel in a buffer containing 89 mM Tris hydrochloride, 89 mM boric acid, and 2 mM disodium ethylenediaminetetra-acetic acid (EDTA) (pH 8.3). After electrophoresis, the DNA fragments were transferred to Hybond-N nylon membranes (Amersham International) by vacuum transfer blotting. The membranes were then hybridized by standard procedures for 18 h at 42~ C with a biotinylated cDNA probe prepared from a 16S and 23S rRNA mixture ofH.pylori NCTC 11638 using Moloney mouse leukaemia virus reverse transcriptase (Gibco-BRL) (28). Biotinylation was achieved by the incorporation of biotin-16-dUTP. The membranes were washed after hybridization and the hybridized probe detected colorimetrically using a nonradioactive detection kit-BIuGENE (Gibco-BRL). The pattern of bands obtained was designated the ribopattern. Full details of the methods of hybridization and band size estimation are given elsewhere (20).
TABLE 1. - Italian strains of H. pylofi used grouped according to patient symptoms. Patient group
No. of isolates (n = 21)
Gastritis (G) moderate (Gm) severe (Gs) erosive (Ge)
2 3 5
Duodenal ulcer (DU) moderate gastritis severe gastritis
2 4
Gastric ulcer (GU) erosive gastritis
4
Duodenal and gastric ulcer (DU and GU) severe gastritis
1
Plasmid DNA extraction and electrophoresis For plasmid DNA extraction, organisms were cultured as described above. Single colonies were picked and plasmid DNA was prepared according to the method of Kado and Liu (12) as modified previously (24). Samples were electrophoresed in a horizontal 0.7% agarose gel for 2 h at 100 V. After staining With ethidium bromide, the gels were viewed on a UV transilluminator and photographed.
Motility testing Isolates were grown on slopes of BHI blood agar with added Isovitalex as described above. After at least two subcultures at 3-4d intervals, the liquid phase culture at the base of the slope of a 2d-old culture was removed and bacterial motility was determined on a microscope slide by standard methods (6). Cultures that were weakly (sluggishly) motile or non-motile were recultured and retested for motility.
RESULTS
16
Enzyme biotypes All the strains ofH. pylofi were enzyme biotype II. They produced leucine arylamidase and
Biotypes and DNA fingerprints of H. pylori
Voh 8, 1992
p h o s p h o h y d r o l a s e b u t n o t esterases C4 and C8 in the API Z y m galleries.
strain was originally isolated is also indicated, b u t n o further distinctions based o n the degree o f gastritis were made. Subset 1 (Mot ~, Tox*) comprised eleven (52%) Italian strains a n d the reference strain ( N C T C 11637) which were actively motile. T h e r e were visible differences b e t w e e n s o m e strains in the degree o f motility, which was expressed in terms o f the p r o p o r t i o n of cells actively moving. Six strains
Motility and cytotoxic activity T h e results o n motility a n d cytotoxic activity are given in T a b l e 2, in which strains are g r o u p e d into three subsets o n the basis o f these two features. T h e s y m p t o m category o f the patient from w h o m the
T A B L E 2. - Biotypic a n d D N A characteristics o f H. pylori from Italy.
Strain no.
Subset 1
Diagnosis group,
Motility
Cytotoxin productionb CHO
Vero
Plasmid present
Genome DNA cut by HaellI
(Mot + Tox §
Gll
DU
(+)
1/2
-
-
-
G27
Ge
+
1/3
-
+
+
G29
Gs
(+)
1/10
I/I0
-
+
G33
Ge
+
1/10
1/10
-
+
G39
DU
+
1/3
1/5
-
+
G89
GU
+
1/3
1/3
-
-
-
-
G103
Gm
(+)
1/5
-
G105
DU+GU
+
1/3
1/5
+
+
G106
DU
(+)
1/2
-
-
+
G126
GU
+
1/3
1/3
-
+
G204G
DU
(+) +
1/5 +
1/5 +
+ -
+ +
NCTC 11637 Subset 2
G (Mot- Tox §
G2
DU
G32
GU
G65
GU
m
G102
Ge
m
G104
Ge
m
G109
Gs
Subset 3
m
D
m
1/2
+
m
1/10
1/10
1/3
1/5
+
+
1/5
+
1/3 1/5
+
1/3
(Mot- Tox-)
G21
Gm
G50
D
m
Ge
D
m
D
G121Y
Gs
m
m
m
G123 c
DU
i
+ +
+
= See Table 1. b = Highest filtrate dilution giving vacuolization on cells. + only, dilution not known. c = Broth culture filtration concentrates contained cytotoxin-asssociated proteins that blotted with serum from subjects infected by cytotoxigenic strains of H. pylori. + = positive result; (+) = weak positive result; - = negative result. 17
Owen R.J. et al.
Eur. J. Epidemiol.
recorded as motile had a higher proportion of definitely motile cells than the five strains with few motile cells that were recorded as weakly motile. The results (Table 2) show that four of the five weakly motile strains were also non-toxigenic for Vero cells. Seven of these eleven strains were from ulcer patients. Subset 2 (Mot, Tox§ contained six strains (29~ which were non-motile and toxigenic for CHO cells, although three strains were non-toxigenic for Vero cells (Table 2). Three of the six strains were from ulcer patients. Subset 3 (Mot-, Tox-) comprised four (19~ nontoxigenic strains that were non-motile after repeated testing. Most (three) of these strains were from patients with gastritis.
HaellI ribopatterns Chromosomal DNA from 15 isolates and NCTC 11637 was completely digested with HaeIII and each gave a unique multiband ribopattern (Fig. 2). No bands were common to every strain ribopattern except for 1.0 -----0.1kb and 1.2 -+ 0.1kb bands occurring singly or as a doublet in patterns of every strain except G102 (subset 2), which contained neither band. DNAs of the following six strains were not cut by HaeIII even
Plasmid content Plasmid DNA (a single band) was detected in three strains of subset 1 and in one strain of subset 2. All four strains were toxigenic for CHO cells and two strains were also toxigenic for Vero cells. No plasmid DNA was detected in any of the other toxigenic strains. HindlII ribopatterns Chromosomal DNA from all 21 isolates and NCTC 11637 was completely or partially digested with HindIII and most gave a unique multiband ribopattern (Fig. 1) with bands between 1 and 10Kb. Nineteen different ribopatterns were detected although strains G21 and G29 had identical patterns as did strains G39 and G106. No bands were common to every ribopattern but a 1.5 _ 0.1kb band was found in all strain patterns except G27, G32 and G103. Most DNA ribopattems also contained a 3.5 + 0.1kb band.
Figure 2. - The rRNA gene patterns (ribopatterns). HaelIl digests of H. pylori (see legend to Fig. 1 for other details). The Siena strain numbers were as follows. Lanes: 2, G2; 3, G21; 4, G27; 5, G29; 6, G32; 7, G33; 8, G39; I1, G50; 12, G102; 13, G104; 14, G105; 15, G106; 16, G123; 17, G126; 18, G204G. Lanes 1 and 10, bacteriophage ;~-HindllI digests. Lane 9, NCTC 11637 - HaelII digest.
Figure 1. - The rRNA gene patterns (ribopattems). HindlII digests of chromosomal DNA probed with biotinylated cDNA from rRNA of H. pylori NCTC 11638, for the following Siena strains of H. pylori. lanes: 1, G2; 2, Gll; 3, G21; 4, G27; 5, G29; 6, G32; 8, G33; 9, G39; 10, G50; 11, G65; 12, G89; 13, G102; 16, G103; 17, G104; 18, G105; 19, G106; 20, G109; 21, G120; 22, G123; 23, G126; 24, G204G. Lanes 7 and 15, bacteriophage h-HindllI digests to provide molecular size markers. Lane 14, NCTC 11637 - HindllI digest.
18
Biotypes and DNA fingerprints of H. pylori
Vol. 8, 1992
after repeated testing on new DNA samples: G11, G89 and G103 (subset 1); G65 and G109 (subset 2) and G120 (subset 3).
rarely reported for H. pylori but members of the species are generally characterised as being actively motile (10). Our results showed an agreement (positive and negative matches) between motility and toxigenicity for 73% of strains (subset 1 and 3). Only for the strains of subset 2 was there no coincidence between these two features. The distribution of the different subset strains in the ulcer and gastritis (nonulcer) patient groups, illustrated in figure 3, indicates that the subset 1 strains were slightly less prevalent in the gastritis only patients, whereas the subset 3 strains were more prevalent in that patient group. The subset 2 strains were present in equal numbers in each group. Little is known about the precise pathogenic role or the virulence mechanisms of H. pylori but it has been speculated that active motility may facilitate colonization of the gastric mucosa and that the cytotoxin may contribute to gastric inflammation and erosion in vivo (17, 30). Our results support the data of Eaton and colleagues (8) who observed an association between motility and cytotoxicity but only a few strains were examined in that study. They found, also, that non-motile strains of H. pylori were less virulent for gnotobiotic piglets than a motile toxigenic strain, and concluded that virulence correlated well with motility but not as well with cytotoxin production. Our data suggest that the subset 1 (Mot" Tox0 strains could possibly be the most virulent (ulcerogenic) as they were predominantly (640/0) from patients with ulcers whereas subset 3 (Mot- Tox-) strains were predominantly (75%) from gastritis-only
DISCUSSION
In 1989, Figura and colleagues (9) demonstrated that cytotoxin producing strains ofH. pylori were more prevalent in patients with ulcers and chronic gastritis than in patients with gastritis only. The association was considered to be statistically significant and was proposed as evidence that a cytotoxin was involved in the development of peptic ulceration in those patients. The aim of the present study was to investigate the strains further to determine if those from ulcer and non-ulcer patients possessed any distinguishing biotypic or DNA features. Associations with enzyme biotype All strains were biotype 2, which was found previously to be the commonest biotype of H. pylori (23). There was no evidence of an association between biotype and cytotoxin production or ulcer presence. Associations with motility Twelve strains (55~ were observed to be consistently motile but it was unexpected that as many as 10 strains (45%) were non-motile. The feature is
Number of strains l 7
[ - ............................................ l
Subset 1
Subset 3
Subset 2
........
6 5 4 ............................................................................................................................................................
~
......
-:.;.;.;.;.:, . .....
O
,
.........-..
Ulcers
Gastritis only
Patient group Figure 3. - Distribution of H. pylori phenotypic subset strains in different patient groups. 19
Owen R.J. et al.
Eur. J. Epidemiol.
patients. We also observed that three subset 1 (gastritis-only) isolates were atypical either in being weakly motile and/or in producing no cytotoxic activity on Vero ceils. This suggested a gradation in virulence between strains within subset 1. However, on the criterion of motility alone, there was no direct correlation with the diagnostic group of the patient. Our study also revealed intermediate strains (subset 2) that were non-motile and CHO toxin producing, although three of these strains were non-toxigenic for Vero cells. There was no clear correlation between the phenotypic characteristics of these subset 2 strains and diagnosis group o f patients.
Associations with ribopatterns
Aclmowledgement
MM was the recipient of a Junior Training Grant (BAP0568-UK) provided within the framework of the CEC Research Action Programme for Biotechnology. The Procter & Gamble Company (Cincinnati) is thanked for financial support. REFERENCES
1. Ascencio F., Guruge J.L., Lungh A., Megraud F., Wei S. and Wadstrom T. (1990): Lectin typing of Helicobacter pylori, In: Malfertheiner P. and Ditschuneit. H. (eds.), Helicobacter pylori gastritis and peptic ulcer. Springer-Verlag, Heidelberg, pp. 104-109. 2. Beji A., Vincent P., Dachis I., Husson M. 0., Cortol A. and Leclerc H. (1982): Evidence of gastritis with several Helicobacter pylori strains. Lancet ii: 14021403.
There was considerable diversity at the genomic level between the strains of H. pylori with most having different HindlII and HaelII ribopatterns. There were matches in HindlII-ribopatterns in two strain pairs but they differed in HaelH ribopatterns. There was no evidence of correlation between total ribopattern or particular bands within the ribopattern, and phenotypic subsets based on motility and cytotoxin production, or site of isolation, or patient diagnosis group. However, the ribopatterns provided a precise fingerprint for individual strain identification as found previously (18) and were clearly more discriminatory than the phenotypic differences observed between H. pylori. We observed that some strain DNAs were not cut by HaelII, and this was thought most likely to be due to some degree of methylation o f the DNA. Strains exhibiting this feature were not associated with any particular phenotypic subset or clinical group and all had different HindlII ribopatterns.
3. Burnie J., Lee W., McNulty C. and Dent J. (1988): Virulence of Campylobacter strains and degree of gastritis. - Lancet i: 302. 4. Burnie J., Lee W., Dent J.C. and McNulty C.A.M. (1988): Immunoblot fingerprinting of Campylobacter pylori. - J. Med. Microbiol. 27: 153-159. 5. Cover T.L., Dooley C.P. and Blaser M.J. (1990): Characterization of and human serologic response to proteins in Helicobacter pylori broth culture supernatants with vacuolizing cytotoxin activity. Infect. Immun. 58: 603-610. 6. Cowan S.T. (1974): Cowan and Steeles' Manual for the identification of medical bacteria, 2nd ed. Cambridge University Press, London. 7. Danielson D., Blomberg B., Jarnerot G. and Kosunen T.U. (1988): Heterogeneity of Campylobacterpylori as demonstrated by co-agglutination testing with rabbit antibodies - Scand. J. Gastroenterol. 23: (suppl. 142) 58-63.
Association with plasmid DNA content The number of strains containing plasmids was relatively small (180/0) and contrasted with earlier studies in which plasmid occurrence rates o f about 45~ were reported (15, 27, 33). The four strains containing plasmids were cytotoxin producers but as the majority o f strains examined in this study did not contain a plasmid, it appeared highly unlikely the property was plasmid-determined in these particular H. pylori. Our study showed that strains of H. pylori from ulcer patients were genomically as heterogeneous as those from patients with gastritis only (no ulcers). The results suggest that motility is closely associated with cytotoxin production in the majority of strains irrespective of the clinical symptoms of the patients. These results were based on a relatively small patient sample, however, and the associations observed need to be investigated on strains from a larger group of patients. As strains o f H . pylori differed consistently in motility, we conclude, however, that it provides a useful additional character for biotyping of isolates.
8. Eaton K.A., Morgan D.R. and Krakowka S. (1989): Campylobacter pylori virulence factors in gnotobiotic piglets. - Infect. Immun. 57:1119-1125. 9. Figura N., Guglielmetti P., Rossolini A. et al. (1989): Cytotoxin production by Campylobacter pylori strains isolated from patients with peptic ulcers and from patients with chronic gastritis only. J. Clin. Microbiol. 27: 225-226. 10. Goodwin S.C. and Armstrong J.A. (1990): Microbiological aspects of Helicobacter pylori (Campylobacter pylori). - Eur. J. Clin. Microbiol. Infect. Dis. 9: 1-13. 11. Hupertz V. and Czinn S. (1988): Demonstration of a cytotoxin from Campylobacter pylori. - Eur. J. Clin. Microbiol. Infect. Dis. 7: 576-578. 12. Kado C.L and Liu S.T. (1981): Rapid procedure for detection and isolation of large and small plasmids. J. Bacteriol. 145: 1365-1373. 20
Biotypes and DNA fingerprints of H. pylori
Vol. 8, 1992
13. Langenberg W., Rauws E.A.J., Widjojokusumo A., Tytgat G.N.J. and Zanen H.C. (1986): Identification of Campylobacter pylori isolates by restriction endonuclease analysis. - J. Clin. Microbiol. 24: 414417.
24. Owen R.J. and Hernandez J. (1990): Occurence of plasmids in "Campylobacter upsaliensis" (catalase negative or weak group) from geographically diverse patients with gastroenteritis or bacteraemia. - Eur. J. Epidemiol. 6: 111-117.
14. Leunk R.D., Johnson P.T., David B.C. et al. (1988): Cytotoxic activity in broth-culture filtrates of Campylobacter pylori. - J. Med. Microbiol. 26: 93-99.
25. Owen R.J., On S.L.W. and Costas M. (1988): Potential transport medium for Campylobacter pylori - J: Clin. Pathol. 41: 1337-1339. 26. Parsonnet J. (1989): The epidemiology of C. pylori. In: Campylobacter pylori gastritis and peptic ulcer disease. Blaser M.J. (Ed.). - Igaku Shoin Medical Publishers, New York, pp. 51-60.
15. MajewsM S.LH. and Goodwin C.S. (1988): Restriction endonuclease analysis of the genome of Campylobacterpylori with a rapid extraction method: evidence for considerable genomic variation. - J. Infect Dis. 157: 465-471.
27. Penfold S.S., Lastovica A.J. and Elisha B.G. (1988): Demonstration of plasmids in Campylobacterpylori, J. Infect. Dis. 157: 850-851.
16. McNulty C.A.M. (1990): Helicobacter pylori (Campylobacter pylori) - its clinical significance. Rev. Med. Microbiol. 1: 76-82.
28. Pitcher D.G., Owen R.J., Dyal P. and Beck A. (1987): Synthesis of a biotinytated DNA probe to detect ribosomal RNA cistrons in Providencia stuartii. FEMS Microbiol. Lett. 48: 283-287.
17. Morgan D.R. and Leunk R.D. (1989): Pathogenesis of infection by C. pylori. In: Campylobacter pylori in gastritis and peptic ulcer disease. Blaser M.J. (Ed.) Igaku Shoin Medical Publishers, New York, pp. 115133.
29. Pitcher D.G.', Saunders N.A. and Owen R.J. (1989) Rapid extraction of bacterial genomic DNA with guanidium thiocyanate. - Lett. Appl. Microbiol. 8: 151-156.
18. Morgan D.D. and Owen R.J. (1990): Use of DNA restriction endonuclease digest and ribosomal RNA gene probe patterns to fingerprint Helicobacter pylori and Helicobacter mustelae isolated from humans and animal hosts. - Mol. Cell. Probes 4: 321-334.
30. Rathbone B.J. and Heatley R.V. (1989): Possible pathogenic mechanisms in Campylobacter pylori infection. In: Campylobacter pylori and gastroduodenal disease. Rathbone B.J. and Heatley R.V. (Eds.). Blackwell Scientific Publications, Oxford, pp. 203-208.
19. Oudbier J.H., Langenberg W., Rauws E.A.J. and BruinMosch C. (1990): Genotypical variation of Campylobacter pylori from gastric mucosa. - J. Clin. Microbiol. 28: 559-565.
31. Reina J. and Alomar P, (1989): Biotypes of Campylobacter pylori isolated in gastroduodenal biopsies. - Europ. Clin. Microbiol. Infect. Dis. 8: 175177.
20. Owen R.J., Costas M. and Dawson C. (1989): Application of different chromosomal DNA restriction digest fingerprints to specific and subspecific identification of Campylobacter isolates. J. Clin. Microbiol. 27: 2338-2343.
32. Shames B., Krajden S., Babida C., Gurgis R.V. Kurjanc~k L. and Penner J.L.(1989): Investigation of Campylobacter pylori strains isolated from a husband and wife. - In: Megraud F. and Lamouliatte H., (eds.), Gastroduodenal Pathology and Campylobacter pylori. - Elsevier, Amsterdam, pp. 439-442.
21. Owen R.J., Costas M., Morgan D.D., On S.L.W., Hill L.R., Pearson A.D. and Morgan D.R. (1989): Onedimensional electrophoretic protein fingerprints for identifying and typing Campylobacter pylori. In: Megraud F. and Lamouliatte H. (eds.), Gastroduodenal pathology and Campylobacterpylori. Elsevier, Amsterdam, pp. 95-98.
33. Simor A.E., Shames B., Drumm B. et. al. (1990): Typing of Campylobacter pylori by bacterial DNA restriction endonuclease analysis and determination of plasmid profile. - J. Clin. Microbiol. 28: 83-86.
-
34. Tjia T.N., Harper W.E.S., Goodwin c.S. and Grubb W.B. (1987): Plasmids in Campylobacter pyloridis. Microbios Lett. 36: 7-11.
22. Owen R.J., Fraser J., Costas M., Morgan D. and Morgan D.R. (1990): Signature patterns of DNA restriction fragments of Helicobacterpylori before and after treatment. - J. Clin. Microbiol. 43: 646-649.
35. Wadstrom T., Guruge J.L., Wei S., Aleljung P. and Ljungh A. (1990): Helicobacter pylori hemagglutinins possible gut mucosa adhesins. - In: Malfertheiner P. and Ditschneit H. eds. Helicobacter pylori gastritis and peptic ulcer. Springer-Verlag, Heidelberg, pp. 96103.
23. Owen R.J. and Desai M. (1990): Preformed enzyme profiling of Helicobacter pylori and Helicobacter mustelae from human and animal sources. - L e t t . Appl. Microbiol. 11: 103-105.
21
