- Email: [email protected]
Conservation and microdiversity of the phospholipase A (pldA) gene of Helicobacter pylori infecting dyspeptics from different countries
FEMS Immunology and Medical Microbiology 32 (2001) 17^25
www.fems-microbiology.org
Conservation and microdiversity of the phospholipase A (pldA) gene of Helicobacter pylori infecting dyspeptics from di¡erent countries Jacqueline Xerry, Robert J. Owen * Helicobacter Reference Unit, Laboratory of Enteric Pathogens, Central Public Health Laboratory, 61 Colindale Avenue, London NW9 5HT, UK Received 19 May 2001 ; received in revised form 24 September 2001; accepted 24 September 2001 First published online 30 October 2001
Abstract Phospholipase activity is important in bacterial pathogenicity and could contribute to the pathogenic role of Helicobacter pylori by degradation of the gastric mucus, and in maintaining long-term colonisation. Our aim was to determine the degree of variation in the phospholipase A gene (pldA) of H. pylori from different geographic locations, and to investigate links between pldA genotype and clinical disease severity, as well as with variation in cagA status and vacA genotypes. PCR-restriction fragment length polymorphism (RFLP) analysis with MboI and HaeIII was used to study 124 isolates from 10 countries that included the two genome-sequenced strains (26695 and J99), as well as Tx30a and NCTC 11637 (type strain). The 925-bp pldA fragment was amplified with a frequency of 90%. The presence of pldA was confirmed in the other strains using an alternative forward primer. Isolates were distinguished by PCR-RFLP analysis with 10 MboI and four HaeIII restriction patterns that combined to give 25 distinct pldA RFLP types. The pldA M2H2 strain genotype was most common (20%) in the UK but similar strains came from several other countries. Microdiversity was evident in pldA sequences of strains representing different RFLP types, and five M2H2 strains each had a distinct pldA sequence type. Intragenic variation was independent of gastric disease severity as well as strain cagA status and vacA genotype, with the exception of eight geographically diverse strains all with the pldA M4H3/cagA+/vacA s1m1 genotype predominantly from peptic ulcer patients. The study indicated a spectrum of genotypic variants and was supportive of a pldA function in H. pylori colonisation and persistence rather than in chronicity of infection. ß 2001 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords : Phospholipase A gene; Sequence diversity; PCR-RFLP; Duodenal ulcer; Dyspeptic; Geographic variation; Helicobacter pylori
1. Introduction Helicobacter pylori is a Gram-negative microaerophilic bacterium that is uniquely adapted to colonise and persist on surfaces of epithelial cells in the human gastric mucosa, particularly in the antral region of the stomach. Since H. pylori was ¢rst cultured in 1982 [1], it has been shown by serological studies to have a global distribution in human populations [2], and has become established as a major risk factor for the development of peptic ulceration, as well as gastric adenocarcinoma or gastric lymphoma in some individuals [3,4]. Possible links, but not direct proven causality, have been demonstrated between infection with H. pylori and several other extragastric disorders such as cardiovascular disease [5,6], gall bladder disease
* Corresponding author. Tel. : +44 (20) 8200 4400/3740; Fax: +44 (20) 8905 9929. E-mail address : [email protected] (R.J. Owen).
[7], sudden infant death syndrome [8], apthous ulcers [9], and idiopathic parkinsonism [10]. Various pathogenicity mechanisms that could be contributing to the role of H. pylori in peptic ulcer disease have been proposed [11^13]. Most widely investigated are vacuolating cytotoxin (vacA) activity [14], lipopolysaccharides (LPS) [15], and factors involved in cytokine (IL-8) induction mediated by genes located on the cag pathogenicity island (PAI) [16]. Degradation of the gastric mucosal barrier by phospholipase (PL) activity is another potential pathogenicity mechanism of H. pylori that may have an important role in establishing and maintaining long-term infection [17]. Bacterial PLs comprise a diverse group of proteins and their action may contribute to virulence of many bacterial pathogens, most notably Clostridium perfringens and Listeria monocytogenes [18]. H. pylori exhibits several PL activities (PLA1 , PLA2 , and PLC), which are thought to have a possible role in the degradation of phospholipid components of the mucosal acid barrier [19]. The H. pylori strain 26695 genome sequence contains an open
0928-8244 / 01 / $20.00 ß 2001 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII: S 0 9 2 8 - 8 2 4 4 ( 0 1 ) 0 0 2 7 5 - 9
FEMSIM 1357 10-12-01
18
J. Xerry, R.J. Owen / FEMS Immunology and Medical Microbiology 32 (2001) 17^25
reading frame (HP0499) encoding a putative protein of V42.5 kDa with homology to Escherichia coli outer membrane PL A (PldA) and encoding an active PL with lecithinase, PLA2 , and haemolytic activities [20]. Further evidence implicating PLs in virulence was indicated by recent work showing that a mutant had reduced ability to colonise gastric tissues in the murine model of disease [20] and that bismuth salts are inhibitors of PL activity [19]. In view of the possible contributory role of PL activity in the pathogenicity of H. pylori, we have investigated the prevalence and microdiversity of the PL A precursor gene (pldA) in isolates from patients resident in the UK and nine other countries in di¡erent continents. We examined links between the di¡erent pldA alleles and severity of clinical infection, as well as their associations with two key markers of pathogenicity (vacA and cagA), which may contribute in some cases to enhanced strain-speci¢c virulence.
assays were obtained from the National Collection of Type Cultures (NCTC, London, UK): NCTC 11637 (type strain; cagA+/vacA s1m1), NCTC 12455 (strain 26695; cagA+/vacA s1m1) [22], NCTC 12908 (strain G50, cagA-/vacA s1m2) [17], and NCTC 13081 (strain Tx30a, cagA-/vacA s2m2) where cagA+ indicates presence of the cag PAI and vacA an allelic type is de¢ned by signal (s) and mid (m) region form. A culture of genome-sequenced strain J99 (cagA+/vacA s1m1) [23] was kindly provided by Professor Diane Taylor (University of Alberta, AB, Canada). 2.2. Preparation of template DNA for PCR
2. Materials and methods
Genomic DNA was extracted from strains of H. pylori grown on 10% CBA and puri¢ed using the cetyltrimethylammonium bromide (CTAB) method [24]. Extracted DNA was dissolved in 50^100 Wl of sterile distilled water and stored at 4³C. DNA concentrations were established by absorbance readings at 230 nm, 260 nm and 280 nm.
2.1. Bacterial strains examined and growth conditions
2.3. PCR ampli¢cation of the H. pylori pldA gene
A set of 124 isolates of H. pylori from 10 di¡erent geographical locations was investigated. The isolates, which were received from various laboratories between 1990 and 1999, were mainly from patients in England (66 isolates) undergoing routine upper gastrointestinal endoscopic investigation for various presentations. Antral biopsy cultures from dyspeptic patients in the following nine other countries were also examined: South Africa (10 isolates), Turkey (9 isolates), Sweden (8 isolates), Canada (7 isolates), Italy (6 isolates), Australia (5 isolates), USA (2 isolates), Peru (1 isolate) and France (1 isolate). Nine isolates from Bangladeshi patients resident in east-London were also included in the study [21]. The strains of H. pylori were categorised in three main groups based on clinical severity of gastrointestinal disease observed during endoscopic diagnosis: the peptic ulcer group (PU, 45 isolates) including cases of duodenal ulcer [40] and/or gastric ulcer [5]; the non-ulcer dyspeptic gastritis group (NUD-gastritis, 38 isolates) including patients diagnosed as having gastritis and/or oesophagitis; and the group of dyspeptics with endoscopically normal mucosa (NUD-normal, 17 isolates). A fourth group (ND, no details) comprised 24 isolates from dyspeptics for whom no further clinical details were available. All isolates of H. pylori were cultured on 10% (v/v) Columbia blood agar (CBA)(Oxoid) and incubated at 37³C under microaerobic conditions (O2 4%, CO2 5%, H2 5%, N2 86%) for 3 days in a variable atmosphere incubator (Don Whitley Scienti¢c Ltd., Yorkshire, UK). Stock cultures were preserved on glass beads in Nutrient Broth (Oxoid) containing 10% (v/v) glycerol over liquid nitrogen or at 380³C. Reference cultures of H. pylori (in lyophilised form) used as controls for the various PCR
Primers for the pldA gene were designed from the HP0499 sequence of strain 26695 (http://www.tigr.org/ docs/tigr-scripts/CMR/report.sp/), and synthesised by MWG-BIOTECH (Milton Keynes, UK). The primers PLA8664 (5P-GTCCAATTCTTGGTATCTCC-3P) and PLA9731 (5P-AGGGTTTAAGCGTATTCCTACC-3P) were used to amplify a 925-bp fragment that represented 87% of the HP0499 sequence. For strain DNAs that failed to amplify pldA using the above primers, an alternative forward primer (PLAEX) was designed from alignments of nine sequences. This primer had the sequence 5PGTTGGCAGAACATTTCTAATGG-3P and when used in combination with PLA9731 gave a 516-bp fragment. Because of its small size this product was unsuitable for restriction fragment length polymorphism (RFLP) analysis. Ampli¢cation reactions were performed in a total volume of 50 Wl containing 100 ng of diluted DNA used as template DNA, 1.5 mM MgCl2 , 0.05 mM of each deoxynucleotide (dATP, dCTP, dGTP and dTTP), 0.4 mM of each oligonucleotide primer, 0.2 Wl (1 U) Taq polymerase (Life Technologies, Paisley, UK) and 5 Wl of 10Ubu¡er provided by the manufacturer. PCR ampli¢cation was performed in a Sprint thermal cycler (Hybaid Ltd, Ashford, UK). The PCR ampli¢cation conditions used comprised one initial cycle at 95³C, 5 min followed by 30 cycles at 94³C, 30 s; 60³C, 1 min; at ¢nally 72³C, 75 s. 2.4. Restriction digest analysis of pldA amplicon Restriction endonuclease digestion with MboI and HaeIII was performed separately on the ampli¢ed DNA product by adding to a ¢nal volume of 20 Wl: 1 Wl of spermidine, 2 Wl of the appropriate bu¡er, 1 Wl of restric-
FEMSIM 1357 10-12-01
J. Xerry, R.J. Owen / FEMS Immunology and Medical Microbiology 32 (2001) 17^25
tion enzyme, 10Wl of the PCR product (containing approximately 20 ng DNA), and sterile distilled water (6 Wl). The mixtures were incubated at 37³C and the digestion was stopped after 4 h by adding 5 Wl of stop-mix solution. A 17-Wl aliquot of the digested PCR product was loaded on a 3% w/v UltraPure agarose gel (Gibco BRL) and stained with ethidium bromide. Bands were viewed under an ultraviolet transilluminator, and DNA fragment sizes were estimated from migration distances using polynomial curve ¢tting functions [25]. 2.5. cagA and vacA genotyping Details of the primers and PCR conditions for the genotyping assays have been described elsewhere but brie£y these were as follows. PCR conditions for the two assays for cagA (a marker for the right-hand end of the cag PAI and for the cagI region) were as described for the F1/B1 primers and the D008/R008 primers [26,27]. Genotyping based on vacA (signal and mid-regions) were also performed as previously described [14]. 2.6. DNA sequence analysis of pldA gene
19
from 124 strains of H. pylori was determined by restriction digest analysis, and 4 and 10 di¡erent RFLP pro¢les were obtained with HaeIII (H-types) and MboI (M-types), respectively. The 516-bp amplicon was not subjected to RFLP analysis because of the small size and the strains were recorded as non-typable in the above scheme. Typical RFLP patterns for the 925-bp amplicon contained between two and ¢ve fragments of 123^772 bp for the Mtypes, and between two and four fragments of 110^634 bp for the H-types. The estimated locations of the di¡erent fragments (ordered from the 5P to the 3P ends) are shown in Fig. 1, as derived from the pldA sequence of NCTC 12455 (strain 26695). The 3P end of the pldA amplicon was highly conserved for recognition sequences of both restriction endonucleases, giving a 291-bp fragment present in pro¢les of all H-types, and a 153-bp fragment present in most (8/10) of the M-types. Greater variation occurred in the middle region and at the 5P end, with di¡erences in one or more bands in the RFLP patterns. To increase discrimination between strains of H. pylori, because of the relatively low degree of variation detected by RFLP analysis within pldA, both sets of RFLP pro¢les were used to de¢ne a total of 25 combined RFLP types, of
pldA amplicons from three isolates representing each of the three most common RFLP types (M2H2, M3H2 and M4H3), and four strains of pldA type M2H2, with di¡erent genotypes (cagA/vacA) and/or geographical origin, were sequenced (MWG-BIOTECH, Milton Keynes, UK). Nucleotide sequences were aligned and analysed with BioEdit (http://www.mbio.ncsu.edu/RNaseP/info/ programs/BIOEDIT/bioedit.html). 3. Results 3.1. Presence of the pldA gene The predicted 925-bp pldA amplicon was obtained with the primer pair PLA8664/PLA9731 from 112 (90%) of the 124 isolates of H. pylori investigated. However, negligible/ no ampli¢cation products were obtained with those primers for DNA from 12 (10%) strains, which included the reference strain NCTC 13081 (strain Tx30a). The results indicated the atypical strains either did not contain pldA or that mutations were present in one or other of the primer sequences. Reduced annealing temperatures and di¡erent concentrations of MgCl2 did not improve the ampli¢cation and yield of product. Ampli¢cations were repeated using an alternative forward primer (PLAEX) and the predicted 516-bp product was obtained for each strain indicating that pldA was present. 3.2. Microvariation within the pldA gene RFLP analysis Intragenic variation within the 925-bp pldA amplicons
Fig. 1. Restriction fragment maps for MboI and HaeIII digests of H. pylori pldA to illustrate each of the main RFLP types. Map locations are based on the pldA sequences for strains 26695 and J99.
FEMSIM 1357 10-12-01
20
J. Xerry, R.J. Owen / FEMS Immunology and Medical Microbiology 32 (2001) 17^25
Table 1 Frequencies of the main pldA genotypes and reference strains of H. pylori Type
No. strains (n = 124)
Reference strain
M2H2 M3H2 M4H3 M4H2 M2H3 M1H3 M1H1 M1H2 M3H3 M4H1 M1H4 M3H1 M4H4 M6H2 M6H3 Uniques Non-typable
20 (16%) 11 (8%) 10 (8%) 9 (7%) 9 (7%) 8 (6%) 7 (5%) 7 (5%) 5 (4%) 4 (3%) 4 (3%) 2 (1%) 2 (1%) 2 (1%) 2 (1%) 10 (8%) 10 (9%)
NCTC NCTC NCTC H185 NCTC H173 NCTC NCTC H180 J99 H406 NCTC NCTC H289 H195 H1412 NCTC
a
cagA cag A
13090 11637 12384 13085 12455 13086
12908 12490
13081 (Tx30a)
14 8 9 6 7 3 4 3 3 4 3 1 2 1 1 7 3
cag A3
vacA type s1m1 s1m2
s2m2
6 3 1 3 2 5 3 4 2 0 1 1 0 1 1 3 9
8 4 8 3 3 2 2 0 3 4 3 0 1 1 0 5a 3a
4 0 1 2 2 3 2 4 1 0 0 0 0 1 0 1 7
8 7 1 4 4 3 3 3 1 0 1 2 1 0 2 4 2
One strain was a mixture of s1m1/s1m2 types.
tries were examined and results showed that the three most common pldA pro¢les (M2H2, M3H2 and M4H3) had a wide geographical distribution and were the main types in seven countries (Table 2). Strains that were non-typable using the PLA8664/PLA9731 primer pair were of diverse origins (¢ve countries), which included the UK and South Africa. Strains from the UK were the largest single set in the study (representing 53%) and 20% were genotype M2H2 although the other isolates showed a diverse range of types including four non-typables. By contrast, type M4H3 was the main isolate infecting the Bangladeshi patients resident in east-London. Strains from South Africa represented the second largest set in the study (10 strains) for which the predominant type was M4H3. The 24 isolates from other Western European Countries (France/ Italy/Sweden) and Turkey had a diverse range of genotypes, in particular the isolates from Italy and from Turkey. The 15 isolates from other parts of the world (USA, Canada, Peru and Australia) also had diverse genotypes although two from Australian were genotype M2H2.
which 12 types were a feature of two or more strains. Table 1 lists the frequencies of strains according to combined MH-type, and a suggested reference strain for each type. The three most frequently occurring combined pldA RFLP pro¢les, representing 33% of strains, were M2H2 (16%), M3H2 (8.8%) and M4H3 (8%). NCTC 11637 (taxonomic type strain of H. pylori) had the second most common pro¢le (M3H2). NCTC 12455 (genome-sequenced strain 26695) had the combined pro¢le M1H1, which was a feature of a small number (5.6%) of strains, and the experimentally observed MboI/HaeIII digest fragment pro¢le was in accordance with that predicted from the HP0499 gene sequence. Some 8% of the strains had unique pldA combined pro¢les. 3.3. Geographical distribution and main features of pldA strain types H. pylori originating predominantly from individuals in the UK with one or more isolates from nine other coun-
Table 2 Distribution of H. pylori pldA strain types by country of origin Country (no. strains)
Main type (no. strains)
UK (66) UK-Bangladeshi (9) South Africa (10) Turkey (9) Sweden (8) Canada (7) Italy (6) Australia (5) USA (2) Peru (1) France (1)
M2H2 (13) M4H3 (3) M4H3 (4) M4H2 (2) M2H2 (2)/M3H2(2) diverse diverse M2H2 (2) M4H1 M3H2 M4H3
No. other types 17 5 2 6 4 5 6 3 ^ ^ ^
FEMSIM 1357 10-12-01
Non-typable isolates (n = 12) 4 ^ 4 1 ^ 2 ^ ^ 1 ^ ^
J. Xerry, R.J. Owen / FEMS Immunology and Medical Microbiology 32 (2001) 17^25
3.4. Associations between pldA strain type and disease Isolates of H. pylori were categorised in three broad patient groups based on severity of gastrointestinal disease reported at endoscopy. Analysis of the strain types in relation to disease shows that overall the di¡erent pldA genotypes were widely distributed across each group (Table 3). The peptic ulcer group (PUD, 45 isolates) contained the highest number of strains (36.3%), with the majority from DU patients. Associated genotypes were diverse with type M4H3 representing the main single genotype (6/40, 15%). A variety of strain genotypes were likewise associated with the non-ulcer dyspepsia patients, but the M2H2 genotype was the most common (18%) in both the gastritis group (7/38 isolates) and the endoscopically normal group (3/17 isolates). 3.5. Associations between pldA type and other genomic markers The cagA status and vacA genotypes were determined for all 124 strains of H. pylori (Table 1). The majority (79/ 124, 64%) of the isolates were cagA+ and these were highly diverse with respect to pldA type. Furthermore, most strains were either vacA s1/m1(40%) or s1/m2 (37%), and both sm-types were also diverse with respect to pldA type. Interestingly, strains with the common M2H2 type were typically cagA + (14/20, 70%) and vacA s1 (16/20, 80%). An association also was evident between vacA s1/m1 and pldA type M4H3 for eight (8/ 10, 80%) strains, and between vacA s1/m2 and pldA type M3H2 for seven (7/11, 63.6%) strains. By contrast, the 12 non-typable strains (including strain Tx30a) were more typically genotype cagA- (9/12, 75%) and vacA m2 (9/12, 75%). When pldA type was used in combination with both cagA and vacA types, two distinct groups of strains could be de¢ned. First, genotype pldA M4H3/cagA+/vacA s1m1 was a feature of eight strains that included NCTC 12384 and seven of these were from patients with DU or GU although they were geographically diverse (BangladeshiUK, France, Italy, South Africa and UK). Second, genotype pldA NT/cagA-/vacA s2m2 was a feature of seven strains that included Tx30a and these were also geographically diverse (UK, Canada, South Africa, USA and Tur-
key). Where details were available, four patients had nonulcer in£ammatory disease, and only one patient had a DU. 3.6. Microvariation within the pldA gene DNA sequence analysis DNA sequences were determined for pldA amplicons of the three H. pylori reference strains representing the three predominant RFLP types (NCTC 13090, type M2H2 ; NCTC 11637, type M3H2; and NCTC 12384, type M4H3). The 925-bp amplicon sequences were obtained for NCTC 12384 and NCTC 13090 whereas only a partial sequence of 553 bp could be obtained for NCTC 11637 due to the presence of secondary structures at the end of a G-run. This problem was overcome by sequencing the product obtained with the PLAEX/PLA9731 primer pair. The published pldA sequence for strain J99 (type M4H1) was also included in the alignments, and all four sequences were compared with the HP0499 sequence for strain 26695 type (M1H1). The various pldA sequences had high homologies (more than 96%) yet each represented a unique allele (data not shown). To investigate relationships amongst strains of H. pylori with the same pldA RFLP type, a further four amplicons were sequenced from four strains (NCTC 11639, H210, H392 and H1431) of the most common type (M2H2), and these were compared with the sequence of the M2H2 reference strain (NCTC 13090). Each of the ¢ve pldA M2H2 sequences represented a unique allelic type. Single nucleotide polymorphisms were present at about 55 di¡erent positions (a frequency of 1 in 17 bases). The sequence of strain H210 terminated at bp position 852, and the strain clustered separately from the other M2H2 strains. Interestingly, all ¢ve M2H2 sequences di¡ered from the M1H1 sequence represented by strain 26695 at seven common locations. These were all single nucleotide polymorphisms except for the GGG to CAA triplet substitution at nucleotide positions 448^450, which was the location of the HaeIII cut-site de¢ning the H1/H2 types. The amino acid alignments (Fig. 2) showed that the latter di¡erences were non-synonymous substitutions altering the amino acid sequence from G (glycine) to Q (glutamine) with an additional 12 randomly distributed amino acid substitutions.
Table 3 Distribution of H. pylori pldA strain types by disease severity Disease group
Main type
PUD (45): DU (40) GU (4) DU+GU (1) NUD-gastritis (38) NUD-normal (17) No details (24)
M4H3 M1H1 M2H3 M2H2 M2H2 M2H2
(6) (2) (7) (3)/M4H2 (3) (5)
21
No. other types 15 2 0 18 9 12
FEMSIM 1357 10-12-01
Non-typable isolates (n = 12) 3 0 0 4 2 3
22
J. Xerry, R.J. Owen / FEMS Immunology and Medical Microbiology 32 (2001) 17^25
Fig. 2. Alignment of nine representative amino acid sequences deduced from 925-nucleotide sequences of H. pylori pldA. To obtain optimum alignments, a hyphen indicates the absence of an amino acid residue. Identical amino acids are indicated by a dot. The deduced amino acid sequence for H. pylori strain 26695 (NCTC 12455) is shown for reference purposes. The pldA genotype is given in parentheses.
4. Discussion The direct role of PL activity in the virulence of H. pylori is unclear but it has been suggested that PLA2 activity may degrade the phospholipid monolayer believed to be the cause of the hydrophobicity of the stomach lining [17]. Furthermore, recent studies demonstrated that the pldA gene from the H. pylori 26695 genome sequence encoded an active PL with lecithinase, PLA2 and haemolytic activities, for which a role was suggested in colonisation of
the gastric mucosa, and possibly in tissue damage after colonisation [20]. As the latter study focussed on just two strains (26695 and SS1) from patients in the UK and Australia respectively, our aim was to obtain a broader understanding of the prevalence and pathogenic role of pldA by investigating its presence and diversity in isolates from UK patients with di¡erent gastric diseases as well as in isolates from other geographical origins. We found pldA was a highly conserved feature of the H. pylori genome irrespective of geographical origin. The re-
FEMSIM 1357 10-12-01
J. Xerry, R.J. Owen / FEMS Immunology and Medical Microbiology 32 (2001) 17^25
sults were consistent with analyses of other putative pathogenicity-associated genes of H. pylori, such as cagA, urease subunit genes (ureA/ureB), the vacuolating cytotoxin gene (vacA), and the £agellin gene (£aA), which also have global distributions in strains from unrelated patients [26^ 32]. The iceA locus was also reported to be present at a high frequency (99%) in a worldwide collection of 321 strains [33]. A subgroup of 12 of our isolates (10%) were divergent within the forward primer region so were nontypable by the present scheme although pldA could be detected in those isolates using a modi¢ed forward primer. Comparisons of the complete genome sequences of two strains (26695 and J99) of H. pylori, provide a wealth of information about organisation, gene order and predicted proteomes [23,34,35]. Although the two sequenced isolates contained between 6 and 7% of genes speci¢c to each strain [23], the potential for inter-strain genome variation within common regions is enormous, as H. pylori appears to have a non-clonal (panmictic) population structure due to frequent recombination [36] and possibly to mutation [37]. Multilocus sequence typing provides only limited evidence of recent clonal grouping [38] and is not feasible for investigating large numbers of strains. In contrast, PCRRFLP analysis is less discriminatory but simpler to perform and provides valuable insights on the degree and location of intragenic variation, particularly when used in conjunction with available sequence data. Overall, our PCR-RFLP analysis indicated only a moderate degree of microdiversity within pldA. Single enzyme digests yielded just four types for HaeIII and ten types for MboI. Even when using combined pro¢les, the low level of diversity within pldA contrasted markedly with the higher level in several other pathogenicity-associated genes. For example, 383 geographically diverse isolates gave 82 distinct ureA/B HaeIII pro¢le types [28]. Intragenic diversity of pldA was more consistent with levels of diversity observed within conserved housekeeping genes, such as the 23S rRNA and atpA genes (Owen, R.J., unpublished results). A low level of intragenic variation would be expected for genes encoding proteins subject to structural or functional constraints, unless they were associated with virulence, in which case, they may be under greater selection pressure due to unpredictable aspects of the host environment [39]. Our ¢ndings point to the pldA locus being under less pressure to respond to changes in the gastric environment, despite challenges presented by a repertoire of host genetic polymorphisms and immune mechanisms. We suggest that the low rate of accumulation of mutations within pldA is perhaps indicative of a role in maintaining colonisation in the gastric mucosa rather than in chronicity of infection. To investigate the e¡ect of H. pylori strain-speci¢c variation in relation to clinical outcome of infection, sequence conservation within pldA was also investigated. As might be predicted from the PCR-RFLP analyses, we found a relatively high degree of sequence conservation (low microdiversity) amongst strains with the same or allied pldA
23
genotypes, even when isolates were from geographically diverse origins and associated with di¡erent clinical presentations. Few deletions and insertions were evident in the nucleotide sequence alignments, and most variation was attributable to single nucleotide polymorphisms, of which the majority represented synonymous mutations in the PldA amino acid sequence. These ¢ndings appeared to be consistent with the 42.5-kDa PldA protein not having a strain-speci¢c role in disease severity. The possible e¡ect of such genotypic changes on the amino acid sequence of PldA and on PL expression and activity is not yet known. However, some variants may be clinically relevant as H. pylori has been shown to have higher PLA2 activity in vitro than other bacteria and there is evidence of greater hydrolysis of phospholipids from analysis of basal gastric aspirates of infected patients [40]. It may be that the expression of pldA is regulated di¡erently in di¡erent strains and an insight into this might be obtained in the future by measurement of PldA activity. To determine if presence and variations within H. pylori pldA were linked to other putative markers of pathogenicity, results on cagA presence and vacA sm genotype, were compared with the pldA PCR-RFLP type. Certain strain types, such as cagA-positive and vacA s1m1, have been associated previously with increased host in£ammatory response. Although such interpretations are controversial [12,41], they support the notion that infection with a cagA-positive isolate increases the risk but does not reliably predict a clinically signi¢cant presentation [42]. Predictions are complicated as the predominant type of H. pylori in circulation may di¡er according to geographical region [43]. For instance, predominant genotypes de¢ned on iceA in combination with cagA status and vacA alleles in the USA were di¡erent from those in Japan/Korea and Columbia [42]. While we found no evidence that pldA, vacA allele and cagA status were either consistently associated or linked with severity of clinical presentation, a set of eight strains from geographically diverse locations with the pldA M4H3/cagA+/vacA s1m1 genotype were commonly associated with peptic ulcer disease. In addition, a second set of seven geographically diverse isolates that included strain Tx30a had an identical genotype (pldA.NT/cagA-/vacA s2m2). Clinical details were not available for all of these isolates but four were associated with non-ulcer dyspepsia, so they could represent a genomic variant at the lower end of the pathogenicity spectrum. In conclusion, pldA appears to be a conserved feature of the H. pylori genome with isolates of the same PCR-RFLP genotype having a widespread geographical distribution. While there was no evidence for strain-speci¢c associations between pldA RFLP or sequence type and gastroduodenal presentation, comparisons with other putative pathogenicity markers indirectly support a role for PldA activity in colonisation and persistence. Despite the ubiquity and intrinsic genomic diversity of H. pylori, our study has high-
FEMSIM 1357 10-12-01
24
J. Xerry, R.J. Owen / FEMS Immunology and Medical Microbiology 32 (2001) 17^25
lighted the existence of strains with common multiple genotypic traits, which could provide a focus for further work on strain-speci¢c pathogenicity determinants.
[20]
Acknowledgements
[21]
We wish to thank Dr A. Lastovica (South Africa), Dr Y. Akyon (Turkey) and Professor T. Wadstrom (Sweden) for providing isolates.
[22]
[23]
References [24] [1] Warren, J.R. and Marshall, B.J. (1983) Unidenti¢ed curved bacilli on gastric epithelium in active chronic gastritis. Lancet 1, 1273^1275. [2] Pounder, R.E. and Ng, D. (1995) The prevalence of Helicobacter pylori infection in di¡erent countries. Aliment. Pharmacol. Ther. 9, 33^39. [3] Blaser, M.J. (1998) Helicobacter pylori and gastric diseases. Br. Med. J. 316, 1507^1510. [4] Dunn, B.E., Cohen, H. and Blaser, M.J. (1997) Helicobacter pylori. Clin. Microbiol. Rev. 10, 720^741. [5] Danesh, J. and Whincup, M. (2000) Low grade in£ammation and coronary heart disease prospective study and updated meta-analysis. Br. Med. J. 321, 199^204. [6] Gasbarrini, A. and Cremonini, F. et al. (1999) The role of Helicobacter pylori in cardiovascular and cerebrovascular diseases. J. Physiol. Pharmacol. 50, 735^742. [7] Lin, T.T., Yeh, C.T., Wu, C.S. and Liaw, Y.F. (1995) Detection and partial sequence analysis of Helicobacter pylori DNA in bile samples. Dig. Dis. Sci. 40, 2214^2219. [8] Kerr, J.R., Al-Khatlaf, A., Barson, A.J. and Burnie, J.P. (2000) An association between sudden infant death syndrome (SIDS) and Helicobacter pylori. Arch. Dis. Child. 83, 429^434. [9] Birec, C., Grandhi, R. and McNeill, K. et al. (1999) Detection of Helicobacter pylori in oral aphthous ulcers. J. Oral Pathol. Med. 28, 197^203. [10] Charlett, A. and Dobbs, R.J. et al. (1999) Parkinsonism: siblings share Helicobacter pylori seropositivity and facets of syndrome. Acta Neurol. Scand. 1, 26^35. [11] Labigne, A. and de Reuse, H. (1996) Determinants of Helicobacter pylori pathogenicity. Infect. Agents Dis. 5, 191^202. [12] Atherton, J. (1998) H. pylori virulence factors. Br. Med. Bull. 54, 105^120. [13] McGee, D.J. and Mobley, H.L. (1999) Mechanisms of Helicobacter pylori infection : bacterial factors. Curr. Top. Microbiol. Immunol. 241, 155^180. [14] Atherton, J.C., Cover, T.L. and Twells, R.J. et al. (1999) Simple and accurate PCR-based typing of Helicobacter pylori vacA.. J. Clin. Microbiol. 37, 2979^2982. [15] Moran, A.P., Appelmelk, B.J. and Aspinall, G.O. (1996) Molecular mimicry of host structures by lipopolysaccharides of Campylobacter and Helicobacter spp: implications in pathogenesis. J. Endotoxin Res. 3, 521^531. [16] Censini, S., Langa, C. and Xiang, Z. et al. (1996) cag, a pathogenicity island of Helicobacter pylori, encodes type I-speci¢c and disease-associated virulence factors. Proc. Natl. Acad. Sci. USA 93, 14648^ 14653. [17] Songer, J.G. (1997) Bacterial phospholipases and their role in virulence. Trends Microbiol. 5, 156^161. [18] Titball, R.W. (1998) Bacterial phospholipases. J. Appl. Microbiol. 84, 1275^1375. [19] Ottlecz, A. and Romero, J.J. et al. (1993) Phospholipase activity of
[25]
[26]
[27]
[28] [29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
Helicobacter pylori and its inhibition by bismuth salts. Biochem. Biophys. Stud. Dig. Dis. Sci. 38, 2071^2080. Dorrell, N., Martino, M.C. and Stabler, R.A. (1999) Characterization of Helicobacter pylori PldA, a phospholipase with a role in colonization of the gastric mucosa. Gastroenterology 117, 1098^1104. Banatvala, N., Davies, G.R. and Abdi, Y. et al. (1994) High prevalence of Helicobacter pylori metronidazole resistance in migrants to East London: relation with previous metronidazole exposure and gastroduodenal disease. Gut 35, 1562^1566. Tombs, J.F., White, O. and Kerlavage, A.R. et al. (1999) The complete genome sequence of the gastric pathogen Helicobacter pylori. Nature 388, 539^547. Alm, R.A., Ling, L-S.L. and Moir, D.T. et al. (1999) Genomic-sequence comparison of two unrelated isolates of the human gastric pathogen Helicobacter pylori. Nature 397, 176^180. Wilson, K. (1987) Preparation of genomic DNA from bacteria, p.2.4.1-2.4.5. In: Current protocols in molecular biology (Ausubel, F.M., Brent, R., Kingston, R.E. D.D. Moore, J.A. Smith, J.G. Seidman, and K. Struhl Eds.). John Wiley and Sons, New York. Lorenz, E., Leeton, S. and Owen, R.J. (1997) A simple method for sizing large fragments of bacterial DNA separated by PFGE. CABIO 13, 485^486. Owen, R.J., Peters, T.M. and Varea, R. et al. (2000) Molecular epidemiology of Helicobacter pylori in England: prevalence of cag pathogenicity island markers and IS605 presence in relation to patient age and severity of gastric disease. FEMS Immunol. Med. Microbiol. 30, 65^71. Slater, E., Owen, R. and Williams, M. et al. (1999) Conservation of the cag pathogenicity island of Helicobacter pylori and associations with vacuolating cytotoxin alleles and IS605. Gastroenterology 117, 1308^1315. Covacci, A. and Telford, J.L. (1999) Helicobacter pylori virulence and genetic geography. Science 284, 1328^1333. Owen, R.J., Slater, E.R. and Xerry, J. et al. (1998) Development of a scheme for genotyping Helicobacter pylori based on allelic variation in urease subunit genes. J. Clin. Microbiol. 36, 3710^3712. van Doorn, L.J., Figueiredo, C. and Megraud, F. et al. (1999) Geographic distribution of vacA allelic types of Helicobacter pylori. Gastroenterology 116, 823^830. Hurtado, A., Owen, R.J. and Desai, M. (1994) Flagellin gene pro¢ling of Helicobacter pylori infecting symptomatic and asymptomatic individuals. Res. Microbiol. 145, 585^594. Peters, T., Owen, R.J. and Teare, L. (1998) Helicobacter pylori with identical vacA genotypes are associated with a variety of gastroduodenal pathologies in dyspeptic patients from one geographical area. Gut 45 (suppl. 111), 02/62. Figuerido, C., Quint, W.G. and Sanna, R. et al. (2000) Genetic organization and heterogeneity of the iceA locus of Helicobacter pylori. Gene 246, 59^68. Doig, P., Jonge, B.L. and Alm, R.A. et al. (1999) Helicobacter pylori physiology predicted from genomic comparison of two strains. Microbiol. Mol. Biol. 63, 675^707. Marais, A., Mendez, G.L. and Hazell, S.L. et al. (1999) Metabolism and genetics of Helicobacter pylori the genome era. Microbiol. Mol. Biol. 63, 642^674. Suerbaum, S., Smith, J.M., Bapumia, K., Morelli, G., Smith, N.H., Kunstmann, E., Dyrek, I. and Achtman, M. (1998) Free recombination within Helicobacter pylori. Proc. Natl. Acad. Sci. USA 95, 12619^12624. Wang, G., Humayun, M.Z. and Taylor, D. (1999) Mutation as an origin of genetic variability in Helicobacter pylori. Trends Microbiol. 7, 488^493. Achtman, M., Azuma, T. and Berg, D.E. et al. (1999) Recombination and clonal groupings within Helicobacter pylori from di¡erent geographical regions. Mol. Microbiol. 32, 459^470. Moxon, R.E., Rainey, P.B., Nawak, M.A. and Lenski, R.E. (1994)
FEMSIM 1357 10-12-01
J. Xerry, R.J. Owen / FEMS Immunology and Medical Microbiology 32 (2001) 17^25 Adaptive evolution of highly mutable loci in pathogenic bacteria. Curr. Biol. 4, 24^33. [40] Langton, S.R. and Cesares, S.D. (1992) Helicobacter pylori associated phospholipase A2 activity: a factor in peptic ulcer production. J. Clin. Pathol. 45, 221^224. [41] Parsonnet, J., Friedman, G.D. and Orentreich, N. et al. (1997) Risk for gastric cancer in people with CagA positive or CagA negative Helicobacter pylori infection. Gut 40, 297^301.
25
[42] Yamaoka, Y., Kodama, T. and Gutierrez, O. et al. (1999) Relationship between Helicobacter pylori iceA, cagA and vacA status and clinical outcome: studies in four di¡erent countries. J. Clin. Microbiol. 37, 2274^2279. [43] van Doorn, L.J., Figueiredo, Sanna, R., Blaser, M.J. and Quint, W.G. (1999) Distinct variants of Helicobacter pylori cagA are associated with vacA subtypes. J. Clin. Microbiol. 37, 2306.
FEMSIM 1357 10-12-01
www.fems-microbiology.org
Conservation and microdiversity of the phospholipase A (pldA) gene of Helicobacter pylori infecting dyspeptics from di¡erent countries Jacqueline Xerry, Robert J. Owen * Helicobacter Reference Unit, Laboratory of Enteric Pathogens, Central Public Health Laboratory, 61 Colindale Avenue, London NW9 5HT, UK Received 19 May 2001 ; received in revised form 24 September 2001; accepted 24 September 2001 First published online 30 October 2001
Abstract Phospholipase activity is important in bacterial pathogenicity and could contribute to the pathogenic role of Helicobacter pylori by degradation of the gastric mucus, and in maintaining long-term colonisation. Our aim was to determine the degree of variation in the phospholipase A gene (pldA) of H. pylori from different geographic locations, and to investigate links between pldA genotype and clinical disease severity, as well as with variation in cagA status and vacA genotypes. PCR-restriction fragment length polymorphism (RFLP) analysis with MboI and HaeIII was used to study 124 isolates from 10 countries that included the two genome-sequenced strains (26695 and J99), as well as Tx30a and NCTC 11637 (type strain). The 925-bp pldA fragment was amplified with a frequency of 90%. The presence of pldA was confirmed in the other strains using an alternative forward primer. Isolates were distinguished by PCR-RFLP analysis with 10 MboI and four HaeIII restriction patterns that combined to give 25 distinct pldA RFLP types. The pldA M2H2 strain genotype was most common (20%) in the UK but similar strains came from several other countries. Microdiversity was evident in pldA sequences of strains representing different RFLP types, and five M2H2 strains each had a distinct pldA sequence type. Intragenic variation was independent of gastric disease severity as well as strain cagA status and vacA genotype, with the exception of eight geographically diverse strains all with the pldA M4H3/cagA+/vacA s1m1 genotype predominantly from peptic ulcer patients. The study indicated a spectrum of genotypic variants and was supportive of a pldA function in H. pylori colonisation and persistence rather than in chronicity of infection. ß 2001 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords : Phospholipase A gene; Sequence diversity; PCR-RFLP; Duodenal ulcer; Dyspeptic; Geographic variation; Helicobacter pylori
1. Introduction Helicobacter pylori is a Gram-negative microaerophilic bacterium that is uniquely adapted to colonise and persist on surfaces of epithelial cells in the human gastric mucosa, particularly in the antral region of the stomach. Since H. pylori was ¢rst cultured in 1982 [1], it has been shown by serological studies to have a global distribution in human populations [2], and has become established as a major risk factor for the development of peptic ulceration, as well as gastric adenocarcinoma or gastric lymphoma in some individuals [3,4]. Possible links, but not direct proven causality, have been demonstrated between infection with H. pylori and several other extragastric disorders such as cardiovascular disease [5,6], gall bladder disease
* Corresponding author. Tel. : +44 (20) 8200 4400/3740; Fax: +44 (20) 8905 9929. E-mail address : [email protected] (R.J. Owen).
[7], sudden infant death syndrome [8], apthous ulcers [9], and idiopathic parkinsonism [10]. Various pathogenicity mechanisms that could be contributing to the role of H. pylori in peptic ulcer disease have been proposed [11^13]. Most widely investigated are vacuolating cytotoxin (vacA) activity [14], lipopolysaccharides (LPS) [15], and factors involved in cytokine (IL-8) induction mediated by genes located on the cag pathogenicity island (PAI) [16]. Degradation of the gastric mucosal barrier by phospholipase (PL) activity is another potential pathogenicity mechanism of H. pylori that may have an important role in establishing and maintaining long-term infection [17]. Bacterial PLs comprise a diverse group of proteins and their action may contribute to virulence of many bacterial pathogens, most notably Clostridium perfringens and Listeria monocytogenes [18]. H. pylori exhibits several PL activities (PLA1 , PLA2 , and PLC), which are thought to have a possible role in the degradation of phospholipid components of the mucosal acid barrier [19]. The H. pylori strain 26695 genome sequence contains an open
0928-8244 / 01 / $20.00 ß 2001 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII: S 0 9 2 8 - 8 2 4 4 ( 0 1 ) 0 0 2 7 5 - 9
FEMSIM 1357 10-12-01
18
J. Xerry, R.J. Owen / FEMS Immunology and Medical Microbiology 32 (2001) 17^25
reading frame (HP0499) encoding a putative protein of V42.5 kDa with homology to Escherichia coli outer membrane PL A (PldA) and encoding an active PL with lecithinase, PLA2 , and haemolytic activities [20]. Further evidence implicating PLs in virulence was indicated by recent work showing that a mutant had reduced ability to colonise gastric tissues in the murine model of disease [20] and that bismuth salts are inhibitors of PL activity [19]. In view of the possible contributory role of PL activity in the pathogenicity of H. pylori, we have investigated the prevalence and microdiversity of the PL A precursor gene (pldA) in isolates from patients resident in the UK and nine other countries in di¡erent continents. We examined links between the di¡erent pldA alleles and severity of clinical infection, as well as their associations with two key markers of pathogenicity (vacA and cagA), which may contribute in some cases to enhanced strain-speci¢c virulence.
assays were obtained from the National Collection of Type Cultures (NCTC, London, UK): NCTC 11637 (type strain; cagA+/vacA s1m1), NCTC 12455 (strain 26695; cagA+/vacA s1m1) [22], NCTC 12908 (strain G50, cagA-/vacA s1m2) [17], and NCTC 13081 (strain Tx30a, cagA-/vacA s2m2) where cagA+ indicates presence of the cag PAI and vacA an allelic type is de¢ned by signal (s) and mid (m) region form. A culture of genome-sequenced strain J99 (cagA+/vacA s1m1) [23] was kindly provided by Professor Diane Taylor (University of Alberta, AB, Canada). 2.2. Preparation of template DNA for PCR
2. Materials and methods
Genomic DNA was extracted from strains of H. pylori grown on 10% CBA and puri¢ed using the cetyltrimethylammonium bromide (CTAB) method [24]. Extracted DNA was dissolved in 50^100 Wl of sterile distilled water and stored at 4³C. DNA concentrations were established by absorbance readings at 230 nm, 260 nm and 280 nm.
2.1. Bacterial strains examined and growth conditions
2.3. PCR ampli¢cation of the H. pylori pldA gene
A set of 124 isolates of H. pylori from 10 di¡erent geographical locations was investigated. The isolates, which were received from various laboratories between 1990 and 1999, were mainly from patients in England (66 isolates) undergoing routine upper gastrointestinal endoscopic investigation for various presentations. Antral biopsy cultures from dyspeptic patients in the following nine other countries were also examined: South Africa (10 isolates), Turkey (9 isolates), Sweden (8 isolates), Canada (7 isolates), Italy (6 isolates), Australia (5 isolates), USA (2 isolates), Peru (1 isolate) and France (1 isolate). Nine isolates from Bangladeshi patients resident in east-London were also included in the study [21]. The strains of H. pylori were categorised in three main groups based on clinical severity of gastrointestinal disease observed during endoscopic diagnosis: the peptic ulcer group (PU, 45 isolates) including cases of duodenal ulcer [40] and/or gastric ulcer [5]; the non-ulcer dyspeptic gastritis group (NUD-gastritis, 38 isolates) including patients diagnosed as having gastritis and/or oesophagitis; and the group of dyspeptics with endoscopically normal mucosa (NUD-normal, 17 isolates). A fourth group (ND, no details) comprised 24 isolates from dyspeptics for whom no further clinical details were available. All isolates of H. pylori were cultured on 10% (v/v) Columbia blood agar (CBA)(Oxoid) and incubated at 37³C under microaerobic conditions (O2 4%, CO2 5%, H2 5%, N2 86%) for 3 days in a variable atmosphere incubator (Don Whitley Scienti¢c Ltd., Yorkshire, UK). Stock cultures were preserved on glass beads in Nutrient Broth (Oxoid) containing 10% (v/v) glycerol over liquid nitrogen or at 380³C. Reference cultures of H. pylori (in lyophilised form) used as controls for the various PCR
Primers for the pldA gene were designed from the HP0499 sequence of strain 26695 (http://www.tigr.org/ docs/tigr-scripts/CMR/report.sp/), and synthesised by MWG-BIOTECH (Milton Keynes, UK). The primers PLA8664 (5P-GTCCAATTCTTGGTATCTCC-3P) and PLA9731 (5P-AGGGTTTAAGCGTATTCCTACC-3P) were used to amplify a 925-bp fragment that represented 87% of the HP0499 sequence. For strain DNAs that failed to amplify pldA using the above primers, an alternative forward primer (PLAEX) was designed from alignments of nine sequences. This primer had the sequence 5PGTTGGCAGAACATTTCTAATGG-3P and when used in combination with PLA9731 gave a 516-bp fragment. Because of its small size this product was unsuitable for restriction fragment length polymorphism (RFLP) analysis. Ampli¢cation reactions were performed in a total volume of 50 Wl containing 100 ng of diluted DNA used as template DNA, 1.5 mM MgCl2 , 0.05 mM of each deoxynucleotide (dATP, dCTP, dGTP and dTTP), 0.4 mM of each oligonucleotide primer, 0.2 Wl (1 U) Taq polymerase (Life Technologies, Paisley, UK) and 5 Wl of 10Ubu¡er provided by the manufacturer. PCR ampli¢cation was performed in a Sprint thermal cycler (Hybaid Ltd, Ashford, UK). The PCR ampli¢cation conditions used comprised one initial cycle at 95³C, 5 min followed by 30 cycles at 94³C, 30 s; 60³C, 1 min; at ¢nally 72³C, 75 s. 2.4. Restriction digest analysis of pldA amplicon Restriction endonuclease digestion with MboI and HaeIII was performed separately on the ampli¢ed DNA product by adding to a ¢nal volume of 20 Wl: 1 Wl of spermidine, 2 Wl of the appropriate bu¡er, 1 Wl of restric-
FEMSIM 1357 10-12-01
J. Xerry, R.J. Owen / FEMS Immunology and Medical Microbiology 32 (2001) 17^25
tion enzyme, 10Wl of the PCR product (containing approximately 20 ng DNA), and sterile distilled water (6 Wl). The mixtures were incubated at 37³C and the digestion was stopped after 4 h by adding 5 Wl of stop-mix solution. A 17-Wl aliquot of the digested PCR product was loaded on a 3% w/v UltraPure agarose gel (Gibco BRL) and stained with ethidium bromide. Bands were viewed under an ultraviolet transilluminator, and DNA fragment sizes were estimated from migration distances using polynomial curve ¢tting functions [25]. 2.5. cagA and vacA genotyping Details of the primers and PCR conditions for the genotyping assays have been described elsewhere but brie£y these were as follows. PCR conditions for the two assays for cagA (a marker for the right-hand end of the cag PAI and for the cagI region) were as described for the F1/B1 primers and the D008/R008 primers [26,27]. Genotyping based on vacA (signal and mid-regions) were also performed as previously described [14]. 2.6. DNA sequence analysis of pldA gene
19
from 124 strains of H. pylori was determined by restriction digest analysis, and 4 and 10 di¡erent RFLP pro¢les were obtained with HaeIII (H-types) and MboI (M-types), respectively. The 516-bp amplicon was not subjected to RFLP analysis because of the small size and the strains were recorded as non-typable in the above scheme. Typical RFLP patterns for the 925-bp amplicon contained between two and ¢ve fragments of 123^772 bp for the Mtypes, and between two and four fragments of 110^634 bp for the H-types. The estimated locations of the di¡erent fragments (ordered from the 5P to the 3P ends) are shown in Fig. 1, as derived from the pldA sequence of NCTC 12455 (strain 26695). The 3P end of the pldA amplicon was highly conserved for recognition sequences of both restriction endonucleases, giving a 291-bp fragment present in pro¢les of all H-types, and a 153-bp fragment present in most (8/10) of the M-types. Greater variation occurred in the middle region and at the 5P end, with di¡erences in one or more bands in the RFLP patterns. To increase discrimination between strains of H. pylori, because of the relatively low degree of variation detected by RFLP analysis within pldA, both sets of RFLP pro¢les were used to de¢ne a total of 25 combined RFLP types, of
pldA amplicons from three isolates representing each of the three most common RFLP types (M2H2, M3H2 and M4H3), and four strains of pldA type M2H2, with di¡erent genotypes (cagA/vacA) and/or geographical origin, were sequenced (MWG-BIOTECH, Milton Keynes, UK). Nucleotide sequences were aligned and analysed with BioEdit (http://www.mbio.ncsu.edu/RNaseP/info/ programs/BIOEDIT/bioedit.html). 3. Results 3.1. Presence of the pldA gene The predicted 925-bp pldA amplicon was obtained with the primer pair PLA8664/PLA9731 from 112 (90%) of the 124 isolates of H. pylori investigated. However, negligible/ no ampli¢cation products were obtained with those primers for DNA from 12 (10%) strains, which included the reference strain NCTC 13081 (strain Tx30a). The results indicated the atypical strains either did not contain pldA or that mutations were present in one or other of the primer sequences. Reduced annealing temperatures and di¡erent concentrations of MgCl2 did not improve the ampli¢cation and yield of product. Ampli¢cations were repeated using an alternative forward primer (PLAEX) and the predicted 516-bp product was obtained for each strain indicating that pldA was present. 3.2. Microvariation within the pldA gene RFLP analysis Intragenic variation within the 925-bp pldA amplicons
Fig. 1. Restriction fragment maps for MboI and HaeIII digests of H. pylori pldA to illustrate each of the main RFLP types. Map locations are based on the pldA sequences for strains 26695 and J99.
FEMSIM 1357 10-12-01
20
J. Xerry, R.J. Owen / FEMS Immunology and Medical Microbiology 32 (2001) 17^25
Table 1 Frequencies of the main pldA genotypes and reference strains of H. pylori Type
No. strains (n = 124)
Reference strain
M2H2 M3H2 M4H3 M4H2 M2H3 M1H3 M1H1 M1H2 M3H3 M4H1 M1H4 M3H1 M4H4 M6H2 M6H3 Uniques Non-typable
20 (16%) 11 (8%) 10 (8%) 9 (7%) 9 (7%) 8 (6%) 7 (5%) 7 (5%) 5 (4%) 4 (3%) 4 (3%) 2 (1%) 2 (1%) 2 (1%) 2 (1%) 10 (8%) 10 (9%)
NCTC NCTC NCTC H185 NCTC H173 NCTC NCTC H180 J99 H406 NCTC NCTC H289 H195 H1412 NCTC
a
cagA cag A
13090 11637 12384 13085 12455 13086
12908 12490
13081 (Tx30a)
14 8 9 6 7 3 4 3 3 4 3 1 2 1 1 7 3
cag A3
vacA type s1m1 s1m2
s2m2
6 3 1 3 2 5 3 4 2 0 1 1 0 1 1 3 9
8 4 8 3 3 2 2 0 3 4 3 0 1 1 0 5a 3a
4 0 1 2 2 3 2 4 1 0 0 0 0 1 0 1 7
8 7 1 4 4 3 3 3 1 0 1 2 1 0 2 4 2
One strain was a mixture of s1m1/s1m2 types.
tries were examined and results showed that the three most common pldA pro¢les (M2H2, M3H2 and M4H3) had a wide geographical distribution and were the main types in seven countries (Table 2). Strains that were non-typable using the PLA8664/PLA9731 primer pair were of diverse origins (¢ve countries), which included the UK and South Africa. Strains from the UK were the largest single set in the study (representing 53%) and 20% were genotype M2H2 although the other isolates showed a diverse range of types including four non-typables. By contrast, type M4H3 was the main isolate infecting the Bangladeshi patients resident in east-London. Strains from South Africa represented the second largest set in the study (10 strains) for which the predominant type was M4H3. The 24 isolates from other Western European Countries (France/ Italy/Sweden) and Turkey had a diverse range of genotypes, in particular the isolates from Italy and from Turkey. The 15 isolates from other parts of the world (USA, Canada, Peru and Australia) also had diverse genotypes although two from Australian were genotype M2H2.
which 12 types were a feature of two or more strains. Table 1 lists the frequencies of strains according to combined MH-type, and a suggested reference strain for each type. The three most frequently occurring combined pldA RFLP pro¢les, representing 33% of strains, were M2H2 (16%), M3H2 (8.8%) and M4H3 (8%). NCTC 11637 (taxonomic type strain of H. pylori) had the second most common pro¢le (M3H2). NCTC 12455 (genome-sequenced strain 26695) had the combined pro¢le M1H1, which was a feature of a small number (5.6%) of strains, and the experimentally observed MboI/HaeIII digest fragment pro¢le was in accordance with that predicted from the HP0499 gene sequence. Some 8% of the strains had unique pldA combined pro¢les. 3.3. Geographical distribution and main features of pldA strain types H. pylori originating predominantly from individuals in the UK with one or more isolates from nine other coun-
Table 2 Distribution of H. pylori pldA strain types by country of origin Country (no. strains)
Main type (no. strains)
UK (66) UK-Bangladeshi (9) South Africa (10) Turkey (9) Sweden (8) Canada (7) Italy (6) Australia (5) USA (2) Peru (1) France (1)
M2H2 (13) M4H3 (3) M4H3 (4) M4H2 (2) M2H2 (2)/M3H2(2) diverse diverse M2H2 (2) M4H1 M3H2 M4H3
No. other types 17 5 2 6 4 5 6 3 ^ ^ ^
FEMSIM 1357 10-12-01
Non-typable isolates (n = 12) 4 ^ 4 1 ^ 2 ^ ^ 1 ^ ^
J. Xerry, R.J. Owen / FEMS Immunology and Medical Microbiology 32 (2001) 17^25
3.4. Associations between pldA strain type and disease Isolates of H. pylori were categorised in three broad patient groups based on severity of gastrointestinal disease reported at endoscopy. Analysis of the strain types in relation to disease shows that overall the di¡erent pldA genotypes were widely distributed across each group (Table 3). The peptic ulcer group (PUD, 45 isolates) contained the highest number of strains (36.3%), with the majority from DU patients. Associated genotypes were diverse with type M4H3 representing the main single genotype (6/40, 15%). A variety of strain genotypes were likewise associated with the non-ulcer dyspepsia patients, but the M2H2 genotype was the most common (18%) in both the gastritis group (7/38 isolates) and the endoscopically normal group (3/17 isolates). 3.5. Associations between pldA type and other genomic markers The cagA status and vacA genotypes were determined for all 124 strains of H. pylori (Table 1). The majority (79/ 124, 64%) of the isolates were cagA+ and these were highly diverse with respect to pldA type. Furthermore, most strains were either vacA s1/m1(40%) or s1/m2 (37%), and both sm-types were also diverse with respect to pldA type. Interestingly, strains with the common M2H2 type were typically cagA + (14/20, 70%) and vacA s1 (16/20, 80%). An association also was evident between vacA s1/m1 and pldA type M4H3 for eight (8/ 10, 80%) strains, and between vacA s1/m2 and pldA type M3H2 for seven (7/11, 63.6%) strains. By contrast, the 12 non-typable strains (including strain Tx30a) were more typically genotype cagA- (9/12, 75%) and vacA m2 (9/12, 75%). When pldA type was used in combination with both cagA and vacA types, two distinct groups of strains could be de¢ned. First, genotype pldA M4H3/cagA+/vacA s1m1 was a feature of eight strains that included NCTC 12384 and seven of these were from patients with DU or GU although they were geographically diverse (BangladeshiUK, France, Italy, South Africa and UK). Second, genotype pldA NT/cagA-/vacA s2m2 was a feature of seven strains that included Tx30a and these were also geographically diverse (UK, Canada, South Africa, USA and Tur-
key). Where details were available, four patients had nonulcer in£ammatory disease, and only one patient had a DU. 3.6. Microvariation within the pldA gene DNA sequence analysis DNA sequences were determined for pldA amplicons of the three H. pylori reference strains representing the three predominant RFLP types (NCTC 13090, type M2H2 ; NCTC 11637, type M3H2; and NCTC 12384, type M4H3). The 925-bp amplicon sequences were obtained for NCTC 12384 and NCTC 13090 whereas only a partial sequence of 553 bp could be obtained for NCTC 11637 due to the presence of secondary structures at the end of a G-run. This problem was overcome by sequencing the product obtained with the PLAEX/PLA9731 primer pair. The published pldA sequence for strain J99 (type M4H1) was also included in the alignments, and all four sequences were compared with the HP0499 sequence for strain 26695 type (M1H1). The various pldA sequences had high homologies (more than 96%) yet each represented a unique allele (data not shown). To investigate relationships amongst strains of H. pylori with the same pldA RFLP type, a further four amplicons were sequenced from four strains (NCTC 11639, H210, H392 and H1431) of the most common type (M2H2), and these were compared with the sequence of the M2H2 reference strain (NCTC 13090). Each of the ¢ve pldA M2H2 sequences represented a unique allelic type. Single nucleotide polymorphisms were present at about 55 di¡erent positions (a frequency of 1 in 17 bases). The sequence of strain H210 terminated at bp position 852, and the strain clustered separately from the other M2H2 strains. Interestingly, all ¢ve M2H2 sequences di¡ered from the M1H1 sequence represented by strain 26695 at seven common locations. These were all single nucleotide polymorphisms except for the GGG to CAA triplet substitution at nucleotide positions 448^450, which was the location of the HaeIII cut-site de¢ning the H1/H2 types. The amino acid alignments (Fig. 2) showed that the latter di¡erences were non-synonymous substitutions altering the amino acid sequence from G (glycine) to Q (glutamine) with an additional 12 randomly distributed amino acid substitutions.
Table 3 Distribution of H. pylori pldA strain types by disease severity Disease group
Main type
PUD (45): DU (40) GU (4) DU+GU (1) NUD-gastritis (38) NUD-normal (17) No details (24)
M4H3 M1H1 M2H3 M2H2 M2H2 M2H2
(6) (2) (7) (3)/M4H2 (3) (5)
21
No. other types 15 2 0 18 9 12
FEMSIM 1357 10-12-01
Non-typable isolates (n = 12) 3 0 0 4 2 3
22
J. Xerry, R.J. Owen / FEMS Immunology and Medical Microbiology 32 (2001) 17^25
Fig. 2. Alignment of nine representative amino acid sequences deduced from 925-nucleotide sequences of H. pylori pldA. To obtain optimum alignments, a hyphen indicates the absence of an amino acid residue. Identical amino acids are indicated by a dot. The deduced amino acid sequence for H. pylori strain 26695 (NCTC 12455) is shown for reference purposes. The pldA genotype is given in parentheses.
4. Discussion The direct role of PL activity in the virulence of H. pylori is unclear but it has been suggested that PLA2 activity may degrade the phospholipid monolayer believed to be the cause of the hydrophobicity of the stomach lining [17]. Furthermore, recent studies demonstrated that the pldA gene from the H. pylori 26695 genome sequence encoded an active PL with lecithinase, PLA2 and haemolytic activities, for which a role was suggested in colonisation of
the gastric mucosa, and possibly in tissue damage after colonisation [20]. As the latter study focussed on just two strains (26695 and SS1) from patients in the UK and Australia respectively, our aim was to obtain a broader understanding of the prevalence and pathogenic role of pldA by investigating its presence and diversity in isolates from UK patients with di¡erent gastric diseases as well as in isolates from other geographical origins. We found pldA was a highly conserved feature of the H. pylori genome irrespective of geographical origin. The re-
FEMSIM 1357 10-12-01
J. Xerry, R.J. Owen / FEMS Immunology and Medical Microbiology 32 (2001) 17^25
sults were consistent with analyses of other putative pathogenicity-associated genes of H. pylori, such as cagA, urease subunit genes (ureA/ureB), the vacuolating cytotoxin gene (vacA), and the £agellin gene (£aA), which also have global distributions in strains from unrelated patients [26^ 32]. The iceA locus was also reported to be present at a high frequency (99%) in a worldwide collection of 321 strains [33]. A subgroup of 12 of our isolates (10%) were divergent within the forward primer region so were nontypable by the present scheme although pldA could be detected in those isolates using a modi¢ed forward primer. Comparisons of the complete genome sequences of two strains (26695 and J99) of H. pylori, provide a wealth of information about organisation, gene order and predicted proteomes [23,34,35]. Although the two sequenced isolates contained between 6 and 7% of genes speci¢c to each strain [23], the potential for inter-strain genome variation within common regions is enormous, as H. pylori appears to have a non-clonal (panmictic) population structure due to frequent recombination [36] and possibly to mutation [37]. Multilocus sequence typing provides only limited evidence of recent clonal grouping [38] and is not feasible for investigating large numbers of strains. In contrast, PCRRFLP analysis is less discriminatory but simpler to perform and provides valuable insights on the degree and location of intragenic variation, particularly when used in conjunction with available sequence data. Overall, our PCR-RFLP analysis indicated only a moderate degree of microdiversity within pldA. Single enzyme digests yielded just four types for HaeIII and ten types for MboI. Even when using combined pro¢les, the low level of diversity within pldA contrasted markedly with the higher level in several other pathogenicity-associated genes. For example, 383 geographically diverse isolates gave 82 distinct ureA/B HaeIII pro¢le types [28]. Intragenic diversity of pldA was more consistent with levels of diversity observed within conserved housekeeping genes, such as the 23S rRNA and atpA genes (Owen, R.J., unpublished results). A low level of intragenic variation would be expected for genes encoding proteins subject to structural or functional constraints, unless they were associated with virulence, in which case, they may be under greater selection pressure due to unpredictable aspects of the host environment [39]. Our ¢ndings point to the pldA locus being under less pressure to respond to changes in the gastric environment, despite challenges presented by a repertoire of host genetic polymorphisms and immune mechanisms. We suggest that the low rate of accumulation of mutations within pldA is perhaps indicative of a role in maintaining colonisation in the gastric mucosa rather than in chronicity of infection. To investigate the e¡ect of H. pylori strain-speci¢c variation in relation to clinical outcome of infection, sequence conservation within pldA was also investigated. As might be predicted from the PCR-RFLP analyses, we found a relatively high degree of sequence conservation (low microdiversity) amongst strains with the same or allied pldA
23
genotypes, even when isolates were from geographically diverse origins and associated with di¡erent clinical presentations. Few deletions and insertions were evident in the nucleotide sequence alignments, and most variation was attributable to single nucleotide polymorphisms, of which the majority represented synonymous mutations in the PldA amino acid sequence. These ¢ndings appeared to be consistent with the 42.5-kDa PldA protein not having a strain-speci¢c role in disease severity. The possible e¡ect of such genotypic changes on the amino acid sequence of PldA and on PL expression and activity is not yet known. However, some variants may be clinically relevant as H. pylori has been shown to have higher PLA2 activity in vitro than other bacteria and there is evidence of greater hydrolysis of phospholipids from analysis of basal gastric aspirates of infected patients [40]. It may be that the expression of pldA is regulated di¡erently in di¡erent strains and an insight into this might be obtained in the future by measurement of PldA activity. To determine if presence and variations within H. pylori pldA were linked to other putative markers of pathogenicity, results on cagA presence and vacA sm genotype, were compared with the pldA PCR-RFLP type. Certain strain types, such as cagA-positive and vacA s1m1, have been associated previously with increased host in£ammatory response. Although such interpretations are controversial [12,41], they support the notion that infection with a cagA-positive isolate increases the risk but does not reliably predict a clinically signi¢cant presentation [42]. Predictions are complicated as the predominant type of H. pylori in circulation may di¡er according to geographical region [43]. For instance, predominant genotypes de¢ned on iceA in combination with cagA status and vacA alleles in the USA were di¡erent from those in Japan/Korea and Columbia [42]. While we found no evidence that pldA, vacA allele and cagA status were either consistently associated or linked with severity of clinical presentation, a set of eight strains from geographically diverse locations with the pldA M4H3/cagA+/vacA s1m1 genotype were commonly associated with peptic ulcer disease. In addition, a second set of seven geographically diverse isolates that included strain Tx30a had an identical genotype (pldA.NT/cagA-/vacA s2m2). Clinical details were not available for all of these isolates but four were associated with non-ulcer dyspepsia, so they could represent a genomic variant at the lower end of the pathogenicity spectrum. In conclusion, pldA appears to be a conserved feature of the H. pylori genome with isolates of the same PCR-RFLP genotype having a widespread geographical distribution. While there was no evidence for strain-speci¢c associations between pldA RFLP or sequence type and gastroduodenal presentation, comparisons with other putative pathogenicity markers indirectly support a role for PldA activity in colonisation and persistence. Despite the ubiquity and intrinsic genomic diversity of H. pylori, our study has high-
FEMSIM 1357 10-12-01
24
J. Xerry, R.J. Owen / FEMS Immunology and Medical Microbiology 32 (2001) 17^25
lighted the existence of strains with common multiple genotypic traits, which could provide a focus for further work on strain-speci¢c pathogenicity determinants.
[20]
Acknowledgements
[21]
We wish to thank Dr A. Lastovica (South Africa), Dr Y. Akyon (Turkey) and Professor T. Wadstrom (Sweden) for providing isolates.
[22]
[23]
References [24] [1] Warren, J.R. and Marshall, B.J. (1983) Unidenti¢ed curved bacilli on gastric epithelium in active chronic gastritis. Lancet 1, 1273^1275. [2] Pounder, R.E. and Ng, D. (1995) The prevalence of Helicobacter pylori infection in di¡erent countries. Aliment. Pharmacol. Ther. 9, 33^39. [3] Blaser, M.J. (1998) Helicobacter pylori and gastric diseases. Br. Med. J. 316, 1507^1510. [4] Dunn, B.E., Cohen, H. and Blaser, M.J. (1997) Helicobacter pylori. Clin. Microbiol. Rev. 10, 720^741. [5] Danesh, J. and Whincup, M. (2000) Low grade in£ammation and coronary heart disease prospective study and updated meta-analysis. Br. Med. J. 321, 199^204. [6] Gasbarrini, A. and Cremonini, F. et al. (1999) The role of Helicobacter pylori in cardiovascular and cerebrovascular diseases. J. Physiol. Pharmacol. 50, 735^742. [7] Lin, T.T., Yeh, C.T., Wu, C.S. and Liaw, Y.F. (1995) Detection and partial sequence analysis of Helicobacter pylori DNA in bile samples. Dig. Dis. Sci. 40, 2214^2219. [8] Kerr, J.R., Al-Khatlaf, A., Barson, A.J. and Burnie, J.P. (2000) An association between sudden infant death syndrome (SIDS) and Helicobacter pylori. Arch. Dis. Child. 83, 429^434. [9] Birec, C., Grandhi, R. and McNeill, K. et al. (1999) Detection of Helicobacter pylori in oral aphthous ulcers. J. Oral Pathol. Med. 28, 197^203. [10] Charlett, A. and Dobbs, R.J. et al. (1999) Parkinsonism: siblings share Helicobacter pylori seropositivity and facets of syndrome. Acta Neurol. Scand. 1, 26^35. [11] Labigne, A. and de Reuse, H. (1996) Determinants of Helicobacter pylori pathogenicity. Infect. Agents Dis. 5, 191^202. [12] Atherton, J. (1998) H. pylori virulence factors. Br. Med. Bull. 54, 105^120. [13] McGee, D.J. and Mobley, H.L. (1999) Mechanisms of Helicobacter pylori infection : bacterial factors. Curr. Top. Microbiol. Immunol. 241, 155^180. [14] Atherton, J.C., Cover, T.L. and Twells, R.J. et al. (1999) Simple and accurate PCR-based typing of Helicobacter pylori vacA.. J. Clin. Microbiol. 37, 2979^2982. [15] Moran, A.P., Appelmelk, B.J. and Aspinall, G.O. (1996) Molecular mimicry of host structures by lipopolysaccharides of Campylobacter and Helicobacter spp: implications in pathogenesis. J. Endotoxin Res. 3, 521^531. [16] Censini, S., Langa, C. and Xiang, Z. et al. (1996) cag, a pathogenicity island of Helicobacter pylori, encodes type I-speci¢c and disease-associated virulence factors. Proc. Natl. Acad. Sci. USA 93, 14648^ 14653. [17] Songer, J.G. (1997) Bacterial phospholipases and their role in virulence. Trends Microbiol. 5, 156^161. [18] Titball, R.W. (1998) Bacterial phospholipases. J. Appl. Microbiol. 84, 1275^1375. [19] Ottlecz, A. and Romero, J.J. et al. (1993) Phospholipase activity of
[25]
[26]
[27]
[28] [29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
Helicobacter pylori and its inhibition by bismuth salts. Biochem. Biophys. Stud. Dig. Dis. Sci. 38, 2071^2080. Dorrell, N., Martino, M.C. and Stabler, R.A. (1999) Characterization of Helicobacter pylori PldA, a phospholipase with a role in colonization of the gastric mucosa. Gastroenterology 117, 1098^1104. Banatvala, N., Davies, G.R. and Abdi, Y. et al. (1994) High prevalence of Helicobacter pylori metronidazole resistance in migrants to East London: relation with previous metronidazole exposure and gastroduodenal disease. Gut 35, 1562^1566. Tombs, J.F., White, O. and Kerlavage, A.R. et al. (1999) The complete genome sequence of the gastric pathogen Helicobacter pylori. Nature 388, 539^547. Alm, R.A., Ling, L-S.L. and Moir, D.T. et al. (1999) Genomic-sequence comparison of two unrelated isolates of the human gastric pathogen Helicobacter pylori. Nature 397, 176^180. Wilson, K. (1987) Preparation of genomic DNA from bacteria, p.2.4.1-2.4.5. In: Current protocols in molecular biology (Ausubel, F.M., Brent, R., Kingston, R.E. D.D. Moore, J.A. Smith, J.G. Seidman, and K. Struhl Eds.). John Wiley and Sons, New York. Lorenz, E., Leeton, S. and Owen, R.J. (1997) A simple method for sizing large fragments of bacterial DNA separated by PFGE. CABIO 13, 485^486. Owen, R.J., Peters, T.M. and Varea, R. et al. (2000) Molecular epidemiology of Helicobacter pylori in England: prevalence of cag pathogenicity island markers and IS605 presence in relation to patient age and severity of gastric disease. FEMS Immunol. Med. Microbiol. 30, 65^71. Slater, E., Owen, R. and Williams, M. et al. (1999) Conservation of the cag pathogenicity island of Helicobacter pylori and associations with vacuolating cytotoxin alleles and IS605. Gastroenterology 117, 1308^1315. Covacci, A. and Telford, J.L. (1999) Helicobacter pylori virulence and genetic geography. Science 284, 1328^1333. Owen, R.J., Slater, E.R. and Xerry, J. et al. (1998) Development of a scheme for genotyping Helicobacter pylori based on allelic variation in urease subunit genes. J. Clin. Microbiol. 36, 3710^3712. van Doorn, L.J., Figueiredo, C. and Megraud, F. et al. (1999) Geographic distribution of vacA allelic types of Helicobacter pylori. Gastroenterology 116, 823^830. Hurtado, A., Owen, R.J. and Desai, M. (1994) Flagellin gene pro¢ling of Helicobacter pylori infecting symptomatic and asymptomatic individuals. Res. Microbiol. 145, 585^594. Peters, T., Owen, R.J. and Teare, L. (1998) Helicobacter pylori with identical vacA genotypes are associated with a variety of gastroduodenal pathologies in dyspeptic patients from one geographical area. Gut 45 (suppl. 111), 02/62. Figuerido, C., Quint, W.G. and Sanna, R. et al. (2000) Genetic organization and heterogeneity of the iceA locus of Helicobacter pylori. Gene 246, 59^68. Doig, P., Jonge, B.L. and Alm, R.A. et al. (1999) Helicobacter pylori physiology predicted from genomic comparison of two strains. Microbiol. Mol. Biol. 63, 675^707. Marais, A., Mendez, G.L. and Hazell, S.L. et al. (1999) Metabolism and genetics of Helicobacter pylori the genome era. Microbiol. Mol. Biol. 63, 642^674. Suerbaum, S., Smith, J.M., Bapumia, K., Morelli, G., Smith, N.H., Kunstmann, E., Dyrek, I. and Achtman, M. (1998) Free recombination within Helicobacter pylori. Proc. Natl. Acad. Sci. USA 95, 12619^12624. Wang, G., Humayun, M.Z. and Taylor, D. (1999) Mutation as an origin of genetic variability in Helicobacter pylori. Trends Microbiol. 7, 488^493. Achtman, M., Azuma, T. and Berg, D.E. et al. (1999) Recombination and clonal groupings within Helicobacter pylori from di¡erent geographical regions. Mol. Microbiol. 32, 459^470. Moxon, R.E., Rainey, P.B., Nawak, M.A. and Lenski, R.E. (1994)
FEMSIM 1357 10-12-01
J. Xerry, R.J. Owen / FEMS Immunology and Medical Microbiology 32 (2001) 17^25 Adaptive evolution of highly mutable loci in pathogenic bacteria. Curr. Biol. 4, 24^33. [40] Langton, S.R. and Cesares, S.D. (1992) Helicobacter pylori associated phospholipase A2 activity: a factor in peptic ulcer production. J. Clin. Pathol. 45, 221^224. [41] Parsonnet, J., Friedman, G.D. and Orentreich, N. et al. (1997) Risk for gastric cancer in people with CagA positive or CagA negative Helicobacter pylori infection. Gut 40, 297^301.
25
[42] Yamaoka, Y., Kodama, T. and Gutierrez, O. et al. (1999) Relationship between Helicobacter pylori iceA, cagA and vacA status and clinical outcome: studies in four di¡erent countries. J. Clin. Microbiol. 37, 2274^2279. [43] van Doorn, L.J., Figueiredo, Sanna, R., Blaser, M.J. and Quint, W.G. (1999) Distinct variants of Helicobacter pylori cagA are associated with vacA subtypes. J. Clin. Microbiol. 37, 2306.
FEMSIM 1357 10-12-01