Chemotactic response of Helicobacter pylori to human ... - CiteSeerX

11 downloads 0 Views 109KB Size Report
1979; Sober et al., 1950) or through disease (Brassinne, 1974;. Jarnum & Jensen ..... transferrin, IgG, IgM) in Menetrier's disease (giant hypertrophic gastritis): ...
Journal of Medical Microbiology (2004), 53, 807–811

DOI 10.1099/jmm.0.45636-0

Chemotactic response of Helicobacter pylori to human plasma and bile Mulugeta L. Worku,1 Q. Najma Karim,1 John Spencer2 and Ramon L. Sidebotham3 Correspondence

1,3

Ramon L. Sidebotham [email protected]

Received 17 February 2004 Accepted 4 March 2004

2

Departments of Medical Microbiology1 and Gastroenterology3 , Imperial College School of Medicine at St Mary’s, Norfolk Place, London W2 1NY, UK Department of Surgery, Imperial College School of Medicine at Hammersmith, DuCane Road, London W12 0NN, UK

To clarify further the role of chemotaxis in Helicobacter pylori colonization, the in vitro bacterium response to human plasma and bile (secretions containing chemoeffector compounds that are present in the gastric mucus layer) was examined. Human plasma, after dilution to 1 % (v/v) with buffer, was found to be a chemoattractant for the motile bacillus. Human gall-bladder bile, after dilution to 2 % (v/v) with buffer, was found to be a chemorepellent, but did not cause the motility of the bacillus to be diminished after prolonged exposure. The basis of the chemoattractant effect of plasma was explored by examining how urea and 12 amino acids found in plasma affected the taxis of H. pylori. Urea and the amino acids histidine, glutamine, glycine and arginine were the strongest chemoattractants. Other amino acids were chemoattractants, with the exceptions of aspartic and glutamic acids, which were chemorepellents. The basis of the chemorepellent effect of bile was explored by examining how the six most abundant conjugated bile acids in human bile affected the taxis of H. pylori. All the bile acids were chemorepellents, with the greatest effects being demonstrated by taurocholic and taurodeoxycholic acids. The implications of these findings for H. pylori colonization of gastric epithelium are discussed.

INTRODUCTION Within the gastric microenvironment, Helicobacter pylori is found to be most frequently congregating at inter-epithelial junctions, and associating with the epithelial surface by cell attachment and mucus entrapment (Hazell et al., 1986; Thomsen et al., 1990). The incentive to identify the factors that control this distinctive pattern of colonization is considerable, since it has been shown that infection density at the epithelial surface correlates directly with the severity of gastritis and risk of peptic ulceration (Khulusi et al., 1995; Talamini et al., 1997). Freter (1980) has indicated that effective colonization of the epithelial surface only occurs when bacteria with normal motility and chemotaxis behaviour are directed towards the mucosa by an appropriate chemotactic stimulus. In the H. pylori-infected stomach, the important chemotactic gradients across the mucus layer, which draw the bacterium into the epithelial surface, have yet to be identified. However, they are likely to be formed from biliary surfactants refluxing into the stomach, and constituents of plasma transuding from gastric mucosa. Abbreviations: CLV, curvilinear velocity; CR, chemotactic response.

45636 & 2004 SGM Printed in Great Britain

The reflux of bile into the stomach post-prandially is a normal occurrence (King et al., 1984; Schindlbeck et al., 1987), and is reported to be increased in patients with gastroduodenal disease (DuPlessis, 1965; Sobala et al., 1993). In the stomach, refluxed bile does not remove (Bell et al., 1985), but diffuses through the protective mucus layer to alter (Dixon et al., 1986; DuPlessis, 1965) and eventually to be absorbed (Davenport, 1967; Ivey et al., 1970) by the underlying mucosa. Thus, H. pylori colonizing the gastric microenvironment (especially in the pyloric antrum) will be exposed to biliary surfactants and, more particularly, should exist within a bile concentration gradient that decreases from the luminal to the epithelial surface of the mucus layer. This gradient might be expected to draw the bacterium into the epithelial surface, since biliary surfactants are reported to act as chemorepellents for other bacteria (Hugdahl et al., 1988). The leakage or transudation of plasma from microcapillaries into interstitial spaces, and then into the lumen of the stomach, is a normal occurrence (Brassinne, 1979; Hollander & Horowitz, 1962), and is increased considerably when gastric mucosa is damaged experimentally (Brassinne, 1979; Sober et al., 1950) or through disease (Brassinne, 1974; Jarnum & Jensen, 1972). Thus, H. pylori colonizing the gastric microenvironment will be exposed to plasma, and 807

M. L. Worku and others

more particularly should exist within a plasma concentration gradient that decreases from the epithelial to the luminal surface of the mucus layer. This gradient might be expected to draw the bacterium into the epithelial surface, since amino acids found in plasma are reported to act as chemoattractants for other bacteria (Hugdahl et al., 1988; Mesibov & Adler, 1972; Moulton & Montie, 1979). In this preliminary communication we have tested these ideas, by examining how the movements of H. pylori in an in vitro chemotaxis system are affected by human plasma and bile and some of their constituents.

METHODS Gall-bladder bile was collected by routine puncture from a prospective (adult, male) surgical patient. The sample was sterilized by filtration through a 0.22 ìm syringe filter (Acrodiscs) and stored at 4 8C until required. Blood plasma was obtained from a healthy volunteer and used immediately. Conjugated bile acids, amino acids and urea were from Sigma. Bacteria. H. pylori was isolated from endoscopic biopsies of patients with duodenal ulcer or non-ulcer dyspepsia. The bacterium was initially plated out on 7 % (v/v) defibrinated horse blood agar, which included H. pylori-selective supplement (Oxoid), and incubated in a microaerobic environment (10 % CO2 , 18 % O2 and 72 % N2 ) within a Campypak gas jar (Oxoid) at 37 8C for 48 h. Colonies from these plates were suspended in sterile saline to a turbidity equivalent to that of McFarland’s No. 4 standard (108 bacteria ml 1 ). From this suspension, 100 ìl was transferred to 2.9 ml brain heart infusion broth (BBL) supplemented with 10 % (v/v) new born calf serum (Sigma) and H. pylori-selective supplement. The broth was then incubated with agitation in a 50 ml capacity loose-capped container (Bibby Sterilin) in a microaerobic atmosphere (see above) at 37 8C for about 48 h (bacilli) or over 72 h (cocci). Aliquots were withdrawn periodically for assessment of bacterial motility and viable count by the method of Miles et al. (1938). Gram stain, oxidase and catalase tests were used to confirm absence of contamination. Motility measurements. The culture (containing bacilli or cocci) was initially diluted with fresh medium to a concentration not exceeding 107 bacteria ml 1 . Aliquots (10 ìl) of the diluted culture were then added to 90 ìl of 100 mM phosphate/citrate buffer pH 6.5, or gallbladder bile [diluted to 2.2 % (v/v) with phosphate/citrate buffer]. The resulting bacterial suspension was drawn by capillary action into an optical microslide (CamLab) of 0.1 mm path length. The microslide was sealed at one end with vinyl plastic putty (Oxford Labware), transferred to the warm stage (37 8C) of a phase-contrast microscope (340 magnification) and allowed to equilibrate for 5 min before observations commenced with a Hobson BacTracker (Hobson Tracking Systems).

In this study, a single representative parameter was used to assess H. pylori motility: curvilinear velocity (CLV); the distance in ìm travelled along the path of the bacterium in each second between two stops. To ascertain the contribution of Brownian movement to bacterial motility, H. pylori was killed (no growth on horse blood agar) by exposure to 10 % (v/v) formalin for 10 min at room temperature, before examination with the BacTracker. Data collected with the BacTracker were analysed by Mann–Whitney U-test with an SPSS (Social Sciences Version 6) statistics package. A detailed description of the Hobson BacTracker and its operation may be found in the manufacturer’s literature. Chemotaxis measurements. The technique used to determine how

808

H. pylori accumulates in gradients formed by bile, plasma or their constituents was a modification of that described by Adler (1973). Microcapillary tubes of 10 ìl capacity (Sigma) were filled with blood plasma or gall-bladder bile, diluted to 1 and 2 % (v/v), respectively, in 100 mM phosphate/citrate buffer pH 6.5, or with blood plasma components (amino acids and urea) or selected bile acids, at concentrations of 1 and 2 mM, respectively, in phosphate/citrate buffer, and then sealed at the upper end with vinyl plastic putty (Oxford Labware). Control microcapillaries, used to assess background accumulations of bacteria, were filled with phosphate/citrate buffer alone. The open end of each microcapillary tube was inserted into an Eppendorf tube of 0.5 ml capacity, containing a suspension of the bacterium (bacilli or cocci) in buffer (300 ìl) at a concentration of 106 cells ml 1 , and incubated horizontally under microaerobic conditions for 45 min at 37 8C. After this time, the contents of the upper two-thirds of the tube were expelled onto a glass slide, and numbers of bacteria that had accumulated into the microcapillary counted with a Hobson BacTracker. Each assay was performed on seven different isolates (unless stated otherwise) and repeated at least six times. To normalize data from different experiments, results were expressed as the chemotactic response (CR), calculated as the number of bacteria in the test microcapillary, divided by the number of bacteria in the control microcapillary. Statistical analyses. The significance of results was tested using the Mann–Whitney U-test. Differences at the P , 0.05 level were considered to be statistically significant.

RESULTS Bacterial motility The motility of the H. pylori bacillus was not significantly diminished (when compared to that in phosphate/citrate buffer), when the bacterium was exposed to diluted gallbladder bile for up to 6 h (Fig. 1). Accumulation of bacilli and cocci in microcapillary tubes Bacillary and coccoidal forms of H. pylori were isolated from the same culture, and their accumulations in microcapillary tubes, containing phosphate/citrate buffer, 1 % blood plasma or 2 % gall-bladder bile, were compared under identical test conditions. Results are summarized in Fig. 2. The accumulation of the motile bacillary form of H. pylori in plasma was found to be significantly greater (P , 0.001) and in gallbladder bile was found to be significantly less (P , 0.01) than in buffer. In contrast, the accumulation of the non-motile, non-flagellate (Worku et al., 1999) coccoidal form of H. pylori was not significantly different in buffer, plasma or bile. Chemotactic response to human plasma, urea and some amino acids Human plasma, diluted to 1 % (v/v) in phosphate/citrate buffer, was found to be a significant chemoattractant for the H. pylori bacillus (Table 1). The responsiveness of the bacillus was also tested towards urea, which is a major constituent of plasma (Snook, 1986), and 12 amino acids that occur in human plasma (Stein & Moore, 1954). Results are sumJournal of Medical Microbiology 53

H. pylori chemotaxis to plasma and bile 40

Table 1. Chemotactic response of H. pylori to human plasma, human gall-bladder bile and selected constituents Chemotaxis was measured under microaerobic conditions with a modified Adler apparatus in which the control microcapillary contained 100 mM phosphate/citrate buffer, pH 6.5. Plasma and bile were diluted in buffer. Amino acids and urea were at a concentration of 1 mM, and bile acids were at a concentration of 2 mM in buffer. Data are results from seven different isolates. Abbreviations: GCA, glycocholic acid; TCA, taurocholic acid; GDCA, glycodeoxycholic acid; TDCA, taurodeoxycholic acid; GCDCA, glycochenodeoxycholic acid; TCDCA, taurochenodeoxycholic acid.

CLV (µm s⫺1)

30

20

Substrate 10

0 Buffer

Bile 1

Bile 2

Bile 3

Fig. 1. Mean CLV of H. pylori in 100 mM phosphate/citrate buffer, pH 6.5 (buffer), and after exposure to human gall-bladder bile diluted to 2 % (v/v) in buffer for 1 h (bile 1), 3 h (bile 2) and 6 h (bile 3) at 37 8C. The histogram shows mean results from six different isolates. Differences in mean CLV were without significance.

8⫻

104

Bacteria (ml⫺1)

6 ⫻ 104

Plasma 1 % (v/v) Bile 2 % (v/v) Plasma constituents Urea L-Alanine L-Arginine L-Aspartic acid L-Asparagine L-Glutamic acid L-Glutamine Glycine L-Histidine L-Leucine L-Proline L-Tyrosine L-Valine Bile acids GCA TCA GDCA TDCA GCDCA TCDCA

CR (mean 6 SD)

P value*

4.6  0.4 0.64  0.2

, 0.001 , 0.01

3.6  0.4 2.3  0.3 4.8  0.5 0.7  0.25 2.3  0.4 0.9  0.2 6.4  0.5 5.2  0.4 6.1  0.7 1.6  0.35 2.2  0.35 1.3  0.2 1.7  0.4

, 0.01 , 0.01 , 0.001 , 0.05 , 0.01

0.65  0.14 0.53  0.13 0.69  0.15 0.34  0.10 0.66  0.11 0.72  0.19

, 0.01 , 0.01 , 0.01 , 0.01 , 0.01 , 0.01

NS

, 0.001 , 0.001 , 0.001 , 0.05 , 0.01 NS

, 0.05

*Mann–Whitney U-test of means compared with control. significant.

NS,

not

4 ⫻ 104

marized in Table 1. Urea and all the amino acids (apart from aspartic acid and glutamic acid) acted as chemoattractants at a concentration of 1 mM in phosphate/citrate buffer.

2 ⫻ 104

Chemotactic response to human bile and some bile acids 0 Buffer

Plasma

Bile

Fig. 2. Accumulation of bacillary (filled bars) and coccoidal forms (open bars) of H. pylori in microcapillary tubes containing 100 mM phosphate/citrate buffer, pH 6.5, 1 % (v/v) plasma in buffer or 2 % (v/v) gall-bladder bile in buffer. Mean  SD CLV 32.1  4.7 ìm s 1 (bacilli) and 5.0  1.0 ìm s 1 (cocci). Mean CLV of bacilli killed by exposure to formalin solution 5.5  1.2 ìm s 1 . The histogram shows mean results from two different isolates. http://jmm.sgmjournals.org

Human gall-bladder bile, diluted to 2 % (v/v) in phosphate/ citrate buffer, was found to be a significant chemorepellent for the H. pylori bacillus (Table 1). The responsiveness of the bacillus was also tested towards the six most abundant conjugated bile acids in human bile (Lentner, 1981). Results are summarized in Table 1. All the bile acids were found to be significant chemorepellents at a concentration of 2 mM in phosphate/citrate buffer, with the strongest chemorepellent being taurodeoxycholic acid. 809

M. L. Worku and others

DISCUSSION H. pylori is a chemotactic bacterium in which sensor proteins on the cell membrane are coupled to the flagellar motors through two CheY response regulator proteins, three CheV proteins (of uncertain function) and a histidine kinase sensor CheA (Foynes et al., 2000; Pittman et al., 2001). The importance of this chemoreceptor complex for effective gastric colonization has been demonstrated in the failure of motile, but non-chemotactic deletion mutants to survive in an animal model (Foynes et al., 2000). In this study, we have sought to clarify further the role of chemotaxis in H. pylori colonization by examining how the bacterium responds in vitro to human plasma and bile, secretions containing chemoeffector compounds that are present in the gastric mucus layer. Comparative accumulation of bacillary and coccoidal forms of H. pylori in microcapillary tubes, containing buffer, plasma and gall-bladder bile, indicated that the motile bacillus responds chemotactically (in our test system) to both human plasma and bile. Highly diluted human plasma was found to be a significant chemoattractant for the bacillus, and highly diluted gall-bladder bile a significant chemorepellent. Bile is reported to be bactericidal for H. pylori, although at higher concentrations than those used in our experiments (Tompkins & West, 1987). Our finding, that the motility of the H. pylori bacillus is not significantly diminished on prolonged exposure to diluted bile, is therefore of importance, because it established that the chemorepellent effect of bile in our experiments was not simply an artefact due to loss of bacterial viability. The basis of the chemoattractant effect of diluted plasma was explored by examining the chemotactic behaviour of urea and several plasma amino acids. Measurements were conducted at a concentration of 1 mM, selected because it lies well below the mean concentration range for free amino acids [2.5–4 mM (Stein & Moore, 1954)] and for urea [5– 10.5 mM (Snook, 1986)] in human plasma, and was the amino acid concentration that most often gave an optimum chemotactic response with other bacteria (Mesibov & Adler, 1972; Moulton & Montie, 1979). Urea and the amino acids histidine, glutamine, glycine and arginine were the most potent chemoattractants amongst the plasma components tested. All the other amino acids were more moderate chemoattractants, except for aspartic and glutamic acids, which were chemorepellents. For most of the amino acids we could discern no relationship between the magnitude of the mean chemotactic response and the nature of the substituent group R in the generalized amino acid formula: R.CH.NH2 .COOH. Exceptions were the chemorepellents aspartic and glutamic acids, where R includes a carboxyl group. Our experiments suggest that the carboxylamide group, -CO.NH2 , might be an important binding structure for H. pylori receptors, because (i) conversion of aspartic and glutamic acids to their respective 810

amides turned chemorepellents into chemoattractants and (ii) this functional group is present in urea, one of the strongest chemoattractants tested. Our finding that arginine and urea are significant attractants for H. pylori agrees with earlier reports (Mizote et al., 1997; Cerda et al., 2003). The response of the bacterium to these two plasma constituents (and to bicarbonate ion) has recently been linked with expression of the HP0099 gene (Cerda et al., 2003). The basis of the chemorepulsive effect of diluted gall-bladder bile was explored by examining the chemotactic behaviour of the six most abundant conjugated bile acids in the secretion (Lentner, 1981). Measurements were conducted at a concentration of 2 mM, reflecting the mean concentration range for combined bile acids in human gall-bladder bile (Arianoff, 1966; Nakayama, 1967) after dilution to 2 % (v/v) in buffer. All the bile acids were significant chemorepellents for the H. pylori bacillus at this concentration, with the greatest effects being demonstrated by taurocholic and taurodeoxycholic acids. We are uncertain of the structural basis of this repellent action, although our findings with aspartic and glutamic acids (see above) suggest that there might be an involvement of free or esterified carboxyl groups in these surfactant molecules. Implications for gastric colonization The manner in which H. pylori colonizes the gastric microenvironment (Hazell et al., 1986; Thomsen et al., 1990) implies that chemotactic gradients, which draw the bacterium into the epithelial surface, are present across the mucus layer. Our in vitro experiments strengthen the view that an important chemotactic gradient across the mucus layer is formed from components of plasma that transude from microcapillaries in the gastric mucosa. They showed, for instance, that plasma is a significant chemoattractant for the H. pylori bacillus at a dilution greater than that which normally occurs in the gastric lumen [between two and four times, when based on a urea marker (Piper et al., 1967; Snook, 1986)], and therefore at a dilution greater than that which must occur at the epithelial surface of the H. pyloriinfected mucus layer. Our in vitro experiments also strengthen the view that an important chemotactic gradient across the mucus layer is formed from components of bile that reflux into the stomach. They showed, for instance, that bile acids are significant chemorepellents for the H. pylori bacillus at a concentration of 2 mM. This concentration is likely to be attained post-prandially by combined bile acids in the pyloric antrum (Han et al., 1996), although less so in the body of the normal acid-secreting stomach (Dixon et al., 1986; Schindlbeck et al., 1987) Bile and plasma chemotactic gradients across the gastric mucus layer may, therefore, be influential in directing Journal of Medical Microbiology 53

H. pylori chemotaxis to plasma and bile

H. pylori to the pyloric antrum, and in enabling this important pathogen to attain high population densities at the epithelial surface of the gastric mucus layer.

ACKNOWLEDGEMENTS We wish to thank Abbott Laboratories and the University of London for their gifts of equipment used in this work.

REFERENCES Adler, J. (1973). A method for measuring chemotaxis and use of the

method to determine optimum conditions for chemotaxis by Escherichia coli. J Gen Microbiol 74, 77–91. Arianoff, A. A. (1966). Etude de la composition de la bile humaine normale et pathologique. T-sels billiares cholestero electrolytes. Acta Gastro-ent Belg 29, 733–738 (in French).

Jarnum, S. & Jensen, K. B. (1972). Plasma protein turnover (albumin, transferrin, IgG, IgM) in Menetrier’s disease (giant hypertrophic gastritis): evidence of non-selective protein loss. Gut 13, 128–137. Khulusi, S., Mendall, M. A., Patel, P., Levy, J., Badve, S. & Northfield, T. C. (1995). Helicobacter pylori infection density and gastric inflamma-

tion in duodenal ulcer and non-ulcer subjects. Gut 37, 319–324. King, P. M., Adam, R. D., Pryde, A., McDicken, W. N. & Heading, R. C. (1984). Relationships of human antroduodenal motility and transpy-

loric fluid movement: non-invasive observations with real-time ultrasound. Gut 25, 1384–1391. Lentner, C. (1981). Bile. In Geigy Scientific Tables: Units of Measurement,

Body Fluids, Composition of the Body, Nutrition, vol. 1, pp. 134–146. Edited by C. Lentner. Basle: Ciba-Geigy. Mesibov, R. & Adler, J. (1972). Chemotaxis toward amino acids in Escherichia coli. J Bacteriol 112, 315–326. Miles, A. A., Misra, S. S. & Irwin, J. O. (1938). The estimation of the

bactericidal power of blood. J Hyg 38, 732–748.

Bell, A. E., Sellers, L. A., Allen, A., Cunliffe, W. J., Morris, E. R. & RossMurphy, S. B. (1985). Properties of gastric and duodenal mucus: effects

Mizote, T., Yoshiyama, H. & Nakazawa, T. (1997). Urease-independent

of proteolysis, disulfide reduction, bile, acid, ethanol and hypertonicity on mucus gel structure. Gastroenterology 88, 269–280.

chemotactic responses of Helicobacter pylori to urea, urease inhibitors, and sodium bicarbonate. Infect Immun 65, 1519–1521.

Brassinne, A. (1974). Gastric clearance of serum albumin in normal

Moulton, R. C. & Montie, T. C. (1979). Chemotaxis by Pseudomonas

man and in certain gastroduodenal disorders. Gut 15, 194–199. Brassinne, A. (1979). Effect of ethanol on plasma protein shedding in

the stomach. Dig Dis Sci 24, 44–47. Cerda, O., Rivas, A. & Toledo, H. (2003). Helicobacter pylori strain

ATCC700392 encodes a methyl-accepting chemotaxis receptor protein (MPC) for arginine and sodium bicarbonate. FEMS Microbiol Lett 224, 175–181. Davenport, H. W. (1967). Absorption of taurocholate-24-14 C through

the canine gastric mucosa. Proc Soc Exp Biol Med 125, 670–673. Dixon, M. F., O’Connor, H. J., Axon, A. T. R., King, R. F. J. G. & Johnson, D. (1986). Reflux gastritis: distinct histopathological entity? J Clin Pathol

39, 524–530. DuPlessis, D. J. (1965). Pathogenesis of gastric ulceration. Lancet i,

974–978.

aeruginosa. J Bacteriol 137, 274–280. Nakayama, F. (1967). Quantitative microanalysis of bile. J Lab Clin Med

69, 594–609. Piper, D. W., Fenton, B. H. & Goodman, L. R. (1967). Lactic, pyruvic,

citric, and uric acid and urea content of human gastric juice. Gastroenterology 53, 42–48. Pittman, M. S., Goodwin, M. & Kelly, D. J. (2001). Chemotaxis in the

human gastric pathogen Helicobacter pylori: different roles for CheW and the three CheV paralogues, and evidence for CheV2 phosphorylation. Microbiology 147, 2493–2504. Schindlbeck, N. E., Heinrich, C., Stellaard, F., Paumgartner, G. & Muller-Lissner, S. A. (1987). Healthy controls have as much bile reflux

as gastric ulcer patients. Gut 28, 1577–1583.

Foynes, S., Dorrell, N., Ward, S. J., Stabler, R. A., McColm, A. A., Rycroft, A. N. & Wren, B. W. (2000). Helicobacter pylori possesses two CheY

Snook, J. (1986). Urea/creatinine ratio and gastrointestinal haemor-

response regulators and a histidine kinase sensor, CheA, which are essential for chemotaxis and colonization of the gastric mucosa. Infect Immun 68, 2016–2023.

Sobala, G. M., O’Connor, H. J., Dewar, E. P., King, R. F. G., Axon, A. T. R. & Dixon, M. F. (1993). Bile reflux and intestinal metaplasia in gastric

Freter, R. (1980). Prospects for preventing the association of harmful

Sober, H. A., Hollander, F. & Sonnenblick, B. T. (1950). Response of

bacteria with host mucosal surfaces. In Bacterial Adherence: Receptors and Recognition, series B, vol. 6, pp. 441–458. Edited by E. H. Beachey. London: Chapman & Hall. Han, S. W., Evans, D. G., el-Zaatari, F. A. K., Go, M. F. & Graham, D. Y. (1996). The interaction of pH, bile and Helicobacter pylori may explain

rhage. Lancet ii, 400.

mucosa. J Clin Pathol 46, 235–240. gastric mucous barrier in pouch dogs to repeated topical application of eugenol. Am J Physiol 162, 120–130. Stein, W. H. & Moore, S. (1954). The free amino acids of human blood plasma. J Biol Chem 211, 915–926.

duodenal ulcer. Am J Gastroenterol 91, 1135–1137.

Talamini, G., Zamboni, G. & Cavallini, G. (1997). Antral mucosal

Hazell, S. L., Lee, A., Brady, L. & Hennessy, W. (1986). Campylobacter

Helicobacter pylori infection density as a risk factor of duodenal ulcer. Digestion 58, 211–217.

pyloridis and gastritis: association with intercellular spaces and adaptation to an environment of mucus as important factors in colonization of the gastric epithelium. J Infect Dis 153, 658–663. Hollander, F. & Horowitz, M. I. (1962). Serum proteins in gastric mucus

Thomsen, L. L., Gavin, J. B. & Tasman-Jones, C. (1990). Relation of

Helicobacter pylori to the human gastric mucosa in chronic gastritis of the antrum. Gut 31, 1230–1236.

and other secretions. Implications in relation to the protein-losing enteropathies. Gastroenterology 43, 75–83.

Tompkins, D. S. & West, A. P. (1987). Campylobacter pylori, acid and

Hugdahl, M. B., Beery, J. T. & Doyle, M. P. (1988). Chemotactic behavior

Worku, M. L., Sidebotham, R. L., Walker, M. M., Keshavarz, T. & Karim, Q. N. (1999). The relationship between Helicobacter pylori motility,

of Campylobacter jejuni. Infect Immun 56, 1560–1566. Ivey, K. J., DenBesten, L. & Bell, S. (1970). Absorption of bile salts from

human gastric mucosa. J Appl Physiol 29, 806–808.

http://jmm.sgmjournals.org

bile. J Clin Pathol 40, 1387.

morphology and phase of growth: implications for gastric colonization and pathology. Microbiology 145, 2803–2811.

811