fulltext

UMEÅ UNIVERSITY ODONTOLOGICAL DISSERTATIONS
No. 82
ISBN 91-7305-537-9 ISSN 0345-7532
Edited by the Dean of the Faculty of Medicine and Odontology
Helicobacter pylori - Bacterial
Adherence and Host Response
Farzad O. Olfat
Department of Odontology,
Umeå University
Sweden 2003
©2003 by Farzad O. Olfat
ISBN 91-7305-537-9
Printed by Print & Media, Umeå, Sweden
In memory of my father
Table of contents
Abstract
List of papers
History
7
8
9
Source and Transmission of H. pylori
10
Host responses & disease outcome
Peptic ulcer
Gastric cancer
11
12
12
H. pylori infection parameters
13
H. pylori colonization factors
Urease
Flagella
13
13
14
H. pylori virulence factors
Vacuolating cytotoxin (VacA)
cag pathogenicity island (cag-PAI)
14
14
15
Adhesion
Adhesins
Blood group Antigen Binding Adhesin (BabA)
Sialic acid Binding Adhesin (SabA)
Adherence-associated LipoProteins (AlpA/AlpB)
H. pylori Adhesin A (HpaA)
16
18
18
19
20
20
Lipopolysaccharide (LPS)
22
Neutrophil activating protein (HP-Nap)
22
Aim of the present investigations
Result and discussion
24
26
Concluding remarks
Acknowledgments
References
32
34
36
Abstract
The gastric pathogen Helicobacter pylori infects more than half of the population
worldwide. H. pylori manage to establish persistent infection, which would be life-long if
not treated. In order to establish such an infection, this pathogen has to deal with the host
immune system. H. pylori has certain characteristics which make the bacteria less
announced to the host immune system. Additionally, for remaining in the harsh and acidic
environment of the stomach with peristaltic movements and a high frequency of turnover of
epithelial cells, H. pylori has developed different binding modes to structures present both
in the mucus and on the surface of gastric cells and also to extracellular matrix proteins.
Evidently, adhesion has a determinant role for a successful colonization by H. pylori. It has
been shown that a small fraction of the H. pylori infection is in intimate contact and
attached to the host epithelium. Despite its small proportion, this group maintains the
persistency of infection. As there is no suitable in vitro system to mimic the human stomach
for studies of H. pylori infection, we have developed the In Vitro Explant Culture technique
(IVEC). By using this model we could show that H. pylori use the Lewis b blood group
antigen to bind to the host gastric mucosa, during experimental conditions most similar to
the in vivo situation. Furthermore, we could show that the host tissue responses to the
bacterial attachment by expression of Interleukin 8 (IL-8), which will guide the
inflammatory processes. Interestingly, by inhibition of bacterial adhesion through receptor
competition i.e., by use of soluble Lewis b antigen, IL-8 production was hampered in the
IVEC system, which further validates the presence of a tight relation between bacterial
adhesion and induction of host immune responses. One of the inflammation signaling
cursors in vivo is the upregulated sialylated Lewis x (sLex) antigen, an inflammation
associated carbohydrate structure well established as a binding site for the selectin family of
adhesion molecules. We could show that during chronic gastric inflammation, which is
actually caused by the persistent H. pylori infection, the bacterial cells adapt their binding
mode, and preferentially bind to sLex, which will provide an even more intimate contact
with the host cells. This interaction is mediated by SabA, the H. pylori adhesin for
sialylated oligosaccharides/glycoconjugates. By employing red blood cells as a model we
could further demonstrate that SabA is identical to the “established” H. pylori
hemagglutinin. We could also show that SabA binds to sialylated glycolipids (gangliosides)
rather than glycoproteins on cell surfaces. Our result also revealed that SabA also binds to
and activates human neutrophils. Such effect was unrelated to BabA and the H. pylori
Neutrophil Activating Protein (HP-NAP), which were not directly involved in the
activation of neutrophils. Furthermore, phagocytosis of bacteria by neutrophils was
demonstrated to be mainly dependent on presence of SabA. Interestingly, HP-NAP showed
a possible role in guiding the bacterial adhesion during conditions of limited sialylation, i.e.
equivalent to mild gastritis, when the tissue would be less inflamed and sialylated. In
conclusion, H. pylori adhesion causes host tissue inflammation, then the bacteria will adapt
to the new condition and bind to epithelial cells in a tighter mode by synergistic activities of
BabA and SabA. Additionally, SabA bind to and activate human neutrophils, which will
exacerbate inflammation responses and cause damage to host tissue. Thus, BabA and SabA
are potential candidates to be targeted for therapeutic strategies against H. pylori and gastric
disease.
7
List of papers in this thesis:
I) Farzad O. Olfat, Erik Näslund, Jacob Freedman, Thomas Borén and Lars Engstrand.
Cultured Human Gastric Explants: A Model for Studies of Bacteria-Host Interaction During
Conditions of Experimental Helicobacter pylori Infection.
J. Infect Diseases 2002;186:423–7
II) Jafar Mahdavi*, Berit Sondén*, Marina Hurtig*, Farzad O. Olfat, Lina Forsberg,
Niamh Roche, Jonas Ångström, Thomas Larsson, Susann Teneberg, Karl-Anders Karlsson,
Siiri Altraja, Torkel Wadström, Dangeruta Kersulyte, Douglas E. Berg, Andre Dubois,
Christoffer Petersson, Karl-Eric Magnusson, Thomas Norberg, Frank Lindh,
Bertil B. Lundskog, Anna Arnqvist, Lennart Hammarström, Thomas Borén.
Helicobacter pylori SabA Adhesin in Persistent Infection and Chronic Inflammation.
Science 2002; 297:573-8
III) Farzad O. Olfat, Niamh Roche, Marina Aspholm, Berit Sondén, Lars Engstrand,
Susann Teneberg, Thomas Borén.
The Neuraminyllactose-Binding Hemagglutinin of Helicobacter pylori is Identical to the
Sialic Acid Binding Adhesin, SabA.
In manuscript
IV) Christoffer Petersson, Maria Forsberg*, Marina Aspholm*, Farzad O. Olfat,
Tony Forslund, Thomas Borén and Karl-Eric Magnusson.
Helicobacter pylori Evokes a Strong Inflammatory Response in Human Neutrophils by SabA
Adhesin Mediated Selectin-Mimicry.
In manuscript
*These authors contributed equally to this work.
Reprints were made with permission from the publishers.
8
History:
Helicobacter pylori was first cultured two decades ago in Robin Warren and Barry
Marshall’s lab by an unintentionally prolonged incubation of bacterial culture
plates during Easter holidays. This microorganism was initially called
Campylobacter pyloridis, but the name was changed to Helicobacter pylori in
1989, as it was shown that this bacterium did not belong to the Campylobacter
genus (4).
Though a newcomer to scientific research field, this organism was observed by
histopathologists as early as 1874 (8) both in human and dog stomachs (6). In
1906, Krienitz reported the presence of spiral shaped bacteria in the stomach tissue
of gastric cancer patients (34). Similarly, in 1938 Doenges reported the presence
of spiral shaped organisms in autopsy tissue from human stomachs (14). However,
these early findings were generally considered to be a result of contamination and
ignored by the scientific community. A few years later studies did support the
presence of a spiral bacterium in the human stomach in fresh gastric biopsies from
patients suffering of gastric ulcer or carcinoma (25). However, due to its acidic
environment the human stomach was still considered to be aseptic. This dilemma
was finally overcome when H. pylori was cultured in Warren and Marshall’s
laboratory in 1982 (40,78). Efforts to establish an animal model for studying the
virulence of H. pylori were unsuccessful and to fulfill Koch’s postulates, Dr.
Marshall voluntarily ingested the H. pylori microorganism. As a result, Marshall
developed acute gastritis and proved that H. pylori not only infects the human
stomach but also induces inflammation in the gastric tissue (41).
9
Source and transmission of H. pylori:
H. pylori is postulated to have followed the human population around the world
during the last 12,000 years (23). Other more recent migration events such as rediscovery of America by the Spanish and the slave traffic between Africa and
America have further contributed to the generation of mixed bacterial populations.
The origin of H. pylori infection is still unclear though it is believed to be a
zoonosis originating from dogs and sheep (16). The main source of transmission
of H. pylori seems to be from person to person, as human and primates are the
only known reservoirs for H. pylori (5, 43). The frequency of infection among
children is high when other family members are infected, especially the mother
who increase the infection rate 5.3 times over the children with uninfected mothers
(17, 39). The possibility that infection is spread from child to child was ruled out
by an epidemiological study in Sweden, where it was shown that H. pylori is not
spread between infected and uninfected children in day-care centers (73).
Parsonnet et al. cultured H. pylori from the vomit and an aerosol diameter of 0.3
meters around all naturally infected adults tested (59) and Varoli et al. showed
presence of viable H. pylori in the gastric juice (77). In addition, Oral hygiene and
tooth loss has been suggested to affect the micro-flora in the gastric region, which
increases the possibility for the development of chronic diseases. The association
of poor dental health and dental plaque formation with esophageal and gastric
cancer has been shown earlier (79). H. pylori has been shown to reside in the oral
cavity and in dental plaque (47,68). Song et al. showed that 100% of the H. pylori
infected patients were positive for H. pylori DNA in their oral cavity.
Unfortunately, the number of subjects included in this study was only 20 (69).
Another study including a large number of patients (n=4500) showed an
association between periodontal disease and sero-prevalence of H. pylori. In
particular, periodontal pockets with a depth of 5 mm or more, strongly correlated
with H. pylori infection (18). Dental plaques are the result of a multitude of
10
bacterial layers that tightly adhere to each other. H. pylori was shown to adhere to
two bacterial species with plaque origin i.e. Fusobacterium nucleatum and
Fusobacterium periodonticum. Interestingly, H. pylori does not adhere to
Fusobacterium mortiferum and Fusobacterium ulcerans, which are two species
not found in plaque isolates (3). This may indicate that H. pylori likely binds to
bacterial species of plaque origin enhancing residence in the oral cavity. Whether
this localization of H. pylori can serve as a source of transmission or re-infection
of the same individual is still not clear. In the developing parts of the world also
water-born transmission of H. pylori infections are evident. In a study on Peruvian
children, a 3-fold higher risk of H. pylori infection was shown among children
limited to external water supply, compared to those with access to internal water
supplies (33). Here, H. pylori DNA could be detected in the drinking water (29),
although, no H. pylori was cultured from the water supplies. The environmental
prerequisites for H. pylori transmission might be more complex and H. pylori
survival was recently demonstrated in a water living amoeba, Acanthamoeba
castellanii, to suggest symbiotic spread of H. pylori by this amoeba (81).
Host responses and disease outcome:
Presence of H. pylori in the stomach has been associated with chronic antral
gastritis and peptic ulcers (58,60). As the bacteria enter the stomach and colonize
it, most of the bacteria are free living in the mucus and only a small portion of the
population attaches to the epithelial cells. The adherent part of the bacterial
population induces host inflammatory responses (28). The H. pylori associated
histological changes begin with acute gastritis, which within a few weeks,
progress to active chronic gastritis (13) and as its consequence, the host starts to
produce antibodies and cytokines in response to the infection. The levels of other
pro-inflammatory cytokines such as Interleukin (IL)-1ȕ, IL-6 and tumor necrosis
11
factor alpha (TNF-Į) are also increased (44,84). One of the main cytokines
produced by gastric epithelium in response to H. pylori infection is IL-8 (11). This
cytokine attracts neutrophils and monocytes to the site of infection. As a result of
cytokine production, the underlying tissue becomes heavily infiltrated by
lymphocytes and plasma cells (13). As this trafficking becomes more intense, the
lymphoid follicles are developed and cause the inflammation in the host tissue.
Peptic ulcer
Before discovery of H. pylori, the causes of peptic ulcer (gastric ulcer and
duodenal ulcer) were believed to be smoking, stress and excessive alcohol
consumption, which increased acid production and caused a painful condition.
This assumption raised the famous phrase: “No acid-No ulcer”. Even if this phrase
was correct per se, it did not explain the main cause of ulcer. The level of acid
production is different in a gastric ulcer versus a duodenal ulcer. In cases of gastric
ulcers, acid production remains normal and is even decreased in some patients. In
contrast, acid production is increased in duodenal ulcer patients (9). It has been
suggested that there are two major links between H. pylori infection and peptic
ulcer disease. First, the majority of peptic ulcer patients are infected by H. pylori
(64). Second, eradication of H. pylori infection heals the ulceration and in most
cases stops relapses of disease (75).
Gastric cancer
Gastric cancer is the third most common cause of cancer-related mortality with an
average of 700,000 deaths per year worldwide (55,56). In 1994 the International
Agency for Research on Cancer (IARC), a branch of the World Health
12
Organization (WHO), classified H. pylori infection as a class I carcinogenic
exposure (1). Most gastric cancer patients are sero-positive for H. pylori with the
majority being CagA-positive (20,80). However, the correlation between cagA and
cancer could not be shown in the Japanese population since 96.3% of all gastric
cancer associated H. pylori strains in Japan are cagA-positive (85). As not all H.
pylori infected individuals develop gastric cancer suggest that the cause of gastric
cancer is a combination of bacterial virulence factors, host genetic factors and
environmental influences.
H. pylori infection parameters
Several different characteristics and products of H. pylori have been described as
virulence factors. Considering the infection scenario, it is possible to distinguish
between colonization and virulence factors. The colonization factors are those that
are necessary for bacterial survival in the human at the very first step/acute
infection when the microbe first enters the stomach. The virulence factors are
those, which aid in persistence of the infection such as toxins and factors causing
tissue inflammation in the host.
H. pylori colonization factors:
Urease
The first and most crucial requirement for the bacterium as it enters the stomach is
to survive in the acidic milieu. For this purpose, H. pylori is equipped with the
urease enzyme and expression of this enzyme is a universal feature among all H.
pylori strains. This enzyme is a 550 kDa protein complex, which catalyzes the
13
hydrolysis of urea to CO2 and ammonia. The ammonia increases the pH in
periplasmic space and protects the bacteria from the acidic milieu. A urease
mutant was shown to be deficient in colonizing gnotobiotic piglets as compared to
the wild type (19) emphasizing the importance of this enzyme for survival and
successful colonization.
Flagella
Motility is important for colonization and is the result of rotation of the polar
structures named flagella. H. pylori has a unipolar set of 3-5 flagella which make
this bacterium highly motile (71). The flagellar filament is covered by a sheath,
which is an extension of the outer membrane and most probably protects the
flagella from acid-induced damage (26). Ottemann et al. showed that motility is
required for colonization of gastric mucosa of mice by H. pylori. A flagellated but
non-motile H. pylori mutant strain in comparison to its wild type strain was less
virulent in the infectious dose of the mouse animal model (53). Motility is not only
required for bacteria to reach the gastric mucus and escape the acidic surrounding,
but is also important for persistence of the H. pylori infection. The protecting
mucus layer on the gastric epithelium is always secreted toward the lumen and
motility is required for movement of the bacteria into the newly formed mucus
while the outer mucus layer is shed from the surface.
H. pylori virulence factors:
Vacuolating cytotoxin (VacA)
One of the major virulence factors of H. pylori is a Vacuolating cytotoxin, VacA.
This toxin causes cytoplasmic vacuolization in gastric epithelial cells (10). The
14
toxin effect on eukaryotic cells was first observed when a cell-free supernatant of
H. pylori growth culture induced vacuolization on cell culture monolayers (36).
The purified toxin from culture supernatants has low to no activity but becomes
activated at pH 5 and below. Furthermore, VacA binds to host cells surface
proteins prior to internalization (42). In a review, Papini et al. pointed out that if
the binding of VacA to the host cells is blocked, its toxic effects are reduced and
in addition, cell lines that do not express the binding sites for VacA are resistant to
this toxin (54). This may suggest that in parallel with low pH the binding site is
required for the toxin to introduce its toxic effects on host cells.
cag pathogenicity island (cag PAI)
One of the best-acknowledged virulence factors in H. pylori is the cytotoxinassociated pathogenicity island, cag-PAI. Expression of this virulence factor is
tightly associated with VacA expression and constitutes the type I strains of H.
pylori. However, even though these two virulence factors are not linked on the
chromosome, cag-PAI negative strains rarely express VacA, which was the reason
why it was named cytotoxin associated-PAI. The GC content of this island
indicates that the origin is not of H. pylori and most likely it has been acquired by
horizontal gene transfer. The cag-PAI is 40 kb in size and composed of 27 open
reading frames, one of which is the cagA. CagA expression has been considered as
a main marker for the presence of cag-PAI and is one of the most immunogenic
proteins of H. pylori. The cag-PAI contains additional immunogenic products,
which have been epidemiologically associated with ulcer disease and gastric
cancer (57). In addition, induction of IL-8 as a response to H. pylori infection by
the host tissue is dependent on the presence of the cag-PAI (11). Despite the
capability of the cag-PAI to induce the host immune responses, the presence or
absence of this island does not favor bacterial resistance to phagocytosis by PMN
15
cells or macrophages (49). The cag-PAI encodes a type IV secretion system,
which injects one or more effector protein/s into the host epithelial cells (50,66).
The only known effector protein so far is CagA (50). Directly after translocation
into the epithelial cells, CagA is phosphorylated on tyrosine residues (70).
Interestingly, Tanaka et al. showed that the location of phosphorylated CagA in
the epithelial cells is directly beneath the site of bacterial attachment (72).
Although cag-PAI is highly correlated to expression of IL-8 in the host tissue (11)
induction of IL-8 is not dependent on translocation of CagA. The cag-PAI genes
involved in translocation of CagA and induction of IL-8 production were
investigated by Fischer et al. (24). Most of the proteins involved in type IV
secretion are required for both translocation of CagA and induction of IL-8
production suggesting that an assembled type IV apparatus could activate the
signaling cascade in the host cells by binding to certain receptors. This activation
could induce the IL-8 production. Validation of this hypothesis remains for future
studies, but if so it would once more emphasize the involvement of bacterial
adhesion in induction of host inflammatory responses.
Adhesion:
In contrast to acute infections, the first step for most pathogens to successfully
establish a persistent infection is to colonize the targeted body site. Adhesion is
one of the determining factors for successful colonization of pathogens. The
importance of adhesion is emphasized when tropism of the pathogen is directed
towards cells in a turbid milieu.
A clear example of such infection is the uropathogenic E. coli, which colonizes
both the lower and upper urinary tracts (30). To establish colonization in this
tissue, the bacteria must first successfully maintain adherence to urinary tract
epithelium here, the binding affinity of E. coli must be greater than the high
16
pressure of fluid floating over the epithelial surface. Another turbulent area in the
human body is the gastrointestinal tract where, the peristalsis movement of the
gastric wall combined with the fast turn over rate of the epithelial cells demands
that the bacteria strongly adhere to the tissue to accomplish colonization. One of
the most common epitopes on the surface of gastro-intestinal epithelium is the
histo-blood group antigen Lewis b. This antigen is not only present on the surface
of mucosal tissue but also expressed in soluble glycoconjugates in saliva and tears
of 80% of the population. Whether binding of H. pylori to Lewis b antigen is a
result of bacterial adaptation or not, does not change the fact that this binding
enables H. pylori to be one of the most highly spread pathogens in humans
worldwide. The relation between the blood groups and development of gastric
disease has been demonstrated half a century ago, where it was described that the
incidence of peptic ulcer is higher among individuals of blood group O phenotype
(2). Moreover, for non-secretor individuals who lack the Lewis b antigen in body
secretions, the incidence of peptic ulcers is higher than for secretor individuals.
This indicates the importance of the Lewis b antigen in H. pylori infection. H.
pylori binds to a variety of different receptors on the host tissue and possibly
adapts its adhesin properties based on the conditions of the host tissue. Bacteria
colonizing the mucus will approach the epithelial lining and adhere to the Lewis b
antigen on the surface of the epithelial cells. This adhesion and interaction with
host cells induces inflammation and leads to sialylation of tissue cells. To further
enhance adhesion, the bacteria adapt to the new environmental conditions in the
tissue and express a new adhesin allowing it to bind to sialylated structures (38).
Such precise and programmed adaptations increase the “phenotypic fitness” of H.
pylori towards successful persistent infection.
17
Adhesins:
Blood group Antigen Binding Adhesin (BabA)
The BabA adhesin was first identified and characterized by use of a novel
technique named Re-Tagging (31). This outer membrane protein with a molecular
mass of 75 kDa is exposed on the surface of the bacterium (31, 61) but is not
present on the flagella sheath (31). By characterizing the N-terminal amino acid
sequence of this protein, the corresponding babA gene was identified.
Furthermore, two copies of gene, babA1 and babA2 were found in the type strain
CCUG 17875 used in this study. Inactivation of babA1 did not affect binding of
the bacteria to the Lewis b antigen, but when babA2 was deleted the bacteria no
longer bound the Lewis b antigen. The sequence of babA1 and babA2 revealed
that babA1 lacks a full translational start site and signal peptide. This finding was
later confirmed by Pride et al. who showed the same result in a worldwide
collection of H. pylori strains (62). Among clinical isolates, Ilver et al.
demonstrated that there is a high correlation between the type I strains (cagA and
VacA positive) and bacterial binding to Lewis b antigen. However, deletion of the
entire cag-PAI did not affect binding to the Lewis b antigen, which suggests that
the expression and activity of BabA is independent of cag-PAI (31). Additional
studies demonstrate that a distinct subpopulation of type I strains of H. pylori also
express BabA and bind to the Lewis b antigen. This group, defined as the “triple
positive strains” are highly associated with development of severe gastric disease
such as peptic ulcer and gastric cancer (27,63). BabA allows H. pylori to firmly
attach to the gastric lining through binding to the fucose residues of the Lewis b
antigen (7). The bacterial-host interaction introduced by BabA-Lewis b antigen
enables H. pylori to have a tight interaction with the epithelial cell surfaces, which
induces inflammatory responses (28). This would make the BabA one of the main
virulence factors of H. pylori for inducing inflammation in the gastric host tissue.
18
Sialic acid Binding Adhesin (SabA)
After characterization of BabA it was shown that a babA2 mutant had lost its
capacity to bind to the Lewis b antigen, but it could still bind to human gastric
mucosa in an in situ binding assay. To characterize this binding in detailed, both
the babA1 and the babA2 gene were deleted to eliminate the possibility of
chromosomal recombination leading to a functional BabA. Unexpectedly, the
combined babA1A2 mutant bound to inflamed human gastric tissue but not healthy
tissue suggesting that a receptor, not related to the Lewis b antigen, is present in
inflamed tissue. In addition, when the babA1A2 mutant and its parent strain were
pre-treated with soluble Lewis b antigen, the binding of the wild type strain was
decreased by 80% but binding of the babA1A2 mutant was not affected. The
binding mode of the babA1A2 mutant was analyzed by thin layer chromatography
overlay. In this system the human gangliosides were separated on silica gel and
overlayed with isotope labeled bacteria. The result showed that the babA1A2
mutant did not bind to Lewis b but to human glycophingolipids (GSLs) in a
sialidase sensitive manner. By using a Rhesus monkey model, the babA1A2
mutant was shown to bind to the gastric tissue in which the inflammatory
responses to infection were pronounced. Healthy gastric tissue of the same animal
treated with antibiotics and thus devoid of H. pylori infection, did not bind the
babA1A2 mutant in situ. Using a chromatogram binding assay, mass spectrometry
and nuclear magnetic resonance (NMR), the receptor for babA1A2 mutant was
identified to be a sialyl-dimeric-Lewis x antigen. Furthermore, when the wild type
and babA1A2 mutant were pre-incubated with soluble sialyl Lewis x antigen, the
binding of babA1A2 mutant to the inflamed tissue was significantly decreased
(90%), whereas the wild type J99 binding was unaffected. As soon as the receptor
was known, the bacterial adhesin for binding sialyl Lewis x antigen was purified
and characterized. The adhesin, SabA, is an outer membrane protein and consists
of 651 amino acids. The sabA gene is common among clinical isolates, but
19
expression of the adhesin is highly subject to phase variation. Taken together,
these results demonstrate that H. pylori infection induces inflammation in the host
tissue, which leads to an upregulation of sialyl Lewis x expression in the tissue.
This sialylation is used by the bacteria to obtain an even more intimate binding to
the host cells.
Adherence-associated LipoProteins (AlpA/AlpB)
The Adherence-associated LipoProteins, AlpA and AlpB, are outer membrane
proteins and are encoded by two different but closely homologous genes alpA and
alpB. These two proteins contain 518 amino acids each and are highly
homologous to each other (51). The genes are co-transcribed from one operon,
however, inactivation of alpB did not affect expression of alpA (51). Both proteins
may be necessary for proper adhesion of H. pylori to gastric tissue in vitro (48).
However, there is no clear data showing that AlpA and AlpB act as receptor
binding adhesins.
H. pylori Adhesin A (HpaA)
HpaA is the first bacterial protein suggested to be involved in H. pylori
hemagglutination activity and has been the subject of many studies due to the
controversial results by different groups. Evans et al. identified the hpaA gene (21)
and showed that the corresponding protein is expressed on the bacterial surface
and involved in the binding to N-acetylneuraminyl-2,3-lactose resulting in
characteristic hemagglutination activity (HA). In contrast, O’Toole and colleagues
proposed that HpaA is not exposed on the bacterial surface since a hpaA knockout mutant did not show any difference in binding to sialic acid compared to its
parent strain (52). A third study by Jones et al. showed by immuno-gold electron
20
microscopy that HpaA is a flagellar sheath protein and that HpaA is not an adhesin
for binding to a gastric cell line (AGS) neither needed for HA. In this study HpaA
was suggested to be a lipoprotein (32). In another effort to clarify these
contradictory results, Lundström and colleagues tested five different strains of H.
pylori for the presence of HpaA using electron microscopy. HpaA was found on
both the flagella and the cell surface of all tested strains (37). In addition, the
localization of HpaA was similar in reference strains CCUG 17874 and CCUG
17875. It is noteworthy that strain CCUG 17875, in contrast to CCUG 17874,
lacks the ability to cause sialic acid dependent hemagglutination (sia-HA) of red
blood cells (RBCs). These results confirm previous results obtained by O'Toole et
al. where the hpaA null mutation did not affect adherence to sialic acid suggesting
that H. pylori binding to sialic acid occurs independent of HpaA.
Schematic illustration of different Lewis antigens.
21
Lipopolysaccharide (LPS)
A common structure in the LPS of most H. pylori strains is fucosylated
oligosaccharides, which mimic the human Lewis antigens. Over 80% of H. pylori
isolates tested express Lewis x and/or Lewis y on their O-antigen chain of LPS (67)
and less than 5% contain Lewis a or Lewis b (82).
Expression of the host antigen on the surface could be mechanism by which the
bacterium evades the host immune responses (76). Wirth et al. suggested a
correlation between H. pylori Lewis expression and host Lewis phenotype (83).
Interestingly, this concept was strengthened by the observation that Helicobacter
mustelae, a natural gastric pathogen of ferrets, expresses blood group antigen A on
its surface in correlation with expression of the same antigen in its host (12).
Muotiala et al. showed that the LPS of H. pylori has much less immunological
activity in mice compared to that of other species. Lethal toxicity of H. pylori LPS
was 500-fold less than LPS from Salmonella (45). Evans et al. observed that H.
pylori LPS does not activate neutrophils (22). Taken together, the low toxicity of
the LPS and the expression of host antigens on the LPS, may enhance a persistent
infection by H. pylori.
Neutrophil activating protein (HP-NAP)
One of the primary host responses towards H. pylori infection is the recruitment of
neutrophils and monocytes to the site of infection. It has been shown that a
bacterial protein named H. pylori’s neutrophil activating protein (HP-NAP) is
chemotactic for the human neutrophils and causes translocation of neutrophils
from the blood stream to the infected stomach mucosa. In addition, HP-NAP binds
to the neutrophils and thereby initiates the activation of these cells. The napA gene
has been shown to be harbored by the majority of tested strains (22).
22
Determination of the size of this protein has yielded different results in different
studies (15,46,74). The definite size and structure of HP-NAP was established
when this protein was crystallized and its folded structure was determined (86).
The amino acid sequence of HP-NAP suggests that this protein is an iron-binding
protein (74) and homologous to the DNA protecting protein (Dps). Interestingly,
no changes in HP-NAP expression were observed when the bacteria were cultured
in iron-enriched or depleted media (74). This suggests that absence or presence of
iron does not regulate expression of HP-NAP. It has been shown that
immunization of mice with HP-NAP protects against H. pylori infection (65).
Furthermore, the frequency of HP-NAP positive strains was higher among clinical
strains isolated from peptic ulcer disease (35). These findings suggest a more
pathogenic role for the HP-NAP than only iron binding.
23
Aim of the present investigations:
Paper I:
The aim of this study was to develop a model for studying the interactions
between H. pylori and viable host gastric tissue. Other available systems for
studying H. pylori infections are various animal models, and cultured primary
gastric cells or cell lineages. However, none of these models are able to closely
mimic the natural environment of the human stomach. The results obtained by
these systems must be interpreted with caution, as the cell lineages have different
glycolysation patterns during their culture condition than in the tissue and most of
them do not produce mucosa and are thereby sensitive to the acidic conditions.
Therefore, we developed an in vitro model system, which closely mimics the in
vivo condition. The In Vitro Explant Culture (IVEC) model employs intact vital
human gastric tissue, which is responsive to H. pylori infection as measured by
induction of the IL-8 cytokine. Furthermore, we studied the effect of bacterial
adhesion on activation of the host responses in this model. We further investigated
the Lewis b mediated adhesion of H. pylori to the gastric explants and its influence
on induction of IL-8 production in the IVEC model.
Paper II:
Adhesion is a crucial factor for many pathogens for establishment of an infection.
H. pylori binds to the Lewis b antigen with high affinity by employing its adhesin
BabA. However, it was shown that a babA1A2 knock-out mutant strain still binds
to gastric tissue in an in situ binding assay. This binding only occurred when the
tissue was inflamed. The aim of this study was to identify and characterize the
24
corresponding adhesin, which mediates binding of H. pylori to the inflamed
gastric tissue.
Paper III:
Hemagglutination of RBCs caused by different microbes has been considered as a
marker that indicates the ability of the pathogen to adhere to host cells. In the case
of H. pylori, HA of RBCs occurs in a sialic acid dependent manner. In our study,
we aimed to characterize the hemagglutination activity of H. pylori and investigate
any relationships between the hemagglutinin activity and the sialic acid binding
adhesin, SabA, to identify the corresponding adhesin i.e. the hemagglutinin.
Paper IV:
H. pylori infection induces host immune responses and local inflammation in the
gastric mucosa. The local inflammation is a result of immigration of leukocytes
from blood stream to the site of infection. This attraction is initiated by different
cytokines produced by the gastric tissue. One class of major white blood cells that
is attracted to the site of infection by IL-8, are the neutrophils. This study was
designed to assess the effects of the H. pylori adhesins BabA and SabA, with or
without involvement of the H. pylori neutrophil activating protein (HP-NAP), on
stimulation of human neutrophils.
25
Results and discussions:
Paper I:
In the IVEC model, H. pylori adherence to the gastric tissue is mediated by the
Lewis b antigen, which is the most well-described receptor for H. pylori.
Interestingly, the bacteria multiplied and increased in number only if the explants
were vital. Furthermore, growth of bacteria was based on interaction with the
gastric tissue, as the cell culture medium alone, used in this system could not
sustain bacterial growth (Fig.1). As colonization is completely dependent of the
gastric explants, the role of the Lewis b antigen present on the gastric explants cell
surfaces in the colonization process was analyzed. To prevent bacteria from
binding to the Lewis b antigen present on the tissue, H. pylori was pre-treated with
soluble Lewis b glycoconjugate prior to contact with gastric explants.
Pre-treatment dramatically decreased the number of recovered colony forming
units (cfu) from gastric explants after 36 hours. In addition, IL-8 production from
these explants also showed a significant decrease indicating the relation between
bacterial load and the size of immune responses. The results also highlight the
importance of adhesion for bacterial colonization and also for induction of host
immune responses. By immuno-histochemistry, the site of IL-8 production was
localized to the epical site of cell surfaces. Interestingly, a patchy pattern of IL-8
distribution was found in the gastric epithelium that was similar to the patchy
mode of H. pylori colonization of the gastric tissue. This observation suggest that
adherence of H. pylori to gastric epithelial cells induces IL-8 production. In
conclusion, in this study we demonstrate that immune responses are induced in the
host tissue by H. pylori adhesion to Lewis b antigens located on the gastric tissue.
Moreover, the IVEC is highly suitable for the analysis of host-pathogen
interactions.
26
0,8
B
A
Abs(600nm)
0,6
0,4
0,2
0
0
3
6
9
12
24
36
48
0
3
6
9
12
24
36
48
Time(hours)
Figure 1: Growth curves of H. pylori in Brucella broth (A), and IVEC cell
culture medium (B), measured in three independent experiments.
Paper II:
In this paper, a Lewis b antigen independent binding mode was observed by strain
17875 babA1A2 deletion mutant. This mutant was further shown to adhere to the
gastric tissue of a H. pylori infected patient with gastritis. In contrast to the parent
strain, the babA1A2 mutant did not bind to healthy non-H. pylori infected gastric
tissue suggesting that attachment by the babA1A2 mutant is not dependent on the
regular BabA-Lewis b interaction. In addition, this BabA-independent binding was
correlated with tissue inflammation. The structure that the babA1A2 mutant binds
to on sialylated glycolipid (GLS) was purified and identified as sialyl-dimericLewis x. Further experiments showed that the binding pattern of the babA1A2
mutant matched the distribution of sialyl Lewis x antigen in sections of gastric
27
tissues. Then, a Rhesus monkey, naturally infected with H. pylori, was treated
with antibiotic to clear the infection. After bacterial eradication, gastritis in the
gastric tissue was decreased to the base line. Six months after eradication the sialyl
Lewis x antigen in the gastric tissue was shown to be localized in the gastric gland
region and absent from the surface of epithelium. The same pattern was observed
when the babA1A2 mutant bound to the same tissue in situ. This monkey was then
experimentally infected with defined H. pylori strains, which induced
inflammation in the gastric tissue by a pronounced lymphocyte infiltration. The H.
pylori-induced inflammation elevated expression of sialyl Lewis x in the tissue. As
a consequence of infection, the sialyl Lewis x was not only maintained in the
gastric glands but also present on the epithelial surface, which at this stage could
serve as attachment sites for the babA1A2 mutant strain. The binding analyses of
96 clinical isolates showed that binding to sialyl Lewis x is highly correlated with
presence of cagA, as 43% of cagA+ strains binds sialyl Lewis x but only 11% of
cagA- strains showed this binding characteristics. However, similar to Lewis b
binding, deletion of cag-PAI did not affect binding to the sialyl Lewis x. These
clinical isolates were shown to exhibit distinct differences in the detailed binding
specificities for various sialylated antigens and blood group antigens where, 43%
of the strains which binds to sialyl Lewis x antigens also binds to the related but
distinctly different sialyl Lewis a antigen (see illustrated Lewis antigens). In
comparison, none of the strains that lack the sialyl Lewis x binding properties,
managed to bind the sialyl Lewis a antigen, which show that these binding
properties are closely related. These data clearly show that H. pylori induces
inflammation in the gastric tissue. As a result of the inflammatory cascade, sialyl
Lewis x was over-expressed and enhances bacterial adhesion to the tissue. Finally,
the adhesin responsible for binding to sialyl Lewis x was identified by the ReTagging technique and named SabA. The obtained data was validated by an
isogenic mutant of the sabA gene, which totally lost its binding activity toward
sialylated antigens. This study not only identified a novel H. pylori adhesin but
28
also showed how this pathogen manipulates the inflammatory responses to benefit
the persistent infection.
Paper III:
Hemagglutination of RBCs has been used to estimate and verify virulence of
different pathogens for decades, HA is the result of binding of the microbe to
glycoprotein and/or glycolipid structures on the surface of RBCs. These structures
are presumably present also on the normally targeted tissues of the host and thus,
HA has been interpreted as a virulence factor. One of the major structures on the
RBCs surface is sialic acid. In sia-HA, the binding of a pathogen to sialic acid is
inhibited by enzymatic treatment of RBCs by neuraminidase, which cleaves the
sialylated residues on the cell surface. However, H. pylori also causes
hemagglutination (HA) independently of sialic acid and some strains even
demonstrate higher HA activity when the sialic acid residues have been
chemically removed. This might be due to exposure of core structures and branch
residues that the bacteria can reach in the absence of sialic acid residues. In this
study, we observed a high correlation between sialyl Lewis x binding and sia-HA
titers among Swedish clinical isolates. This observation attracted our attention as
no conclusive data has been provided to identify the H. pylori hemagglutinin. A
J99sabA mutant (see paper II) was shown to be devoid of sia-HA properties
compared to its parent strain. In parallel, a J99babA2 mutant did not show any
distinct changes in sia-HA activity compared to the wild type. This suggests that
BabA-Lewis b interactions do not support binding to RBCs and do not cause HA.
In addition, SabA could be isolated from solubilized outer membrane protein
extracts of H. pylori by absorption to untreated RBCs while, neuraminidase
treatment abolished such affinity purification. To further characterize the nature of
sialic acid-SabA interaction, we showed that this interaction occurs mainly
29
between gangliosides and SabA and that binding is enhanced by complex
gangliosides with an extended core chain. Collectively, we have identified SabA
as the true H. pylori hemagglutinin. We emphasize the importance of this study
since the identity of the H. pylori hemagglutinin has been disputed for a decade.
Paper IV:
H. pylori is well established to activate human neutrophils. In this study, we
analyzed how this phenomenon occurs and which bacterial proteins in particular
are involved. It has been postulated that H. pylori Nap (HP-NAP) is the direct
cause of this activation. We used bacterial strains J99 and the isogenic J99sabA,
J99babA, J99sabAbabA, and J99napA mutants and investigated the binding ability
of all strains to
125
I labeled Lewis b and sialyl Lewis x glycoconjugates. The
J99napA mutant bound to Lewis b and sialyl Lewis x to the same extent as the
wild type, indicating that HP-NAP has no binding activity toward Lewis b and
sialyl Lewis x. Furthermore, to study the neutrophil responses to the strains used
in this study, the total and intracellular production of reactive oxygen species
(ROS) in neutrophils was measured. Interestingly, the J99sabA and J99sabAbabA
mutant strains caused the lowest induction of this response, while the neutrophil
responses toward J99napA were at a higher level compared to the wild type. The
increase in the ROS integral values by J99napA may suggest that this protein may
moderate the neutrophil toxic releases. In parallel, it was shown that the J99 strain,
J99napA and J99babA mutant strains, but not J99sabA and J99sabAbabA strains,
caused the aggregation of neutrophils (similar to the HA by the RBCs), which
mainly depends on bacterial binding to sialylated antigens on the surface of
neutrophils. To further study the binding ability of these strains to sialylated
structures on the host cells, a sia-HA assay was employed. Interestingly, we could
show that when the RBCs was treated with neuraminidase in low concentrations, a
30
several fold drop of sia-HA activity was found for the J99napA mutant compared
to the less sensitive strains J99 and J99babA mutant. This result suggests that HPNAP may assists the SabA mediated binding of bacteria to sialylated antigens. It is
also likely that in conditions where the sialyl Lewis x availability is limited, HPNAP would facilitate and guide the bacterial attachment to host cells. Finally, it
was concluded that SabA is the main factor for H. pylori adhesion to and
activation of human neutrophils.
31
Concluding remarks
Detailed investigations on the molecular interactions between human pathogens
and host are often limited to the availability of representative model systems. The
ideal situation would be a system that mimics the natural conditions of the
infections in humans. However, studies are often performed in other hosts, such as
rodents, or on transformed cell lines that exhibit many limitations.
H. pylori colonizes a unique niche, i.e. the human stomach, which has a complex
histological composition and an acidic milieu. By establishment of the IVEC
system, we could study the interactions between H. pylori and the human gastric
tissue in high resolution. IVEC was shown to be reliable for studying adhesion,
infection, host immune responses. We could link H. pylori adhesion to human
gastric tissue immune responses and also demonstrated that H. pylori attachment
to the Leb antigen induces the host immune system and cause a significant
cytokine production.
To conclude, we have used several different models and techniques to address
questions like; why a BabA mutant that does not bind healthy tissues still binds to
inflamed tissues? Which host receptor does H. pylori bind to in the inflamed
gastric tissue, and which adhesin is mediating this interaction? We first
characterized the host receptor as sialyl Lewis x, which is an inflammation marker
in the host tissue. By employing a Rhesus monkey model, we clearly showed that
H. pylori infection induces extensive inflammation in the host gastric tissue
followed by an upregulation of sialyl Lewis x expression on the epithelial cells.
Furthermore, we discovered a novel bacterial adhesin SabA, which is responsible
for the binding to sialyl Lewis x and described its characteristics in details. When
the bacteria enter the human stomach, the tissue is healthy and does not express
sialyl Lewis x. In this phase of infection H. pylori takes advantage by binding to
the Lewis b antigen on the surface of the gastric epithelium, an interaction
mediated by the BabA adhesin. The close attachment of H. pylori to the tissue
32
causes inflammation in the host and thus, sialyl Lewis x become available in
excess. The bacteria will then be able to achieve an even more intimate attachment
to the host cells by employing another adhesin SabA. During the course of
colonization, IL-8 production is induced, which strongly attracts neutrophils to the
site of infection. As neutrophils reach the infected tissue, the bacteria detach,
probably as a result of down regulation of SabA, as expression of this adhesin is
subject to a high frequency of phase variation. In addition, we showed that SabA is
the main bacterial protein that activates human neutrophils. Activation of
neutrophils causes release of toxic components, which leads to extensive host
tissue damage. As the integrity of the tissue is disturbed, the gastric cells will leak
and release their contents. This is an important source of nutrition for H. pylori
and contributes to the persistence of infection in the tissue. In our studies, not only
a new H. pylori adhesin was discovered and characterized, but our results also
highlighted how an H. pylori induced inflammation serves this bacteria to persist
the infection. Although H. pylori hemagglutination has been described as early as
1988, the identity of the corresponding hemagglutinin has not been known. Here
we show that the uncharacterized H. pylori hemagglutinin is the SabA adhesin.
Collectively, our results have contributed to a better understanding of the
mechanism behind the H. pylori persistent infection and in addition presented a
novel in vitro model for studying the H. pylori-host interactions.
33
Suggestions/ perspectives for future research:
By employing the IVEC model, response/activation due to the interaction
between bacteria and host could now be analyzed by techniques such as DNA arrays.
Thus, gastric epithelium infected by H. pylori could be analyzed for transcriptional
activation in relation to non-infected explants, i.e., to reveal which genes are regulated in
the gastric mucosa due to H. pylori infection. This could be various defensins or Tollreceptors for the direct protection of host tissue, but could also be the activation of
glycosyltransferases associated with the inflammation activation responses of the local
tissue.
Furthermore, the bacterial genome itself could be globally investigated in the
IVEC model for detailed studies of the gene activation of the microbes themselves, i.e.,
which bacterial genes are induced due to bacterial adhesion to the host cells. This is of
particular interest since the Swedish CDC Centre (Smittskyddsinstitutet) has established
techniques for through-put spotting of microbial array-chips. In addition, a panel of
strains harboring different candidate genes and corresponding (adhesin) knock-out
mutant would contribute to better understanding of H. pylori infection and
characterization of virulence genes.
The initial screening of inflammation responses by the mutants used in the IVEC
model could then be applied to analyzes of inflammation and colonization patterns in the
Rhesus monkey model of H. pylori infection, i.e., the model used in Paper II. These
analyzes would also tightly relate to colonization efficiency and establishment of
persistent infection, which are already ongoing projects with respect to beneficiary effect
of probiotics in modulating the infectious load or activity of H. pylori.
The mechanism behind the contact dependent indentation phenomenon on the
RBC cell surfaces by H. pylori needs to be further investigated, since the disruption of
RBCs could be due to the activities of SabA in the manipulation of the host cell
cytoskeleton.
Taken together, our results has brought new insights to the mechanisms by which
the gastric tissue is damaged by the H. pylori pathogen, and in addition, provides
opportunities for future detailed analyzes of the host response activated due to microbial
adhesion.
Acknowledgments:
During all these years I have come cross many wonderful people, but I have no chance to
name them all. So forgive me for not mentioning everybody’s name but I love you all.
First of all I would like to express my gratitude to Thomas Borén who introduced me to
the field of H. pylori and support me with his scientific skills and for all the bright ideas
during these years.
Next I want to thank Lars Engstrand for accepting me in his group at SMI (CDC)
Stockholm, and providing me with possibility to do all my practical work in his lab and
for being such a relaxed boss.
I am grateful to Debra Milton for her critical reading of my thesis with a short notice.
I would also like to thank:
Pylori-pinglorna, Anna, Annelie, Britta, Christinorna; both Nilsson and Persson, Helen,
Karin, Kristina, Lena, Maria, Martin (sorry, grabben) and Sandra for creating such a
wonderful atmosphere in the group. For all their attention and caring during these years.
And specially for their support, coming up to Umeå for my defence, Thank you all,
And in particular: Lena for her kind support in the lab whenever I needed it.
Christina N, for being my best friend. We started working on H. pylori in Umeå as fresh
students at the very same day, sitting in the corridor waiting for Thomas! I am also
grateful for your critical comments on my research and thesis and also for all our nonscientific discussions. Tramakasee Maay ferrennd!!!
Christina P, Cryspan, for her friendship and her wonderful energy for arranging so many
good activities and also for being my punching bag!
Sophie, for all chatting moments in the corridors when everybody else was gone home.
Britt Marie for her 14:30 work-outs. Sture, Tomas, Ville, Annika, Elisabeth, Mona, and
all other staff in CDC centre (SMI). Ingegerd for her kind assistance during the last 3
years.
My group in Umeå, Anna A, Carina, Jafar, Inger and Rolf. For being such a nice people.
Susanne V, for her wonderful sense of humour.
Marina A, for our interesting scientific (and less scientific) discussions.
The Stockholm gänget:
Andreas for all get away in Stockholm nights, Barbara for being such a patient smashing
target in the badminton matches, Bugra for being such a good politician, despite he just
become a physician. Dilek for being a caring and warm-hearted friend, Janet for her
everyday mail with all the funny stuff, Joakim for sharing the interest of fast vehicles,
Liam for letting a mouse be resident in his flat a few days.
34
Christian Serrano (Killen) for inlines tours, tough badminton matches and his skilful help
with graphic design of the thesis and finally for feeling so much free to fall in sleep
everywhere (especially in the middle of parties).
Christina F. for being so serious and funny at the same time during 7-8 pm every
Monday.
Natasha, for giving me a new dimension of the game by her professional skills but more
importantly, by her kind and nice personality turning the badminton to a goodminton.
Carlos, for we always can count on you and Anna Karin for the demanding badminton
matches and teaching me dancing Bug!
Bio-klubben; Sven, Anna, Björn, Ulrika, and others, but most importantly Cryspan for
arranging those nice evenings.
Helena for being such a dear friend and for all our discussions about life with or without
H. pylori in it.
Dr. Annette for not only being a good friend but also a remarkable physician taking good
care of me, when needed.
Marjan for her kind concern about my work, my everyday life and for her valuable
friendship.
Johanna Björkman, you will never be forgotten. May your soul rest in peace.
Well, when I look back I feel how lucky I am to have such a good friends who make the
life much more fun. Thank you guys!
And now to my loving mother, for nothing of this had been possible if it was not for her
efforts. For all her caring, and Love. Simply being a mother, a wonderful one.
And also my younger brother and his family, for their support.
My dear grandmother for all her good thoughts and love.
And to my father, my only wish is that you were here with me today. I’m still missing
you a lot.
Finally to my wonderful daughter, Afsoon, who has given me the energy and motivation
to carry on. With her cute comments, calling Helicobacter, helicopter when she was
younger. Now in the school telling people that her father is a bakterie-jobbare!!
35
References:
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
1994. IARC working group on the evaluation of carcinogenic risks to humans:
schistosomes, liver flukes and Helicobacter pylori.
Aird, I., H. H. Bentall, J. A. Mehigan, and J. A. Roberts. 1954. The blood
groups in relation to peptic ulceration and carcinoma of colon, rectum, breast, and
bronchus; an association between the ABO groups and peptic ulceration. Br Med
J 4883:315-21.
Andersen, R. N., N. Ganeshkumar, and P. E. Kolenbrander. 1998.
Helicobacter pylori adheres selectively to Fusobacterium spp. Oral Microbiol
Immunol 13:51-4.
Anonymous. 1989. Campylobacter pylori becomes Helicobacter pylori. Lancet
2:1019-20.
Berkowicz, J., and A. Lee. 1987. Person-to-person transmission of
Campylobacter pylori. Lancet 2:680-1.
Bizzozero. 1892. Sulle ghiandale tubulari del tube gastroenterico e sui rapporti
dellero coll epithelio de rivestimento dellamucosa. Giornale dell'Accademia di
Medicina di Torino, anno LV 40: 205.
Borén, T., P. Falk, K. A. Roth, G. Larson, and S. Normark. 1993. Attachment
of Helicobacter pylori to human gastric epithelium mediated by blood group
antigens. Science 262:1892-5.
Bottcher, G. 1874.
Cover, T. L. 2001. H. pylori pathogenesis, p. 509-558. In E. A. Groisman (ed.),
Priciples of bacterial pathogenesis. Academic Press, New York.
Cover, T. L. 1996. The vacuolating cytotoxin of Helicobacter pylori. Mol
Microbiol 20:241-6.
Crabtree, J. E., J. I. Wyatt, L. K. Trejdosiewicz, P. Peichl, P. H. Nichols, N.
Ramsay, J. N. Primrose, and I. J. Lindley. 1994. Interleukin-8 expression in
Helicobacter pylori infected, normal, and neoplastic gastroduodenal mucosa. J
Clin Pathol 47:61-6.
Croinin, T. O., M. Clyne, and B. Drumm. 1998. Molecular mimicry of ferret
gastric epithelial blood group antigen A by Helicobacter mustelae.
Gastroenterology 114:690-6.
Dixon, M. F. 2001. Helicobacter pylori: Physiology and Genetic. ASM Press,
Washington D.C.
Doenges, J. L. 1938. Spirochetes in gastric glands of macacus rhesus and humans
without definite history of related disease. Proc Soc Exp Biol Med 38:536-538.
Doig, P., J. W. Austin, and T. J. Trust. 1993. The Helicobacter pylori 19.6kilodalton protein is an iron-containing protein resembling ferritin. J Bacteriol
175:557-60.
Dore, M. P., A. R. Sepulveda, H. El-Zimaity, Y. Yamaoka, M. S. Osato, K.
Mototsugu, A. M. Nieddu, G. Realdi, and D. Y. Graham. 2001. Isolation of
Helicobacter pylori from sheep-implications for transmission to humans. Am J
Gastroenterol 96:1396-401.
36
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
Drumm, B., G. I. Perez-Perez, M. J. Blaser, and P. M. Sherman. 1990.
Intrafamilial clustering of Helicobacter pylori infection. N Engl J Med 322:35963.
Dye, B. A., D. Kruszon-Moran, and G. McQuillan. 2002. The relationship
between periodontal disease attributes and Helicobacter pylori infection among
adults in the United States. Am J Public Health 92:1809-15.
Eaton, K. A., C. L. Brooks, D. R. Morgan, and S. Krakowka. 1991. Essential
role of urease in pathogenesis of gastritis induced by Helicobacter pylori in
gnotobiotic piglets. Infect Immun 59:2470-5.
Ekstrom, A. M., M. Held, L. E. Hansson, L. Engstrand, and O. Nyren. 2001.
Helicobacter pylori in gastric cancer established by CagA immunoblot as a
marker of past infection. Gastroenterology 121:784-91.
Evans, D. G., T. K. Karjalainen, D. J. Evans, Jr., D. Y. Graham, and C. H.
Lee. 1993. Cloning, nucleotide sequence, and expression of a gene encoding an
adhesin subunit protein of Helicobacter pylori. J Bacteriol 175:674-83.
Evans, D. J., Jr., D. G. Evans, T. Takemura, H. Nakano, H. C. Lampert, D.
Y. Graham, D. N. Granger, and P. R. Kvietys. 1995. Characterization of a
Helicobacter pylori neutrophil-activating protein. Infect Immun 63:2213-20.
Falush, D., T. Wirth, B. Linz, J. K. Pritchard, M. Stephens, M. Kidd, M. J.
Blaser, D. Y. Graham, S. Vacher, G. I. Perez-Perez, Y. Yamaoka, F.
Megraud, K. Otto, U. Reichard, E. Katzowitsch, X. Wang, M. Achtman, and
S. Suerbaum. 2003. Traces of human migrations in Helicobacter pylori
populations. Science 299:1582-5.
Fischer, W., J. Puls, R. Buhrdorf, B. Gebert, S. Odenbreit, and R. Haas.
2001. Systematic mutagenesis of the Helicobacter pylori cag pathogenicity island:
essential genes for CagA translocation in host cells and induction of interleukin-8.
Mol Microbiol 42:1337-48.
Freedburg, A. S., and L. E. Barron. 1940. The presence of spirochetes in
human gastric mucosa. Am J Dig Dis 7:443-445.
Geis, G., S. Suerbaum, B. Forsthoff, H. Leying, and W. Opferkuch. 1993.
Ultrastructure and biochemical studies of the flagellar sheath of Helicobacter
pylori. J Med Microbiol 38:371-7.
Gerhard, M., N. Lehn, N. Neumayer, T. Borén, R. Rad, W. Schepp, S.
Miehlke, M. Classen, and C. Prinz. 1999. Clinical relevance of the Helicobacter
pylori gene for blood-group antigen-binding adhesin. Proc Natl Acad Sci U S A
96:12778-83.
Guruge, J. L., P. G. Falk, R. G. Lorenz, M. Dans, H. P. Wirth, M. J. Blaser,
D. E. Berg, and J. I. Gordon. 1998. Epithelial attachment alters the outcome of
Helicobacter pylori infection. Proc Natl Acad Sci U S A 95:3925-30.
Hulten, K., S. W. Han, H. Enroth, P. D. Klein, A. R. Opekun, R. H. Gilman,
D. G. Evans, L. Engstrand, D. Y. Graham, and F. A. El-Zaatari. 1996.
Helicobacter pylori in the drinking water in Peru. Gastroenterology 110:1031-5.
Hultgren, S. J., S. Abraham, M. Caparon, P. Falk, J. W. St Geme, 3rd, and S.
Normark. 1993. Pilus and nonpilus bacterial adhesins: assembly and function in
cell recognition. Cell 73:887-901.
37
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
Ilver, D., A. Arnqvist, J. Ögren, I. M. Frick, D. Kersulyte, E. T. Incecik, D. E.
Berg, A. Covacci, L. Engstrand, and T. Borén. 1998. Helicobacter pylori
adhesin binding fucosylated histo-blood group antigens revealed by retagging.
Science 279:373-7.
Jones, A. C., R. P. Logan, S. Foynes, A. Cockayne, B. W. Wren, and C. W.
Penn. 1997. A flagellar sheath protein of Helicobacter pylori is identical to
HpaA, a putative N-acetylneuraminyllactose-binding hemagglutinin, but is not an
adhesin for AGS cells. J Bacteriol 179:5643-7.
Klein, P. D., D. Y. Graham, A. Gaillour, A. R. Opekun, and E. O. Smith.
1991. Water source as risk factor for Helicobacter pylori infection in Peruvian
children. Gastrointestinal Physiology Working Group. Lancet 337:1503-6.
Krienitz, W. 1906. Ueber das Auftreten von Spirochaetne vershiegener Form im
Magen-inhalt bei Carcinoma ventriculi. Dtsch Med Wochebschr 28:872-889.
Leakey, A., J. La Brooy, and R. Hirst. 2000. The ability of Helicobacter pylori
to activate neutrophils is determined by factors other than H. pylori neutrophilactivating protein. J Infect Dis 182:1749-55.
Leunk, R. D., P. T. Johnson, B. C. David, W. G. Kraft, and D. R. Morgan.
1988. Cytotoxic activity in broth-culture filtrates of Campylobacter pylori. J Med
Microbiol 26:93-9.
Lundstrom, A. M., K. Blom, V. Sundaeus, and I. Bolin. 2001. HpaA shows
variable surface localization but the gene expression is similar in different
Helicobacter pylori strains. Microb Pathog 31:243-53.
Mahdavi, J., B. Sonden, M. Hurtig, F. O. Olfat, L. Forsberg, N. Roche, J.
Angström, T. Larsson, S. Teneberg, K. A. Karlsson, S. Altraja, T. Wadström,
D. Kersulyte, D. E. Berg, A. Dubois, C. Petersson, K. E. Magnusson, T.
Norberg, F. Lindh, B. B. Lundskog, A. Arnqvist, L. Hammarström, and T.
Borén. 2002. Helicobacter pylori SabA adhesin in persistent infection and
chronic inflammation. Science 297:573-8.
Malaty, H. M., T. Kumagai, E. Tanaka, H. Ota, K. Kiyosawa, D. Y. Graham,
and T. Katsuyama. 2000. Evidence from a nine-year birth cohort study in Japan
of transmission pathways of Helicobacter pylori infection. J Clin Microbiol
38:1971-3.
Marshall. 1983. Unidentified Curved Bacilli on Gastric Epithelium in Active
Chronic Gastritis. The Lancet:1273.
Marshall, B. J., J. A. Armstrong, D. B. McGechie, and R. J. Glancy. 1985.
Attempt to fulfil Koch's postulates for pyloric Campylobacter. Med J Aust
142:436-9.
Massari, P., R. Manetti, D. Burroni, S. Nuti, N. Norais, R. Rappuoli, and J.
L. Telford. 1998. Binding of the Helicobacter pylori vacuolating cytotoxin to
target cells. Infect Immun 66:3981-4.
Mitchell, H. M., A. Lee, and J. Carrick. 1989. Increased incidence of
Campylobacter pylori infection in gastroenterologists: further evidence to support
person-to-person transmission of C. pylori. Scand J Gastroenterol 24:396-400.
Moss, S. F., S. Legon, J. Davies, and J. Calam. 1994. Cytokine gene expression
in Helicobacter pylori associated antral gastritis. Gut 35:1567-70.
38
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
Muotiala, A., I. M. Helander, L. Pyhala, T. U. Kosunen, and A. P. Moran.
1992. Low biological activity of Helicobacter pylori lipopolysaccharide. Infect
Immun 60:1714-6.
Namavar, F., M. Sparrius, E. C. Veerman, B. J. Appelmelk, and C. M.
Vandenbroucke-Grauls. 1998. Neutrophil-activating protein mediates adhesion
of Helicobacter pylori to sulfated carbohydrates on high-molecular-weight
salivary mucin. Infect Immun 66:444-7.
Nguyen, A. M., L. Engstrand, R. M. Genta, D. Y. Graham, and F. A. elZaatari. 1993. Detection of Helicobacter pylori in dental plaque by reverse
transcription-polymerase chain reaction. J Clin Microbiol 31:783-7.
Odenbreit, S., G. Faller, and R. Haas. 2002. Role of the alpA/B proteins and
lipopolysaccharide in adhesion of Helicobacter pylori to human gastric tissue. Int
J Med Microbiol 292:247-56.
Odenbreit, S., B. Gebert, J. Puls, W. Fischer, and R. Haas. 2001. Interaction
of Helicobacter pylori with professional phagocytes: role of the cag pathogenicity
island and translocation, phosphorylation and processing of CagA. Cell Microbiol
3:21-31.
Odenbreit, S., J. Puls, B. Sedlmaier, E. Gerland, W. Fischer, and R. Haas.
2000. Translocation of Helicobacter pylori CagA into gastric epithelial cells by
type IV secretion. Science 287:1497-500.
Odenbreit, S., M. Till, D. Hofreuter, G. Faller, and R. Haas. 1999. Genetic
and functional characterization of the alpA/B gene locus essential for the adhesion
of Helicobacter pylori to human gastric tissue. Mol Microbiol 31:1537-48.
O'Toole, P. W., L. Janzon, P. Doig, J. Huang, M. Kostrzynska, and T. J.
Trust. 1995. The putative neuraminyllactose-binding hemagglutinin HpaA of
Helicobacter pylori CCUG 17874 is a lipoprotein. J Bacteriol 177:6049-57.
Ottemann, K. M., and A. C. Lowenthal. 2002. Helicobacter pylori uses motility
for initial colonization and to attain robust infection. Infect Immun 70:1984-90.
Papini, E., M. Zoratti, and T. L. Cover. 2001. In search of the Helicobacter
pylori VacA mechanism of action. Toxicon 39:1757-67.
Parkin, D. M., F. I. Bray, and S. S. Devesa. 2001. Cancer burden in the year
2000. The global picture. Eur J Cancer 37 Suppl 8:S4-66.
Parkin, D. M., P. Pisani, and J. Ferlay. 1993. Estimates of the worldwide
incidence of eighteen major cancers in 1985. Int J Cancer 54:594-606.
Parsonnet, J., G. D. Friedman, N. Orentreich, and H. Vogelman. 1997. Risk
for gastric cancer in people with CagA positive or CagA negative Helicobacter
pylori infection. Gut 40:297-301.
Parsonnet, J., S. Hansen, L. Rodriguez, A. B. Gelb, R. A. Warnke, E. Jellum,
N. Orentreich, J. H. Vogelman, and G. D. Friedman. 1994. Helicobacter pylori
infection and gastric lymphoma. N Engl J Med 330:1267-71.
Parsonnet, J., H. Shmuely, and T. Haggerty. 1999. Fecal and oral shedding of
Helicobacter pylori from healthy infected adults. Jama 282:2240-5.
Peek, R. M., Jr., and M. J. Blaser. 1997. Pathophysiology of Helicobacter
pylori-induced gastritis and peptic ulcer disease. Am J Med 102:200-7.
Petersson, C., B. Larsson, J. Mahdavi, T. Boren, and K. E. Magnusson. 2000.
A new method to visualize the Helicobacter pylori-associated Lewis(b)-binding
39
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
adhesin utilizing SDS-digested freeze-fracture replica labeling. J Histochem
Cytochem 48:877-83.
Pride, D. T., R. J. Meinersmann, and M. J. Blaser. 2001. Allelic Variation
within Helicobacter pylori babA and babB. Infect Immun 69:1160-71.
Rad, R., M. Gerhard, R. Lang, M. Schoniger, T. Rosch, W. Schepp, I.
Becker, H. Wagner, and C. Prinz. 2002. The Helicobacter pylori blood group
antigen-binding adhesin facilitates bacterial colonization and augments a
nonspecific immune response. J Immunol 168:3033-41.
Rauws, E. A., W. Langenberg, H. J. Houthoff, H. C. Zanen, and G. N.
Tytgat. 1988. Campylobacter pyloridis-associated chronic active antral gastritis.
A prospective study of its prevalence and the effects of antibacterial and antiulcer
treatment. Gastroenterology 94:33-40.
Satin, B., G. Del Giudice, V. Della Bianca, S. Dusi, C. Laudanna, F. Tonello,
D. Kelleher, R. Rappuoli, C. Montecucco, and F. Rossi. 2000. The neutrophilactivating protein (HP-NAP) of Helicobacter pylori is a protective antigen and a
major virulence factor. J Exp Med 191:1467-76.
Segal, E. D., J. Cha, J. Lo, S. Falkow, and L. S. Tompkins. 1999. Altered
states: involvement of phosphorylated CagA in the induction of host cellular
growth changes by Helicobacter pylori. Proc Natl Acad Sci U S A 96:14559-64.
Simoons-Smit, I. M., B. J. Appelmelk, T. Verboom, R. Negrini, J. L. Penner,
G. O. Aspinall, A. P. Moran, S. F. Fei, B. S. Shi, W. Rudnica, A. Savio, and J.
de Graaff. 1996. Typing of Helicobacter pylori with monoclonal antibodies
against Lewis antigens in lipopolysaccharide. J Clin Microbiol 34:2196-200.
Song, Q., T. Lange, A. Spahr, G. Adler, and G. Bode. 2000. Characteristic
distribution pattern of Helicobacter pylori in dental plaque and saliva detected
with nested PCR. J Med Microbiol 49:349-53.
Song, Q., A. Spahr, R. M. Schmid, G. Adler, and G. Bode. 2000. Helicobacter
pylori in the oral cavity: high prevalence and great DNA diversity. Dig Dis Sci
45:2162-7.
Stein, M., R. Rappuoli, and A. Covacci. 2000. Tyrosine phosphorylation of the
Helicobacter pylori CagA antigen after cag-driven host cell translocation. Proc
Natl Acad Sci U S A 97:1263-8.
Suerbaum, S. 1995. The complex flagella of gastric Helicobacter species. Trends
Microbiol 3:168-70; discussion 170-1.
Tanaka, J., T. Suzuki, H. Mimuro, and C. Sasakawa. 2003. Structural
definition on the surface of Helicobacter pylori type IV secretion apparatus. Cell
Microbiol 5:395-404.
Tindberg, Y., C. Bengtsson, F. Granath, M. Blennow, O. Nyren, and M.
Granstrom. 2001. Helicobacter pylori infection in Swedish school children: lack
of evidence of child-to-child transmission outside the family. Gastroenterology
121:310-6.
Tonello, F., W. G. Dundon, B. Satin, M. Molinari, G. Tognon, G. Grandi, G.
Del Giudice, R. Rappuoli, and C. Montecucco. 1999. The Helicobacter pylori
neutrophil-activating protein is an iron-binding protein with dodecameric
structure. Mol Microbiol 34:238-46.
40
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
Van der Hulst, R. W., E. A. Rauws, B. Koycu, J. J. Keller, M. J. Bruno, J. G.
Tijssen, and G. N. Tytgat. 1997. Prevention of ulcer recurrence after eradication
of Helicobacter pylori: a prospective long-term follow-up study.
Gastroenterology 113:1082-6.
Wang, G., Z. Ge, D. A. Rasko, and D. E. Taylor. 2000. Lewis antigens in
Helicobacter pylori: biosynthesis and phase variation. Mol Microbiol 36:1187-96.
Varoli, O., M. P. Landini, M. LaPlaca, A. Tucci, R. Corinaldesi, G. F.
Paparo, V. Stanghellini, and L. Barbara. 1991. Presence of Helicobacter pylori
in gastric juice. Am J Gastroenterol 86:249.
Warren, R. J. 1983. Unidentified Curved Bacilli on Gastric Epithelium in Active
Chronic Gastritis. The Lancet:1273.
Watabe, K., M. Nishi, H. Miyake, and K. Hirata. 1998. Lifestyle and gastric
cancer: a case-control study. Oncol Rep 5:1191-4.
Webb, P. M., J. E. Crabtree, and D. Forman. 1999. Gastric cancer, cytotoxinassociated gene A-positive Helicobacter pylori, and serum pepsinogens: an
international study. The Eurogst Study Group. Gastroenterology 116:269-76.
Winiecka-Krusnell, J., K. Wreiber, A. von Euler, L. Engstrand, and E.
Linder. 2002. Free-living amoebae promote growth and survival of Helicobacter
pylori. Scand J Infect Dis 34:253-6.
Wirth, H. P., M. Yang, M. Karita, and M. J. Blaser. 1996. Expression of the
human cell surface glycoconjugates Lewis x and Lewis y by Helicobacter pylori
isolates is related to cagA status. Infect Immun 64:4598-605.
Wirth, H. P., M. Yang, R. M. Peek, Jr., K. T. Tham, and M. J. Blaser. 1997.
Helicobacter pylori Lewis expression is related to the host Lewis phenotype.
Gastroenterology 113:1091-8.
Yamaoka, Y., M. Kita, T. Kodama, N. Sawai, and J. Imanishi. 1996.
Helicobacter pylori cagA gene and expression of cytokine messenger RNA in
gastric mucosa. Gastroenterology 110:1744-52.
Yamaoka, Y., T. Kodama, M. Kita, J. Imanishi, K. Kashima, and D. Y.
Graham. 1998. Relationship of vacA genotypes of Helicobacter pylori to cagA
status, cytotoxin production, and clinical outcome. Helicobacter 3:241-53.
Zanotti, G., E. Papinutto, W. Dundon, R. Battistutta, M. Seveso, G. Giudice,
R. Rappuoli, and C. Montecucco. 2002. Structure of the neutrophil-activating
protein from Helicobacter pylori. J Mol Biol 323:125-30.
41
I
423
CONCISE COMMUNICATION
Cultured Human Gastric Explants: A Model for Studies of Bacteria-Host
Interaction during Conditions of Experimental Helicobacter pylori Infection
Farzad O. Olfat,1,2 Erik Näslund,3 Jacob Freedman,3
Thomas Borén,2,a and Lars Engstrand1,a
1
Swedish Institute for Infectious Disease Control, Solna, 2Department
of Odontology and Oral Microbiology, Umeå University, Umeå,
and 3Division of Surgery, Danderyd Hospital, Karolinska Institutet,
Stockholm, Sweden
The unique environment of the human stomach makes it difficult to establish representative
in vitro models for Helicobacter pylori that mimic the natural infection. The in vitro explant
culture (IVEC) technique is based on coculture of human gastric explants with H. pylori,
where bacteria-host interaction is studied on the basis of interleukin (IL)–8 secretion of the
explant tissue in response to infection. In this study, it was shown that H. pylori attachment
to gastric epithelial tissue was required for induction of representative inflammatory responses,
assessed here by IL-8 production. Furthermore, IL-8 production by the explant tissue in
response to H. pylori infection demonstrated that the explants were adequately responsive.
The IVEC technique for studies of the interplay between H. pylori and the human gastric
mucosa during conditions of experimental infections in vitro could add knowledge to our
understanding of the complex bacteria-host cross-talk in vivo.
Helicobacter pylori is a gram-negative, microaerophilic, rodshaped bacterium that colonizes the human gastric mucosa and
establishes a persistent infection. H. pylori is the causative factor of chronic gastritis and peptic ulcer disease, and a strong
association between H. pylori and the development of gastric
cancer has been shown elsewhere [1, 2]. The natural niche for
H. pylori infection (i.e., the human gastric mucosa) is difficult
to mimic in primary cell cultures. Several cell lineages have
been used to study H. pylori infection [3]. However, none of
them adequately reproduces the local conditions of the human
stomach (e.g., acidic environment, mucus production, and complex glycosylation patterns of the gastric epithelial cells).
The possibility of culturing human gastric mucosa in vitro
is an interesting approach to mimic in vivo conditions [4, 5].
In the human gastric mucosa, H. pylori has the opportunity to
Received 10 December 2001; revised 17 March 2002; electronically published 17 July 2002.
Presented in part: 4th International Workshop on Pathogenesis and Host
Response in Helicobacter Infections, Helsingør, Denmark, 6–9 July 2000
(abstract C-84).
Informed consent was obtained from all patients participating in this
study. Human experimentation guidelines of the Karolinska Institute Research Ethical Board were followed, and written approval of the study protocol was obtained from the board.
Financial support: Swedish Research Council (11218 [to T.B.], 10848 [to
L.E.], and 13150 [to E.N.]); Swedish Cancer Society.
a
T.B. and L.E. are senior authors of two research groups that made equal
contributions to this research.
Reprints or correspondence: Dr. Lars Engstrand, Swedish Institute for
Infectious Disease Control, SE-171 82 Solna, Sweden (Lars.Engstrand@
smi.ki.se).
The Journal of Infectious Diseases 2002; 186:423–7
䉷 2002 by the Infectious Diseases Society of America. All rights reserved.
0022-1899/2002/18603-0017$15.00
access its natural binding receptors, such as the fucosylated
blood group antigens [6], which allow the bacteria to adhere
to the gastric epithelial cells and thereby trigger host inflammatory responses [7]. The ability to adhere to host cells is a
major virulence factor, and a correlation between the presence
of the expressed form of the blood group antigen binding adhesin gene in H. pylori and disease outcome has been suggested
elsewhere [8]. In parallel, clinical studies have demonstrated
increased interleukin (IL)–8 concentration in gastric biopsy
specimens obtained from H. pylori–infected individuals [9]. In
addition, H. pylori infection has been shown to up-regulate the
expression of IL-8 in cultured cells [10]. In the present study,
the H. pylori reference strain, CCUG 17875, was studied for
its ability to colonize and induce IL-8 in a human gastric explant model. The adherence and colonization abilities of H.
pylori strain CCUG 17875 were compared with bacterial cells
for which binding to the gastric epithelium lining was competitively inhibited by pretreatment with the natural receptor,
soluble Lewis b blood group antigen. It was shown that attachment was required for induction of IL-8 production, here
representing the inflammatory response to H. pylori infection.
The in vitro explant culture (IVEC) technique may become a
useful model for studies of bacteria-host interactions.
Materials and Methods
Processing of biopsy specimens. Two tissue samples were obtained from the corpus region of the stomachs of patients undergoing a vertical banded gastroplasty operation because of morbid
obesity. The patients included in this study were all women, 25–43
years old. The doughnut-shaped tissue samples were 2.5 cm in
424
Olfat et al.
diameter (figure 1A). Samples were transferred to 0.9% NaCl solution immediately after removal and sent to the laboratory at the
Swedish Institute for Infectious Disease Control (Solna). The underlying tissue layers were first removed, and then ∼60 explants
(diameter, 2 mm) were punched out from the epithelium. The presence of H. pylori in the tissue material was analyzed by urease test,
where 4 explants were applied to 2% urea-phenol red solution. In
parallel, 4 explants were homogenized and cultured on chocolate
agar (CA) plates (Colombia agar, 8.5% horse blood, and 10% horse
serum) under microaerophilic conditions (5% O2, 10% CO2, and
85% N2) for 5 days. Only biopsy specimens from H. pylori–negative
patients were used in this study.
Culture of human gastric explants. Four explants were applied
to each insert cup (Transwell; Costar) with the mucosal side facing
up. Auturp medium [4], supplemented with 2 mg/L Nalidixic acid
(Sigma-Aldrich) and 4 mL/L Skirrows supplement (Oxoid), was
added to the wells to allow contact with the bottom of the insert
cup. Thus, a capillary contact between the explants and the culture
medium from below was established (figure 1B). The insert cup’s
membrane pore size (0.4 mm) prevented H. pylori from passing
through the insert into the medium below. The explants were infected with H. pylori after 6 h of incubation at 37⬚C in 5% CO2.
Inoculation of human gastric explants. H. pylori strain CCUG
17875 was cultured on CA plates for 48 h, harvested, washed, and
resuspended in cell culture medium. A bacterial inoculum of
2 ⫻ 10 6 cfu suspended in 200 mL of modified Auturp medium was
used in all experimental series. The bacterial cell suspension was
applied to each insert cup to cover the explants. A set of 4 unin-
JID 2002;186 (1 August)
fected explants was included in each plate as a control. After coincubation for 1 h at 37⬚C, the H. pylori suspension was removed,
and the explants were washed 3 times with 400 mL of cell culture
medium, before start of the IVEC analysis. The culture medium
underneath the insert cups was replaced with fresh medium at each
time point, as given in figure 2D, and stored at ⫺20⬚C for subsequent IL-8 quantification analyzes.
Quantitative analyzes of colony-forming units in the explants.
The sets of explants from single insert cups were harvested and
homogenized at different time points and cultured on CA plates
in 2 dilutions. The mean numbers were calculated after 3 days.
Growth of H. pylori in explant culture medium versus broth. H.
pylori strain CCUG 17875 (106 cells) was suspended in 9 mL of
explant culture medium and in parallel in Brucella broth supplemented with 5% fetal calf serum and incubated at 37⬚C in 5% CO2.
The bacterial growth was followed over time by measuring the
optical density (OD) at 600 nm. Gram staining and urease test were
performed to confirm growth of H. pylori.
Analyzes of correlation between concentration of IL-8 and duration of explants culture. Thirty explants were infected after 6 h,
as described above. The concentration of IL-8 was measured at
18–96 h, where the concentration refers to IL-8 produced between
each medium replacement. Additional explants (n p 30 ) were infected at different time points, from 6 to 96 h. To standardize each
inoculation, bacterial aliquots were stored at ⫺70⬚C in freezing
medium containing 2 g of Casamino acid (Difco), 2 g of pepton
(Acumedia), 0.4 g of yeast extract (Acumedia), 0.32 g of bacteriologic agar (Acumedia), 0.04 g of L-cystein, 0.2 g of glucose, 28
Figure 1. A, Tissue sample (2.5 cm) from a human stomach before removal of underlying tissues and punching out the explants. B, Assembly
of the insert cup (Transwell; Costar) and culturing of the human gastric explants. C, Interleukin (IL)–8 immunostaining of an explant tissue
section 24 h after inoculation with Helicobacter pylori. The arrows show the location of IL-8. D, IL-8 immunostaining of noninoculated tissue.
Figure 2. A, Growth curves of Helicobacter pylori strain CCUG 17875 in gastric explants from 4 different individuals. For each time point during
the period of 12 to 48 h, 4 explants were harvested for viable count. B, Growth curves in Brucella broth supplemented with 5% fetal calf serum
(䡬) and in cell culture medium used for culturing of the explants (䢇), measured by optical density (absorbance [Abs] at 600 nm). C, Production
of interleukin (IL)–8 from explants cultured and inoculated at the same time points. The IL-8 concentrations were analyzed at different time
points from 18 h to 4 days (light gray bars). Dark gray bars show noninoculated controls. D, IL-8 production as a function of age and viability
of the explants. All culture explants were started simultaneously, but each group of 4 explants was infected at different time points. IL-8 concentration
was analyzed 24 h after inoculation with H. pylori (light gray bars). Noninoculated controls are shown in dark gray bars. Each bar represents
mean of IL-8 concentration from 4 explants measured by ELISA in 3 different dilutions and double wells. E, H. pylori strain CCUG 17875 cfu/
mg of explant tissue over 36 h. The bacterial suspension in the first group of bars was pretreated with cell medium only. The second group was
preincubated with Lewis b blood group antigen. The third group was pretreated with H2 blood group antigen. F, IL-8 concentration in noninoculated
explants (dark gray bars), IL-8 induced by strain 17875 during coculture (white bars), preincubated bacteria with H2 conjugate (medium gray
bars), and response of IL-8 production from explants infected with bacteria pretreated with Lewis b antigen (light gray bars).
426
Olfat et al.
mL of glycerol, 1 g of NaCl, and 240 mL of Milli-Q–filtered water
(pH, 7.0 Ⳳ 0.2; Millipore). The bacterial aliquot was washed twice,
and the OD was adjusted to 0.1 at 600 nm. A bacterial suspension
containing 2 ⫻ 10 6 bacteria cells was added to each insert cup. The
number of viable cells after washes and OD adjustment was measured by viable counts and remained unchanged after storage. The
explants were analyzed 24 h postinfection for the concentration of
IL-8 secreted into the medium of the lower chamber, using double
wells and triple dilutions from each time point by ELISA (Eli pair;
Diaclone). Viable counts were performed in parallel.
Inhibition of adherence of H. pylori by treatment with soluble
Lewis b antigen. Five mg/mL Lewis b antigen or H2 neoglycoconjugates (IsoSep AB) was mixed with 2 ⫻ 10 6 cells of H. pylori
strain CCUG 17875 and incubated for 1 h. Alternatively, the bacterial cells were incubated without glycoconjugate. After washing,
the bacterial cells were finally applied to the explant cultures, as
described above.
Immunohistochemistry and IL-8 staining. Two sets of infected
and uninfected explants were fixed in 10% formalin overnight, and
5-mm tissue sections were obtained after paraffin imbedding. After rehydration, the tissue sections were incubated with mouse
anti–human IL-8 antibodies (HyCult Biotechnology), followed by
StrepABComplex (DAKO A/C).
Results
Immunohistochemistry and IL-8 staining.
Sections incubated with anti–IL-8 antibodies showed a clear staining of the
apical surface of epithelial cells in H. pylori–inoculated tissue
(figure 1C). Sections from control tissue showed no staining
(figure 1D). No staining was observed with secondary antibodies or staining reagent only.
Bacterial growth in explants and in IVEC medium. The log
growth phase started 24 h after inoculation, and the stationary
growth phase was reached 12 h later in the explants (figure 2A).
The culture medium alone did not promote growth, compared
with Brucella broth. The log phase in Brucella broth was
reached after 24 h (figure 2B).
IL-8 production in explants after H. pylori infection. Increase of IL-8 production was demonstrated in response to H.
pylori and showed a high activity after 24–36 h before it rapidly
declined (figure 2C). To further analyze the time-dependent IL8 production as a function of age of the explants, the explants
were infected at different time points. IL-8 production was highest in young cultures (18–24 h) and declined in older cultures
(24–96 h), showing that explants are responsive during 24 h of
culture, whereas prolonged culture of explants reduced the IL8 production (figure 2D).
Inhibition of H. pylori infection and IL-8 production by pretreatment with soluble Lewis b antigen. H. pylori strain CCUG
17875 was pretreated with soluble Lewis b conjugate to inhibit
binding to the explants. The lower level of colonization was
observed at all 3 time points by viable counts. At 36 h postinfection, viable counts from untreated bacteria was 50% higher,
compared with H. pylori cells that had been preincubated with
JID 2002;186 (1 August)
Lewis b conjugate. No reduction was observed when H. pylori
had been preincubated with soluble H2 conjugate (figure 2E).
Similar to the reduction in the bacterial load, IL-8 production
was decreased when bacterial adherence was reduced by soluble
Lewis b conjugate (figure 2F).
Discussion
Because the natural niche for H. pylori infection is the rough
environment of the human stomach, it has been of interest to
develop an in vitro model that can mimic the human gastric
mucosa. The cell culture lineages, both cancer and primary, do
not mimic the in vivo infection conditions (i.e., mucus production and complex glycosylation patterns). Animal models
have been used widely to study H. pylori infection. However,
there are major differences between animal and human gastric
mucosa. Therefore, the development of models such as IVEC
is of interest when trying to mimic the human infection. The
present explant adhesion assay demonstrates the potential of
the IVEC technique for studies of H. pylori infection.
The ability of H. pylori to bind to the Lewis b blood group
antigen for targeted adherence to the gastric human tissue has
been shown previously in formalin-fixed tissue [7]. In the present study, the binding characteristics were analyzed by the
IVEC technique. Pretreatment of H. pylori with the Lewis b
antigen efficiently inhibited colonization of the gastric explants.
The bacterial loads (colony-forming units per milligram) in
these explants were reduced by 150% at the 36-h time point.
However, when H. pylori was pretreated with the nonbinding
H2 blood group antigen, no reduction in the number of colonyforming units per milligram was demonstrated, which suggests
a specific effect of Lewis b antigen–mediated binding. In parallel, the growth of H. pylori in the explant culture medium
showed that bacterial growth is not supported by this medium.
This suggests that viable explants are required for bacterial
growth in the IVEC model (i.e., nutritional agents crucial for
growth of bacteria are provided by the viable tissue, as prolonged culture of explants showed to decrease the number of
bacteria; data not shown). Biochemical and genomic analysis
have demonstrated that several amino acids are needed to promote growth of H. pylori [11].
Lewis b antigen–mediated adhesion of H. pylori activates
various inflammatory responses [7]. It has been demonstrated
that the cag pathogenicity island up-regulates IL-8 production
[12]. In the present model, the production of IL-8 was at a
maximum in the interval of 24–36 h (figure 2C), similar to the
active growth phase (figure 2A), where H. pylori enters the
stationary growth phase after ∼36 h of culture. The response
of the explants to the infection became weaker because of both
age of the explants and bacterial growth phase. In the explant
culture model, background production of IL-8 was consistently
detected in explants not infected with H. pylori. However, the
levels were increased after inoculation with H. pylori (figure
JID 2002;186 (1 August)
IVEC Technique and H. pylori Infection
2F). The IL-8 response to H. pylori infection indicates that the
tissue is viable, because IL-8 production is an active process
that involves phosphorylation-dependent signal transduction
pathways [13]. A patchy distribution of H. pylori in vivo has
been shown elsewhere [14]. Interestingly, the uneven distribution of IL-8 demonstrated in the explants tissue, by anti–IL-8
antibodies, could correlate with a patchy distribution of the
bacteria, but we were unable to visualize a colocalization of
bacteria and IL-8. The relationship between tissue viability and
production of IL-8 was also analyzed, where the highest IL-8
concentration was produced in fresh explants (figure 2D). The
preincubation of bacteria with Lewis b antigen, which results
in decreased IL-8 production, would be an interesting subject
for further studies of H. pylori infection, because it has been
shown that cytokines produced in vivo by human gastric tissue
induces mucosal damage [15]. The IVEC technique is an attractive model for studying bacteria-host interactions and for
detailed analysis of inflammation. Furthermore, gene expression and regulation patterns can be analyzed with techniques
such as microarray and quantitative reverse-transcriptase polymerase chain reaction.
References
1. Dixon MF. Helicobacter pylori and peptic ulceration histopathological aspects. J Gastroenterol Hepatol 1991; 6:125–30.
2. Ekstrom AM, Held M, Hansson LE, Engstrand L, Nyren O. Helicobacter
pylori in gastric cancer established by CagA immunoblot as a marker of
past infection. Gastroenterology 2001; 121:784–91.
3. McClain MS, Schraw W, Ricci V, Boquet P, Cover TL. Acid activation of
427
Helicobacter pylori vacuolating cytotoxin (VacA) results in toxin internalization by eukaryotic cells. Mol Microbiol 2000; 37:433–42.
4. Rosberg K, Hubinette R, Nygård G, Berglindh T, Rolfsen W. Studies of
Helicobacter pylori in a gastric mucosa in vitro animal model. Scand J
Gastroenterol 1991; 21:43–8.
5. Lindholm C, Quiding-Jarbrink M, Lönroth H, Svennerholm AM. Induction
of chemokine and cytokine responses by Helicobacter pylori in human
stomach explants. Scand J Gastroenterol 2001; 36:1022–9.
6. Borén T, Falk K, Roth KA, Larson G, Normark S. Attachment of Helicobacter pylori to human gastric epithelium mediated by blood group
antigens. Science 1993; 262:1892.
7. Guruge JL, Falk PG, Lorenz RG, et al. Epithelial attachment alters the
outcome of Helicobacter pylori infection. PNAS 1998; 95:3925–30.
8. Gerhard M, Lehn N, Neumayer N, et al. Clinical relevance of the Helicobacter pylori gene for blood-group antigen-binding adhesin. PNAS 1999;
96:12778–83.
9. Rieder G, Einsiedl W, Hatz RA, Stolte M, Enders GA, Walz A. Comparison
of CXC chemokines ENA-78 and interleukin-8 expression in Helicobacter
pylori–associated gastritis. Infect Immun 2001; 69:81–8.
10. Sharma SA, Tummuru MK, Blaser MJ, Kerr LD. Activation of IL-8 gene
expression by Helicobacter pylori is regulated by transcription factor nuclear factor–kB in gastric epithelial cells. J Immunol 1998; 160:2401–7.
11. Doig P, de Jonge BL, Alm RA, et al. Helicobacter pylori physiology predicted
from genomic comparison of two strains. Microbiol Mol Biol Rev 1999;63:
675–707.
12. Audibert C, Burucoa C, Janvier B, Fauchere JL. Implication of the structure
of the Helicobacter pylori cag pathogenicity island in induction of interleukin-8 secretion. Infect Immun 2001; 69:1625–9.
13. Yu Y, De Waele C, Chadee K. Calcium-dependent interleukin-8 gene expression
in T84 human colonic epithelial cells. Inflamm Res 2001; 50:220–6.
14. Engstrand L, Rosberg K, Hubinette R, Berglindh T, Rolfsen W, Gustavsson
S. Topographic mapping of Helicobacter pylori colonization in long-term–
infected pigs. Infect Immun 1992; 60:653–6.
15. Crabtree JE. Role of cytokines in pathogenesis of Helicobacter pylori–induced
mucosal damage. Dig Dis Sci 1998; 43(Suppl 9):46S–55.
II
RESEARCH ARTICLE
Helicobacter pylori SabA Adhesin
in Persistent Infection
and Chronic Inflammation
Jafar Mahdavi,1* Berit Sondén,1*† Marina Hurtig,1*
Farzad O. Olfat,1,2 Lina Forsberg,1 Niamh Roche,3
Jonas Ångström,3 Thomas Larsson,3 Susann Teneberg,3
Karl-Anders Karlsson,3 Siiri Altraja,4 Torkel Wadström,5
Dangeruta Kersulyte,6 Douglas E. Berg,6 Andre Dubois,7
Christoffer Petersson,8 Karl-Eric Magnusson,8 Thomas Norberg,9
Frank Lindh,10 Bertil B. Lundskog,11 Anna Arnqvist,1,12
Lennart Hammarström,13 Thomas Borén1‡
Helicobacter pylori adherence in the human gastric mucosa involves specific
bacterial adhesins and cognate host receptors. Here, we identify sialyl-dimericLewis x glycosphingolipid as a receptor for H. pylori and show that H. pylori
infection induced formation of sialyl-Lewis x antigens in gastric epithelium in
humans and in a Rhesus monkey. The corresponding sialic acid– binding adhesin
(SabA) was isolated with the “retagging” method, and the underlying sabA gene
( JHP662/HP0725) was identified. The ability of many H. pylori strains to adhere
to sialylated glycoconjugates expressed during chronic inflammation might
thus contribute to virulence and the extraordinary chronicity of H. pylori
infection.
Helicobacter pylori persistently infects the
gastric mucosa of more than half of all
people worldwide, causes peptic ulcer disease, and is an early risk factor for gastric
cancer (1). Many H. pylori strains express
adhesin proteins that bind to specific hostcell macromolecule receptors (2). This adherence may be advantageous to H. pylori
by helping to stabilize it against mucosal
1
Department of Odontology/Oral Microbiology,
Umeå University, SE-901 87 Umeå, Sweden. 2The
Swedish Institute for Infectious Disease Control, SE171 82 Solna, Sweden. 3Institute of Medical Biochemistry, Göteborg University, Box 440, SE-405 30 Göteborg, Sweden. 4Institute of Molecular and Cell Biology, Tartu University, EE-51010 Tartu, Estonia. 5Department of Infectious Diseases and Medical
Microbiology, Lund University, SE-223 62 Lund, Sweden. 6Department of Molecular Microbiology, Washington University Medical School, St. Louis, MO
63110, USA. 7Laboratory of Gastrointestinal and Liver
Studies, Department of Medicine, USUHS, Bethesda,
MD 20814 – 4799, USA. 8Department of Molecular
and Clinical Medicine, Division of Medical Microbiology, Linköping University, SE-581 85 Linköping, Sweden. 9Department of Chemistry, Swedish University
of Agricultural Sciences, SE-750 07 Uppsala, Sweden.
10
IsoSep AB, Dalkärrsv. 11, SE-146 36 Tullinge, Sweden. 11Department of Medical Biosciences/Clinical
Cytology, Umeå University, SE-901 87 Umeå, Sweden. 12Department of Molecular Biology, Umeå University, SE-901 87 Umeå, Sweden. 13Center for Biotechnology, Karolinska Institute, Novum, SE-141 57
Huddinge, Sweden.
*These authors contributed equally to this work.
†Present address: Unité de Génétique Mycobactérienne, Institut Pasteur, 25 rue du Dr Roux, 75724 Paris
Cedex 15, France.
‡To whom correspondence should be addressed. Email: Thomas.Boren@odont.umu.se
shedding into the gastric lumen and ensuring good access to nourishing exudate from
gastric epithelium that has been damaged
by the infection. The best defined H. pylori
adhesin-receptor interaction found to date
is that between the Leb blood group antigen
binding adhesin, BabA, a member of a
family of H. pylori outer membrane proteins, and the H, Lewis b (Leb), and related
ABO antigens (3–5). These fucose-containing blood group antigens are found on red
blood cells and in the gastrointestinal mucosa (6 ). Blood group–O individuals suffer
disproportionately from peptic ulcer disease, suggesting that bacterial adherence to
the H and Leb antigens affects the severity
of infection (7 ). Additional H. pylori– host
macromolecule interactions that do not involve Leb-type antigens have also been
reported (8).
H. pylori is a genetically diverse species,
with strains differing markedly in virulence.
Strains from persons with overt disease generally carry the cag-pathogenicity island (cagPAI) (9, 10), which mediates translocation of
CagA into host cells, where it is tyrosine phosphorylated and affects host cell signaling (11).
Leb antigen binding is most prevalent among
cag⫹ strains from persons with overt disease (4,
12). Separate studies using transgenic mice that
express the normally absent Leb antigen suggest that H. pylori adherence exacerbates inflammatory responses in this model (13). Taken
together, these results point to the pivotal role of
H. pylori adherence in development of severe
disease.
Leb antigen–independent binding. Earlier studies identified nearly identical babA
genes at different H. pylori chromosomal
loci, each potentially encoding BabA. The
babA2 gene encodes the complete adhesin,
whereas babA1 is defective because sequences encoding the translational start and signal
peptide are missing (4). Our experiments began with analyses of a babA2– knockout mutant derivative of the reference strain
CCUG17875 (hereafter referred to as 17875).
Unexpectedly, this 17875 babA2 mutant
bound to gastric mucosa from an H. pylori–
infected patient with gastritis.
A 17875 derivative with both babA genes
inactivated (babA1A2) was constructed (14).
This babA1A2 mutant also adhered (Fig. 1, B
and C), which showed that adherence was not
due to recombination to link the silent babA1
gene with a functional translational start and
signal sequence. Pretreatment with soluble Leb
antigen (structures in table S1) resulted in
⬎80% lower adherence by the 17875 parent
strain (Figs. 1E and 3C) but did not affect
adherence by its babA1A2 derivative (Figs. 1F
and 3C).
In contrast to binding to infected gastric
mucosa (Fig. 1, A to I), the babA1A2 mutant did not bind to healthy gastric mucosa
from a person not infected with H. pylori
(Fig. 1, J to M), whereas the 17875 parent
strain bound avidly (Fig. 1, M versus L).
These results implicated another adhesin
that recognizes a receptor distinct from the
Leb antigen and possibly associated with
mucosal inflammation.
Adherence was also studied in tissue
from a special transgenic mouse that produces Leb antigen in the gastric mucosa,
the consequence of expression of a humanderived ␣1,3/4 fucosyltransferase (FT)
(15). Strain 17875 and the babA1A2 mutant
each adhered to Leb mouse gastric epithelium (Fig. 2A, ii and iii), whereas binding
of each strain to the mucosa of nontransgenic (FVB/N) mice was poor and was
limited to the luminal mucus. Thus, Leb
mice express additional oligosaccharide
chains (glycans), possibly fucosylated, but
distinct from the Leb antigen that H. pylori
could exploit as a receptor.
sdiLex antigen–mediated binding. To
search for another receptor, thin-layer chromatography (TLC)–separated glycosphingolipids (GSLs) were overlaid with H. pylori
cells and monoclonal antibodies (mAbs), as
appropriate. These tests (i) showed that the
babA1A2 mutant bound acid GSLs (Fig. 2B,
iv, lanes 2 and 4), (ii) confirmed that it did not
bind Leb GSL (lane 9), (iii) showed that its
binding was abrogated by desialylation (lanes
3 and 5), and (iv) revealed that it did not bind
sialylated GSLs of nonhuman origin (lane 1)
[table S1, numbers 2 to 5, in (16)] (indicating
that sialylation per se is not sufficient for
www.sciencemag.org SCIENCE VOL 297 26 JULY 2002
573
RESEARCH ARTICLE
adherence). In addition, the binding pattern of
the babA1A2 mutant matched that of the mAb
against sLex (Fig. 2B, ii versus iv), except
that sLex-mono-GSL was bound more weakly by the babA1A2 mutant than by the mAb
(Fig. 2B, iv) and no binding to sialyl-Lewis
a-GSL was detected (table S1). Thus, the
babA1A2 mutant preferably binds sialylated
gangliosides, possibly with multiple Lex (fucose-containing) motifs in the core chain
(thus, slower migration on TLC).
The babA1A2-mutant strain was next used
to purify a high-affinity binding GSL from
human adenocarcinoma tissue (14). The H.
pylori– binding GSL was identified by mass
spectrometry and 1H nuclear magnetic resonance (NMR) as the sialyl-dimeric-Lewis x
antigen, abbreviated as the sdiLex antigen
(Fig. 2, B and C) (14).
Further tests with soluble glycoconjugates
showed that the 17875-parent strain bound both
sLex and Leb antigens, whereas its derivative
bound only sLex (Fig. 3A). Pretreatment of the
babA1A2 mutant with sLex conjugate reduced
its in situ adherence by more than 90% (Figs. 1I
and 3C) (14) but did not affect adherence of the
17875 parent strain (Figs. 1H and 3C). Similarly, pretreatment of tissue sections with an mAb
that recognizes sdiLex reduced adherence by
the babA1A2 mutant by 72% (Fig. 3E).
Titration experiments showed that the
babA1A2 mutant exhibited high affinity for the
sdiLex GSLs, with a level of detection of 1
pmol (Fig. 2B, iv) (table S1, number 8). At least
2000-fold more (2 nmol) of the shorter sialyl(mono)-Lewis x GSL was needed for binding
(table S1, number 7). In contrast, its affinity
(Ka) for soluble conjugates was similar for
Fig. 1. The sLex antigen confers adherence of H. pylori to the epithelium of H. pylori–infected
(strain WU12) human gastric mucosa (Fig. 3A). H/E staining reveals mucosal inflammation (A). The
17875 parent strain (B) and the babA1A2 mutant (C) both adhere to the gastric epithelium. The
surface epithelium stains positive (arrows) with both the Leb mAb (D) and sLex mAb (G) [AIS
described in (14)]. The 17875 strain and babA1A2 mutant responded differently after pretreatment
(inhibition) with soluble Leb antigen [(E) and (F), respectively)], or with soluble sLex antigen [(H)
and (I), respectively)]. In conclusion, the Leb antigen blocked binding of the 17875 strain, whereas
the sLex antigen blocked binding of the babA1A2 mutant. H/E-stained biopsy with no H. pylori
infection (J). No staining was detected with the sLex mAb (K). Here, strain 17875 adhered (L), in
contrast to the babA1A2 mutant (M), because noninflamed gastric mucosa is low in sialylation (K).
574
mono and dimeric forms of sLex, 1 ⫻ 108 M⫺1
and 2 ⫻ 108 M⫺1, respectively (Fig. 3B) (16).
These patterns suggest that the sialylated binding sites are best presented at the termini of
extended core chains containing multiple Lewis
x motifs, such as GSLs in cell membranes.
Such optimization of steric presentation would
be less important for soluble receptors.
The Scatchard analyses (16) also estimated that 700 sLex-conjugate molecules were
bound per babA1A2-mutant bacterial cell
(Fig. 3B), a number similar to that of Leb
conjugates bound per cell of strain 17875.
Clinical isolates. A panel of 95 European clinical isolates was analyzed for sLex
binding (14 ). Thirty-three of the 77 cagA⫹
strains (43%) bound sLex, but only 11% (2
out of 18) of cagA– strains (P ⬍ 0.000).
However, deletion of the cagPAI from strain
G27 did not affect sLex binding, as had also
been seen in studies of cagA and Leb-antigen
binding. Out of the Swedish clinical isolates,
39% (35 out of 89) bound sLex, and the great
majority of these sLex-binding isolates, 28
out of 35 (80%), also bound the Leb antigen.
Fifteen of the 35 (43%) sLex-binding isolates
also bound the related sialyl-Lewis a antigen
(sLea) (table S1, number 6), whereas none of
the remaining 54 sLex-nonbinding strains
could bind to sLea. The clinical isolates
SMI65 and WU12 (the isolate from the infected patient in Fig. 1, A to M) illustrate
such combinations of binding modes (Fig.
3A). Of the two strains with sequenced genomes, strain J99 (17 ) bound sLex, sLea, and
Leb antigens, whereas strain 26695 (18)
bound none of them (Fig. 3A).
Binding to inflamed tissue. Gastric
tissue inflammation and malignant transformation each promote synthesis of sialylated
glycoconjugates (19), which are rare in
healthy human stomachs (20). Immunohistochemical analysis showed that the sialylated
antigens were located to the apical surfaces of
the surface epithelial cells (fig. S2). To study
gastric mucosal sialylation in an H. pylori
context, gastric biopsies from 29 endoscopy
patients were scored for binding by the
babA1A2 mutant and by the 17875 strain in
situ, and for several markers of inflammation
(table S2A) (14 ). Substantial correlations
were found between babA1A2-mutant adherence and the following parameters: (i) levels
of neutrophil (PMN) infiltration, 0.47 (P ⬍
0.011); (ii) lymphocyte/plasma cell infiltration, 0.46 (P ⬍ 0.012); (iii) mAb staining for
sLex in surface epithelial cells and in gastric
pit regions, 0.52 (P ⬍ 0.004); and (iv) histological gastritis score, 0.40 (P ⬍ 0.034). In
contrast, there was no significant correlation
between babA1A2-mutant binding in situ and
H. pylori density in biopsies from natural
infection (0.14, P ⬍ 0.47), nor between any
inflammatory parameter and in situ adherence of strain 17875 (Leb and sLex binding).
26 JULY 2002 VOL 297 SCIENCE www.sciencemag.org
RESEARCH ARTICLE
The series of 29 patient biopsies was then
compared to a series of six biopsies of H.
pylori–noninfected individuals, and a considerable difference was found to be due to
lower adherence of the babA1A2 mutant (P ⬍
0.000) (table S2B), whereas strain 17875
showed no difference.
Binding to infected tissue. Biopsy material from a Rhesus monkey was used to
directly test the view that H. pylori infection
stimulates expression of sialylated epithelial
glycosylation patterns that can then be exploited by H. pylori for adherence [monkey
biopsies in (14 )]. This monkey (21) had been
cleared of its natural H. pylori infection, and
gastritis declined to baseline. Gastric biopsies
taken at 6 months post-eradication showed
expression of sLex in the gastric gland region
(Fig. 4A) and no expression in the surface
epithelium (Fig. 4A). In situ adherence of the
babA1A2 mutant was limited to gastric
glands and was closely matched to the sLex
expression pattern (Fig. 4B). No specific adherence to the surface epithelium was seen
(Fig. 4B).
At 6 months post-therapy, this gastritis-free
animal had been experimentally infected with a
cocktail of H. pylori strains, of which the J166
strain (a cagA-positive, Leb and sLex binding
isolate) became predominant a few months later. This led to inflammation, infiltration by
lymphocytes [Fig. 4C, bluish due to hematoxyline/eosine (H/E)–staining], and microscopic
detection of H. pylori infection (Fig. 4, E and F)
(21). The virulent H. pylori infection led to
strong sLex-antigen expression in the surface
epithelium (Fig. 4C) and maintained expression
in the deeper gastric glands (Fig. 4C); thus, a
bi-layered expression mode supported strong
binding of the babA1A2 mutant to both regions
(Fig. 4D). Bacterial pretreatment with sLex
conjugate eliminated surface epithelial adherence and reduced gastric gland adherence by
88%. Thus, persistent H. pylori infection upregulates expression of sLex antigens, which H.
pylori can exploit for adherence to the surface
epithelium.
Binding to Leb transgenic mouse gastric epithelium. We analyzed gastric mucosa of Leb mice for sLex antigen–dependent H. pylori adherence in situ [AIS, in
(14 )]. Pretreatment with the sLex conjugate
reduced binding by babA1A2-mutant bacteria by more than 90% (Figs. 2A, vi, and
3D) but did not affect binding by its 17875
parent (Figs. 2A, v, and 3D). In comparison, pretreatment with soluble Leb antigen
decreased adherence by ⬎80% of the
17875 strain (Leb and sLex binding),
whereas binding by the babA1A2 mutant
was not affected. mAb tests demonstrated
sLex antigen in the gastric surface epithelium and pits of Leb mice (Fig. 2A, iv).
That is, these mice are unusual in producing sialyl as well as Leb glycoconjugates
(even without infection), which each serve
as receptors for H. pylori.
A finding that H. pylori pretreatment with
Leb conjugate blocked its adherence to Leb
mouse tissue, whereas sLex pretreatment did
not, had been interpreted as indicating that H.
pylori adherence is mediated solely by Leb
antigen (22). However, because strain 17875
and its babA1A2 mutant bound similar levels of
soluble sLex conjugate (Fig. 3A), soluble Leb
seems to interfere sterically with interactions
between sLex-specific H. pylori adhesins and
sLex receptors in host tissue (see also Fig. 1, E
versus F). Because an excess of soluble sLex
did not affect H. pylori binding to Leb receptors
[Figs. 1H and 3C (in humans) and Figs. 2A, v,
and 3D (in Leb-mice)], the steric hindrance is
not reciprocal. Further tests showed that liquid
phase binding is distinct. A 10-fold excess of
soluble unlabeled Leb conjugate (3 ␮g) along
with 300 ng of 125I-sLex conjugate did not
affect strain 17875 binding to soluble sLex
glycoconjugate.
Thus, soluble Leb conjugates interfere
with sLex-mediated H. pylori binding specifically when the sLex moieties are constrained on surfaces. The lack of reciproc-
ity in these Leb-sLex interference interactions contributes to our model of receptor
positioning on cell surfaces.
SabA identified. We identified the sLexbinding adhesin by “retagging.” This technique exploits a receptor-bound multifunctional biotinylated crosslinker, and ultraviolet
(UV) irradiation to mediate transfer of the
biotin tag to the bound adhesin (4). Here, we
added the crosslinker to sLex conjugate and
used more UV exposure (14) than had been
used to isolate the BabA adhesin (4) to compensate for the lower affinity of the sLexthan the Leb-specific adhesin for cognate soluble receptors (Fig. 3B). Strain J99 was used,
and a 66-kDa protein was recovered (Fig.
5A). Four peptides identified by mass spectrometry (MS)–matched peptides encoded by
gene JHP662 in strain J99 (17) (gene
HP0725 in strain 26695) (18). Two of the
four peptides also matched those from the
related gene JHP659 (HP0722) (86% protein
level similarity to JHP662) (fig. S1).
To critically test if JHP662 or JHP659 encodes SabA, we generated camR insertion alleles of each gene in strain J99. Using radiolabeled glycoconjugates, we found that both sLex
Fig. 2. (A) The sLex antigen confers adherence of H. pylori to the
transgenic Leb mouse gastric
mucosa. H/E-stained biopsy of
Leb mouse gastric mucosa (i).
The 17875 strain (ii) and
babA1A2 mutant (iii) both adhered to the surface epithelium,
which stained positive with the
sLex mAb (iv). Adherence by the
babA1A2 mutant was lost (vi),
while adherence by the 17875
strain was unperturbed (v), after
pretreatment of the bacteria with soluble sLex antigen. (B) H. pylori binds to fucosylated
gangliosides such as the sLex GSL. GSLs were separated on TLC and chemically stained (i) (14).
Chromatograms were probed with sLex mAb (ii), the 17875 strain (iii), and the babA1A2 mutant
(iv). Lanes contained acid (i.e., sialylated) GSLs of calf brain, 40 ␮g; acid GSLs of human neutrophil
granulocytes, 40 ␮g; sample number 2 after desialylation, 40 ␮g; acid GSLs of adenocarcinoma, 40
␮g; sample number 4 after desialylation, 40 ␮g; sdiLex GSL, 1 ␮g; sLea GSL, 4 ␮g; s(mono)Lex GSL,
4 ␮g; and Leb GSL, 4 ␮g. Binding results are summarized in table S1. (C) The high-affinity H. pylori
GSL receptor was structurally identified by negative ion fast atom bombardment mass spectrometry. The molecular ion (M-H⫹)– at m/z 2174 indicates a GSL with one NeuAc, two fucoses, two
N-acetylhexosamines, four hexoses, and d18 :1-16:0; and the sequence NeuAcHex(Fuc)HexNAcHex(Fuc)HexNAcHexHex was deduced. 1H-NMR spectroscopy resolved the sdiLex antigen;
NeuAc␣2.3Gal␤1.4(Fuc␣1.3)GlcNAc␤1.3Gal␤1.4(Fuc␣1.3) GlcNAc␤1.3Gal␤1.4Glc␤1Cer (14).
www.sciencemag.org SCIENCE VOL 297 26 JULY 2002
575
RESEARCH ARTICLE
and sLea antigen–binding activity was abolished in the JHP662 (sabA) mutant, but not in
the JHP659 (sabB) mutant. Thus, the SabA
adhesin is encoded by JHP662 (HP0725). This
gene encodes a 651-aa protein (70 kDa) and
belongs to the large hop family of H. pylori
outer membrane protein genes, including babA
(17, 18). The sabA gene was then identified by
PCR in six sLex-binding and six non–sLexbinding Swedish isolates, which suggests that
sabA is present in the majority of H. pylori
isolates (14).
Parallel studies indicated that the sabA
inactivation did not affect adherence mediated by the BabA adhesin (Fig. 5B, i) and that
pretreatment of the J99sabA mutant with soluble Leb antigen prevented its binding to
gastric epithelium (Fig. 5B, ii). This implies
that the SabA and BabA adhesins are organized and expressed as independent units.
Nevertheless, Leb conjugate pretreatment of
J99 (BabA⫹ and SabA⫹) might have interfered with sLex antigen–mediated adherence
(see Fig. 1E). To determine if this was a steric
effect of the bulky glycoconjugate on exposure of SabA adhesin, single babA and sabA
mutant J99 derivatives were used to further
analyze Leb and sLex adherence (Fig. 5C, i to
iv). Both single mutants adhered to the inflamed gastric epithelial samples, whereas
the babAsabA (double) mutant was unable to
bind this same tissue.
Instability of sLex binding. When
screening for SabA mutants, we also analyzed single colony isolates from cultures of
parent strain J99 (which binds both sLex and
Leb antigens) (Fig. 3A), which indicated that
1% of colonies had spontaneously lost the
ability to bind sLex. Similar results were
obtained with strain 17875 (Fig. 3A) with an
Fig. 3. H. pylori binds
sialylated antigens. (A)
H. pylori strains and
mutants (14) were analyzed for binding to
different 125I-labeled
soluble fucosylated and
sialylated (Lewis) antigen-conjugates [RIA in
(14)]. The bars give
bacterial binding, and
conjugates used are
given in the diagram.
(B) For affinity analyses
(16), the sLex conjugate was added in titration
series.
The
babA1A2 mutant was
incubated for 3 hours
with the s(mono)Lex
conjugate to allow for
equilibrium in binding,
which demonstrates an
affinity (Ka) of 1 ⫻ 108
M⫺1. (C) Adherence of
strain 17875 (sLex and
Leb-antigen binding)
and babA1A2 mutant
to biopsy with inflammation and H. pylori infection, as scored by
the number of bound
bacteria after pretreatment with soluble Leb
(Fig. 1, E and F) or sLex
antigen (Fig. 1, H and I)
(14). The Leb antigen
reduced adherence of
strain 17875 by ⬎80%,
whereas the sLex antigen abolished adherence of the babA1A2
mutant. (D) Adherence
of strain 17875 and
babA1A2 mutant to biopsy of Leb mouse gastric mucosa was scored by the number of bound bacteria
after pretreatment with sLex conjugate. Adherence by the babA1A2 mutant was abolished (Fig. 2A, vi),
while adherence by strain 17875 was unperturbed (Fig. 2A, v). (E) Adherence of spontaneous phase
variant strain 17875/Leb (Leb-antigen binding only in Fig. 3A) and babA1A2 mutant was analyzed after
pretreatment of histo-sections of human gastric mucosa with mAbs recognizing the Leb antigen or the
sdiLex antigen (FH6), with efficient inhibition of adherence of both strain 17875/Leb and the babA1A2
mutant, respectively. Value P ⬍ 0.001 (***), value P ⬍ 0.01 (**), value P ⬍ 0.05 (*).
576
OFF (non–sLex-binding) variant called
17875/Leb. In contrast, each of several hundred isolates tested retained Leb antigen–
binding capacity.
Upstream and within the start of the sabA
gene are poly T/CT tracts that should be
hotspots for ON/OFF frameshift regulation
[see HP0725 in (18)], which might underlie
the observed instability of sLex-binding activity. Both strains’ genome sequences,
26695 and J99, demonstrate CT repeats that
suggest sabA to be out of frame (six and nine
CTs, respectively) (17, 18). Four sLex-binding and four non–sLex-binding Swedish isolates, and in addition strain J99, were analyzed by PCR for CT repeats, and differences
in length were found between strains. Ten CT
repeats [as compared to nine in (17)] were
Fig. 4. Infection induces expression of sLex
antigens that serve as binding sites for H. pylori
adherence. (A) and (B) came from a Rhesus
monkey with healthy gastric mucosa. (C) to (F)
came from the same Rhesus monkey 9 months
after established reinfection with virulent clinical H. pylori isolates. Sections were analyzed
with the sLex mAb (A) and (C) and by adherence in situ with the babA1A2 mutant (B) and
(D). In the healthy animal, both the sLex mAb
staining and adherence of the babA1A2 mutant
is restricted to the deeper gastric glands [(A)
and (B), arrows] [i.e., no mAb-staining of sLex
antigens and no babA1A2-mutant bacteria
present in the surface epithelium (A) and (B),
arrowhead]. In contrast, in the H. pylori–infected and inflamed gastric mucosa, the sLex
antigen is heavily expressed in the surface epithelium [(C), arrowhead with brownish immunostaining], in addition to the deeper gastric
glands [(C), arrow]. The up-regulated sLex antigen now supports massive colocalized adherence of the babA1A2 mutant in situ [(D), arrowhead] in addition to the constant binding to
the deeper gastric glands [(B) and (D), arrow].
The established infection was visualized by
Genta stain, where the H. pylori microbes
were present in the surface epithelium [(E)
and (F), arrowhead; (F) is at a higher magnification], while no microbes were detected in
the deeper gastric glands [(E), arrow].
26 JULY 2002 VOL 297 SCIENCE www.sciencemag.org
RESEARCH ARTICLE
found in the sLex-binding J99 strain, which
puts this ORF in frame, and could thus explain the ON bindings. These results further
support the possibility for a flexible locus that
confers ON/OFF binding properties (23).
Dynamics of sialylation during health
and disease. Our analyses of H. pylori adherence provide insight into human responses to persistent infections, where gastritis and inflammation elicit appearance of
sdiLex antigens and related sialylated carbohydrates in the stomach mucosa, which
cag⫹ (virulent) H. pylori strains by adaptive mechanisms exploit as receptors in
concert with the higher affinity binding to
Leb. These two adherence modes may each
benefit H. pylori by improving access to
nutrients leached from damaged host tissues, even while increasing the risk of bactericidal damage by these same host defenses (Fig. 6). In the endothelial lining, sialylated Lewis-glycans serve as receptors for
selectin cell adhesion proteins that help
guide leukocyte migration and thus regulate strength of response to infection or
injury (24 ). For complementary attachment, the neutrophils themselves also express sialylated Lewis glycans, and such
neutrophil glycans allow binding and infection by human granulocytic erlichiosis (25).
However, sialylated glycoconjugates are
low in healthy gastric mucosa but are expressed during gastritis. This sialylation
was correlated with the capacity for SabAdependent, but not BabA-dependent, H. pylori binding in situ. Our separate tests on
Rhesus monkeys for experimental H. pylori
infection confirmed that gastric epithelial
sialylation is induced by H. pylori infection. In accord with this, high levels of
sialylated glycoconjugates have been found
in H. pylori–infected persons, which decreased after eradication of infection and
resolution of gastritis (26 ). Thus, a sialy-
Fig. 5. Retagging of SabA and identification of the corresponding gene, sabA.
After contact-dependent retagging (of
the babA1A2 mutant) with crosslinker
labeled sLex conjugate, the 66-kDa biotin-tagged adhesin SabA [(A), lane 1]
was identified by SDS–polyacrylamide
gel
electrophoresis/steptavidin-blot,
magnetic-bead-purified, and analyzed
by MS for peptide masses (14). As a
control, the Leb conjugate was used to
retag 17875 bacteria, which visualized
the 75-kDa BabA adhesin [(A), lane 2]
(4). (B) The J99/JHP662::cam (sabA)
mutant does not bind the sLex antigen
but adheres to human gastric epithelium [(B), i], due to the BabA-adhesin,
because pretreatment with soluble Leb
antigen inhibited binding to the epithelium [(B), ii]. (C) The J99 wild-type
strain (which binds both Leb and sLex
antigens) [(C), i] and the J99 sabA mutant (which binds Leb antigen only) [(C),
ii] both exhibit strong adherence to the
gastric mucosa. In comparison, the J99
babA mutant binds by lower-affinity interactions [(C), iii], while the J99 sabAbabA mutant, which
is devoid of both adherence properties, does not bind to the gastric mucosa [(C), iv].
Fig. 6. H. pylori adherence in health and disease. This figure illustrates the proficiency
of H. pylori for adaptive multistep mediated attachment. (A) H.
pylori (in green) adherence is mediated by
the Leb blood group
antigen expressed in glycoproteins (blue chains) in the gastric surface
epithelium (the lower surface) (3, 32). H. pylori uses BabA (green Y’s) for
strong and specific recognition of the Leb antigen (4). Most of the sLexbinding isolates also bind the Leb antigen (SabA, in red Y’s). (B) During
persistent infection and chronic inflammation (gastritis), H. pylori triggers
the host tissue to retailor the gastric mucosal glycosylation patterns to
up-regulate the inflammation-associated sLex antigens (red host, triangles).
Then, SabA (red Y structures) performs Selectin-mimicry by binding
the sialyl-(di)-Lewis x/a glycosphingolipids, for membrane close
lated carbohydrate used to signal infection
and inflammation and to guide defense responses can be co-opted by H. pylori as a
receptor for intimate adherence.
High levels of sialylated glycoconjugates
are associated with severe gastric disease,
including dysplasia and cancer (27, 28). Sialylated glycoconjugates were similarly abundant in parietal cell– deficient mice (22). sLex
was also present in the gastric mucosa of
transgenic Leb mice (Fig. 2A). Whether this
reflects a previously unrecognized pathology
stemming from the abnormal (for mice) gastric synthesis of ␣1.3/4FT or Leb antigen, or
from fucosylation of already sialylated carbohydrates, is not known.
Persons with blood group O and “nonsecretor” phenotypes (lacking the ABO blood
group–antigen synthesis in secretions such as
saliva and milk) are relatively common (e.g.,
⬃45 and ⬃15%, respectively, in Europe), and
each group is at increased risk for peptic ulcer
disease (29). The H1 and Leb antigens are
abundant in the gastric mucosa of secretors (of
blood group O) (6), but not in nonsecretors,
where instead the sLex and sLea antigens are
found (30). The blood group O– disease association was postulated to reflect the adherence
of most cag⫹ H. pylori strains to H1 and Leb
antigens (3). We now suggest that H. pylori
adherence to sialylated glycoconjugates contributes similarly to the increased risk of peptic
ulcer disease in nonsecretor individuals.
Adaptive and multistep–mediated attachment modes of H. pylori. Our findings
that the SabA adhesin mediates binding to the
structurally related sialyl-Lewis a antigen
(sLea, in table S1) is noteworthy because
sLea is an established tumor antigen (31) and
marker of gastric dysplasia (27), which may
further illustrate H. pylori capacity to exploit
a full range of host responses to epithelial
damage. The H. pylori BabA adhesin binds
Leb antigen on glycoproteins (32), whereas
its SabA adhesin binds sLex antigen in membrane glycolipids, which may protrude less
from the cell surface. Thus, H. pylori adher-
attachment and apposition. (C) At sites of vigorous local inflammatory response, as illustrated by the recruited activated white blood
cell (orange “bleb”), those H. pylori subclones that have lost sLexbinding capacity due to ON/OFF frameshift mutation might have
gained local advantage in the prepared escaping of intimate contact
with (sialylated) lymphocytes or other defensive cells. Such adaptation of bacterial adherence properties and subsequent inflammation
pressure could be major contributors to the extraordinary chronicity
of H. pylori infection in human gastric mucosa.
www.sciencemag.org SCIENCE VOL 297 26 JULY 2002
577
RESEARCH ARTICLE
ence during chronic infection might involve
two separate receptor-ligand, interactions— one
at “arm’s length” mediated by Leb, and another, more intimate, weaker, and sLex-mediated
adherence. The weakness of the sLex-mediated
adherence, and its metastable ON/OFF switching, may benefit H. pylori by allowing escape
from sites where bactericidal host defense responses are most vigorous (Fig. 6C). In summary, we found that H. pylori infection elicits
gastric mucosal sialylation as part of the chronic inflammatory response and that many virulent strains can exploit Selectin mimicry and
thus “home in” on inflammation-activated domains of sialylated epithelium, complementing
the baseline level of Leb receptors. The spectrum of H. pylori adhesin–receptor interactions
is complex and can be viewed as adaptive,
contributing to the extraordinary chronicity of
H. pylori infection in billions of people worldwide, despite human genetic diversity and host
defenses.
References and Notes
1. T. L. Cover et al., in Principles of Bacterial Pathogenesis, E. A. Groisman, Ed. (Academic Press, New York,
2001), pp. 509 –558.
2. M. Gerhard et al., in Helicobacter pylori: Molecular and
Cellular Biology, S. Suerbaum and M. Achtman, Eds.
(Horizon Scientific Press, Norfolk, UK, 2001), chap. 12.
3. T. Borén et al., Science 262, 1892 (1993).
4. D. Ilver et al., Science 279, 373 (1998).
5. M. Hurtig, T. Borén, in preparation.
6. H. Clausen, S. i. Hakomori, Vox Sang 56, 1 (1989).
7. T. Borén, P. Falk, Science 264, 1387 (1994).
8. K. A. Karlsson, Mol. Microbiol. 29, 1 (1998).
9. S. Censini et al., Proc. Natl. Acad. Sci. U.S.A. 93, 14648
(1996).
10. N. S. Akopyants et al., Mol. Microbiol. 28, 37 (1998).
11. E. D. Segal et al., Proc. Natl. Acad. Sci. U.S.A. 96,
14559 (1999).
12. M. Gerhard et al., Proc. Natl. Acad. Sci. U.S.A. 96,
12778 (1999).
13. J. L. Guruge et al., Proc. Natl. Acad. Sci. U.S.A. 95,
3925 (1998).
14. Strains and culture, adherence in situ (AIS), H. pylori
overlay to TLC, isolation and identification of the
s-di-Lex GSL, apical localization of sLex antigen expression (fig. S2), retagging and identification of
SabA/sabA, alignment of JHP662/JHP659 (fig. S1),
construction of adhesin gene mutants, analyses of
sabA by PCR, RIA and Scatchard analyses, gastric
biopsies from patients and monkeys analyzed for H.
pylori binding activity and inflammation, and a summary of H. pylori binding to glycosphingolipids (table
S1) are available as supporting online material.
15. P. G. Falk et al., Proc. Natl. Acad. Sci. U.S.A. 92, 1515
(1995).
16. G. Scatchard, Ann. N.Y. Acad. Sci. 51, 660 (1949).
17. R. A. Alm et al., Nature 397, 176 (1999).
18. J. F. Tomb et al., Nature 388, 539 (1997).
19. S. i. Hakomori, in Gangliosides and Cancer, H. F.
Oettgen, Ed. (VCH Publishers, New York, 1989), pp.
58 – 68.
20. J. F. Madrid et al., Histochemistry 95, 179 (1990).
21. A. Dubois et al., Gastroenterology 116, 90 (1999).
22. A. J. Syder et al., Mol. Cell 3, 263 (1999).
23. A. Arnqvist et al., in preparation.
24. J. Alper, Science 291, 2338 (2001).
25. M. J. Herron et al., Science 288, 1653 (2000).
26. H. Ota et al., Virchows Arch. 433, 419 (1998).
27. P. Sipponen, J. Lindgren, Acta Pathol. Microbiol. Immunol. Scand. 94, 305 (1986).
28. M. Amado et al., Gastroenterology 114, 462 (1998).
29. P. Sipponen et al., Scand. J. Gastroenterol. 24, 581
(1989).
30. J. Sakamoto et al., Cancer Res. 49, 745 (1989).
31. J. L. Magnani et al., Science 212, 55 (1981).
32. P. Falk et al., Proc. Natl. Acad. Sci. U.S.A. 90, 2035
(1993).
33. We thank R. Clouse, H.M.T. El-Zimaity, D. Graham, and I.
Anan for human biopsy material; J. Gordon and P. Falk
for sections of Leb and FVB/N mouse stomach; L. Engstrand and A. Covacci for H. pylori strains; H. Clausen for
the FH6 mAb; the Swedish NMR Centre, Göteborg University; P. Martin and Ö. Furberg for digital art/movie
work; and J. Carlsson for critical reading of the manuscript. Supported by the Umeå University Biotechnology
Fund, Swedish Society of Medicine/Bengt Ihre’s Fund,
Swedish Society for Medical Research, Lion’s Cancer
Research Foundation at Umeå University, County Council of Västerbotten, Neose Glycoscience Research Award
Grant ( T.B.), Swedish Medical Research Council [11218
( T.B.), 12628 (S.T.), 3967 and 10435 (K.-A.K.), 05975
(L.H.), and 04723 ( T.W.)], Swedish Cancer Society
[4101-B00-03XAB ( T.B.) and 4128-B99-02XAB (S.T.)],
SSF programs “Glycoconjugates in Biological Systems”
( T.B., S.T., and K.-A.K.), and “Infection and Vaccinology”
( T.B. and K.-A.K.), J. C. Kempe Memorial Foundation
(J. M., L. F., and M.H.), Wallenberg Foundation (S.T. and
K.-A.K), Lundberg Foundation (K.-A.K.), ALF grant from
Lund University Hospital (T.W.), and grants from the
NIH [RO1 AI38166, RO3 AI49161, RO1 DK53727 (D.B.)]
and from Washington University (P30 DK52574). These
experiments were conducted according to the principles
in the “Guide for the Care and Use of Laboratory Animals,” Institute of Laboratory Animal Resources, NRC,
HHS/NIH Pub. No. 85–23. The opinions and assertions
herein are private ones of the authors and are not to be
construed as official or reflecting the views of the DOD,
the Uniformed Services University of the Health Sciences, or the Defense Nuclear Agency.
Supporting Online Material
www.sciencemag.org/cgi/content/full/297/5581/573/DC1
Materials and Methods
Figs. S1 and S2
Tables S1 and S2
17 December 2001; accepted 17 June 2002
REPORTS
An Aligned Stream of
Low-Metallicity Clusters in the
Halo of the Milky Way
Suk-Jin Yoon* and Young-Wook Lee
One of the long-standing problems in modern astronomy is the curious division
of Galactic globular clusters, the “Oosterhoff dichotomy,” according to the
properties of their RR Lyrae stars. Here, we find that most of the lowest
metallicity ([Fe/H] ⬍ –2.0) clusters, which are essential to an understanding of
this phenomenon, display a planar alignment in the outer halo. This alignment,
combined with evidence from kinematics and stellar population, indicates a
captured origin from a satellite galaxy. We show that, together with the
horizontal-branch evolutionary effect, the factor producing the dichotomy
could be a small time gap between the cluster-formation epochs in the Milky
Way and the satellite. The results oppose the traditional view that the metalpoorest clusters represent the indigenous and oldest population of the Galaxy.
More than 60 years ago, Oosterhoff (1) discovered that Galactic globular clusters could
be divided into two distinct groups according
Center for Space Astrophysics, Yonsei University,
Seoul 120 –749, Korea.
*To whom correspondence should be addressed. Email: sjyoon@csa.yonsei.ac.kr
578
to the mean period of type ab RR Lyrae
variables (具Pab 典). This dichotomy was one of
the earliest indications of systematic difference among globular clusters, whose reality
has been strengthened by subsequent investigations (2). Given that most characteristics of
Galactic globular clusters appear to be distributed in a continuous way, it is unusual
that a quantity used to characterize variable
stars falls into two rather well-defined classes. Moreover, the two groups are known to
differ in metal abundance (3) and kinematic
properties (4), which may indicate distinct
origins. Whatever the reasons for the dichotomy, the question of whether the two groups
originated under fundamentally different conditions is of considerable interest regarding
the formation scenarios of the Galactic halo.
Despite many efforts during the last decades,
the origin of this phenomenon still lacks a
convincing explanation.
Figure 1 shows the Oosterhoff dichotomy.
Clusters belong to groups I and II if their
values of 具Pab 典 fall near 0.55 and 0.65 days,
respectively (5). Group I is more metal-rich
than group II (3). Based on our horizontalbranch (HB) population models (6, 7), we
have found that the presence of the relatively
metal-rich (–1.9 ⬍ [Fe/H] ⬍ –1.6) clusters in
group II (hereafter group II-a) can be understood by the sudden increase in 具Pab 典 at
[Fe/H] ⬇ –1.6 (indicated by a blue line in
Fig. 1). As [Fe/H] decreases, HB stars get
hotter (i.e., the HB morphology gets bluer),
and there is a certain point at which the
zero-age portion of the HB just crosses the
26 JULY 2002 VOL 297 SCIENCE www.sciencemag.org
III
The Neuraminyllactose-Binding Hemagglutinin
of Helicobacter pylori is Identical to the
Sialic Acid Binding Adhesin, SabA
Farzad O. Olfat1,2, Marina Aspholm1,
Niamh Roche3, Berit Sondén1, Lars Engstrand2,
Susann Teneberg3, Thomas Borén1§
1Department of Odontology/Oral microbiology, Umeå University, SE-901 87
Umeå, Sweden
2The Swedish Institute for Infectious Disease Control, SE-171 82 Solna,
Sweden
3Institute of Medical Biochemistry, Göteborg University, Box 440, SE-405 30
Göteborg, Sweden
§To whom correspondence should be addressed:
Dept. of Odontology/Oral microbiology
Umeå University
S-901 87, Umeå
Sweden
Tel: +46-90-785 60 36
Fax: +46-90-785 60 37
ABSTRACT
Adherence of Helicobacter pylori to the inflamed and infected gastric mucosa is dependent on
the epithelial and membrane-close sialylated/fucosylated gangliosides and, the cognate sialic
acid binding adhesin, SabA. Here we show that SabA is identical with the recognized H.
pylori sia-hemagglutinin, sia-HA. Thus, SabA replaces the disputed “H. pylori adhesin A”,
HpaA as the true identity of sia-HA. The logic conclusion was based on defined adhesindeletions mutants and, in addition, by analyzes of bacterial binding specificities and affinities
for defined sialylated oligosaccharides (glycans) and blood group antigens. Here, a series of
clinical H. pylori isolates were analyzed for neuraminidase-sensitive (sialyl-dependent)
hemagglutination (sia-HA), and the sia-HA titers were positively correlated to the bacterial
affinity for sLex binding. Most convincing were the findings that the sabA deletion mutant
lost all sia-HA properties. In addition, the topographic presentation of the sialyl-dependent
binding site for SabA, on the cell surfaces, was biochemically mapped to sialylated
glycosphingolipids with extended core chains. Among the clinical isolates, 3 distinct sialyldependent adherence traits were found, which suggest that SabA adhesins demonstrate
distinct differences in detailed recognition of sialyl-glycan epitopes. The SabA affinity for
specific sialylated glycans might specialize the microbe for detailed host tropism such as local
areas of inflamed and/or dysplastic tissue, i.e. host responses that develop during chronic
inflammation.
1
INTRODUCTION
Adherence to host tissue is often regarded as a virulence factor for microorganisms that
colonize turbid niches, such as the urinary tract infectious E. coli that express different
adhesive pili-structures and, intestinal pathogens such as Vibrio cholera that in addition,
express the adhesive CTB-toxin. The gastric pathogen Helicobacter pylori is no exception
since it is found in tight association with the gastric mucosa and epithelium in human- and
primate populations worldwide. Here, H. pylori is exposed to a demanding environment with
rapid turnover and fast shedding of the surface cells in the epithelial lining and mucus layer.
Fortunately for H. pylori the epithelium is protected from the luminal acidic gastric juice by
the thin slimy mucus layer, which the H. pylori bacteria has taken to its advantage. Out in this
mucus layer, the H. pylori microorganisms are found either associated with the heavily
glycosylated mucin molecules, or rapidly swimming as a result of their rapid flagellar
motility. The direction of motility, towards the gastric epithelium, is aided by chemotactic
mechanisms that sense local gradients of bicarbonate, urea and pH. Only a minor part of the
infection is actually found adherent to the epithelium (17), which probably reflects a
compromise for the microbe in the adhesion processes. H. pylori has a rather limited genome
(1)(49), which requires H. pylori to scavenge the host cells for essential nutrients lacking in
its metabolism. Thus, adhesive traits wills provide the bacteria with better access to nutrients
that are released from cells damaged by the local inflammation, but the adhesive traits will
also expose the microbes for the immune response such as local and reactive neutrophils.
However, as a unique microorganism in the stomach, H. pylori has managed to take these
difficulties to its advantage, for a life-style in persistent infection and chronic mucosal
inflammation (8)(23). In the majority of infected individuals the H. pylori infection will go
unnoticed, since the molecular cross-talk between the bacteria and host immune system is
well synchronized. However, about 10% of infected individuals will develop symptoms such
as chronic gastritis and peptic ulcer disease (47) and of critical concern is the high odds-ratio
of H. pylori infection for development of gastric cancer (10).
H. pylori has developed a series of adhesive activities towards different carbohydrates
e.g. sialylated (12), sulfated (44) and fucosylated (4) glycoproteins and glycolipids, which
confer bacterial adherence to the host mucosa. The LPS of H. pylori might also enable
bacterial adherence, possibly due to mechanisms of homotypic interaction between Lex
structures expressed on both host cells and in the H. pylori LPS (9)(32). Adherence mediated
by the Lewis b blood group antigen (Leb) (highly expressed in individuals of blood group O
phenotype) is among the best described (4). The fucosylated blood group antigens are by
definition found in red blood cell (RBC) glycoconjugates, but for the perspective of H. pylori
adhesion, they are also expressed both in the surface epithelium and, in addition, in the slimy
mucus-layer lining almost the entire oral-gastro-intestinal (GI)-tract (40). The cognate
adhesin, the blood group antigen binding adhesin, BabA, is prevalent among the virulent type1 strains and shown to be highly associated with severe gastric disease (15). Thus, the tight
bacterial adhesion to host tissue mediated by the BabA-Leb interaction could partly explain
why individuals of blood group O phenotype are at higher risk for peptic ulcer disease (5).
Recently, a H. pylori babA deletion mutant, which does not bind Leb (since it is devoid
of the BabA adhesin), nevertheless demonstrated tropism for the target tissue, i.e., the gastric
epithelium and surface mucosa. Further analyzes showed that the babA mutant adheres
preferentially to inflamed gastric mucosa, where adhesion occurs via sialylated antigens,
which are expressed in association with inflammation, in particular the sialylated Lewis x/a
antigens (sLex, sLea). Structural analyzes revealed that H. pylori binds with highest affinity to
an extended form of the sLex antigen (sdimericLex), i.e. a ganglioside with repetitive Lex
motifs, and fully 40% of the sLex binding strains can, in addition to sLex, bind the related
sialyl-Lewis a antigen (sLea) (33). Similar to Leb binding (19), the sLex binding properties
are most prevalent among the virulent type-1 class of H. pylori strains (33). The sLex and
2
sLea oligosaccharides are well characterized binding sites for the E-, L-, and P-selectin
molecules i.e., a family of cell adhesion molecules expressed primarily in response to
activation of local tissue inflammation. Thus, H. pylori was suggested to exploit mechanisms
of “selectin-mimicry” to “home in” on inflammation-activated gastric tissue, mediated by
binding to sLex and sLea (33). This is actually similar to the P-selectin receptors mediated
adherence and invasion mode of neutrophils by human granulocytic Ehrlichiosis (16). By
analyzes of experimentally infected Rhesus monkeys and of gastric biopsy material of
individuals with acid peptic disease, high correlations were found between expression of
sialylated Lewis antigen, gastritis, and H. pylori infection (33). The combination of targeted
high-affinity binding to the fucosylated blood group antigens, such as Lewis b, in the
epithelial lining, followed by the membrane-closer and inflammation associated sialylated
antigens, suggests tightly synchronized mechanisms of programmed and multi-step dependent
adherence for optimal adaptation to the host tissue and local environment during both
conditions of health and disease. Here the fucosylated blood group antigens would
continuously target H. pylori for adherence to the gastric epithelium. During inflammation,
adherence mediated by sialylated glycolipids (gangliosides) would then guide H. pylori to
intimate contact with the host cell membrane, which would benefit the bacteria by improving
access to nutrients leached from damaged host tissues (33).
Adherence through binding to sialylated glycans is prevalent among virulent
microorganisms such as Gram-negative bacteria (E.coli) (28), Gram-positive bacteria
(Actinomyces) (27), influenza and adenoviruses (3)(52), and also by parasites (38). Most
commonly, the requirements for sialylated glycoconjugates have been first described by
analyzes of adherence to RBCs and subsequent aggregation (hemagglutination). Similarly,
soon after discovery of H. pylori, the bacterial cells were shown to agglutinate RBCs, i.e., H.
pylori exhibited hemagglutination activity (11). Since the hemagglutination reactivity was lost
by sialidase (neuraminidase) treatment and soluble N-acetylneuraminyl(sialyl)-D2.3lactose
could inhibit hemagglutination (sia-HA) the hemagglutination activity was suggested to be
sialic acid dependent (12). Neuraminidase sensitive hemagglutination was demonstrated by
one third of clinical isolates (26). The H. pylori adhesin A, HpaA, was originally affinity
purified by use of fetuin (a sialylated protein), and identified as the Nacetylneuraminyllactose binding hemagglutinin (13). However, an isogenic hpaA deletion
mutant in strain CCUG17874 did not demonstrate any reduction in hemagglutination activity
(37). Later, another lab confirmed that HpaA is not involved in hemagglutination and could
show by immuno-gold localization that HpaA is most likely a flagellar sheath protein (20).
Recently, the inflammation associated sLex/a binding adhesin was identified as the sialic acid
binding adhesin, SabA, from H. pylori and a sabA-deletion mutant had lost both sLex and
sLea binding, completely (33). However, the sabA-deletion did not affect the BabA dependent
binding to Leb, which suggests that SabA and BabA are organized and expressed as
independent functional units.
Here, clinical H. pylori isolates were analyzed for neuraminidase-sensitive (sialyldependent) hemagglutination (sia-HA), and sia-HA titers were found to strongly correlate
with bacterial affinity for sLex binding. A sabA deletion mutant was shown to be totally
devoid of sia-HA properties, which suggests that SabA is essential for sia-HA. In addition, the
binding site for SabA on the cell surfaces was topographically mapped to the sialylated
glycosphingolipids (gangliosides) with extended core chains. Among clinical isolates, 3
distinct adherence traits were found, which suggest that the SabA adhesin has the capacity to
adapt its recognition of glycan epitopes to local differences in sialylation-patterns and
inflammation-responses. The series of results demonstrate that the true identity of the H.
pylori N-acetylneuraminyllactose binding hemagglutinin activity functionally resides in the
sialic acid binding adhesin, SabA.
3
MATERIALS AND METHODS
Bacterial strains. The H. pylori strains used in this study were CCUG17875,
CCUG17874 (19), J99 (1), the double mutant 17875babA1::kan babA2::cam (referred to as
the babA1A2-mutant), the 17875/Leb strain which binds Leb but negates sialylated antigens,
J99sabA(JHP662)::cam (J99sabA), and J99babA::cam (J99babA) (33). In addition, the panel
of 80 clinical H. pylori isolates came from the University Hospital in Uppsala, Sweden, and
were described in (19, 33). Bacteria were grown on Brucella agar, supplemented with 10%
bovine blood and 1% IsoVitalex (Becton Dickinson, US) for 35-48 hours at 37qC in 10 %
CO2 and 5% O2 before harvest in PBS, 0.05% Tween 20, 1% BSA (PBS-BB). E.coli strains
J96, HB101-pPAP5, HB101-pAZZ50, HB101- pSH2 (46) were cultured over night on Luria
broth plates supplemented with chloramphenicol and tetracycline.
Hemagglutination conditions and reagents. Fresh human blood from a healthy donor of
blood group A phenotype was used for the hemagglutination assays. The blood cells were
washed twice with phosphate-buffered saline (PBS). A 20% RBC suspension was treated with
trypsin from bovine pancreas (Sigma-Aldrich, St. Louis, MO, US) for 2 hours in neutral pH,
in 28°C, with or without (negative control) 0.05 mg or 0.1mg/ml trypsin (6) followed by 1mM
phenylmethylsulfonylfluoride (PMSF) (Sigma-Aldrich) for 15 minutes in a final
concentration of 1mM. The RBCs were then gently washed 5 times. A 4% suspension of
trypsin treated RBC’s were incubated with neuraminidase type VI from Clostridium
perfringens (Sigma-Aldrich) in the concentration of 0.1 unit/ml according to Hirmo et.al.
(18). The RBC’s were washed several times and a mixture of 0.75% RBC’s was then mixed
with the bacterial suspension in a round bottom ELISA plate and the aggregation of RBC’s
was determined visually after one-hour incubation at room temperature.
Affinity purification of SabA using RBC´s. Bacteria were grown over night, harvested
from 1 agar plate, washed once with PBS and treated with 500 Pl of deionized water for 15
min at room temperature. After centrifugation (30 min, 10 000 g, 4qC) 1/10 vol. of 10xPBS
was added to the supernatant. 200Pl of RBC suspension (4%) (with or without prior
neuraminidase treatment) was mixed with 200 Pl of bacterial cell-surface protein extract for
75 min at room temperature. The RBCs were washed 3 times with 1500 Pl of PBS, suspended
in SDS-PAGE-sample buffer and heated for 5 min at 95qC. The electrophoresis was
performed on a 7,5% Tris-HCl Ready gel (Bio-Rad Laboratories, Hercules, CA). For
immunoblot analyzes, the proteins were transferred from the SDS-PAGE to Immunoblot
PVDF membranes (Bio-Rad). The membranes were blocked for 1 hr in PBS containing
0,05% Tween 20 (PBS-Tween) and 5% non-fat dried milk and probed with rabbit antibodies
raised against the full length recombinant SabA (HP0662), similar to the rabbit antibodies
raised against BabA, as described by Yamaoka et al., (53). The immunoblot analyzes was
then followed by a second incubation with horseradish peroxidase-conjugated goat anti rabbit
antibodies (Dako, Glostrup, Denmark). The antibodies were diluted in PBS-Tween and 1%
non-fat dried milk. The blot was washed as before and developed with the SuperSignal West
Pico Chemiluminiscent Substrate (Pierce, Rockford, CA).
H. pylori Binding to defined sialylated and fucosylated neoglycoconjugate in Radio
Immuno Assay (RIA). The sLex and 3´-sialyl lactose (IsoSep AB, Tullinge, Sweden) and 3´sialyl-N-Acetyllactosamine (3-atom and 14-atom spacer) neoglycoconjugates (Dextra
Laboratories, Reading, UK) were 125I-labeledby the Chloramine-T method (19). H. pylori
were incubated with a mixture of approx. 20.000 CPM of labeled conjugate in 300 ng
unlabeled conjugate/ per assay (denoted the a "conjugate cocktail"). After one hour of
incubation in room temperature (RT), the bacteria were pelleted and the 125I-labeled conjugate
bound to the pellet was measured by gamma scintillation counter (Wallac, Oy, Finland) (19).
Binding experiments were carried out 2-3 times in duplicates.
4
Glycosphingolipids. Total acid glycosphingolipid fractions from human erythrocytes
were obtained by standard procedures (22). The individual glycosphingolipids were isolated
by acetylation of the total glycosphingolipid fractions and repeated chromatography on silicic
acid columns. The acid glycosphingolipid fractions were separated by DEAE-Sepharose
chromatography, followed by repeated silicic acid chromatography. The final separation was
achieved by use of HPLC. The identity of the purified glycosphingolipids was confirmed by
mass spectrometry (41), proton NMR spectroscopy (24), and degradation studies (45)(54). A
detailed description of the isolation and characterization of the H. pylori-binding gangliosides
will be given elsewhere (Roche N. et al., Helicobacter pylori and Complex Gangliosides,
manuscript in preparation).
Thin-Layer Chromatography. Thin-layer chromatography was performed on glass- or
aluminum-backed silica gel 60 HPTLC plates (Merck, Darmstadt, Germany), using
chloroform/metanol/0.25% KCl in water (50:40:10, by volume) as solvent system. Chemical
detection was accomplished by anisaldehyde (50).
35
Chromatogram Binding Assay. The conditions used for culture and S-labeling of the
bacteria have been described previously (2). For binding assays, the bacteria were suspended
to 1x10 8CFU/ml in PBS, pH 7.4. The specific activities of the suspensions were
approximately 1 cpm per 100 H. pylori organisms. The chromatogram binding assays were
done as described (2). Mixtures of glycosphingolipids (40 µg/lane) or pure compounds (1-4
µg/lane) were separated on aluminum-backed silica gel plates. The dried chromatograms were
soaked for 1 min in diethylether/n-hexane (1:5, by volume) containing 0.5% (w/v)
polyisobutylmethacrylate (Aldrich Chem. Comp. Inc., Milwaukee, WI). After drying, the
chromatograms were coated with 2% bovine serum albumin (w/v) and 0.1% (w/v) Tween 20
35
in PBS for 2 h at room temperature. Thereafter, a suspension of S-labeled bacteria (diluted
in PBS to 1 x 108 CFU/ml and 1-5 x 106 cpm/ml) was gently sprinkled over the
chromatograms and incubated for 2 h at room temperature. After washing six times with PBS,
and drying, the thin-layer plates were autoradiographed for 3-120 h using XAR-5 x-ray films
(Eastman Kodak, Rochester, NY).
Analyzes of affinity in BabA/Leb-binding. For affinity analyzes (42), the 125I-labeled
sialylated and or fucosylated glycoconjugate were diluted with the corresponding unlabeled
conjugate in series ranging from 0.2-5 ng/ml of substrate. For each assay, 1200 CPM of 125I labeled conjugate was combined with 1 ml of bacteria of 0.001 OD A600. Strain CCUG17874
that does not bind Leb, or strain 17875/Leb that does not bind sialylated conjugates was added
in 0.1 OD A600 as combined carrier for the binding analyzes and as blocking agent to reduce
non-specific binding. Bacteria were incubated in room temperature (RT), 23ºC, for 15 h, in
PBS-BB-buffer, and binding was analyzed as described above by gamma scintillation.
Electron microscopy of H.pylori binding to red blood cells.
Fresh blood was treated with trypsin as mentioned earlier and mixed with strain J99. The
mixture was incubated at room temperature for one hour, then the mixture was gently washed
two times. The pellet was then fixed in glutaraldehyde/paraformaldehyde, post-fixed in
osmium tetroxide, dehydrated in a graded series of alcohol, infiltrated in Agar 100 (Agar
Scientific Ltd), and polymerized at 60ºC for 40 hours. Ultra-thin sections were viewed in a
Philips CM12 transmission electron microscope at 80 kV.
5
RESULTS
Correlation between sialyl-dependent hemagglutination and bacterial binding to
soluble sLex/a antigen. A series of 80 Swedish clinical isolates were analyzed for sialyldependent hemagglutination (sia-HA). Sia-HA was assessed by glycosidic cleavage and
removal of sialic acid derivatives from the RBC surfaces, which reduce the hemagglutination
titers. Neuraminidase from Clostridium perfringens was used; an enzyme that with broad
substrate specificity cleaves D2.3, D2.6 or D2.8 linked sialic acids bound to oligosaccharide
chains of glycoproteins and glycolipids. Hemagglutination titers that are dependent on sialic
acid derivatives would then be lowered, due to reduced bacterial aggregation. For the series of
80 clinical isolates, the resulting series of shift in titers ranged from –2 to •2. Here, in 22
strains (28%) lowered sia-HA titers were found (difference 1-•2). In contrast, 31 strains
(39%) instead increased their HA titers, due to the sialic acid removal (presented as negative
values for difference in sia-dependent-HA titers). The remaining 27 (34%) isolates displayed
no changes in sia-HA titers by removal of the RBC sialic acid residues (Fig.1). A total of 35
out of the 80 strains (44%) were shown to bind the sLex glycoconjugate, which is similar to
the prevalence described by Mahdavi et al (33). The sia-HA titers were then compared with
the binding to the soluble sLex conjugate, where 19 out of the 35 (54%) strains that bind sLex
caused sia-HA, while only 3 strains (7%) out of the 45 strains, which did not bind to sLex,
could still promote sia-HA. Therefore, the 31 isolates, which instead increased their sia-HA
titers after neuraminidase treatment, were similarly most modest in sLex binding. Thus, a high
correlation (0.614, p= 0.000) was found between bacterial binding to sLex antigen (rank
sLex) and, the difference in HA titers after neuraminidase treatment of the RBCs (rank HA)
(Fig.1), i.e., strains that are reduced in hemagglutination activity by neuraminidase treatment
belong to the group of strain that bind sLex the best.
Prevalence of sLex binding among clinical isolates. The panel of clinical isolates was
analyzed for binding to sialylated and soluble 125I-labeled glycoconjugates where 16 (45%) of
the 35 sLex binding strains also bind the closely related sLea antigen, in accordance with
recent results from Mahdavi et. al., 2002 (33). Here, the strains were further analyzed for their
different structural requirements of fucosylated residues in the core chains for binding.
Distinct differences were here found close to 90% of the strains that can bind sLea also
recognized the sialyl-lactosamine conjugate. In comparison, only 2 (10%) out of the 19 strains
that bind sLex but not sLea are able to bind the sialyl-lactosamine conjugates (Fig.2).
The sialic acid binding adhesin, SabA, identified as the sia-hemagglutinin by deletion
mutants. The reference strains CCUG17874 (19) and J99 (1) were analyzed for sia-HA,
where neuraminidase treatment resulted in complete loss of hemagglutination titers. This
result demonstrates that additional binding properties described for H. pylori are of less
importance for sia-hemagglutination of H. pylori. A J99sabA deletion mutant recently showed
a complete loss of binding properties for both sLex and sLea antigens (33). The strong
correlation found between sLex binding and sia-HA titers (Fig.1), suggest that the siahemagglutinin characterized here could be identical to the SabA adhesin. Thus the H. pylori
J99sabA deletion mutant and the J99 wild-type strain were compared for sia-HA properties
(in Table 1). The J99sabA mutant demonstrated a total loss of sialyl-dependent HA properties,
which concludes that the SabA adhesin is the correct identity for the H. pylori sialylhemagglutinin. In comparison, the H. pylori J99babA deletion mutant show no reduction in
hemagglutination activity, which argues for that BabA interactions with fucosylated blood
group antigens support neither RBC binding, nor hemagglutination activity.
Affinity protein purification identified SabA as the sia-hemagglutinin. The specificity
in the interaction of SabA to RBCs was analyzed by affinity purification. Here, bacterial cellsurface protein extract from strains CCUG17874, and 17875:'babA1A2 (which both bind
sLex) were mixed with naive and neuraminidase treated RBCs, respectively. Immunoblot
analyzes showed that SabA could only be affinity-absorbed/purified by RBCs with sialylated
6
surface glycosylation patterns, since no SabA was purified by use of neuraminidase treated
RBCs, i.e., devoid of sialylated glycoconjugates (Fig.3).
Topographic localization of the sialylated binding sites on red blood cells that confer siahemagglutinin by H. pylori. For topographic mapping of sialylated binding sites i.e.,
neuraminidase sensitive hemagglutination, RBCs were first protease treated with trypsin to
remove glycoproteins from the RBC surfaces. Then the RBCs were treated with
neuraminidase to analyze for remaining sialylated-binding sites, i.e. glycoconjugates within
the freshly exposed surfaces that could confer sia-HA properties. The processed RBCs were
then analyzed for sia-HA by the wild-type strain J99. Surprisingly, removal of the RBC
glycoproteins did not abolish or even reduce the sia-HA titers. To verify the efficacy and
stringency of the RBC protease treatment, urinary tract infectious P-fimbriated E.coli that
recognize GalD1.4Gal-presenting glycolipids, were analyzed as a reference for comparison.
Here, the protease treated RBC surfaces conferred stronger hemagglutination due to increased
(steric) accessibility of the adhesive P-fimbriae for the GalD1.4Gal-receptor epitopes
presented in the membrane-close glycolipids (GSLs). In contrast, limited protease treatment
of the RBCs fully and rapidly abrogated sia-HA by S-fimbriated E.coli, known to
hemagglutinate through adherence to (abundant) surface presented sialylated RBC
glycoproteins, such as Glycophorin A (39). Since the HA-responses by the P- and Sfimbriated E.coli were conclusive, the results were no reduction in sia-HA was demonstrated
by H. pylori, as a consequence of protease treatment, strongly suggest that sia-HA of H.
pylori is less dependent on glycoproteins for membrane attachment and aggregation. Instead,
cellular binding by SabA is mediated primarily by binding of SabA to the RBC sialylated
glycolipids i.e., gangliosides.
Preferential binding of SabA to human erythrocyte gangliosides with extended core
chains. The major RBC glycolipid constituents are the GM3 ganglioside and the NeuAcD2.3neolactotetraosylceramide (known as sialyl-paragloboside). In addition, a number of complex
gangliosides have been described. The purified gangliosides are listed in Table 2. Here,
strains CCUG17874 (a reference strain which does not bind Leb, but binds sLex), and J99
demonstrated weak binding to the NeuAcD2.3-neolactotetraosylceramide region (the band
chemically stained in lane 3, Fig 4A), while neither the GM3 region nor the more slowmigrating gangliosides in the total acid fraction of human RBCs stained positive for H. pylori
binding (Fig.4B and 4C; lane 2). In comparison, binding to the total acid fraction of human
neutrophil granulocytes conferred strong binding to the complex and thus slow-migrating
gangliosides as previously described (33). This could in part be due to different levels of
available receptor structures, and after purification of the human erythrocyte gangliosides
extract used in lane 2, additional binding patterns emerged (lanes 3-7). While binding to
NeuAcD2.3-neolactotetraosylceramide (lane 3) was still weak (as described above), and no
interaction with NeuAcD2.6-neolactotetraosylceramide was obtained (lane 4), distinct
binding of H. pylori to complex and extended (thus slow-migrating) gangliosides was found
(lanes 5-7). Thus, both H. pylori strains bound NeuAcD2.3-neolactohexaosylceramide and to
the G9-B ganglioside of similar core chain length (in Table 2). The TLC analyzes showed
distinct binding of H. pylori to linear and branched N-actyl-lactosamine based monosialogangliosides with terminal NeuAcD3Galß4GlcNAcß. The only disialo-ganglioside
recognized was G-10 ganglioside, which carries two NeuAcD2.3Galß1.4GlcNAcßҏglycans. As
anticipated, the J99sabA deletion mutant did not recognize any of the gangliosides bound by
the wild type strains (Fig.4D) since the J99sabA mutant has lost its sialyl-dependent binding
properties.
Characterization of the detailed H. pylori binding specificities by soluble sialylglycans of various complexities. By use of defined sialylated glycoconjugates, the influence
of both length of glycan core-chain and also the complexity in fucosylation for H. pylori
binding was investigated by radio-immuno assays (RIA). The bacterial binding to soluble
sLex conjugate was found to range from 30% down to background levels among clinical
7
isolates (Fig. 1), This result suggests that the SabA adhesins exhibit a spectra of different
affinities for sialylated receptors, which could then relate to distinct differences in the detailed
receptor recognition of SabA among clinical isolates. Thus, SabA might display distinct
polymorphism in the carbohydrate binding domains to support detailed recognition of
distinctly different sialyl-glycan epitopes. This possibility was analyzed by a series o
sialylated glycoconjugates based on the simple sialyl-lactose structure conjugated to human
serum albumin (where the glucose ring had been opened by reductive amination to leave only
the NeuAcD3Gal-di-saccharide). Binding of NeuAcD3Gal by strain CCUG17874, the
babA1A2-mutant and strain J99 was low, in particular compared to the Swedish clinical
isolates SMI65 and SMI27 (Fig.5). The reduced binding of sialyl-lactose by the reference
strains (compared to the SMI65/27 isolates) could have been influenced by the short core
chain. Alternatively, binding could be hampered by less efficient presentation of the sialic
acid residue, in part due to the lack of the D1.3 fucose residue(s), as compared to the highaffinity sLex glycan. To discriminate between these two possibilities, sialyl-lactosamine
conjugates with an intact tri-saccharide, i.e., 3'-sialyl-N-Acetyl-lactosamine attached to
albumin by either 3-atom or 14-atom spacers structures were analyzed for H. pylori binding
and for any possible gain in presentation of the sialylated binding epitope. Both the babA1A2
mutant and the CCUG17874 strains responded to both the 3-atom and the flexible 14-atom
spacer by several-fold increases in binding (Fig.5). Interestingly, the spacer-length related
gain in binding is not a universal trait among H. pylori strains, since the clinical isolates
SMI65 and SMI27 did not increase in binding, Or even reduced their binding compared to the
sialyl-lactose (di)-saccharide. Similarly to previous results, strains 26695 and MO19 lack
binding properties for Leb, sLex and sLea (19)(33) and similarly, they bind to neither of the
sialyl-lactose/lactosamine conjugates used here (data not shown).
Analyzes of binding affinities for sLex, sLea, and sialyl-lactose. The affinities of strain
SMI65 for sialylated conjugates of different complexities were analyzed according to
Scatchard (42). Strain SMI65 demonstrated the affinity of 1.08x109M-1 for sLex, and
6.5x108M-1 for sialyl-lactose. In comparison, strain CCUG17875 binds sLex with the 10-fold
lower affinity of 1x108M-1 (33). The modest binding of CCUG17875 to sialyl-lactose
provides an affinity value in the order of 1x107M-1. Thus, the rather high-affinity interaction
as demonstrated by SabA varies by several 10-folds among clinical isolates and between
sialylated antigens of different complexities. Nevertheless, the SabA-sialyl glycan interaction
is still several orders lower compared to the exceedingly high-affinity of 1x1010-11M-1
demonstrated by BabA for Leb (19)( Aspholm et. al., in manuscript).
H. pylori adheres to red blood cells by membrane indentations. The binding of H.
pylori to RBCs was visualized by EM, with/without prior protease treatment of the RBCs cell
surfaces. There was no visible difference in H. pylori attachment to treated or untreated
RBCs, but almost all sites of contact between bacteria and RBCs show an indentation of the
cell membrane (Fig.6).
8
DISCUSSION
Microbial adherence mediated by sialylated glycoconjugates has in most cases initially been
analyzed by sialic acid dependent hemagglutination, i.e. sia-HA. Similarly, the gastric
pathogen H. pylori was soon shown to bind various glycoconjugates, where one third of fresh
clinical H. pylori isolates cause sia-hemagglutination (26).
Bacterial binding to the sialylated host tissue could confer dynamic traits such as homingin onto local areas of neutrophils invasion and inflammation, where the microorganisms could
utilize the nutrients released from damaged cells. For H. pylori the sia-HA phenotype has not
been defined as a virulence property since the prevalence of sia-HA has not been shown to
correlate with gastric disease (48). However, the true prevalence of sialic acid binding
affinities among clinical isolates could have been underestimated, since; 1) H. pylori exhibit a
high level of ON/OFF phase variation in binding to sialylated glycoconjugates, to suggest that
the sia-HA properties could rapidly be lost in strains with high phase-variation characteristics
(33); 2) almost half of the strains that bind sLex still did not cause sia-HA, which suggests
that many strains exhibit requirements for complex sialylated glycans for high-affinity
binding, structures that are not available on the RBCs; 3) in addition, culture conditions have
also been shown to influence binding specificity and affinity for various sialylated
glycoconjugates (34).
In this report, we could show that the clinical isolates exhibit three major
hemagglutination-patterns in response to neuraminidase treatment; 1) reduced HA-titers, 2)
no change (unaltered) HA titers or 3) increased HA titers (Fig.1). The first pattern would be
the expected shift in HA titers, where the reduced titers relate to the sialidase dependent loss
of sialic acid. The third pattern, which comes with increased HA titers suggests that the
removal of the sialic acid from the glycan chain, might have exposed (cryptic) binding sites
within the core-chain, i.e., sterically less accessible residues or oligosaccharide epitopes
would come out better presented for the microbes, with the removal of the charged sialic acid
at the glycan terminus as a consequence of neuraminidase treatment. The second pattern, with
no change in HA titers was demonstrated by a combination of strains that either lack binding
properties, or alternatively by strains with combinations of adhesins, where the outcome of
neuraminidase treatment would be leveled out. Thus, H. pylori demonstrates a strong
correlation between binding to sLex mediated by SabA and sia-HA.
The prerequisite for optimal recognition of the sialylated epitopes was further revealed by
the difference in bacterial affinities for soluble sialylated antigens of various complexities.
Two main modes of bacterial binding were revealed, where strains CCUG17874/5 bind sLex
with highest affinity, while the affinity for sialyl-lactose was lower. In comparison, the
clinical isolates SMI65 and SMI27 demonstrated less stringent binding properties and bind in
addition to sLex the structurally related sLea, and the sialyl-lactose glycans. The results
suggest that H. pylori strains bind the sialylated glycans differently, as has been described for
the GalD1.4Gal-binding P-fimbriated E.coli, where length and complexity of the glycolipid
binding sites have organized the G-adhesins into different functional classes/subtypes (46).
To investigate the influence of length and flexibility of the glycan core chain, the trisaccharide sialyl-lactosamine conjugated to albumin with either 3-atom or 14 atom spacers
were analyzed for binding efficiency. Strains CCUG17874/5 demonstrated an increased
binding to sialyl-lactosamine with the 14-atom spacer (compared to the 3-atom spacer)
compared to the much shorter sialyl-lactose (here presented as a di-saccharide). In contrast,
the SMI65/27 strains did not respond in terms of binding abilities/affinities to the increased
spacer lengths. Similarly, among the panel of clinical isolates, three rather distinct sialic acid
dependent adherence traits could be deduced; (a) strains that are most specific for the sLex
antigen (the strains clustered in the lower left part of Fig. 2), (b) almost half of the isolates
that bind sLex also bind the sLea antigen and this slightly promiscuous binding mode also
includes the shorter sialyl-lactosamine glycan (such as strains SMI65/27), (c) strains that bind
9
the sLex antigen and the sialyl-lactosamine structure while binding to sLea is low, i.e., strains
with a “V-shaped” binding mode in Fig 2 (such as strains CCUG17874/5). The binding
results, which also reflect the SabA binding affinity for the different isolates, suggest that
strains proficient in sLea binding in general exhibit higher affinity for sLex binding (Fig.2).
These different subtypes of adherence modes, might relate to detailed bacterial tropism
such as targeting of inflamed tissue versus dysplastic tissue. Strains that are associated with
severe gastric disease express the established virulence markers, VacA, and CagA (the type 1
strains) most often express the BabA and SabA adhesins (19)(15)(33). In this respect it is of
particular interest that strains which demonstrate promiscuous binding to sialylated antigens
bind the sLea antigen, because sLea is an established tumor antigen (31) and marker of gastric
dysplasia (43). In the inflamed gastric mucosa extended sialylated gangliosides with
fucosylated repeat motifs such as the sialyl-dimeric-Lewis x glycan could be the main targets
for SabA mediated binding (33). To characterize the topographic localization of the sialic acid
residues the surface of the RBCs were treated with trypsin in combination with
neuraminidase. The results of these treatments demonstrated that the sialic acid glycans to
which H. pylori binds are located in the glycolipid layer and not in the surface displayed
glycoproteins. In the RBC model for bacterial binding, the extended sialylated gangliosides
exhibit the highest affinities (Fig.4 and Table 2).
The tight binding and apposition of H. pylori to the RBC membrane might be a
consequence of bacterial attachment to the sialic acid residues in the gangliosides. During the
adherence process, the RBCs membrane in close contact with the bacteria outer membrane
surfaces is probably locally enriched for gangliosides due to re-organization of the spatial
distribution of areas with gangliosides versus glycoproteins. This, might result in the
activation of the cytoskeleton molecules such as actin filaments, and the formation of local
indentations. By this report, the SabA adhesin replaces the H. pylori adhesin A, HpaA, as the
H. pylori sia-hemagglutinin, as concluded both by defined gene deletions and by protein
affinity-purification analyzes. Although HpaA was first characterized by Evans et.al., already
a decade ago (13), the correct identity of HpaA has since then been critically disputed
(37)(20)(30). Though, correct identification of proteins with binding properties is most often
complicated. Among other H. pylori proteins that have been assigned adhesive properties is
the 63kDa adhesin candidate that binds to phospholipid phosphatidylethanolamine (PE) which
is integrated in the cell membrane and could possibly facilitate signal transduction (29).
However, this protein was later shown to be a catalase (36) and seems unlikely to be a true
adhesin. In addition, chromatography by fetuin-based affinity-columns resulted in the
purification of the H. pylori ferritin protein Pfr (later identified as an iron-binding protein)
(14). However, immobilized sialylated and sulfated oligosaccharide compounds such as fetuin
and heparan sulfate, and glycolipids are highly charged and often result in unspecific binding,
i.e., in addition to the specific ligand-receptor binding. Therefore, the genetic tools to make
defined gene deletion mutants are of major importance for interpretation of complex binding
results, especially for H. pylori where the key-(adhesive)-players are most likely all members
of the highly conserved H. pylori family of HOP proteins.
10
REFERENCES
1. Alm, R. A., L. S. Ling, D. T. Moir, B. L. King, E. D. Brown, P. C. Doig, D. R.
Smith, B. Noonan, B. C. Guild, B. L. deJonge, G. Carmel, P. J. Tummino, A. Caruso, M.
Uria-Nickelsen, D. M. Mills, C. Ives, R. Gibson, D. Merberg, S. D. Mills, Q. Jiang, D. E.
Taylor, G. F. Vovis, and T. J. Trust. 1999. Genomic-sequence comparison of two unrelated
isolates of the human gastric pathogen Helicobacter pylori. Nature 397:176-80.
2.
Angstrom, J., S. Teneberg, M. A. Milh, T. Larsson, I. Leonardsson, B. M. Olsson,
M. O. Halvarsson, D. Danielsson, I. Naslund, A. Ljungh, T. Wadstrom, and K. A.
Karlsson. 1998. The lactosylceramide binding specificity of Helicobacter pylori.
Glycobiology 8:297-309.
3.
Arnberg, N., K. Edlund, A. H. Kidd, and G. Wadell. 2000. Adenovirus type 37 uses
sialic acid as a cellular receptor. J Virol 74:42-8.
4.
Borén, T., P. Falk, K. A. Roth, G. Larson, and S. Normark. 1993. Attachment of
Helicobacter pylori to human gastric epithelium mediated by blood group antigens. Science
262:1892-5.
5.
Borén, T., S. Normark, and P. Falk. 1994. Helicobacter pylori: molecular basis for
host recognition and bacterial adherence. Trends Microbiol 2:221-8.
6.
Borén, T., T. Wadström, S. Normark, J. I. Gordon, and P. G. Falk. 1997.
Helicobacter pylori protocols. Humana Press, Totowa, New Jersey.
7.
Chmiela, M., B. Paziak-Domanska, and T. Wadstrom. 1995. Attachment, ingestion
and intracellular killing of Helicobacter pylori by human peripheral blood mononuclear
leukocytes and mouse peritoneal inflammatory macrophages. FEMS Immunol Med Microbiol
10:307-16.
8.
Cover, T. L., C. P. Dooley, and M. J. Blaser. 1990. Characterization of and human
serologic response to proteins in Helicobacter pylori broth culture supernatants with
vacuolizing cytotoxin activity. Infect Immun 58:603-10.
9.
Edwards, N. J., M. A. Monteiro, G. Faller, E. J. Walsh, A. P. Moran, I. S. Roberts,
and N. J. High. 2000. Lewis X structures in the O antigen side-chain promote adhesion of
Helicobacter pylori to the gastric epithelium. Mol Microbiol 35:1530-9.
10. Ekstrom, A. M., M. Held, L. E. Hansson, L. Engstrand, and O. Nyren. 2001.
Helicobacter pylori in gastric cancer established by CagA immunoblot as a marker of past
infection. Gastroenterology 121:784-91.
11. Emody, L., A. Carlsson, A. Ljungh, and T. Wadstrom. 1988. Mannose-resistant
haemagglutination by Campylobacter pylori. Scand J Infect Dis 20:353-4.
12. Evans, D. G., D. J. Evans, Jr., J. J. Moulds, and D. Y. Graham. 1988. Nacetylneuraminyllactose-binding fibrillar hemagglutinin of Campylobacter pylori: a putative
colonization factor antigen. Infect Immun 56:2896-906.
13. Evans, D. G., T. K. Karjalainen, D. J. Evans, Jr., D. Y. Graham, and C. H. Lee.
1993. Cloning, nucleotide sequence, and expression of a gene encoding an adhesin subunit
protein of Helicobacter pylori. J Bacteriol 175:674-83.
14. Frazier, B. A., J. D. Pfeifer, D. G. Russell, P. Falk, A. N. Olsen, M. Hammar, T. U.
Westblom, and S. J. Normark. 1993. Paracrystalline inclusions of a novel ferritin containing
nonheme iron, produced by the human gastric pathogen Helicobacter pylori: evidence for a
third class of ferritins. J Bacteriol 175:966-72.
15. Gerhard, M., N. Lehn, N. Neumayer, T. Borèn, R. Rad, W. Schepp, S. Miehlke, M.
Classen, and C. Prinz. 1999. Clinical relevance of the Helicobacter pylori gene for bloodgroup antigen-binding adhesin. Proc Natl Acad Sci U S A 96:12778-83.
16. Herron, M. J., C. M. Nelson, J. Larson, K. R. Snapp, G. S. Kansas, and J. L.
Goodman. 2000. Intracellular parasitism by the human granulocytic ehrlichiosis bacterium
through the P-selectin ligand, PSGL-1. Science 288:1653-6.
11
17. Hessey, S. J., J. Spencer, J. I. Wyatt, G. Sobala, B. J. Rathbone, A. T. Axon, and M.
F. Dixon. 1990. Bacterial adhesion and disease activity in Helicobacter associated chronic
gastritis. Gut 31:134-8.
18. Hirmo, S., S. Kelm, R. Schauer, B. Nilsson, and T. Wadström. 1996. Adhesion of
Helicobacter pylori strains to alpha-2,3-linked sialic acids. Glycoconj J 13:1005-11.
19. Ilver, D., A. Arnqvist, J. Ögren, I. M. Frick, D. Kersulyte, E. T. Incecik, D. E. Berg,
A. Covacci, L. Engstrand, and T. Borén. 1998. Helicobacter pylori adhesin binding
fucosylated histo-blood group antigens revealed by retagging. Science 279:373-7.
20. Jones, A. C., R. P. Logan, S. Foynes, A. Cockayne, B. W. Wren, and C. W. Penn.
1997. A flagellar sheath protein of Helicobacter pylori is identical to HpaA, a putative Nacetylneuraminyllactose-binding hemagglutinin, but is not an adhesin for AGS cells. J
Bacteriol 179:5643-7.
21. Kannagi, R., D. Roelcke, K. A. Peterson, Y. Okada, S. B. Levery, and S. Hakomori.
1983. Characterization of an epitope (determinant) structure in a developmentally regulated
glycolipid antigen defined by a cold agglutinin Fl, recognition of alpha-sialosyl and alpha-Lfucosyl groups in a branched structure. Carbohydr Res 120:143-57.
22. Karlsson, K. A. 1987. Preparation of total nonacid glycolipids for overlay analysis of
receptors for bacteria and viruses and for other studies. Methods Enzymol 138:212-20.
23. Kirschner, D. E., and M. J. Blaser. 1995. The dynamics of Helicobacter pylori
infection of the human stomach. J Theor Biol 176:281-90.
24. Koerner, T. A., Jr., J. H. Prestegard, P. C. Demou, and R. K. Yu. 1983. Highresolution proton NMR studies of gangliosides. 1. Use of homonuclear two-dimensional spinecho J-correlated spectroscopy for determination of residue composition and anomeric
configurations. Biochemistry 22:2676-87.
25. Kundu, S. K., B. E. Samuelsson, I. Pascher, and D. M. Marcus. 1983. New
gangliosides from human erythrocytes. J Biol Chem 258:13857-66.
26. Lelwala-Guruge, J., Å. Ljungh, and T. Wadström. 1992. Haemagglutination patterns
of Helicobacter pylori. Frequency of sialic acid-specific and non-sialic acid-specific
haemagglutinins. Apmis 100:908-13.
27. Li, T., I. Johansson, D. I. Hay, and N. Strömberg. 1999. Strains of Actinomyces
naeslundii and Actinomyces viscosus exhibit structurally variant fimbrial subunit proteins and
bind to different peptide motifs in salivary proteins. Infect Immun 67:2053-9.
28. Lindahl, M., and T. Wadstrom. 1984. K99 surface haemagglutinin of enterotoxigenic
E. coli recognize terminal N-acetylgalactosamine and sialic acid residues of glycophorin and
other complex glycoconjugates. Vet Microbiol 9:249-57.
29. Lingwood, C. A., M. Huesca, and A. Kuksis. 1992. The glycerolipid receptor for
Helicobacter pylori (and exoenzyme S) is phosphatidylethanolamine. Infect Immun 60:24704.
30. Lundstrom, A. M., K. Blom, V. Sundaeus, and I. Bolin. 2001. HpaA shows variable
surface localization but the gene expression is similar in different Helicobacter pylori strains.
Microb Pathog 31:243-53.
31. Magnani, J. L., M. Brockhaus, D. F. Smith, V. Ginsburg, M. Blaszczyk, K. F.
Mitchell, Z. Steplewski, and H. Koprowski. 1981. A monosialoganglioside is a monoclonal
antibody-defined antigen of colon carcinoma. Science 212:55-6.
32. Mahdavi, J., T. Boren, C. Vandenbroucke-Grauls, and B. J. Appelmelk. 2003.
Limited role of lipopolysaccharide Lewis antigens in adherence of Helicobacter pylori to the
human gastric epithelium. Infect Immun 71:2876-80.
33. Mahdavi, J., B. Sonden, M. Hurtig, F. O. Olfat, L. Forsberg, N. Roche, J.
Angström, T. Larsson, S. Teneberg, K. A. Karlsson, S. Altraja, T. Wadström, D.
Kersulyte, D. E. Berg, A. Dubois, C. Petersson, K. E. Magnusson, T. Norberg, F. Lindh,
B. B. Lundskog, A. Arnqvist, L. Hammarström, and T. Borén. 2002. Helicobacter pylori
SabA adhesin in persistent infection and chronic inflammation. Science 297:573-8.
12
34. Miller-Podraza, H., J. Bergstrom, M. A. Milh, and K. A. Karlsson. 1997.
Recognition of glycoconjugates by Helicobacter pylori. Comparison of two sialic aciddependent specificities based on haemagglutination and binding to human erythrocyte
glycoconjugates. Glycoconj J 14:467-71.
35. Miller-Podraza, H., J. Bergstrom, S. Teneberg, M. A. Milh, M. Longard, B. M.
Olsson, L. Uggla, and K. A. Karlsson. 1999. Helicobacter pylori and neutrophils: sialic
acid-dependent binding to various isolated glycoconjugates. Infect Immun 67:6309-13.
36. Odenbreit, S., B. Wieland, and R. Haas. 1996. Cloning and genetic characterization of
Helicobacter pylori catalase and construction of a catalase-deficient mutant strain. J Bacteriol
178:6960-7.
37. O'Toole, P. W., L. Janzon, P. Doig, J. Huang, M. Kostrzynska, and T. J. Trust.
1995. The putative neuraminyllactose-binding hemagglutinin HpaA of Helicobacter pylori
CCUG 17874 is a lipoprotein. J Bacteriol 177:6049-57.
38. Pandey, K., S. Singh, P. Pattnaik, C. Pillai, U. Pillai, A. Lynn, S. Jain, and C.
Chitnis. 2002. Bacterially expressed and refolded receptor binding domain of Plasmodium
falciparum EBA-175 elicits invasion inhibitory antibodies. Mol Biochem Parasitol 123:23.
39. Parkkinen, J., G. N. Rogers, T. Korhonen, W. Dahr, and J. Finne. 1986.
Identification of the O-linked sialyloligosaccharides of glycophorin A as the erythrocyte
receptors for S-fimbriated Escherichia coli. Infect Immun 54:37-42.
40. Sakamoto, S., T. Watanabe, T. Tokumaru, H. Takagi, H. Nakazato, and K. O.
Lloyd. 1989. Expression of Lewisa, Lewisb, Lewisx, Lewisy, siayl-Lewisa, and sialyl-Lewisx
blood group antigens in human gastric carcinoma and in normal gastric tissue. Cancer Res
49:745-52.
41. Samuelsson, B. E., W. Pimlott, and K. A. Karlsson. 1990. Mass spectrometry of
mixtures of intact glycosphingolipids. Methods Enzymol 193:623-46.
42. Scatchard, G. 1949. Ann. N.Y. Acad. Sci. 51:660.
43. Sipponen, P., and J. Lindgren. 1986. Sialylated Lewis determinant CA 19-9 in benign
and malignant gastric tissue. Acta Pathol Microbiol Immunol Scand [A] 94:305-11.
44. Slomiany, B. L., J. Piotrowski, A. Samanta, K. VanHorn, V. L. Murty, and A.
Slomiany. 1989. Campylobacter pylori colonization factor shows specificity for
lactosylceramide sulfate and GM3 ganglioside. Biochem Int 19:929-36.
45. Stellner, K., H. Saito, and S. I. Hakomori. 1973. Determination of aminosugar
linkages in glycolipids by methylation. Aminosugar linkages of ceramide pentasaccharides of
rabbit erythrocytes and of Forssman antigen. Arch Biochem Biophys 155:464-72.
46. Stromberg, N., P. G. Nyholm, I. Pascher, and S. Normark. 1991. Saccharide
orientation at the cell surface affects glycolipid receptor function. Proc Natl Acad Sci U S A
88:9340-4.
47. T.L. Cover, e. a. 2001. in Principles of Bacterial Pathogenesis,. Academic Press, New
York.
48. Taylor, N. S., A. T. Hasubski, J. G. Fox, and A. Lee. 1992. Haemagglutination
profiles of Helicobacter species that cause gastritis in man and animals. J Med Microbiol
37:299-303.
49. Tomb, J. F., O. White, A. R. Kerlavage, R. A. Clayton, G. G. Sutton, R. D.
Fleischmann, K. A. Ketchum, H. P. Klenk, S. Gill, B. A. Dougherty, K. Nelson, J.
Quackenbush, L. Zhou, E. F. Kirkness, S. Peterson, B. Loftus, D. Richardson, R.
Dodson, H. G. Khalak, A. Glodek, K. McKenney, L. M. Fitzegerald, N. Lee, M. D.
Adams, J. C. Venter, and et al. 1997. The complete genome sequence of the gastric
pathogen Helicobacter pylori. Nature 388:539-47.
50. Waldi, D. 1962. Sprühreagentien für die dünnschicht-chromatographie, p. 496-515.
In E. Stahl. (ed.), Dünnschicht-Chromatographie. Springer-Verlag, Berlin.
51. Watanabe, K., M. Powell, and S. Hakomori. 1978. Isolation and characterization of a
novel fucoganglioside of human erythrocyte membranes. J Biol Chem 253:8962-7.
13
52. Wiley, D. C., and J. J. Skehel. 1987. The structure and function of the hemagglutinin
membrane glycoprotein of influenza virus. Annu Rev Biochem 56:365-94.
53. Yamaoka, Y., J. Souchek, S. Odenbreit, R. Haas, A. Arnqvist, T. Boren, T.
Kodama, M. S. Osato, O. Gutierrez, J. G. Kim, and D. Y. Graham. 2002. Discrimination
between cases of duodenal ulcer and gastritis on the basis of putative virulence factors of
Helicobacter pylori. J Clin Microbiol 40:2244-6.
54. Yang, H. J., and S. I. Hakomori. 1971. A sphingolipid having a novel type of
ceramide and lacto-N-fucopentaose 3. J Biol Chem 246:1192-200.
14
ACKNOWLEDGEMENTS
We thank Anne von Euler Matell for skilful assistance of electron microscopy imaging. We
are also grateful for professional technical support by Rolf Sjöström. This study was
supported by grants from the Foundation for Strategic Research, SSF; Program
“Glycoconjugates in Biological Systems” (B.S./T.B and N.H./S.T) and “Infection &
Vaccinology” (M.A./TB.), the Swedish Medical Research Council (11218/T.B., 10840/LE,
12628/S.T.), Umeå University Biotechnology Fund (T.B), County Council of Västerbotten
(T.B), and Swedish Cancer Society, Project No 4101-B00-03XAB (T.B. and 4128-B9902XAB (S.T.), Wallenberg Foundation (S.T.), and J. C. Kempe Memorial Foundation (M.A.).
15
FIGURE LEGENDS
Fig. 1
Bacterial binding to soluble sLex antigen and sia-hemagglutination
Correlation between sLex conjugate binding in percent of total added conjugate (shown on
the X-axis) and sia-hemagglutination titers. The numbers on the Y-axis show the number of
strains in each group.
Fig.2
Prevalence of sLex antigen binding among clinical isolates
Clinical isolates were analyzed for binding capacity for sialylated and soluble 125I-labeled
glycoconjugates. The Y-axis shows the percent of total CPM i.e. 100% and the X-axis shows
the different glycoconjugates tested.
Fig.3
Affinity purification of SabA by use of RBCs
Bacterial cell-surface proteins were extracted from strain 17874 and strain 17875 babA1A2respectively. By using RBCs SabA was were affinity purified from these extracts. The
immuno staining confirms the presence of SabA as a result of binding to sialic acid residues
on the RBC surface. Lanes 1 and 3 show the detected SabA. Lanes 2 and 4 demonstrate the
lack of affinity purification of SabA when the RBCs are neuraminidase treated. Molecular
weigh markers (in kilo Dalton) are indicated on the left.
Fig.4
Binding of SabA to extended erythrocyte gangliosides
Comparison of binding of H. pylori strains CCUG17874, J99 and J99sabA- to human
erythrocyte gangliosides. (A) Chemical detection by anisaldehyde and (B-D) autoradiograms
35
obtained by binding of S-labeled H. pylori strains CCUG17874, J99 and J99sabA-. The
gangliosides were separated on aluminum-backed silica gel plates, using
chloroform/metanol/0.25% KCl in water (50:40:10, by volume) as solvent system, and the
binding assays were performed as described in the "Materials and methods" section. The lanes
contain: Lane 1, reference gangliosides of human neutrophil granulocytes, 40 µg; Lane 2,
gangliosides of human erythrocytes, 40 µg; Lane 3, NeuAcD3-neolactotetraocylceramide
(NeuAcD3Galß4GlcNAcß3Galß4Glcß1Cer),1µg;Lane4,
NeuAcD6neolactotetraocylceramide (NeuAcD6Galß4GlcNAcß3Galß4Glcß1Cer), 1 µg; Lane 5,
NeuAc D 3-neolactohexaocylceramide
(NeuAcD3Galß4GlcNAcß3Galß4GlcNAcß3Galß4Glcß1Cer), 1 µg; Lane 6, G-10 ganglioside
(NeuAcD3Galß4GlcNAcß6 (NeuAcD3Galß4GlcNAcß3)Galß4GlcNAcß3Galß4Glcß1Cer), 1
µg;
Lane
7,
G9-B
ganglioside
(GalD3(FucD2)Galß4GlcNAcß6
(NeuAcD3Galß4GlcNAcß3)Galß4GlcNAcß3Galß4Glcß1Cer), 1 µg; Lane 8, reference
gangliotetraosylceramide (Galß3GalNAcß4Galß4Glcß1Cer) of mouse feces, 4 µg.
Autoradiography was for 12 h.
16
Fig. 5
Binding of H. pylori to sialyl-glycans of various complexity
Binding characteristics of several H. pylori strains to sialyl-glycans was studied. Here strains
CCUG17875 and the babA1A2 knock out mutant, CCUG17874, 26695, J99, MO19, P466 and
the clinical isolates SMI65 and SMI27 were analyzed by using 125I labelled Lewis b (dark
blue bars), sialyl-Lewis x (brown bars), sialyl-Lewis a (yellow bars), and 3´sialyl-lactose
glycoconjugates (green bars), 3´sialyl-lactosamine-3C (light blue bars) and 3´sialyllactosamin-14C (red bars).
Fig. 6
Interaction of H. pylori strain J99 with human red blood cells. The series of images were
prepared by the Philips CM12 transmission electron microscope at 80 kV. Arrows point to the
indentations of the RBCs surface at the site of attachment. Size scale is given at the right
bottom of each image and RBCs are labeled to distinguish the bacteria from the RBCs.
17
Table 1: Bacterial dilutions/titers and strength of sialic acid dependent hemagglutination of
differently enzymatically treated RBCs, as illustrate by + for HA and - for lack of HA.
J99
x2
x4
x8
x16
x32
x64
x128
J99ǻsabA
x2
x4
x8
x16
x32
x64
x128
J99ǻbabA
x2
x4
x8
x16
x32
x64
x128
J99ǻnap
x2
x4
x8
x16
x32
x64
x128
E.coli_P-fimbria
x2
x4
x8
x16
x32
x64
x128
E.coli_S-fimbria
x2
x4
x8
x16
x32
x64
x128
0mg
Trypsin
0U Neua
0mg Trypsin
0.1U Neu
0.05mg Trypsin
0U Neu
0.05mg Trypsin
0.1U Neu
0.1mg Trypsin
0U Neu
0.1mg Trypsin
0.1U Neu
+
+
+
+
(+)
-
-
+
+
+
+
+
-
(+)
-
+
+
+
+
(+)
-
+
(+)
-
-
-
-
-
-
-
+
+
+
+
(+)
-
-
+
+
+
+
+
-
(+)
-
+
+
+
+
-
(+)
-
+
+
+
+
(+)
-
-
+
+
+
+
+
-
(+)
-
+
+
+
+
(+)
-
(+)
-
+
(+)
-
+
+
+
(+)
-
+
+
+
(+)
-
+
+
+
+
(+)
-
+
+
+
(+)
-
+
+
+
+
(+)
-
+
+
(+)
-
-
-
-
-
-
a) Neuraminidase treatment, activity is presented by units (U).
Table 2:
18
NeuAcα3SPG
NeuAcα6SPG
NeuAcα3-nLc6
2. G2
3. G4
4. G6
GD1b
10.
-
-
+++
-
-
-
-
-
-
-
-
-
-
-
-
J99/SabA-
(15)
(25)
(25)
(25)
(25)
(21)
(51)
(51)
(51)
(51)
Ref.
a) Binding is defined as follows: +++ denotes a binding when less than 0.5 µg of the glycosphingolipid was applied on the thin-layer chromatogram, while + denotes an
occasional binding at 0.5 µg, and - denotes no binding even at 4 µg.
Galβ3GalNAcβ4(NeuAcα8NeuAcα3)Galβ4Glcβ1Cer
NeuAcα Galβ3GalNAcβ4(NeuAcα3)Galβ4Glcβ1Cer
GD1a
9.
NeuAcα8NeuAcα3Galβ4GlcNAcβ3Galβ4Glcβ1Cer
NeuAcα3Galβ4GlcNAcβ6(NeuAcα3Galβ4GlcNAcβ3)Galβ4GlcNAcβ3Galβ4Glcβ1Cer
DPG
7. DG3
NeuAcα8NeuAcα3Galβ4Glcβ1Cer
+++
+++
NeuAcα3Galβ4GlcNAcβ3Galβ4GlcNAcβ3Galβ4Glcβ1Cer
Galα3(Fucα2)Galβ4GlcNAcβ6(NeuAcα3Galβ4GlcNAcβ3)Galβ4GlcNAcβ3Galβ4Glcβ1Cer
-
+
-a
CCUG 17874
and J99
NeuAcα6Galβ4GlcNAcβ3Galβ4Glcβ1Cer
NeuAcα3Galβ4GlcNAcβ3Galβ4Glcβ1Cer
NeuAcα3Galβ4Glcβ1Cer
Structure
8. G-10/DG6
GD3
6.
II. Disialo-gangliosides
5. G9-B
NeuAc-GM3
1. G1
I. Monosialo-gangliosides
No. Abbreviation Trivial name
Human erythrocyte gangliosides and results from binding of Helicobacter pylori
Table 2.
Figure 1:
20
Figure 2:
100
% bound
80
60
40
20
0
sLe x
sLe a
21
s-lactosamine
Figure 3:
1
2
75
50
22
3
4
Figure 4:
23
Figure 5:
Lew is b
Sialyl-Lew is x
100
Sialyl-Lew is a
90
3´-Sialyllactose
80
3´-Sialyl-N -Acetyllactosam ine (3C)
3´-Sialyl-N -Acetyllactosam ine (14C)
Binding (%)
70
60
50
40
30
20
10
0
CCUG17875
babA1A2mutant
CCUG17874
26695
J99
H. py lori strain
24
M019
P466
SMI65
SMI27
Figure 6:
25
IV
Title:
Helicobacter pylori evokes a strong inflammatory response in human neutrophils by SabA
adhesin mediated selectin-mimicry
Running title:
The SabA adhesin and activation of neutrophils
Keywords:
Helicobacter pylori, human neutrophils, Sialic acid-binding adhesin (SabA), selectinmimicry, sialyl-Lewis x, phagosytosis
Authors:
Christoffer Petersson1, Maria Forsberg1*, Marina Aspholm2*, Farzad O. Olfat2,3, Tony
Forslund1, Thomas Borén2 and Karl-Eric Magnusson1
*These authors contributed equally to this work.
For correspondence:
Christoffer Petersson, Division of Medical Microbiology, Faculty of Health Sciences,
Linköpings universitet, SE-581 85 Linköping, Sweden. E-mail:chrpe@imk.liu.se, Tel. (+46)
13 222093, Fax: (+46) 13 224789
Footnote:
1
Department of Molecular and Clinical Medicine, Division of Medical Microbiology, Faculty
of Health Sciences, Linköpings universitet, SE-581 85 Linköping, Sweden.
1
2
Department of Odontology/Oral microbiology, Umeå University, SE-901 87 Umeå, Sweden.
3
The Swedish Institute for Infectious Disease Control, SE-171 82 Solna, Sweden.
2
Summary
The human pathogen Helicobacter pylori expresses two dominant adhesins; the Lewis b
blood group antigen binding adhesin, BabA, and the sialic acid-binding adhesin, SabA. These
adhesins recognize specific carbohydrate moieties of the gastric epithelium, i.e. the Lewis b
antigen, Leb, and the sialyl-Lewis x antigen, sLex, respectively, that sustain infection and
promote inflammatory processes in the gastroduodenal tract.
To assess specific effects of BabA, SabA and the neutrophil activating protein, HP-NAP, in
local inflammation, we investigated the traits of different H. pylori mutants in their capacity
to interact with and stimulate human neutrophils. We found that the SabA adhesin was a key
regulator of induction of phagocytosis and oxidative metabolism, as evaluated by flow
cytometry, fluorescence microscopy and luminol-enhanced chemiluminescence. Furthermore,
the HP-NAP deletion resulted in maintained release of reactive oxygen species, ROS, and
impaired adherence to the host cells. Obtained effects were influenced by the relative
bacteria-to-neutrophil ratios tested (10:1, 25:1, 100:1). It is concluded that the SabA adhesin
stimulates human neutrophils through selectin-mimicry, and that HP-NAP facilitates the
recognition and binding of the sialylated antigens during conditions of limited sialylation,
which tunes the inflammation processes to balance in the chronic H. pylori infected gastric
mucosa.
3
Introduction
Infection by Helicobacter pylori is now well established as the major cause of chronic
gastritis, peptic ulcer disease, gastric lymphoma, and gastric carcinoma (Ota and Genta, 1997;
Kuipers, 1997; Graham, 2000; Feldman, 2001). Infections with H. pylori are spread across the
world. The rate of colonization in different populations range from 25 to 90%; highest
infection rates, i.e. 50% and above have predominantly been shown in studies carried out in
countries belonging to the developing world (Logan and Hirschl, 1996; Thomas et al., 1999;
Marshall, 2002). In industrialized countries the infections often result in a symptom-free form
of gastritis (type B gastritis). However, 10-20% of theses carriers develop more severe
diseases, such as: peptic ulcer, gastric cancer, and gastric lymphoma of mucosa-associated
lymphoid tissue type, also called MALTomas (Cotran et al., 1999; Karlsson, 2000; Suerbaum,
2000; Feldman, 2001). Though, H. pylori is one of the most common infectious agents in
human seen worldwide, the pathogenic mechanisms that confer gastric disease to the host is
still poorly understood.
H. pylori is commonly considered to be a non-invasive enteropathogenic bacterium.
Nonetheless, intracellular H. pylori have on several occasions been spotted in both cultured
epithelial cells and human gastric biopsy specimens (Wyle et al., 1990; Noach et al., 1994a;
el-Shoura, 1995; Ko et al., 1999; Björkholm et al., 2000; Kwok et al., 2002; Amieva et al.,
2002; Semino-Mora et al., 2003). H. pylori infections are characteristically associated with a
dense infiltration of mainly neutrophils into the epithelial surface layer (Sobala et al., 1991;
Dixon et al, 1996; Ernst and Gold, 2000). There is furthermore a good correlation between
accumulation of polymorphonuclear granulocytes and the severity of mucosal damage, and
subsequent development of gastroduodenal diseases (Warren & Marshall, 1983; Westblom et
al., 1999). However, a direct physical interaction in vivo in between H. pylori and leukocytes
4
is disputed (e.g., Steer, 1975; Noach et al., 1994a; el-Shoura, 1995; Zu et al., 2000). In
addition, the exact mechanisms for the recruitment of neutrophils to the H. pylori-infected
gastric mucosa are poorly understood, and a clear-cut interpretation of the interactions
involved has not yet been fully delineated (Crabtree, 2001). Clearly identifiable is, however,
that mucosal chemokine responses promote the attraction of neutrophils (Baggiolini et al,
1997). In the case with H. pylori the neutrophilic recruitment has been linked particularly to
the cytokine C-X-C subfamily, and especially to the interleukin 8, IL-8 (Sharma et al., 1995).
Tight binding of H. pylori to the epithelium triggers the secretion of IL-8, which in extension
elicits chemotaxis and activation of neutrophils (Noach et al., 1994b; Crabtree et al., 1995;
Shimoyama et al., 1998; Hofman et al., 2000). Despite the question marks surrounding this
issue, there are strong implications that carbohydrate molecule interactions play a crucial role
(Ångström et al., 1998; Karlsson, 2000; Teneberg et al., 2000; Teneberg et al., 2002). As
such, bacterial adhesion mediated by BabA adhesin to Leb in the human gastric epithelium
was just previously shown to give strong IL-8 release, as evaluated by the in vitro explant
culture, IVEC, technique (Olfat et al., 2002). Recently, we also reported that sLex is upregulated in inflamed gastric epithelium and used by H. pylori for tight adhesion to the
membrane-close sLex-gangliosides (Mahdavi et al., 2002). Thus, the BabA and SabA adhesins
of H. pylori act synergistically to furnish binding to the human gastric mucosa in both health
and disease. The present study was undertaken to assess whether the SabA adhesin bind to
sialyl-Lewisx glycoconjugates found on the human neutrophils (Stocks et al., 1990; Stocks
and Kerr, 1993), and thus mediates bacterial adhesion and initiate inflammatory responses.
Previous reports suggest that the sLex on leukocytes most likely facilitates neutrophil homing
of H. pylori (Miller-Podraza et al., 1999), and might even so conduce to the modulation of a
prolonged and sustained inflammation (Karlsson, 2000). By use of specific H. pylori mutants,
we have been able to study the highly specific and prompt interaction between the adhesins on
5
intact H. pylori and human neutrophils. In comparison, mutants in the BabA adhesin and in
the H. pylori neutrophil-activating protein, HP-NAP, were used to establish the SabA
adhesin’s significance. HP-NAP is a multimeric 200 kDa virulence factor, which affect
neutrophil recruitment and is considered to be highly involved in H. pylori-associated disease
processes (Dundon et al., 2002; Zanotti et al., 2002). HP-NAP was co-investigated as the
protein has been described to bind to sialylated and sulphated glycolipids and glycans
(Teneberg et al., 1997). Our results unequivocally suggest that the SabA adhesin is the key
molecule in the activation of human neutrophils, and likewise in the onset of persistent gastric
inflammation.
Results
Analysis of expression of the different adhesins by H. pylori J99 wild type and mutants
strains
Strain J99 and the series of mutants were analyzed for their binding properties by use of
soluble neoglycoconjugates, i.e. 125I-labeled semi synthetic glycoproteins based on albumin
were used in a RIA-assay (IIver et al., 1998; Mahdavi et al., 2002). Fig. 1A illustrates that the
J99babA-mutant and sabAbabA-mutant do not recognize the Leb antigen at all, since they are
devoid of the BabA adhesin. Similarly, in Fig. 1B, the J99sabA-mutant and J99sabAbabAmutant do not bind their cognate receptor, the sLex antigen. The parent strain J99 recognizes
both antigens, although the association is much stronger to Leb due to the high affinity
binding properties of BabA (Mahdavi et al., 2002). Interestingly, the J99 napA-mutant binds
both Leb and sLex similar to the J99 parent strain, which suggest that the HP-NAP protein
does not constitute a part of the alleged BabA or SabA adhesin complexes.
6
Figure 1A-B. Bacterial binding to soluble neoglycoconjugates. The H. pylori strains were
incubated with
125
I-labeled Lewis b (Fig. 1A) and sialyl-dimeric-Lewis X glycoconjugates
(Fig. 1B), reflecting relative adhesion by wild-type and mutant H. pylori bacteria. Data are
presented as mean percent of bound conjugate added.
7
Effect of multiplicity of infection (MOI) on the kinetics of generation of reactive oxygen
species (ROS)
By analyzing the pattern of luminol-enhanced chemiluminescence (CL) in the absence and
presence of extracellular superoxide dismutase (SOD) and catalase, we can distinguish
between total and intracellular production of ROS in the neutrophils. Fig. 2 illustrates the
rapid kinetics of total CL generation for wild-type H. pylori J99. The kinetics is less sustained
when the MOI is increased, which indicates expeditious degranulation and activation of the
NADPH-oxidase. In contrast, negligible effects were obtained with the supernatants of the
bacterial cultures, when used for priming, to induce hyper-responsiveness. This suggests low
impact from hyper-reactive chemoattractants, such as N-formylated bacterial peptides.
Although, some small initial peaks were seen in all instances within the first few minutes of
the interaction, superimposed on the general strong CL-signals (Fig. 2).
8
Figure 2. Kinetics of the ROS production in human neutrophils challenged with strain H.
pylori J99 at indicated bacteria-to-neutrophil ratios, alternatively to supernatants from strain
J99 and the full series of mutants. The ROS production was measured using a luminolenhanced chemiluminescence (CL) system. The cells (1 x 106) were incubated for 10 min at
37 qC prior to adding supernatants or respective bacteria. The supernatants were obtained as
described in experimental procedures. One representative experiment out of at least five
independent experiments is displayed.
Total ROS production by H. pylori J99 and corresponding series of mutants
Total response of ROS extends to both intracellular release and extracellular discharge of
ROS. In general, either the peak or integral value of CL-response is taken as a quantitative
measure of ROS production. They presumably reflect two distinct capacities of the
neutrophils, i.e. to mount a rapid and strong response or a sustained reaction. Either aspect
may be important for killing of intruding microorganisms. Fig. 3A, which displays the
integral value, indicates several interesting features. First, the J99sabA-mutant and the
“double mutant” J99sabAbabA produce the lowest responses, while the J99babA-mutant is
slightly more effective than the J99 wild-type (that binds both Leb and sLex) in activation of
9
the leukocytes. Interesting, for the J99 napA mutant (devoid of HP-NAP expression) the
response pattern was seen to be even more enhanced, which suggests that HP-NAP might
instead exert a “quenching”-type of activity on neutrophilic CL-generation. The subsequent
analysis of peak values for ROS production (Fig. 3B) follows essentially the same pattern for
the J99 sabA-mutant and the J99sabAbabA-mutant. However, here the J99napA-mutant did
not differ in CL-reactivity from the J99 wild-type strain. The similar response between the
parent strain and the HP-NAP mutant that expresses the SabA adhesin, indicate that the
adhesin-glycan interaction trigger an initial fast response event from the neurophils, followed
by a slower release mechanism. This implies that the HP-NAP could be a mediator in downmodulating and quenching a too excessive ROS production. So, worth noting is that H. pylori
might have dual systems for initiation and termination of leukocytic activity, thereby
optimizing its interaction with labile and sensitive immunological cells in the local and
inflamed mucosa. Other MOIs were also tested in the CL-experiments, but only a
representative range is displayed.
10
Figure 3. The production of ROS in human neutrophils in response to activation with strain
H. pylori J99 and different mutants. Neutrophils (1 x 106) were preincubated at 37 qC prior to
stimulation with bacteria, at a bacteria-to-cell ratio of 100:1. The ROS-production was
measured by luminol-amplified CL. The bars represent either the integral values (Fig. A.) or
peak values (Fig. B.) in percent of the wild-type ROS-production. Data are given as mean ±
SEM (standard error of the mean) of at least five experiments. Statistical significance vs
control (i.e. wild-type J99): **, p<0,001; *, p<0,01, by unpaired, two-tailed Student’s t-test.
11
Intracellular production of ROS by wild type and mutant H. pylori J99 bacteria
By selective scavenging of the extracellularly released ROS intracellular generation of
respiratory burst products could be analyzed. Only small differences between the bacterial
strains regarding the “integral” and “peak” CL of intracellular generated ROS were detected
(Table 1). As for the intracellular response, the SabA adhesin stands out as significant, both
for integral and peak values. This compared to the J99 sabA-mutant, which indicate strongly
impaired responses when the bacterial lectin is absent. In comparison, the J99 napA mutant
also demonstrate decreased reactivity, though restricted to the integral value, illustrating that
the intracellular release is reduced, while the previous analyses in Fig. 3A suggest that the
CL-responses in total are elevated. Taken together, these results putatively suggest that the
J99 napA-mutant specifically releases enhanced cellular levels of ROS, when compared to
parent J99 strain. Hence, HP-NAP exhibit an intrinsic role in the bacterial-host balancing
release of ROS generated during periods of chronic inflammation and gastritis.
12
Infection
Integral value
Peak value
(Mean r SEM)
(Mean r SEM)
J99
87 r 2
90r2
J99sabA
63 r 12
70r12
J99babA
83 r 3
79r6
J99sabAbabA
80 r 10
93r3
J99napA
63 r 4
90r2
Table 1. Intracellular total production of ROS after activation of human neutrophils with
wild-type and the series of H. pylori J99 mutants. ROS-production was measured using CL as
described elsewhere, with the exception that SOD and catalase were used to measure the
intracellular ROS-production. The infection ratio was 100:1. Data are presented as mean ±
SEM of the intracellular ROS-production expressed as percentage of total ROS production of
at least three separate experiments.
Phagocytosis of H. pylori wild-type and mutants by neutrophils analyzed by flow
cytometry
The interaction between the bacteria and neutrophils was studied at three relative infection
ratios, i.e. at MOI:s of 10:1, 25:1 and 100:1. In all three cases the J99sabA-mutant, and the
combined J99sabAbabA-mutant showed the lowest relative values (Table 2). This observation
underscores that SabA is a critical mediator of H. pylori-host cell-contact dependent
activation of human neutrophils, hence decisive in the generation of ROS (Fig. 4).
Pretreatment of the neutrophils with cytochalasin B or D, at final concentrations of 10 PM,
substantially lowered the total neutrophil-associated fluorescence recorded at all three
13
investigated ratios (data not shown). These results verify the expected inhibition of
phagocytosis, and thereby the accuracy of the chosen method. Worth noting is that the effects
conferred by the mutant phenotypes were most pronounced at the highest MOI, where the J99
napA-mutant displayed a distinct behavior regarding both adherence and ingestion (Table 2).
At the lowest MOI (10:1) the J99 napA-mutant on the contrary had a stimulating effect. These
results suggest that the functional phenotype of the series of adhesin mutants, including the
napA-mutant, is best displayed during conditions of bacterial competition for their tropic
binding sites, i.e. when approaching MOI ratios of 50:1 to 100:1, while less discriminating
binding modes tend to level out the typical binding characteristics and related CL- responses
for the mutant phenotypes. The relative errors were in a large range, probably due to the
individual variation among blood donors, in part owing to variation in phenotypes and
excitation states.
Figure 4. Internalization of the wild-type H. pylori J99 strain (thick curve) and the mutant
strain J99sabA (thin curve) analyzed by flow cytometry. The phagocytosis curves
representing the relative value of neutrophils ingesting FITC-labelled bacteria after quenching
with ethidium bromide. A shift to the right indicates higher total mean fluorescent value
(FL1). The graph shows results of a single experiment out of four separate experiments
performed.
14
Bacteria:Neutrophil
Adherence %;
Ingestion %;
Total interaction;
Ratio 10:1 (n =4)
Mean ± SEM
Mean ± SEM
MFV
J99
100
100
106
J99babA
86±46
109±3
64
J99sabA
61±24
66±16
37
J99sabAbabA
75±56
74±32
38
J99napA
120±62
105±52
50
Bacteria:Neutrophil
Adherence %;
Ingestion %;
Total interaction;
Ratio 25:1 (n =4)
Mean ± SEM
Mean ± SEM
MFV
J99
100
100
351
J99babA
51±8
66±10
176
J99sabA
27±5
36±8
101
J99sabAbabA
14±4
22±3
68
J99napA
43±18
49±17
116
Bacteria:Neutropihl
Adherence %;
Ingestion %;
Total interaction;
Ratio 100:1 (n =4)
Mean ± SEM
Mean ± SEM
MFV
J99
100
100
1673
J99babA
43±12
54±12
870
J99sabA
28±9
28±6
548
J99sabAbabA
21±4
28±3
469
J99napA
37±12
40±6
521
Table 2. Adherence, ingestion and total interaction of H. pylori J99 and its specific deletion
mutants with human neutrophils. Results are shown as relative percentage versus the wild-
15
type strain, and “n” equals the number of independent experiments with new batches of
neutrophils derived from different blood donors. MFV=total mean fluorescent value. Values
are given as mean ± SEM.
Phagocytosis of H. pylori wild-type and mutants by neutrophils assessed with
fluorescence microscopy
The strong difference between the H. pylori J99 parent strain, wild-type, and the J99 sabAmutant in direct interaction with isolated neutrophils is here visualized (Fig. 5). It is
interesting to notice that in all cases except for the J99 sabA-mutant, and the J99sabAbabAmutant, there was a massive aggregation of bacteria and neutrophils, especially at the highest
MOI of 100:1. The phagocytic index clearly distinguishes the role of the SabA adhesin in the
bacteria-cell interaction (Table 3). The equal results receive from the three strains expressing
the SabA adhesin just fairly reflects the difficulties with the microscopy method in detecting
subtle distinctions between traits in the samples tested. Thus, the SabA molecule displays a
key role for targeting of bacterial cells to the neutrophils, where the sialylated receptors might
also be up regulated on the neutrophil membrane due to the tight interaction. Such signaling
does not tend to require viable bacterial cells, since both thawed and freshly prepared FITClabelled bacteria yield the same effects, i.e. tight neutrophil interaction and aggregation.
16
Figure 5. Fluorescence microscopy of human neutrophils after phagocytosis of the wild-type
H. pylori J99 strain (A) and the J99 sabA-mutant (B). FITC-labelled bacteria and nonphagocytosed bacteria were quenched with ethidium bromide or trypan blue, which may give
some red background staining at the cell membranes. Note the large sialic acid-mediated
aggregates between the wild-type strain and the neutrophils (A) and in contrast the solitary
neutrophils associated with the lectin-deficient J99 sabA-mutant (B). White arrowheads
indicating ingested H. pylori. Bars = 10 Pm.
17
Bacteria
Fraction of phagocytosing neutrophils
(Mean r SEM)
J99
0,70 r 0,05
J99sabA
0,04 r 0,03
J99sabAbabA
0,08 r 0,03
J99babA
0,70 r 0,05
J99napA
0,70 r 0,05
Table 3. Phagocytosis of wild-type and mutant strains of H. pylori. Neutrophils were
prewarmed for 5 min at 37°C prior to addition of either wild-type H. pylori J99 or the series
of J99 mutants (sabA, sabAbabA, babA or napA), at a bacteria-to-cell ratio of 100:1, the
phagocytosis was allowed to proceed for 10 or 30 min. Samples were put on ice and evaluated
by fluorescence microscopy as described in experimental procedures. A phagocytic index is
given, obtained from the fraction of phagocytosing neutrophils of four separate experiments ±
SEM.
Analyzes of the influence of HP-NAP on bacterial binding to host cells by use of
neuraminidase and protease treated red blood cells
ROS production proposes a plausible way to assess the contribution of HP-NAP in bacterial
interaction with neutrophils. To analyze if the protein positively or negatively modifies the
efficiency of the adhesins to bind and interact with the host cell, a modified hemagglutination
method was used. Since the SabA adhesin binds to sialylated antigens, in particular to the
sLex glycans, neuraminidase treatment was employed to reduce or remove the surface
presented sialylated residues from the human red blood cells (RBC), thereby abolishing the
sialic acid-dependent hemagglutination (sia-HA). The regular hemagglutination activity was
strong for the J99 wild-type strain, and so for the J99napA- and J99babA-mutants (Table 5). It
18
was in all instances very sensitive to neuraminidase treatment, where sia-HA was already
eliminated at concentrations of 0.1 U/ml, which points to sialic acid-dependent binding mode.
In harmony with this, the J99 sabA-mutant exhibited no hemagglutinating capacity at all,
which confirms that the SabA adhesin recognize sialylated glycoconjugates on the host cell
surface. In contrast to the neuraminidase treatment, limited protease treatment of the RBCs
did not affect the sia-HA titres of the J99 strain, implying that the sialic acid residues
responsible for aggregation of RBCs are preferentially localized in the sialylated glycolipids
(gangliosides) and not within the glycoproteins (Olfat et al., in preparation). The sia-HA was
further analyzed by titration of the neuraminidase treatment, where low concentrations was
used (0,02U/ml). This treatment merely reduced the sia-HA titres from 1:8 till 1:4, i.e.
conferred a 2-fold reduction in sia-HA for the J99 wild-type and the babA-mutant. In contrast,
a full 8-fold reduction in sia-HA was found for the napA-mutant; results concluding that the
HP-NAP either facilitates SabA-mediated recognition of sialylated antigens on the host cells
surfaces, or assists in subsequent binding to sialylated antigens. Since protease treatments of
RBCs did not interfere with sialic acid dependent binding, the HP-NAP was similarly
analyzed for the influence of protease treatment and sia-HA. Interesting here is that already
limited protease treatment fully regained the sia-HA properties for the napA-mutant, which
suggests a cooperative synergistic binding property of the HP-NAP. In terms of a limited host
cell sialylation, as in the case with a low degree of mucosal inflammation, this trait could
facilitate the SabA-mediated binding of H. pylori. Speculatively, it indicates that HP-NAP
posses a modulating effect, alternatively a quenching mechanism that could down-modulate a
sustained signal event, as suggested by the results in Table 1 and Figure 3.
19
RBC
Bacterial strain
Treatment
J99
J99napA
J99babA
0 mg trypsin/
0 U Neu
1:8
1:8
1:8
0 mg trypsin/
0,02 U Neu
1:4
1:1
1:4
0 mg trypsin/
0,1 U Neu
0
0
1:1
0 mg trypsin/
0,2 U Neu
0
0
1:1
0,05 mg trypsin/
0 U Neu
1:8
1:8
1:8
0,05 mg trypsin/
0,02U Neu
1:4
1:4
1:4
0,05 mg trypsin/
0,1U Neu
1:1
1:1
1:1
0,05 mg trypsin/
0,2U Neu
0
0
1:1
Table 4. Hemagglutination assay titres of RBC by the J99 strain and the corresponding napAmutant and babA-mutant. The erythrocytes were pre-treated with either trypsin or
neuraminidase, or a combination there-of as displayed in the checkerboard. “0” refers to no
hemagglutination, while the titres are given for the bacterial dilution steps where no losses of
agglutination were shown.
20
Discussion
Our working hypothesis has been that the SabA adhesin is a key-element in the non-opsonic
local immune response in the gastric epithelium. This response is mainly characterized by the
recruitment, accumulation and infiltration of activated neutrophils and monocytes in the
gastric mucosa. Undoubtedly, this engagement of the leukocytes is an essential step in the
initiation and maintenance of the gastric inflammatory process (Ernst and Gold, 2000). In the
present investigation, we have also tried to disclose the relative contribution of another
virulence trait of the H. pylori, i.e. the neutrophil activating protein (HP-NAP). This molecule
is well-characterized and studied, and known to be involved in the interaction and activation
of human neutrophils (Evans et al., 1995; Satin et al., 2000; Nishioka et al., 2003), as well as
in the scavenging of reactive oxygen species (Cooksley et al., 2003). How the combined
expressions of the virulence proteins help H. pylori to survive in the epithelium lining in the
stomach has still not been resolved. The HP-NAP is considered to be chemotactic, proinflammatory and promoting release of nutrients from the mucosa, thus supporting the growth
of the bacteria (Namavar et al., 1998; Dundon et al., 2001). Recent studies have shown
convincingly that H. pylori adhesins BabA and SabA are critical for the establishment of H.
pylori locally in normal and inflamed tissue, respectively (Ilver et al., 1998; Mahdavi et al.,
2002). We now demonstrate that SabA, but not BabA alone (Figs. 2; Tables 2 and 3) is
essential for attachment and activation of human neutrophils by the H. pylori strain J99. In the
present investigation the HP-NAP did seemingly not, as a component in a water-soluble
extract, promote ROS generation in neutrophils (Fig.1), neither did it as bacteria-associated
impel the activation of the oxidative metabolism (Figs. 2A and 2B). Rather, it modified the
pattern of generation of respiratory burst products. Careful analysis of the experimental
21
conditions is also needed for a final assessment of the role of HP-NAP, perhaps being both
important as chemoattractant and ROS-protectant (Teneberg et al., 1997; Dahlgren &
Karlsson, 1999; Bylund & Dahlgren, 2002; Cooksley et al., 2003).
The ability of the different bacterial strains to elicit an oxidative burst in the neutrophils, is
clearly reflected in the total chemiluminescence production, comprising both intracellular and
extracellular formation of reactive oxygen species, whether judged by peak or integral values
(Figs. 3A and 3B). Deserving attention here is also the negligible contribution of the
supernatants (Fig. 2). In contrast to our results, Kim et al. (2001) have previously reported
activation of the neutrophils by crude water supernatant preparations. On the other hand, this
could point towards the highly variable ability of H. pylori-extracts to stimulate neutrophils
(Leakey et al., 2000). To study phagocytosis of the bacteria, we used both fluorescence
microscopy and flow cytometry. By modifying the method described by Heinzelmann et al.
(1999), we were able to consolidate quantitatively the trends in ingestion and adherence
between the different bacterial traits (Table 2). At all ratios between bacteria and neutrophils
the SabA adhesin seems to promote the interaction further than any of the other bacterial
proteins investigated here. Flow cytometry was used as an objective measure of phagocytosis,
and to reduce the drawbacks with traditional microscopy, which is laborious and sensitive to
variations in observer interpretations. As previously used by Potter and Harding (2001), we
tested FITC-labelled heat-killed bacteria in our systems. However, in our case heat-killed
bacteria failed to give any useful information; no differences between the traits tested could
be distinguished. Likewise, attention has been drawn to the problems with FITC-labelling,
claiming that the fluorophore itself interfere in the phagocytosis, giving misleading results
(Weingart et al., 1999). Therefore, it has been recommended that a better choice is Auramine
O stain (Hartmann et al., 2001). When trying Auramine O-labelling, we found that the uptake
between the different H. pylori mutants was completely leveled out.
22
Thus, the Auramine O labeling evidently changed the surface properties, and thereby
promoted phagocytic uptake of all strains. Envisioning that the MOI, i.e. the relative number
of intruding H. pylori and host defense cells, reflects conditions that can occur in inflamed
tissue, its effect of the magnitude and rate of ROS generation was specifically addressed (Fig.
2 and Table 2). The finding of the HP-NAP-mutant phenotype is of principal interest since
both bacterial adherence and ingestion by the leukocytes decreased with higher proportion of
bacteria (Table 2). We suggest this is a specific consequence of the loss of the HP-NAP and
not due to altered expression of another surface appendage, e.g. SabA, since the HP-NAPmutant and the wild-type strain display similar results in the regular sia-HA analysis, where
unlimited neuraminidase was used. The binding results from white blood cells (Table 2) and
red blood cells during limited sialic-acid levels (Table 4) strengthen the conception that HPNAP participated in adherence to host cells during conditions of limited sialylation, such as
initial and low-grade gastritis, or in situations of competition between bacteria, as modeled in
the chase with high MOI series in the flow cytometry analysis (Table 2). The present results
propose a central role of the SabA adhesin in provoking an inflammatory response, which is
dependent on the bacterial load. This strengthens previous findings of the expression of SabA
binding sites and bacterial association with inflamed gastric tissue (Mahdavi et al., 2002).
Present and previous findings speak for the following model of H. pylori colonization and
disease promotion. The BabA adhesion is a prerequisite for the initial colonization (Ilver et
al., 1998; Linden et al., 2002), the HP-NAP helps recruit inflammatory cells (Dundon et al.,
2001; Namavar et al., 1998; Satin et al., 2000) and protects bacteria from reactive oxygen
species (Cooksley et al., 2003), and SabA, together with the bacterial load determine the
magnitude of the inflammatory response. Whether it is for good or bad with a strong initial
reaction remains to be revealed.
23
Experimental procedures
Reagents
Polymorphprep• and Lymphoprep• was obtained from Axis-Shield PoC AS (Oslo,
Norway). Horseradish peroxidase (HRP) was purchased from Boeringer-Mannheim
(Mannheim, Germany). Superoxide dismutase (SOD) and catalase were obtained from Roch
Diagnostics (Mannheim, Germany). Unless otherwise noted, other reagents and chemicals
were purchased from Sigma-Aldrich, St Louis, MO.
Bacterial strains and growth conditions
The bacterial mutant strains of H. pylori used in this study were based on the reference strain
J99 (Alm et al., 1999). The latter belongs to the special group of disease-promoting strains,
the so-called “type I strains”. These harbor the virulence-associated genes: vacuolating
cytotoxin activity (vacA), and the cytotoxin-associated gene A (cagA). Four mutant strains
were constructed from this strain, previously described in detail (Mahdavi et al., 2002).
Briefly, these four J99 derivatives had the following characteristics: the SabA adhesin
knockout mutant, denoted J99 sabA; the BabA adhesin mutant, denoted J99babA; the doublemutant in SabA and BabA adhesins, here denoted J99sabAbabA; and a mutant in the
neutrophil-activating protein of H. pylori (HP-NAP), here abbreviated J99napA. In the
construction of the J99napA::kan mutant, the napA gene was amplified by PCR, using the
napA1F (forward) and napA1R (reverse) primers. The PCR fragment was cloned into the
pBluescriptSK+/- EcoRV site (Stratagene, La Jolla, CA). The resulting plasmid was
linearized with primers napA2F (forward) and napA2R (reverse), ligated with the kanamycin
24
resistance (KanR) cassette from pILL600 (Suerbaum et al., 1993), and then used to transform
the J99 strain. For transformation, strains were grown for 24 hrs on agar plates before addition
of 2Pg of plasmid DNA. After transformation they were cultivated on non-selective plates for
48 h to allow unrestrictive growth, and then transferred to kanamycin-containing plates. The
transformants were analyzed by PCR using primers napA3F and napA4R which verified that
the KanR cassette was inserted into the napA gene. Western blot analysis of napA mutants
using anti-NapA antibodies show that the mutant strains were devoid of HP-NAP expression.
The following oligonucleotides were used for PCR:
napA1F: 5´ TCAAGCCATAGCGGATAAGCT 3´
napA1R: 5´TTGATAATGGCAAGGAAGTGGA 3´
napA2F: 5´ CACGATCGCATCCGCTTGCA 3´
napA2R: 5´TTACCGTAACTTATGCGGATGAT 3´
napA3F: 5´TGGTGTAGGATAGCGATCAAG 3´
napA4R: 5´TAATGTCATTCCACTTGTCTAAG 3´
The H. pylori strains were maintained as frozen glycerol stocks and were cultured on GC II
agar base plates (Becton-Dickinson; Mountain View, CA), supplemented with: 2.5% (w/v)
Lab M agar no. 2 (Topley House; Bury, UK), 7% (v/v) horse blood citrate, 8% (v/v) heatinactivated horse serum, 0.03% (v/v) Iso-Vitalex, 12 Pg/ml Vancomycin, 2 Pg/ml
Amphotericin B, 20 Pg/ml nalidixic acid, and a final concentration of HCl to 2.5 mM. For the
growth of the sabA and babA strains, the plates were also supplemented with 20 Pg/ml
chloramphenicol. For the napA-mutant strain the plates were supplemented with 25 Pg/ml
kanamycin, and for the sabAbabA-mutant the plates were supplemented with both 20 Pg/ml
chloramphenicol and 25 Pg/ml kanamycin, to secure the selective growth of the specific
25
mutants. The bacteria were incubated at 37°C, 5% O2, 85% N2 and 10% CO2, respectively, for
2 days prior to use.
Western blotting
Bacterial protein extracts were separated on 7.5 % acrylamide gels (Bio-Rad Laboratories,
Hercules, CA) and transferred to PVDF membranes (Bio-Rad). The blots were incubated with
rabbit antiserum raised against NapA (kindly provided by Cesare Montecucco, Venetian
Institute of Molecular Medicine, Padova, Italy) and the bound antibody was visualized by a
peroxidase-coupled goat anti-rabbit antibody (DAKO, Glostrup, Denmark) and Super signal
reagents (PIERCE, Rockford, IL).
Preparation of human neutrophils
Peripheral venous blood was drawn from non-medicated healthy blood donors at Linköping
University Hospital, and anti-coagulated with 5 U/ml heparin. Neutrophils were separated and
isolated as previously described elsewhere (Boyum, 1968; Ferrante and Thong, 1980). In
brief, the blood was layered over a density gradient consisting of LymphoprepTM layered over
PolymorphprepTM, and centrifuged at 480 x g for 40 min at room temperature. The band rich
in polymorphonuclear cells was collected, and contaminating erythrocytes removed by
hypotonic lysis. Neutrophils of about 95% purity were resuspended in Krebs-Ringer
phosphate buffer containing 120 mM NaCl, 4.9 mM MgSO4 x 7H2O, 1,7 mM KH2PO4, 8.3
mM Na2HPO4 x 2 H2O, 1 mM CaCl2, and 10 mM glucose (KRG, pH 7.3). Total cell count
was calculated in a Channelyzer 256 (Beckman Coulter Inc., Fullerton, CA). Viability of the
cells was >98% as assessed by Trypan blue exclusion. The cells were kept on ice until use.
26
FITC-labeling of H. pylori
The different derivatives of H. pylori J99 were grown as described above, and harvested from
the agar plate to PBS pH 7.3 supplemented with 0.05% Tween 20. After centrifugation (3500
x g, 4 min) and resuspension three times in PBS pH 7.3 with 0.05% Tween 20, the optical
density (OD) was calibrated to 2.0 at a wavelength of 650 nm using PBS, pH 7.3. This OD
corresponded to a concentration of bacteria to 5x108/ml, verified by optical quantification in a
counting chamber. The suspension was centrifuged (3500 x g, 4 min), and after discarding the
supernatant the pellet was resuspended in 2 ml bicarbonate buffer (0.1 M NaHCO3, pH 9.6)
containing 0.1mg/ml FITC. The solution was incubated for 1 hr at RT under constant rotation
in dark. After three resuspension and washing steps in PBS, pH 7.3, bacteria were
resuspended in KRG, pH 7.3, and divided into aliquots. Before frezzing, the tubes were
sonicated at 80 W for 3 x 30 sec to disperse clumps of H. pylori. Aliquoted bacteria were
stored frozen at –80°C until used.
Phagocytosis measurement with fluorescence microscopy
A fluorescence-quenching method (Hed, 1977) that distinguishes between extracellular and
intracellular located bacteria was used to determine phagocytosis. Neutrophils and bacteria
were prepared as previously described. Prior to use the thawed vials of H. pylori strains were
sonicated to disrupt clusters of bacteria. FITC-labeled H. pylori were incubated with
neutrophils at various bacteria-to-PMN ratios; 10:1, 25:1, 50:1 and 100:1 for 10 or 30 min at
37 qC. One drop of the neutrophil-bacteria mixture and one drop of Trypan blue (2 mg/ml in
25 mM citrate-phosphate buffer and 25 mM NaCl, pH 7.4) were placed on a microscopic
slide. The procedure was also supplementary performed with ethidium bromide (10mg/ml) as
the quenching agent (Heinzelmann et al., 1999).
27
The fraction of cells containing fluorescent bacteria, i.e. ingested bacteria, was determined
and used as a measure of phagocytosis (phagocytic index). Usually 80 to 100 cells in each
sample were counted, this out of a randomized field of vision in the fluorescence microscope.
Fluorescence microscopy was performed using a Zeiss Axioskop microscope equipped with a
50 W mercury vapour lamp fitted with standard filter sets for viewing FITC-fluorescence i.e.
with excitation and emission maximum at 488 and 523 nm, correspondingly (Carl Zeiss, Jena,
Germany).
Phagocytosis measured with flow cytometry
To further quantify the phagocytosis of bacteria, we also used FACS technology. The method
applied was a modification of a previously described protocol (Heinzelmann et al., 1999;
Hartmann et al., 2001). FITC-labeled strains of H. pylori were thawed at room temperature
and declumped by sonication. Freshly prepared neutrophils (1x107/ml) were used. The assay
was carried out in 2 ml Eppendorf tubes with a final volume of 500 Pl. The samples were
incubated in KRG supplemented with 0.1% human serum albumin (HSA). Phagocytosis was
allowed to proceed in dark for 30 min at 37°C with periodic agitation. Placing the tubes on ice
terminated phagocytosis. Different multiplicities of infection, MOI, were used; 25:1, 50:1 and
100:1 (bacteria:neutrophil). The suspension was centrifuged (400 x g, 4 min), and after
discarding the supernatant, the pellet was resuspended in ice-cold KRG, pH 7.3. Samples
were then in addition resuspended three times, centrifuged and washed with ice-cold KRG to
remove excess of FITC-labelled bacteria. Finally, the pellet of washed cells was resuspended
in 500 Pl of ice-cold KRG and split into two. One half of the suspension was used for
measuring phagocytosis, the other half for quantifying the quenching capacity of sample. To
make a distinction between extracellulary adhering bacteria cells and those that had been
phagocytosed, we used ethidium bromide, at a final concentration of 50 Pg/ml, as quenching
28
agent for the FITC. To verify efficient quenching of adherent bacteria, both 10 PM
cytochalasin D and 10 PM cytochalasin B were separately used to inhibit phagocytosis.
Between the sequential steps, the test tubes were maintained on ice, and kept in dark. All
measurements were done with a FACSCalibur (Becton Dickinson, San Jose, CA, USA)
equipped with an argon laser operating at an excitation wavelength of 488 nm. Evaluation of
collected data was made using the CellQuest version 3.1f software (Becton Dickinson). For
each sample 10 000 events were collected. Events were gated to discard signals that did not
qualify as mammalian cells, represented dead cells (high red fluorescence), or were big
aggregates of bacteria and cells. The average number of cells analysed were 4000. The
percentage of mean fluorescence intensity for FITC (green) was used as measure for the
activity. Percent ingestion was determined as 100 x [(FL1 after EtBr for mutant strain)/(FL1
after EtBr for wild type strain)]. The quantified percent adhesion was determined as 100 x
[(FL1-before EtBr for mutant strain)-(FL1 after EtBr for mutant strain)/(FL1-before EtBr for
wild type strain)-(FL1 after EtBr for wild type strain)].
The neutrophil respiratory burst activity
The respiratory burst in neutrophils was determined by luminol-enhanced chemiluminescence
(CL) that allows measurment of released reactive oxygen species (ROS). Both extra- and
intracellular CL was analyzed in a six-channel Bioluminat LB 9505 instrument
(Berthold Co., Wildbad, Germany), using disposable 4 ml polypropene tubes. Neutrophils (1
x 106) were allowed to equilibrate for 5 min at 37qC in KRG with calcium, luminol (20 PM),
HRP (4 U/ml), thereafter the light emission was recorded continuously for 60 min. After
achieving a baseline, the various strains of H. pylori was added in different ratios
(25:1, 50:1, 100:1) to the neutrophils. To measure intracellularly produced ROS only, tubes
containing SOD (200 U/ml) and catalase (2000 U/ml) were used instead of HRP to scavenge
29
the extracellularly released superoxide anion and hydrogen peroxide (Dahlgren & Karlsson,
1999; Karlsson &Dahlgren, 2002). To rule out the possibility that a soluble component in the
supernatant gave the observed results, neutrophils were also incubated with 200 Pl of the
supernatant from each mutant remaining after the first washing step in the FITC-labelling
procedure of the bacterium.
Agglutination of RBCs by SabA and HP-NAP mutants.
To test for sialic acid-specific binding characteristics of the bacteria, human red blood cells
(RBC) were prepared in the following way. They were treated with bovine pancreas trypsin at
a concentration of 0.05 mg/ml (Borén et al., 1997), or with PBS buffer only, pH 7.4. After this
treatment the cells were incubated with 1 mM phenylmetylsulfonylfluoride (PMSF), followed
by a three-fold wash in PBS, pH 7.4, and then an incubation with different concentrations
(0.02, 0.1, or 0.2 U/ml) of Clostridium perfringens neuraminidase type VI. In parallel,
agglutination without neuraminidase treatment was also tested.
The basal density of bacteria used in the assay was 1x108 bacteria/ml. A 0.75-% RBC
suspension was used for the hemagglutination assay and for assessment of the
protease/neuraminidase effect. Twenty-five Pl of bacterial suspension was mixed with the
same volume of 0.75% RBCs in round-bottom ELISA plates, and the state of aggregation was
visually read off after 1 h incubation in RT.
Measurement of Binding Activity of the strains of H. pylori J99
The binding assay was performed as previously described (Falk et al., 1994; Ilver et al., 1998)
with minor modifications. The Lewis b and sialyl-dimeric-Lewis x glycoconjugates (IsoSep
30
AB, Tullinge, Sweden) were labeled with
125
I by the chloramine T method. Bacteria were
harvested from agar plate in PBS, pH 7.3. One ml of bacteria suspension (optical density at
600 nm = 0.1) was incubated with 300 ng of 125I-labeled conjugate for 2 h in PBS containing
1% albumin and 0.05% Tween 20. The bacteria were pelleted by centrifugation and the
amount of bound radio-labeled conjugates bound to the bacterial pellet was measured by
gamma scintillation (1282 Compugamma CS; LKB Wallac, Oy, Finland). Binding
experiments were performed in duplicates.
Acknowledgments
This work was supported by grants from the Program of Infection & Vaccinology (ChP/KEM
and MH/TB) and the Program of Inflammation (MF), Foundation for Strategic Research,
SSF; the ELFA Research Foundation (ChP); the Swedish Society for Medical Research
(ChP); the Swedish Research Council, Project Nos. 621-2001-3570, 521-2001-6565, 11218
(KEM/TB), the King Gustav V 80-Year Foundation (KEM); Swedish Cancer Society, Project
No 4101-B00-03XAB (TB); Umeå University Biotechnology Fund (TB), and County Council
of Västerbotten (TB).
We wish to thank Andreas Lundgren for critical linguistic reading of the manuscript.
31
References
Alm, R.A., Ling, L.S., Moir, D.T., King, B.L., Brown, E.D., Doig, P.C., Smith, D.R. et al.
(1999) Genomic-sequence comparison of two unrelated isolates of the human gastric
pathogen Helicobacter pylori. Nature 14:176-80. Erratum in: Nature 25:719.
Amieva, M.R., Salama, N.R., Tompkins, L.S., Falkow, S. (2002) Helicobacter pylori enter
and survive within multivesicular vacuoles of epithelial cells. Cell Microbiol 10:677-690.
Ångström J., Teneberg S., Milh M.A., Larsson T., Leonardsson I., Olsson B.M., et al. (1998)
The lactosylceramide binding specificity of Helicobacter pylori. Glycobiology 4:297-309.
Baggiolini, M., Dewald, B., Moser, B. (1997) Human chemokines: an update. Annu Rev
Immunol 15:675-705.
Björkholm, B., Zhukhovitsky, V., Löfman, C., Hulten, K., Enroth, H., Block, M., Rigo, R.,
Falk, P., Engstrand, L. (2000) Helicobacter pylori entry into human gastric epithelial cells: A
potential determinant of virulence, persistence, and treatment failures. Helicobacter 3:148154.
Borén, T., Wadström, T., Normark S., Gordon J.I., Falk P.G., (1997). Methods for the
Identification of H. pylori Host Receptors. In: Helicobacter pylori protocols. Clayton, C.L.,
and Mobley, H.L.T. (eds.) New Jersey:Humana Press Inc., pp 205-224
32
Bylund J., Dahlgren C. (2002) Problems in identifying microbial-derived neutrophil
activators, focusing on Helicobacter pylori. Trends Microbiol 10:12-14
Cooksley C., Jenks P.J., Green A., Cockayne A., Logan R.P.H., Hardie K.R. (2003) NapA
protects Helicobacter pylori from oxidative stress damage, and its production is influenced by
the ferric uptake regulator. J Med Microbil 52:461-469.
Cotran, R.S., Kumar, V., Collins, T., eds. (1999) Robbins Pathologic Basis of Disease. 6th ed.
London: W.B. Saunders Company, pp. 315; 787-802.
Crabtree, J.E. (2001) Cytokine responses in Helicobacter pylori infection. In: Helicobacter
pylori: Molecular and Cellular Biology. Achtman, M., and Suerbaum, S. (eds.) Norfolk:
Horizon Scientific Press, pp 63-83.
Crabtree, J.E., Covacci, A., Farmery, S.M., Xiang, Z., Tompkins, D.S., Perry, S. et al. (1995)
Helicobacter pylori-induced interleukin-8 expression in gastric epithelial cells is associated
with CagA positive phenotype. J Clin Pathol 1:41-45.
Dahlgren, C., Karlsson, A. (1999) Respiratory burst in human neutrophils. J Immunol
Methods 232:1-14
Dixon, M.F. (1996) Pathological consequences of Helicobacter pylori infection. Scand J
Gastroenterol Suppl 215:21.
33
Dixon M.F., Genta R.M., Yardley J.H., Correa P. (1994) Classification and grading of
gastritis. The updated Sydney System. International Workshop on the Histopathology of
Gastritis, Houston 1994. Am J Surg Pathol 20:1161-81
Dundon, W.G., Nishioka, H., Polenghi, A., Papinutto, E., Zanotti, G., Montemurro, P. et al.
(2002) The neutrophil-activating protein of Helicobacter pylori. Int J Med Microbiol 6-7:
545-550.
el-Shoura S.M. (1995) Helicobacter pylori: I. Ultrastructural sequences of adherence,
attachment, and penetration into the gastric mucosa. Ultrastruct Pathol. 19:323-333.
Ernst, P.B., and Gold, B.D. (2000) The disease spectrum of Helicobacter pylori: the
immunopathogenesis of gastroduodenal ulcer and gastric cancer. Annu Rev Microbiol.
54:615-640.
Evans D.J., Evans D.G., Takemura T., Nakano H., Lampert H.C., Graham D.Y. et al. (1995)
Characterization of a Helicobacter pylori neutrophil-activating protein. Infect Immun
63:2213-2220.
Falk, P., Borén, T., Haslam, D., Caparon, M.G. (1994) Bacterial adhesion
and colonisation assays. Methods Cell Biol 45:165–192
Feldman, R.A. (2001) Epidemiologic observation and open questions about disease and
infection caused by Helicobacter pylori. In: Helicobacter pylori: Molecular and Cellular
Biology. Achtman, M., and Suerbaum, S. (eds.) Norfolk: Horizon Scientific Press, pp. 29-51.
34
Graham, D.Y. (2000) Helicobacter pylori infection is the primary cause of gastric cancer. J
Gastroenterol. 35: 90-97.
Hartmann P., Becker R., Franzen C., Schell-Frederick E., Romer J., Jacobs M., Fatkenheuer
G., Plum G. (2001) Phagocytosis and killing of Mycobacterium avium complex by human
neutrophils. J Leukoc Biol 69: 397-404
Hed, J. (1977) The extinction of fluorescence by crystal violet and its use to differentiate
between attached and ingested microorganisms in phagocytosis. FEMS Microbiol Letters
1:357-361
Heinzelmann M., Gardner S.A., Mercer-Jones M., Roll A.J., Polk H.C. (1999) Quantification
of phagocytosis in human neutrophils by flow cytometry. Microbiol Immunol 43: 505-512
Hirmo S., Kelm S., Schauer R., Nilsson B., Wadström T. (1996) Adhesion of Helicobacter
pylori strains to alpha-2,3-linked sialic acids. Glycoconj J. 13:1005-1011.
Hofman, V., Ricci, V., Galmiche, A., Brest, P., Auberger, P., Rossi, B. et al. (2000) Effect of
Helicobacter pylori on polymorphonuclear leukocyte migration across polarized T84
epithelial cell monolayers: role of vacuolating toxin VacA and cag pathogenicity island. Infect
Immun 9:5225-5233.
Ilver, D., Arnqvist, A., Ögren, J., Frick, I.M., Kersulyte, D., Incecik, E.T. et al. (1998)
Helicobacter pylori adhesin binding fucosylated histo-blood group antigens revealed by
retagging. Science 279:373-377.
35
Karlsson A., Dahlgren C. (2002) Assembly and activation of the neutrophil
NADPH oxidase in granule membranes. Antioxid Redox Signal 4:49-60.
Karlsson, K.A. (2000) The human gastric colonizer Helicobacter pylori: a challenge for hostparasite glycobiology. Glycobiology 8:761-771.
Kim J.S., Jung H.C., Kim J.M., Song I.S., Kim C.Y. (2000) Helicobacter pylori water-soluble
surface proteins activate human neutrophils and up-regulate expression of CXC chemokines.
Dig Dis Sci 45:83-92
Ko, G.H., Kang, S.M., Kim, Y.K., Lee, J.H., Park, C.K., Youn, H.S., Baik, S.C. et al. (1999)
Invasiveness of Helicobacter pylori into human gastric mucosa. Helicobacter 2:77-81.
Kuipers, E.J. (1997) Helicobacter pylori and the risk and management of associated diseases:
gastritis, ulcer disease, atrophic gastritis and gastric cancer. Aliment Pharmacol Ther 1:71-88.
Kwok, T., Backert, S., Schwarz, H., Berger, J., Meyer, T.F. (2002) Specific entry of
Helicobacter pylori into cultured gastric epithelial cells via a zipper-like mechanism. Infect
Immun 4:2108-2120.
Leakey, A., La Brooy, J., Hirst, R. (2000) The ability of Helicobacter pylori to activate
neutrophils is determined by factors other than H. pylori neutrophil-activating protein. J Infect
Dis 182:1749-55.
36
Lindén, S., Nordman, H., Hedenbro, J., Hurtig, M., Borén, T., Carlstedt, I. (2002) Strain- and
blood group-dependent binding of Helicobacter pylori to human gastric MUC5AC
glycoforms. Gastroenterology 123:1923-1930.
Logan, R.P.H., and Hirschl A.M. (1996) Epidemiology of Helicobacter pylori infection. Curr
Opin Gastroenterol 12:1-5.
Mahdavi, J., Sonden, B., Hurtig, M., Olfat, F.O., Forsberg, L., Roche, N. et al. (2002)
Helicobacter pylori SabA adhesin in persistent infection and chronic inflammation. Science
297:573-578.
Marshall B.J., Warren J.R. (1984) Unidentified curved bacilli in the stomach of patients with
gastritis and peptic ulceration. Lancet 1:1311-1315.
Marshall, B.J. (2002) Helicobacter in the year 2000. [WWW document] URL
http://www.barryjmarshall.com/helicobacter/hpy2k/frMain.htm
Miller-Podraza, H., Bergström J., Teneberg S, Milh M.A., Löngård M., Olsson, B.M., Uggla,
L. et al. (1999) Helicobacter pylori and neutrophils: sialic acid-dependent binding to various
isolated glycoconjugates. Infect Immun 12:6309-6313.
Namavar, F., Sparrius, M., Veerman, E.C., Appelmelk, B.J., Vandenbroucke-Grauls, C.M.
(1998) Neutrophil-activating protein mediates adhesion of Helicobacter pylori to sulfated
carbohydrates on high-molecular-weight salivary mucin. Infect Immun 66:444-477.
37
Nishioka H., Baesso I., Semenzato G., Trentin L., Rappuoli R., Giudice G.D., Montecucco C.
(2003) The neutrophil-activating protein of Helicobacter pylori (HP-NAP) activates the
MAPK pathway in human neutrophils. Eur J Immunol 33: 840-849.
Noach, L.A., Bosma, N.B., Jansen, J., Hoek, F.J., van Deventer, S.J., Tytgat, G.N. (1994b)
Mucosal tumor necrosis factor-alpha, interleukin-1 beta, and interleukin-8 production in
patients with Helicobacter pylori infection. Scand J Gastroenterol 5:425-429.
Noach, L.A., Rolf, T.M., Tytgat, G.N. (1994a) Electron microscopic study of association
between Helicobacter pylori and gastric and duodenal mucosa. J Clin Pathol. 47:699-704.
Ota, A., and Genta R.M. (1997) Morphological characterisation of the gastric mucosa during
infection with H. pylori. In: The Immunbiology of H. pylori: from Pathogenesis to
Prevention, (Ernst, P. B., Michetti, P. & Smith, P. D., eds.) Philadelphia: Lippincott-Raven
Publishers, pp. 15-27.
Parkkinen, J., Rogers, G.N., Korhonen, T., Dahr, W., Finne, J. (1986) Identification of the Olinked sialyloligosaccharides of glycophorin A as the erythrocyte receptors for S-fimbriated
Escherichia coli. Infect Immun 54:37-42.
Potter, N.S., and Harding, C.V. (2001) Neutrophils process exogenous bacteria via an
alternate class I MHC processing pathway for presentation of peptides to T lymphocytes. J
Immunol 167: 2538-2546
38
Satin, B., Del Giudice, G., Della Bianca, V., Dusi, S., Laudanna, C., Tonello, F. (2000) The
neutrophil-activating protein (HP-NAP) of Helicobacter pylori is a protective antigen and a
major virulence factor. J Exp Med 191:1467-1476.
Semino-Mora C, Doi SQ, Marty A, Simko V, Carlstedt I, Dubois A (2003) Intracellular and
interstitial expression of Helicobacter pylori virulence genes in gastric precancerous intestinal
metaplasia and adenocarcinoma. J Infect Dis 187:1165-1177.
Sharma, S.A., Tummuru, M.K., Miller, G.G., Blaser, M.J. (1995) Interleukin-8 response of
gastric epithelial cell lines to Helicobacter pylori stimulation in vitro. Infect Immun 5:16811687.
Shimoyama, T., Everett, S.M., Dixon, M.F., Axon, A.T., Crabtree, J.E. (1998) Chemokine
mRNA expression in gastric mucosa is associated with Helicobacter pylori cagA positivity
and severity of gastritis. J Clin Pathol 10:765-770.
Sobala, G.M., Crabtree, J.E., Dixon, M.F., Schorah, C.J., Taylor, J.D., Rathbone B.J.,
Heatley, R.V., Axon, A.T. (1991) Acute Helicobacter pylori infection: clinical features, local
and systemic immune response, gastric mucosal histology, and gastric juice ascorbic acid
concentrations. Gut 11:1415-1418.
Steer, H.W. (1975) Ultrastructure of cell migration throught the gastric epithelium and its
relationship to bacteria. J Clin Pathol 8:639-846.
39
Stocks, S.C., Albrechtsen, M., Kerr, M.A. (1990) Expression of the CD15 differentiation
antigen (3-fucosyl-N-acetyl-lactosamine, LeX) on putative neutrophil adhesion molecules
CR3 and NCA-160. Biochem J 2:275-280.
Stocks, S.C., and Kerr, M.A. (1993) Neutrophil NCA-160 (CD66) is the major protein carrier
of selectin binding carbohydrate groups LewisX and sialyl lewisX. Biochem Biophys Res
Commun 1:478-483.
Suerbaum, S. (2000) Genetic variability within Helicobacter pylori. Int J Med Microbiol.
2:175-181.
Suerbaum S., Josenhans C., Labigne A. (1993) Cloning and genetic characterization of the
Helicobacter pylori and Helicobacter mustelae flaB flagellin genes and construction of H.
pylori flaA- and flaB-negative mutants by electroporation-mediated allelic exchange. J
Bacteriol 175:3278-88
Teneberg, S., Jurstrand, M., Karlsson, K.A., Danielsson, D. (2000) Inhibition of nonopsonic
Helicobacter pylori-induced activation of human neutrophils by sialylated oligosaccharides.
Glycobiology 10:1171-1181.
Teneberg S., Leonardsson I., Karlsson H., Jovall P.A., Ångstrom J., Danielsson D., Näslund I.
et al. (2002) Lactotetraosylceramide, a novel glycosphingolipid receptor for Helicobacter
pylori, present in human gastric epithelium. J Biol Chem. 22:19709-19719.
40
Teneberg, S., Miller-Podraza, H., Lampert, H.C., Evans, D.J., Evans, D.G., Danielsson, D. et
al. (1997) Carbohydrate binding specificity of the neutrophil-activating protein of
Helicobacter pylori. J Biol Chem 272:19067-19071.
Thomas, J.E., Dale, A., Harding, M., Coward, W.A., Cole, T.J., Weaver, L.T. (1999)
Helicobacter pylori colonization in early life. Pediatr Res 1999 2:218-223.
Warren, J.R., and Marshall, B.J. (1983) Unidentified curved bacilli on gastric epithelium in
active chronic gastritis. Lancet 1:1273-1275.
Westblom T.U., Czinn S.J., Nedrud, J.G., eds. (1999) Gastroduodenal Disease and
Helicobacter pylori. Pathophysiology, Diagnosis and Treatment. Curr. Top. Microbiol.
Immunol. 241. Berlin: Springer-Verlag
Wyle, F.A., Tarnawski, A., Schulman, D., Dabros, W. (1990) Evidence for gastric mucosal
cell invasion by C. pylori: an ultrastructural study. J Clin Gastroenterol. 12:92-98.
Zanotti, G., Papinutto, E., Dundon, W., Battistutta, R., Seveso, M., Giudice, G. et al. (2002)
Structure of the Neutrophil-activating protein from Helicobacter pylori. J Mol Biol 323:125130.
Zu, Y., Cassai, N.D., Sidhu, G.S. (2000) Light microscopic and ultrastructural evidence of in
vivo phagocytosis of Helicobacter pylori by neutrophils. Ultrastruct Pathol 5:319-323.
41