Adherence of Helicobacter pylori to Abiotic Surfaces Is Influenced by Serum

Abstract

Helicobacter pylori bacteria cultured in a chemically defined medium without serum readily adhere to a variety of abiotic surfaces. Growth produces microcolonies that spread to cover the entire surface, along with a planktonic subpopulation. Serum inhibits adherence. Initial attachment is protein mediated, but other molecules are responsible for more permanent attachment.

In vitro experiments with the gastric pathogen Helicobacter pylori have typically been performed with complex media containing blood or serum; however, such media complicate studies of bacterial metabolism and may not accurately represent the in vivo environment. Additionally, a growing body of evidence suggests that contaminated well water or seawater could be a source of H. pylori (4, 6, 7, 16). The data presented may be relevant for these environments. H. pylori grows in the chemically defined, protein-free medium Ham's F-12 (14) and develops interesting changes in bacterial morphology, motility, and adherence, which are reversed by the addition of serum.

H. pylori bacteria cultured microaerobically (5% O₂, 10% CO₂) in Ham's F-12 medium are longer and have a more filamentous morphology. When inoculated at low densities, isolated adherent bacteria divide to form microcolonies (Fig. 1A) that eventually form a nearly confluent layer on the surface of the culture vessel. The nonadherent subpopulation is largely nonmotile. When 1% fetal bovine serum (FBS) is added to the medium, H. pylori bacteria form large autoaggregates that do not adhere to the culture vessel (Fig. 1B). A larger proportion of the serum-grown bacteria are motile. Identical adherence and growth properties were seen in polystyrene cultureware and in borosilicate glass slide chambers used for differential interference contrast imaging of cultures.

FIG. 1.

Microscopic characteristics of H. pylori grown without (A) or with (B) 1% FBS in borosilicate slide chambers. (A) H. pylori 26695m microcolonies growing in the absence of serum. (B) Floating clumps of bacteria formed during growth with serum. Differential interference contrast imaging was carried out at a magnification of ×1,000.

We explored whether the nonmotile bacteria cultured without serum expressed flagella by using scanning electron microscopy of H. pylori cultured on Thermanox coverslips. Bacteria were fixed and coated with gold-palladium using standard methods similar to those of Tsang et al. (15). Some of the adherent bacteria possess flagella, and thicker, irregular filaments were observed (Fig. 2A). These filaments do not appear to be pili and might be extracellular matrix, as suggested by subsequent experiments. The rare serum-grown organisms adhering to coverslips (Fig. 2B) were short and had a bacillary morphology, features that are both typical of motile serum-grown organisms.

FIG. 2.

Scanning electron micrographs of adherent H. pylori 26695m following growth on coverslips in the (A) absence or (B) presence of 5% FBS. (A) Bacteria grown without serum have a filamentous morphology, and flagella can be seen on several of the bacteria. (B) Morphology typical of motile H. pylori grown in the presence of serum. Microscopy was performed with a Philips XL20 microscope. Magn, magnification; Det SE, secondary electron detection system; WD, working distance.

Without serum, spiral or filamentous bacteria initially dominate (Fig. 3, 24-h panel), but they are replaced by U forms and coccoid bacteria as the culture ages (72- and 96-h panels). Viability staining of biofilms (Fig. 3) (BacLight; Invitrogen Corp., Carlsbad, CA) indicates concomitant loss of viability. Most organisms stained green (viable) at 24 h; this was followed by increasing numbers of viable U forms and red (nonviable) coccoid bacteria (Fig. 3, 72-h panel). Viability could be prolonged by diluting the growth medium or replacing it completely with saline (data not shown). Filamentous bacteria viewed by light microscopy or scanning electron microscopy do not appear to be in the process of dividing; however, a close examination of fluorescence images reveals chains of short rods (Fig. 3, 24-h panel inset). The fluorophores used in viability staining bind nucleic acids, not cell wall elements. Thus, some form of internal septation may be present in filamentous forms.

FIG. 3.

Viability staining of biofilms after various amounts of time in culture. Replicate wells were inoculated with H. pylori 26695m in F-12 without serum and incubated for the stated amount of time before staining. The enlarged inset in the 24-h panel shows chains of short rods, which appear as individual bacteria in the absence of staining. The enlarged inset in the 72-h panel shows conversion to U forms and coccoid morphology with some loss of viability. Fluorescence microscopy was performed at ×400 magnification.

Swimming bacteria are uncommon in the absence of serum; however, a few bacteria within microcolonies spin and twitch, as though tethered at one end. For an example of this, see Fig. S1A in the supplemental material. If these cells are in the process of dividing, they may become motile upon completion of cell division, providing an important means of dispersal.

To determine whether adherence in the absence of serum is limited to polystyrene or whether it is a more general phenomenon, H. pylori was cultured with coupons of various materials. For quantification of adherent bacteria, coupons were washed three times with phosphate-buffered saline. Adherent bacteria were lysed with a compatible lysis buffer (25 mM Tris phosphate, pH 7.8; 2 mM dithiothreitol; 2 mM 1,2-diaminocyclohexane-N,N,N′,N′-tetra-acetic acid monohydrate; 10% glycerol; 1% Triton X-100 [J. Bessetti, Promega, personal communication]). A portion of the lysate was then assayed for ATP with CellTiter-Glo reagents (Promega). We detected significant adherence on glass, polycarbonate, polyvinylchloride, and aluminum (data not shown). The toxicity of stainless steel and copper prevented the detection of adherence by ATP content. Previous studies have shown the adherence of water-stressed H. pylori to a variety of surfaces (1, 2); however, our results suggest that an absence of serum rather than osmotic stress is required for adherence. We found comparable levels of adherence to untreated polystyrene (i.e., not “tissue culture treated”) and to the untreated side of Thermanox coverslips (data not shown). Thus, there appears to be no correlation between H. pylori adherence and surface hydrophobicity.

Experiments with serum concentrations ranging from 0.1% to 5% show that serum decreases adherence in a dose-dependent manner up to a concentration of 1% (Fig. 4). Adherence is also inhibited in wells that had been precoated with serum and washed to remove unbound protein. This observation eliminates the hypothesis that lack of adherence is due solely to H. pylori itself becoming coated with serum proteins. Rather, serum prevented H. pylori from contacting the well surface either by steric hindrance or by blocking the relevant adhesin(s). Therefore, it appears that H. pylori binds more avidly to polystyrene than it does to the serum proteins coating the well.

FIG. 4.

Effect of serum concentration on percentage of H. pylori bacteria adhering to culture surface. H. pylori was inoculated into triplicate wells containing F-12 with 0 to 10% serum. Percent adherence of viable bacteria was assessed after overnight growth by the following formula: 100(ATP content of adherent bacteria/total ATP). Error bars represent means ± standard deviations. The graph is a representative example of numerous similar experiments. ctrl, no serum.

To ensure that adherence in the absence of serum was not a strain-specific phenomenon, we tested a number of different H. pylori strains (Table 1). All strains tested showed significant adherence in the absence of serum (data not shown). The flgR mutant strain, which does not express flagella (12), a 26695 luxS quorum-sensing mutant, and an isogenic complemented strain (5) were equally able to form biofilms in the absence of serum (data not shown). Thus, adherence does not require flagella or quorum sensing.

TABLE 1.

H. pylori strains used in this study

H. pylori strain	Source or reference
26695m (motile 26695)	13
UMAB41	11
HpDJM17 (clinical isolate)	10
43504	ATCC
SS1	A. Lee
26695 luxS	9
26685 luxS+ (complemented)	9
43504 flgR	3

Experiments following parallel triplicate cultures from 8 to 64 h postinoculation showed more adherence (around 50%) during early logarithmic growth than during stationary phase (around 25%), but a significant planktonic population was always present (data not shown). The decline in percent adherence during stationary phase is likely due to a reduction of available surface area, since the bacteria appear to bind only to the surface and not to each other.

The planktonic subpopulation is capable of adhering to fresh culture vessels within a few hours, or sooner if the culture vessel is centrifuged, indicating that the protein(s) or molecule(s) responsible for adherence is expressed in this population. Short-time-interval experiments using late-log-phase cultures explored the nature of the adhesion process. Trypsin added at the time of inoculation inhibits the adhesion of H. pylori to multiwell plates (Fig. 5A) but is unable to remove organisms that had been allowed to adhere overnight (data not shown). The effect of trypsin is abrogated by the protease inhibitor leupeptin. When trypsin is added at various times after the adherence of H. pylori, it becomes evident that attachment is initially protease sensitive but becomes progressively more resistant to trypsin over the course of several hours (Fig. 5B). The degradation of the CellTiter-Glo luciferase reagent during assay of trypsin-containing samples was prevented by the addition of leupeptin. Chloramphenicol treatment does not prolong susceptibility to trypsin removal, suggesting that protein synthesis is not required for the second phase of attachment. Formalin-fixed H. pylori bacteria also adhere in the absence of serum but can be removed by trypsin regardless of the length of incubation prior to trypsin addition (data not shown). Thus, bacteria do not need to be alive for the initial, protein-mediated phase of adherence.

FIG. 5.

(A) Effect of trypsin (50 μg/ml) on adherence of H. pylori in the absence of serum. H. pylori bacteria from a culture grown without serum were allowed to adhere to a 24-well plate for 4 h in the presence or absence of trypsin and/or 100 μM leupeptin. (B) Time course of sensitivity of adherent H. pylori 26695m to removal by trypsin. Bacteria were added to wells and allowed to adhere. Trypsin was added at the time points indicated. Samples were assayed for ATP content 2 h after the last trypsin addition. Percent adherence of viable bacteria was assessed by the following formula: 100(ATP content of adherent bacteria/total ATP). Error bars represent means ± standard deviations. ctrl, control.

We conclude that serum profoundly influences H. pylori motility, morphology, and adherence. Adherence properties of H. pylori have potential in vitro and in vivo relevance. Epidemiological data suggest contaminated water supplies as a potential reservoir for H. pylori, and the closely related organism Campylobacter jejuni is suggested to spread among poultry flocks via contaminated watering devices (8, 17). Mucus serum concentrations in newly infected hosts are certain to be low, and serum levels may remain low even in mucus overlying inflamed tissue. Therefore, an examination of adherence to cells or host molecules in a variety of serum concentrations is warranted.