• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of jbacterPermissionsJournals.ASM.orgJournalJB ArticleJournal InfoAuthorsReviewers
J Bacteriol. Mar 2007; 189(5): 1794–1802.
Published online Dec 22, 2006. doi:  10.1128/JB.00899-06
PMCID: PMC1855705

Weak Rolling Adhesion Enhances Bacterial Surface Colonization[down-pointing small open triangle]


Bacterial adhesion to and subsequent colonization of surfaces are the first steps toward forming biofilms, which are a major concern for implanted medical devices and in many diseases. It has generally been assumed that strong irreversible adhesion is a necessary step for biofilm formation. However, some bacteria, such as Escherichia coli when binding to mannosylated surfaces via the adhesive protein FimH, adhere weakly in a mode that allows them to roll across the surface. Since single-point mutations or even increased shear stress can switch this FimH-mediated adhesion to a strong stationary mode, the FimH system offers a unique opportunity to investigate the role of the strength of adhesion independently from the many other factors that may affect surface colonization. Here we compare levels of surface colonization by E. coli strains that differ in the strength of adhesion as a result of flow conditions or point mutations in FimH. We show that the weak rolling mode of surface adhesion can allow a more rapid spreading during growth on a surface in the presence of fluid flow. Indeed, an attempt to inhibit the adhesion of strongly adherent bacteria by blocking mannose receptors with a soluble inhibitor actually increased the rate of surface colonization by allowing the bacteria to roll. This work suggests that (i) a physiological advantage to the weak adhesion demonstrated by commensal variants of FimH bacteria may be to allow rapid surface colonization and (ii) antiadhesive therapies intended to prevent biofilm formation can have the unintended effect of enhancing the rate of surface colonization.

Biofilms consist of surface-associated colonies of bacteria (11, 45) and are a major concern for implanted medical devices and in many diseases. They are the dominant mode of bacterial life in nature and exist on biological as well as abiotic surfaces (11, 45). Eradication of biofilms is more problematic than that of bacteria in the planktonic mode of growth, since biofilms are resistant to innate host defenses (21, 32, 51), mechanical removal, and antibiotic treatments (6, 52). Therefore, a more promising strategy that has been proposed is to prevent biofilm formation through interference with the earliest steps of formation (11, 38, 45). Surface adhesion, defined as the binding of a planktonic bacterium to a surface, is the first step, and it is generally assumed that strong irreversible adhesion is necessary for biofilm formation (9, 14). The next step is surface colonization, defined as the spread of adherent bacteria across a surface through division. It has been suggested that biofilms can be prevented by restricting these early stages of colonization by blocking specific receptor-ligand interactions with soluble inhibitors or antibodies that block adhesion rather than prevent bacterial growth (39).

Escherichia coli binding via the protein FimH provides a model system for studying surface adhesion and colonization for two reasons. First, E. coli is the most common cause of both urinary tract infections (22, 23) and biofilms forming on urinary catheters (10, 20, 33), which can lead to bacteremia and increased mortality (29, 40). FimH is the most common adhesin (adhesive protein) in E. coli and other enteric bacteria. It is expressed on the tip of type 1 fimbriae and binds to glycoproteins via N-linked oligosaccharides that terminate in single or multiple mannose residues. Natural ligands for FimH include uroplakins on urinary epithelial cells in urinary bladders (4, 19, 25, 30, 31) and immunoglobulin A or mucin in intestines (1, 35) and lungs. FimH has also been shown to mediate adhesion to abiotic surfaces (46). However, while many of the studies of biofilms have used abiotic surfaces, implanted biomaterials are rapidly coated with glycoproteins that are deposited by bodily fluids and mediate bacterial adhesion (34). For example, urine contains mannose-containing glycoproteins such as Tamm-Horsfall protein (47), which can be expected to coat urinary catheters. Thus, the adhesion of E. coli via FimH to mannosylated surfaces provides a relevant model for the formation of biofilms on both urinary catheters and natural tissues in many physiological compartments.

Second, FimH mediates a range of adhesive behaviors that may be important for surface colonization. Most commensal strains of E. coli express FimH variants that bind to mannose-terminated glycoproteins weakly in static or low-shear-stress conditions so that they roll across the surface or even detach (37, 54). However, increased shear stress causes the bacteria to bind instead in a strong stationary fashion, a phenomenon termed shear-dependent stick-and-roll adhesion (37, 54) which involves a reversible change in behavior rather than a selection of subpopulations. This likely occurs because the drag force on the bacteria at high shear stress induces a conformational change in FimH (reference 55 and P. Aprikian, V. Tchesnokova, B. Kidd, E. Trinchina, O. Yakovenko, V. Vogel, W. Thomas, and E. Sokurenko, submitted for publication) that activates it to form longer-lived bonds (53, 54), a phenomenon known as catch bonds (18, 53, 59). This shear-enhanced stick-and-roll adhesion is observed not only when FimH binds to model ligands that terminate in monomannose residues but even when it binds to model ligands such as bovine RNase B (RNaseB) which have high-mannose oligomannose carbohydrate modifications (3M), although the bacteria roll more slowly upon and detach less frequently from the latter surface (37). A high proportion of pathogenic bacteria contain point mutations that enhance static adhesion to various extents (49, 50). Since the mutations do not increase high-shear adhesion, some engineered point mutations can cause E. coli to bind so strongly at low shear stress that binding is shear inhibited (37, 54), similar to what has been shown or assumed for many other adhesins. While it is generally assumed that strong adhesion is advantageous for bacteria, the high conservation and evolutionary dominance of weak FimH adhesion in commensal E. coli (49, 50) suggests that weak shear-enhanced adhesion has advantages to the bacteria. Finally, E. coli FimH is not the only adhesin to display shear-enhanced stick-and-roll adhesive behavior (37). Thus, E. coli binding via FimH can serve as a model system for two common modes of surface adhesion: shear-enhanced stick-and-roll adhesion and strong shear-independent adhesion.

Here we ask how rolling adhesion, which is associated with weak binding to a surface in the presence of fluid flow, affects the early stages of biofilm formation. In order to compare the evolutionarily dominant weak-binding phenotype with stronger-binding phenotypes, we compare three E. coli strains—one with a typical weak shear-enhanced FimH, one with an engineered constitutively strong binding FimH, and one with a naturally occurring FimH that displays an intermediate level of adhesion relative to the other two. By analyzing videos of colonization by sparsely seeded bacteria, we observed that weak rolling adhesion enhances the rate of early-stage E. coli surface colonization and results in a more uniform surface coverage, while the strong adhesion that is normally assumed to be advantageous results in slower surface coverage and the formation of tight microcolonies. These results suggest that care must be taken when trying to prevent biofilm formation by inhibiting adhesion. Indeed, we found that the addition of a soluble inhibitor to adhesion, methyl α-d-mannopyranoside (αMM), actually enhanced the rate of surface colonization for the strong-binding FimH by allowing the bacteria to roll.


Preparation of bacterial strains.

In order to distinguish between the abilities of bacteria to roll versus stick and other effects of shear stress, we used bacteria that were genetically engineered to express a “FimH-hi” variant that binds strongly even at low shear stress, as well as a “FimH-med” variant that shows an intermediate level of binding. The E. coli strains expressing FimH-wt (a typical commensal FimH variant), FimH-med, and FimH-hi were constructed as described previously by transforming an E. coli K-12 derivative, AAEC191A, from which the fim operon had been deleted (reference 55 and Aprikian et al., submitted), with the recombinant plasmid pPKL114, containing the fim operon with a stop codon that prevents fimH expression, and finally with a pGB2-24-based plasmid containing the indicated variant of FimH. FimH-wt has a sequence identical to that of FimH-f18. FimH-med is identical to FimH-f18, but differs by an S62A mutation and displays binding characteristics that are intermediate in comparison to the two other variants (24, 41, 57). FimH-hi is equivalent to FimH-j96 with a mutation from alanine to aspartic acid in residue 188 (A188D), which removes the regulatory mechanism that normally keeps the adhesin in a low-binding state in static conditions (Aprikian et al., submitted). (FimH-j96 is slightly higher binding than FimH-f18 due to the V27A substitution [48].) Other than the differences in FimH, the three E. coli strains are genetically identical. Bacteria were grown in SuperBroth (AMRESCO Inc.) overnight, washed, and brought to 1 × 108 CFU/ml in 0.2% bovine serum albumin (BSA) in phosphate-buffered saline (PBS) prior to experiments. All bacteria showed essentially the same planktonic growth rates when grown in shaking baths after this preparation: the FimH-wt, FimH-med, and FimH-hi E. coli strains doubled every 22.5 ± 0.5, 22.7 ± 1.1, and 22.6 ± 1.5 min, respectively.

Coating of plates.

Glycoprotein plates were prepared by placing 100 μl of 50 μg/ml RNaseB (9001-99-4; Sigma) in 0.02 M bicarbonate on glass coverslips for approximately 18 h at room temperature, followed by three successive washes in 0.2% BSA-PBS prior to use in each experiment.

Measuring IC50 of inhibitor.

IC50 is the concentration of inhibitor required to inhibit adhesion by 50% as a result of blocking specific receptor-ligand binding interactions. Bacterial adhesion to RNaseB-coated wells was measured with a radionuclide adhesion assay (48) in the presence of various concentrations of αMM. The number of bacteria prelabeled with radioactive thymidine was determined after a 45-minute incubation. αMM does not affect growth rates of planktonic bacteria.

Measuring bacterial adhesion in dynamic conditions.

All adhesion (not biofilm growth) experiments were performed at room temperature in PBS-BSA buffer on RNaseB plates in a once-through, nongrowth, shear-controlled flow system. The protein-coated glass coverslips formed the bottom of a parallel-plate flow chamber (2.5 cm [length] × 0.25 cm [width] × 250 μm [height]) (GlycoTech). Other than the differences in the surfaces, these assays were performed as previously described (36). The bound bacteria were recorded in time-lapse digital videos with the camera shutter held open just long enough to blur out free-floating bacteria. To measure the number of bacteria accumulating, a solution of 108 CFU/ml was washed through the flow chamber for 5 minutes at the indicated level of shear stress. To measure the fractions of bacteria that roll at each shear, bacteria were first loaded at 0.01 Pa (0.021 ml/min) for approximately 2 minutes until approximately 100 bacteria were present in order to bind enough bacteria to get statistically significant data even at shear levels that prevent initial attachment. The shear stress was then changed to the indicated level, and after 20 seconds, a 1-minute 1-frame-per-second time-lapse video was recorded. To measure the fractions of rolling bacteria, the positions of bacteria 20 seconds apart were compared, and the bacteria were designated as rolling if they moved more than 1 bacterial radius (approximately 1 μm). Essentially no bacteria were observed to detach at any of the measured shears.

Shear-stress-controlled biofilm reactor.

A heated parallel-plate flow chamber, an in-line heater, a temperature controller (all from Warner Instruments), and a syringe pump were used to grow bacteria. Figure Figure11 shows a schematic of the system. Shear stress was varied by adjusting the volumetric flow rate with the syringe pump. The syringe pump was used instead of a peristaltic pump to avoid oscillations in the flow rate, and continual flow was made possible by using two syringes in an alternating-infusion-and-withdrawal mode attached to valves to create a once-through flow environment continuously drawing new bacterial growth medium into the flow cell and expelling used medium for collection in a waste container. To smooth out the high-frequency pulses in flow that still resulted from a stepping motor syringe pump, compliant silicone tubing was used, and the syringe size was chosen so that the step motor on the syringe pump could be set at a high rate (>1,000 steps per second). A temperature of approximately 37°C was maintained with the in-line heater, heated parallel-plate flow chamber, and temperature controller. An RNaseB-coated glass coverslip was used as the bottom of the flow chamber (prepared as described above). SuperBroth was used as the bacterial growth medium with 1 μg/ml ampicillin and chloramphenicol antibiotics, to which the bacterial strains used in this work were resistant. The initial inoculum concentration of 106 to 108 CFU/ml was delivered into the flow chamber, and the flow was switched to a low-shear or static condition. When the desired number of adherent bacteria per field of view was achieved (these number were later normalized for comparison between experiments), the source was switched to fresh SuperBroth, and the unattached bacteria were washed out at the shear stress used for that experiment. This seeding protocol typically took 5 minutes. The volumetric flow rate in the Warner Instruments heated parallel-plate flow chamber was 0.38 ml/min for 0.2 Pa, 0.95 ml/min for 0.5 Pa, and 3.8 ml/min for 2 Pa.

FIG. 1.
Once-through, shear- and temperature-controlled parallel-plate flow chamber system.

Data analysis.

The rates of surface colonization were quantified by applying a minimum intensity threshold to the images in MetaMorph to distinguish the bacteria from the surface and then calculating the total area covered by bacteria. The increase (n-fold) was obtained by dividing the total area at each time by that at the start of bacterial growth. The perimeters of the colonies were traced using MetaMorph to record the area and x and y dimensions.


Bacterial adhesion in dynamic conditions.

Before measuring the effect of the adhesion on surface colonization, it is helpful to understand how the different bacterial strains bind to surfaces. E. coli bacteria expressing a typical commensal FimH variant (FimH-wt) were studied binding to the 3M ligand, RNaseB, at 37°C. A large number of bacteria bound at low shear, but increased shear stress above 0.05 Pa reduced the number of E. coli cells binding from solution to the 3M surface (Fig. (Fig.2,2, upper panel). This experiment was also performed with bacteria expressing FimH-med, the strongest-binding phenotype found for clinical isolates of E. coli (24, 41, 57), and FimH-hi, a variant that is genetically engineered to mediate strong adhesion (Aprikian et al., submitted). Bacteria expressing these stronger-binding FimH variants showed only small differences in the numbers of bacteria adhering in these experiments compared to bacteria expressing FimH-wt, with shear stress inhibiting accumulation for all variants tested.

FIG. 2.
Adhesion of E. coli to the 3M surface. (Upper panel) Increased shear stress reduced the number of bacteria that transfer from solution to the surface after 5 min for FimH-wt E. coli and FimH-hi E. coli. (Lower panel) When already adherent bacteria were ...

To see how shear stress and mutations affect the ability of surface-bound bacteria to detach, roll, or bind firmly, bacteria were initially bound at low shear (0.01 Pa) to allow similar numbers to bind in each test, and then they were subjected to various levels of shear stress from 0.01 to 2 Pa. For all FimH variants, none of the bacteria detached within 20 seconds at any shear stress (not shown). At low shear stress, half of the FimH-wt E. coli cells were observed to bind in a rolling fashion in which they moved forward at least 1 bacterial radius over 20 seconds, while the remainder were stationary during this time frame. When the flow was turned up, some of the rolling FimH-wt E. coli cells stopped and bound in a stationary manner (Fig. (Fig.2,2, lower panel). The fraction that remained rolling decreased with shear stress, and all variants stopped rolling completely if the shear was switched to 2 Pa. Thus, while increased shear prevented FimH-wt E. coli cells from initially binding to the 3M surface (Fig. (Fig.2,2, upper panel), it increased the strength of binding for FimH-wt E. coli cells that were already bound (Fig. (Fig.2,2, lower panel). FimH-hi E. coli cells showed reduced rolling at all shears relative to FimH-wt cells and stopped rolling completely at and above 0.1 Pa. FimH-med cells displayed an intermediate behavior in that they rolled more than FimH-hi cells but less than FimH-wt cells.

Effect of rolling adhesion on patterns of surface colonization.

In separate experiments, the surface of a biofilm reactor was seeded sparsely with E. coli cells at low shear stress and transferred to nutrient medium at the indicated flow conditions. After 4 h of growth at 0.2 Pa, approximately 20% of FimH-wt E. coli cells rolled at least 1 bacterial diameter during any 1-min period over the entire time course of the video. By the end of 3 h, FimH-wt E. coli cells binding at 0.2 Pa colonized the surface throughout the field of view (Fig. (Fig.3A;3A; also see Video S1 in the supplemental material). In contrast, when the same number of initially adherent bacteria were allowed to colonize the surface at 2.0 Pa, the bacteria remained in tight microcolonies rather than spreading out as they divided (Fig. (Fig.3B;3B; also see Video S2 in the supplemental material).

FIG. 3.
Surface colonization after 3 h. (A) FimH-wt E. coli cells grown at moderate shear stress (0.2 Pa). (B) FimH-wt E. coli cells grown at 2 Pa, which causes a switch to stationary adhesion. (C) FimH-hi E. coli cells grown at 0.2 Pa. (D) FimH-med E. coli cells ...

Since FimH-hi E. coli cells mediate stationary adhesion at 0.2 Pa, they allow a test of the effect of stationary adhesion on surface colonization without the other complicating effects of increased shear stress. FimH-hi E. coli cells formed tightly clumped microcolonies similar to those of FimH-wt cells at high shear stress (Fig. (Fig.3C;3C; also see Video S3 in the supplemental material). FimH-med E. coli, which demonstrated an intermediate strength of adhesion as shown in the lower panel in Fig. Fig.2,2, was also grown at 0.2 Pa. After three hours of growth, FimH-med E. coli displayed both tight microcolonies and the movement of individual bacteria outside these colonies, spreading more than FimH-hi E. coli but less than FimH-wt E. coli (Fig. (Fig.3D).3D). Together, these experiments demonstrate that whether strong stationary adhesion was caused by point mutations in FimH or by increased shear stress, E. coli formed tight microcolonies and failed to colonize the bulk of the bare surface. In contrast, weak rolling adhesion allowed rapid and fairly uniform colonization of the surface.

Effect of rolling adhesion on the rate of surface colonization.

Although the numbers of bacteria binding at the start of the experiments were similar in all cases, the different patterns of colonization appeared to be associated with a difference in the amount as well as in the pattern of colonization by the end of the 3 hours. In order to quantify the difference, the rate of surface colonization was measured by calculating the percentage of surface area covered as a function of time while the bacteria proliferated and colonized the surface. These measurements were then normalized to show the increase (n-fold) in surface colonization, since the experiments had slightly different numbers of initially adherent bacteria. After the first 120 min, a clear difference was observed between the weakly and strongly adherent bacteria, and this difference grew to up to threefold by just 180 min. This was true whether strong adhesion was caused by increased shear stress or by point mutations in FimH (Fig. (Fig.3E).3E). These experiments suggest that weak rolling adhesion offers a quantitative advantage over strong stationary adhesion when it comes to colonizing a surface.

Dynamics of colonization.

In order to determine how rolling contributed to the different rates of surface colonization, it was necessary to observe the movement and division of individual cells—that is, the dynamics of colonization—during the time course of the experiment. To do this, we grew the biofilms in the same conditions but this time recorded the time-lapse videos at a higher magnification. Figure Figure4A4A shows images taken every 20 min of the colonization of the surface in a region containing one of the initially seeded cells. As these weakly adherent bacteria divided, they moved slowly in the direction of fluid flow, so that many bacteria detached from or rolled out of the region but others remained and some even rolled in from upstream. By 120 min later, there were about 16 cells in the region. Because of this rolling, a different region that initially contained no bacteria soon became indistinguishable from the seeded region in terms of the number of colonizing bacteria (Fig. (Fig.4B).4B). In contrast, FimH-hi E. coli cells moved very little when they divided (Fig. (Fig.4C).4C). After 120 min, the region around the initially seeded bacteria was densely populated with a confluent layer of around 50 to 100 bacteria. However, the regions that did not initially contain any of the initially seeded bacteria remained empty (Fig. (Fig.4D4D).

FIG. 4.
High-resolution images of dynamic changes in colonization. FimH-wt (A and B) and FimH-hi (C and D) E. coli bacteria at 0.2 Pa. Surfaces contain one initially adherent bacteria (A and C) or none (B and D). FimH-wt bacteria spread diffusely, while FimH-hi ...

The effect of adhesive strength on the rate of surface colonization raises the question as to whether adhesive strength affects the rate of division for adherent bacteria in some unknown manner. To test this, the time was measured between divisions for individual FimH-wt and FimH-hi E. coli cells grown on the surface at moderate shear stress (0.2 Pa). Both variants showed an initial lag before the first bacteria divided that may reflect a slow response to the presence of growth medium after overnight growth to the stationary phase followed by prolonged exposure to PBS. However, after this time, bacteria expressing FimH-wt divided at a mean time of 20.2 ± 0.9 min (n = 30, standard error of the mean reported), while bacteria expressing FimH-hi divided every 21.0 ± 0.6 min (n = 28). That is, the rates of division were the same within statistical significance and were essentially unchanged from the rates of division in the planktonic state (see Materials and Methods).

In order to understand what might limit the rate of surface colonization of strongly adherent bacteria, we measured the sizes of several microcolonies of FimH-hi bacteria growing at 0.2 Pa. The growth in the area of the microcolonies is shown in Fig. Fig.5A,5A, starting at the time when the initial cell began to divide; the predicted exponential growth given the measured doubling time is indicated. While the microcolonies at first grow as predicted, they fall below this growth rate by approximately 4 doubling times, when they have about 16 cells (Fig. (Fig.5A).5A). At this time, some of the bacteria are completely surrounded on the surface by other bacteria, and as they divide daughter cells can be seen to rise up above the surface and either detach or create a three-dimensional microcolony. While the phase-contrast videos are not well designed to distinguish between these two fates, either situation means that the newly divided cells that are not on the colony periphery are not contributing to the total surface area that is colonized. The radii of the microcolonies increased linearly once they contained about 16 cells (Fig. (Fig.5B5B).

FIG. 5.
Quantification of FimH-hi microcolony growth. (A) The symbols show the average areas of the tight microcolonies formed by FimH-hi (n = 4 for each symbol). The solid line shows an exponential expansion in area, with the time constant taken directly ...

Longer periods of time.

To see whether the observations about the rates and patterns of colonization extend to longer periods of time, FimH-wt E. coli cells and FimH-hi E. coli cells were grown for 8 h. In these experiments, the field of view was seeded with exactly one bacterium. A FimH-wt biofilm was grown for 8 h at the moderate shear stress of 0.5 Pa. The entire field of view was colonized through steady spreading (Fig. (Fig.6A).6A). Under the same conditions, the FimH-hi E. coli strain produced a very different biofilm. Observation of many fields of view showed that each initial bacterium grew into a single large microcolony over time (Fig. (Fig.6B).6B). As the microcolonies became three-dimensional, the cells which were not in contact with the surface sometimes rolled off in clumps and either disappeared downstream or contributed to the size of the microcolony by attaching to the surface at the downstream edge of the microcolony. At the end of 8 h, the microcolonies were longer in the direction of flow than in the cross-flow direction. New bacteria, presumably detaching from other microcolonies, occasionally bound in the field of view and began to grow additional microcolonies. Nevertheless, even after 8 h, this strain had not colonized the entire surface and remained in microcolonies.

FIG. 6.
Surface colonization after 8 h. (A) FimH-wt E. coli variant grown at 0.5 Pa. (B) FimH-hi E. coli grown at 0.5 Pa. (C) Quantification of the increase (n-fold) in surface area colonized for FimH-hi and FimH-wt E. coli. In each case, one bacterium was seeded ...

Effect of inhibiting adhesion.

To investigate whether a soluble inhibitor can reduce surface colonization, methyl α-d-mannopyranoside was added to the growth medium for both the FimH-wt and FimH-hi strains after the initial bacteria were seeded on the surface. When the shear-enhanced strain was grown at 0.2 Pa with 5 mM αMM in the bacterial growth medium, the rate of colonization decreased (Fig. (Fig.7A).7A). When bacteria divided in the experiment with inhibitor, they would often disappear from the field of view and presumably were washed out of the flow chamber, and the resulting surface colonization was sparse (compare Fig. Fig.5B5B to Fig. Fig.3A).3A). This is consistent with the expectation that inhibitors of adhesion will cause bacteria to detach or fail to attach and thus inhibit surface colonization and biofilm formation.

FIG. 7.
Effect of inhibitor on surface colonization. (A) 5 mM αMM inhibitor (gray symbols) was added to the medium during growth of FimH-wt or FimH-hi E. coli cells for 3 hours at 0.2 Pa. The fraction of surface area covered was calculated and normalized ...

In contrast, the rate of colonization for FimH-hi E. coli actually increased when inhibitor was added (Fig. (Fig.7A.)7A.) In the presence of inhibitor, this strain showed a somewhat uniform coverage (Fig. (Fig.7C)7C) that was similar to that observed with the weak rolling adhesion of the FimH-wt strain. Moreover, individual FimH-hi E. coli cells could be observed to roll on the surface during growth in the presence of inhibitors. This result is in contrast to the expectation that inhibiting adhesion will decrease surface colonization and shows that the partial inhibition of adhesion can induce a rolling behavior that can actually enhance the rate of colonization.

To determine whether the different effects of inhibitors could be due to different IC50s for the two variants, we tested the inhibition of E. coli adhesion to 3 M in static conditions in simple nongrowth assays. We found that the IC50 values for αMM were 15.9 mM for the FimH-wt strain and 1.9 mM for the FimH-hi strain. While there is a difference in IC50 values for the two strains, it is opposite from what would be expected to cause the different effects on colonization—that is, even a sub-IC50 concentration of inhibitor was sufficient to lower colonization by FimH-wt bacteria, while even super-IC50 concentrations of inhibitor were not sufficient to lower colonization by FimH-hi bacteria and instead increased the rate of colonization.


We observed that the strength of adhesion affects the manner and rate at which bacteria colonize a surface in flow conditions. FimH-hi bacteria, which have a high binding strength even in the absence of flow, formed tight microcolonies and showed a limited ability to spread into areas of the surface that were initially bare of bacteria. In contrast, FimH-wt bacteria passively rolled across the surface in the presence of flowing fluid, distributing over large areas fairly evenly, while FimH-med bacteria showed intermediate behavior. Since the environmental conditions and genetic backgrounds of the bacteria in these experiments were identical, these differences in the patterns of colonization were entirely due to the difference in FimH structure. Similar differences were observed when adhesion was strengthened by increasing shear stress instead of by altering FimH structure.

The strength of adhesion affected the rate as well as the pattern of colonization. It was shown that the strength of adhesion did not affect the rate of cell division, which remained the same within statistical error for all variants in planktonic or surface growth modes. It could also be observed that the strength of adhesion had no apparent effect on the size of the adherent bacteria. Thus, adhesive strength must alter the rate of surface colonization via a mechanism other than the rate of growth or division of E. coli cells. It is possible that the growth of strongly adherent bacteria is limited because they spread across the surface primarily through growth and division, while weakly adherent cells can roll to find uncolonized surface.

The ability of a bacterium to translocate across a surface in the presence of fluid flow is a passive form of mobility which relies on external force provided by fluid flow. Active forms of surface mobility, including type IV pili-mediated twitching motility (27) and flagellar motility (12, 39, 42, 44, 56), have been shown to be important for biofilm formation for E. coli (12, 42, 58) as well as for other bacteria, such as the well-studied Pseudomonas aeruginosa (27, 39, 44, 56). In many cases, motility was necessary neither for initial attachment nor for growth of the three-dimensional biofilm but rather for the spread of the bacteria across the surface (27, 42, 56). In the presence of fluid flow, it may become increasingly difficult for bacteria to move under their own power and increasingly convenient to utilize external forces, so that passive mobility could be most useful in these conditions. Other forms of passive mobility, including sliding motility on agar (2, 16) and slippery attachment on abiotic stainless steel (28), have also been reported to increase the rate of colony formation. However, our study is unique in showing weak receptor-mediated adhesion to a surface coated with biomolecules in flow and thus is relevant for medical biofilm formation. The advantages of weak rolling adhesion should be applicable to any receptor-ligand pair and should not be limited to shear-enhanced stick-and-roll adhesion via FimH. While it remains to be determined how the dramatic differences seen in the initial patterns of growth will affect the structures of mature biofilms, these other motility factors have been shown to affect the final structure of biofilms as well (3, 26, 27, 58).

It has recently been shown that entire microcolonies can tear away from biofilms in fluid flow and roll slowly along a surface without detaching from it (43). Like the rolling described here, microcolony rolling would also provide a mechanism for nonmotile bacteria to spread across a surface more rapidly than is possible with linear radial growth. The rolling adhesion we demonstrate here requires weak adhesion and would be relevant in the early stages of biofilm formation. The rolling of microcolonies would be relevant after the formation of mature biofilms and apparently requires high shear stress together with substantial microcolony size to apply drag force sufficient to overcome adhesive interactions. Indeed, we observed some rolling of large pieces of FimH-hi microcolonies between 4 and 8 h that contributed to colony spread. The cellular and microcolony rolling appear to be distinct in terms of mechanism, but together they underlie the importance of rolling in the spread of biofilms.

The two modes of surface colonization—formation of tight microcolonies via strong adhesion and rapid spreading via weak rolling adhesion—may each be advantageous to the bacteria in different conditions. If the initial infection rate is high, as in most traditional biofilm experiments, the distance between initially adherent bacteria will be small and there would be no requirement for rapid spreading over large areas. In this case, strong adhesion is likely to be advantageous to maintain attachment. However, biofilms forming in vivo are likely to start with a small number of initially adherent bacteria that contaminated the surface prior to implantation or that evaded the immune response long enough to bind a surface. In addition, many implants and tissues are exposed to high levels of fluid flow (0.1 to 0.2 Pa in veins, 1 to 5 Pa in arteries [7, 15], 0.3 to 0.5 Pa in the urethra during urination, and high shear due to a highly viscous and moving mucous layer in the intestines [5]) that can prevent spreading via the reattachment of free-floating bacteria (references 8 and 54; also W. E. Thomas, Proceedings of the Third International Conference on Microchannels and Minichannels, Toronto, Ontario, Canada, 13 to 15 June 2005) For these reasons, bacteria in vivo may benefit if they can spread rapidly by rolling over the surface. It has been pointed out that there is a “race to the surface” between bacteria and host cells (13).

Antiadhesive therapy—that is, methods to prevent adhesion—has been proposed as a way to combat bacterial infections, one that should be particularly useful as bacteria become increasingly resistant to antibiotics (38). We observed that antiadhesive treatment could increase or decrease the rate of surface colonization. The antiadhesive treatment used in this work, αMM, is not metabolized by E. coli, nor is it toxic, and so it does not affect the rate of growth of E. coli in planktonic conditions and must act by altering adhesion. We have shown previously that soluble inhibitor is effective in detaching rolling bacteria (36). Consistent with this, we observed here that αMM, even at a concentration below the IC50, was able to reduce colonization by FimH-wt E. coli. This result supports the conventional wisdom that inhibiting adhesion will limit biofilm formation. However, we also show here that the addition of a soluble inhibitor could enhance the rate of surface colonization for strong-binding bacteria. We found earlier that αMM was not very effective in detaching bacteria bound in a strong stationary mode of adhesion (36). Instead of causing the FimH-hi E. coli cells to detach, the inhibitor appears to have caused them to roll slowly. This might be expected if the inhibitor greatly reduces the number of bonds that can form between bacteria and the surface, allowing the bacteria to creep forward as each force-bearing bond breaks, while in the absence of inhibitor, the many additional bonds restrict any movement from occurring when a bond breaks. This demonstrates that antiadhesion inhibitors can enhance the rate of surface colonization for strongly adherent bacteria. This study involved levels of inhibitor that were sufficient to inhibit most but not all initial adhesion in static conditions and shows that subinhibitory concentrations of antiadhesives pose risks just as sublethal concentrations of antibiotics do (17). This work demonstrates that antiadhesive therapy must be approached with a quantitative understanding to avoid doing more harm than good. Because this observation was made for the constitutively active FimH-hi E. coli, the lesson is likely to apply universally to all strong receptor-ligand interactions.

Supplementary Material

[Supplemental material]


This work was supported by R01 grant A150940 from the National Institute of Allergy and Infectious Diseases.


[down-pointing small open triangle]Published ahead of print on 22 December 2006.

Supplemental material for this article may be found at http://jb.asm.org/.


1. Bollinger, R. R., M. L. Everett, D. Palestrant, S. D. Love, S. S. Lin, and W. Parker. 2003. Human secretory immunoglobulin A may contribute to biofilm formation in the gut. Immunology 109:580-587. [PMC free article] [PubMed]
2. Brown, I. I., and C. C. Hase. 2001. Flagellum-independent surface migration of Vibrio cholerae and Escherichia coli. J. Bacteriol. 183:3784-3790. [PMC free article] [PubMed]
3. Chiang, P., and L. L. Burrows. 2003. Biofilm formation by hyperpiliated mutants of Pseudomonas aeruginosa. J. Bacteriol. 185:2374-2378. [PMC free article] [PubMed]
4. Connell, I., W. Agace, P. Klemm, M. Schembri, S. Marild, and C. Svanborg. 1996. Type 1 fimbrial expression enhances Escherichia coli virulence for the urinary tract. Proc. Natl. Acad. Sci. USA 93:9827-9832. [PMC free article] [PubMed]
5. Costerton, J. W., K. R. Rozee, and K. J. Cheng. 1983. Colonization of particulates, mucous, and intestinal tissue. Prog. Food Nutr. Sci. 7:91-105. [PubMed]
6. Costerton, J. W., P. S. Stewart, and E. P. Greenberg. 1999. Bacterial biofilms: a common cause of persistent infections. Science 284:1318-1322. [PubMed]
7. Davies, P. F. 1995. Flow-mediated endothelial mechanotransduction. Physiol. Rev. 75:519-560. [PMC free article] [PubMed]
8. Dickinson, R. B., J. A. Nagel, D. McDevitt, T. J. Foster, R. A. Proctor, and S. L. Cooper. 1995. Quantitative comparison of clumping factor- and coagulase-mediated Staphylococcus aureus adhesion to surface-bound fibrinogen under flow. Infect. Immun. 63:3143-3150. [PMC free article] [PubMed]
9. Donlan, R. M. 2001. Biofilm formation: a clinically relevant microbiological process. Clin. Infect. Dis. 33:1387-1392. [PubMed]
10. Donovan, W. H., R. Hull, and C. D. Rossi. 1996. Analysis of gram negative recolonization of the neuropathic bladder among patients with spinal cord injuries. Spinal Cord 34:587-591. [PubMed]
11. Fux, C. A., J. W. Costerton, P. S. Stewart, and P. Stoodley. 2005. Survival strategies of infectious biofilms. Trends Microbiol. 13:34-40. [PubMed]
12. Genevaux, P., P. Bauda, M. S. DuBow, and B. Oudega. 1999. Identification of Tn10 insertions in the dsbA gene affecting Escherichia coli biofilm formation. FEMS Microbiol. Lett. 173:403-409. [PubMed]
13. Gristina, A. 2004. Biomaterial-centered infection: microbial adhesion versus tissue integration. 1987. Clin. Orthop. Relat. Res. 2004:4-12. [PubMed]
14. Gristina, A. G. 1987. Biomaterial-centered infection: microbial adhesion versus tissue integration. Science 237:1588-1595. [PubMed]
15. Guo, Z., M. Moreau, D. W. Rickey, P. A. Picot, and A. Fenster. 1995. Quantitative investigation of in vitro flow using three-dimensional colour Doppler ultrasound. Ultrasound Med. Biol. 21:807-816. [PubMed]
16. Henrichsen, J. 1972. Bacterial surface translocation: a survey and a classification. Bacteriol. Rev. 36:478-503. [PMC free article] [PubMed]
17. Hoffmann, N., T. B. Rasmussen, P. O. Jensen, C. Stub, M. Hentzer, S. Molin, O. Ciofu, M. Givskov, H. K. Johansen, and N. Hoiby. 2005. Novel mouse model of chronic Pseudomonas aeruginosa lung infection mimicking cystic fibrosis. Infect. Immun. 73:2504-2514. [PMC free article] [PubMed]
18. Isberg, R. R., and P. Barnes. 2002. Dancing with the host; flow-dependent bacterial adhesion. Cell 110(1):1-4. [PubMed]
19. Iwahi, T., Y. Abe, M. Nakao, A. Imada, and K. Tsuchiya. 1983. Role of type 1 fimbriae in the pathogenesis of ascending urinary tract infection induced by Escherichia coli in mice. Infect. Immun. 39:1307-1315. [PMC free article] [PubMed]
20. Jarvis, W. R., and W. J. Martone. 1992. Predominant pathogens in hospital infections. J. Antimicrob. Chemother. 29(Suppl. A):19-24. [PubMed]
21. Jesaitis, A. J., M. J. Franklin, D. Berglund, M. Sasaki, C. I. Lord, J. B. Bleazard, J. E. Duffy, H. Beyenal, and Z. Lewandowski. 2003. Compromised host defense on Pseudomonas aeruginosa biofilms: characterization of neutrophil and biofilm interactions. J. Immunol. 171:4329-4339. [PubMed]
22. Johnson, J. R. 1991. Virulence factors in Escherichia coli urinary tract infection. Clin. Microbiol. Rev. 4:80-128. [PMC free article] [PubMed]
23. Johnson, J. R., N. Kaster, M. A. Kuskowski, and G. V. Ling. 2003. Identification of urovirulence traits in Escherichia coli by comparison of urinary and rectal E. coli isolates from dogs with urinary tract infection. J. Clin. Microbiol. 41:337-345. [PMC free article] [PubMed]
24. Johnson, J. R., S. J. Weissman, A. L. Stell, E. Trintchina, D. E. Dykhuizen, and E. V. Sokurenko. 2001. Clonal and pathotypic analysis of archetypal Escherichia coli cystitis isolate NU14. J. Infect. Dis. 184:1556-1565. [PubMed]
25. Keith, B. R., L. Maurer, P. A. Spears, and P. E. Orndorff. 1986. Receptor-binding function of type 1 pili effects bladder colonization by a clinical isolate of Escherichia coli. Infect. Immun. 53:693-696. [PMC free article] [PubMed]
26. Klausen, M., A. Aaes-Jorgensen, S. Molin, and T. Tolker-Nielsen. 2003. Involvement of bacterial migration in the development of complex multicellular structures in Pseudomonas aeruginosa biofilms. Mol. Microbiol. 50:61-68. [PubMed]
27. Klausen, M., A. Heydorn, P. Ragas, L. Lambertsen, A. Aaes-Jorgensen, S. Molin, and T. Tolker-Nielsen. 2003. Biofilm formation by Pseudomonas aeruginosa wild type, flagella and type IV pili mutants. Mol. Microbiol. 48:1511-1524. [PubMed]
28. Kolari, M., U. Schmidt, E. Kuismanen, and M. S. Salkinoja-Salonen. 2002. Firm but slippery attachment of Deinococcus geothermalis. J. Bacteriol. 184:2473-2480. [PMC free article] [PubMed]
29. Kunin, C. M., S. Douthitt, J. Dancing, J. Anderson, and M. Moeschberger. 1992. The association between the use of urinary catheters and morbidity and mortality among elderly patients in nursing homes. Am. J. Epidemiol. 135:291-301. [PubMed]
30. Langermann, S., R. Mollby, J. E. Burlein, S. R. Palaszynski, C. G. Auguste, A. DeFusco, R. Strouse, M. A. Schenerman, S. J. Hultgren, J. S. Pinkner, J. Winberg, L. Guldevall, M. Soderhall, K. Ishikawa, S. Normark, and S. Koenig. 2000. Vaccination with FimH adhesin protects cynomolgus monkeys from colonization and infection by uropathogenic Escherichia coli. J. Infect. Dis. 181:774-778. [PubMed]
31. Langermann, S., S. Palaszynski, M. Barnhart, G. Auguste, J. S. Pinkner, J. Burlein, P. Barren, S. Koenig, S. Leath, C. H. Jones, and S. J. Hultgren. 1997. Prevention of mucosal Escherichia coli infection by FimH-adhesin-based systemic vaccination. Science 276:607-611. [PubMed]
32. Leid, J. G., M. E. Shirtliff, J. W. Costerton, and A. P. Stoodley. 2002. Human leukocytes adhere to, penetrate, and respond to Staphylococcus aureus biofilms. Infect. Immun. 70:6339-6345. [PMC free article] [PubMed]
33. Maki, D. G., and P. A. Tambyah. 2001. Engineering out the risk for infection with urinary catheters. Emerg. Infect. Dis. 7:342-347. [PMC free article] [PubMed]
34. Mittelman, M. 1995. Adhesion to biomaterials, p. 89-127. In M. Fletcher (ed.), Bacterial adhesion: molecular and ecological diversity. Wiley-Liss, Inc, New York, NY.
35. Moshier, A., M. S. Reddy, and F. A. Scannapieco. 1996. Role of type 1 fimbriae in the adhesion of Escherichia coli to salivary mucin and secretory immunoglobulin A. Curr. Microbiol. 33:200-208. [PubMed]
36. Nilsson, L. M., W. E. Thomas, E. V. Sokurenko, and V. Vogel. 2006. Elevated shear stress protects Escherichia coli cells adhering to surfaces via catch bonds from detachment by soluble inhibitors. Appl. Environ. Microbiol. 72:3005-3010. [PMC free article] [PubMed]
37. Nilsson, L. M., W. E. Thomas, E. Trintchina, V. Vogel, and E. V. Sokurenko. 2006. Catch bond-mediated adhesion without a shear threshold: trimannose versus monomannose interactions with the FimH adhesin of Escherichia coli. J. Biol. Chem. 281:16656-166663. [PubMed]
38. Ofek, I., D. L. Hasty, and N. Sharon. 2003. Anti-adhesion therapy of bacterial diseases: prospects and problems. FEMS Immunol. Med. Microbiol. 38:181-191. [PubMed]
39. O'Toole, G. A., and R. Kolter. 1998. Flagellar and twitching motility are necessary for Pseudomonas aeruginosa biofilm development. Mol. Microbiol. 30:295-304. [PubMed]
40. Platt, R., B. F. Polk, B. Murdock, and B. Rosner. 1982. Mortality associated with nosocomial urinary-tract infection. N. Engl. J. Med. 307:637-642. [PubMed]
41. Pouttu, R., T. Puustinen, R. Virkola, J. Hacker, P. Klemm, and T. K. Korhonen. 1999. Amino acid residue Ala-62 in the FimH fimbrial adhesin is critical for the adhesiveness of meningitis-associated Escherichia coli to collagens. Mol. Microbiol. 31:1747-1757. [PubMed]
42. Pratt, L. A., and R. Kolter. 1998. Genetic analysis of Escherichia coli biofilm formation: roles of flagella, motility, chemotaxis and type I pili. Mol. Microbiol. 30:285-293. [PubMed]
43. Rupp, C. J., C. A. Fux, and P. Stoodley. 2005. Viscoelasticity of Staphylococcus aureus biofilms in response to fluid shear allows resistance to detachment and facilitates rolling migration. Appl. Environ. Microbiol. 71:2175-2178. [PMC free article] [PubMed]
44. Sauer, K., A. K. Camper, G. D. Ehrlich, J. W. Costerton, and D. G. Davies. 2002. Pseudomonas aeruginosa displays multiple phenotypes during development as a biofilm. J. Bacteriol. 184:1140-1154. [PMC free article] [PubMed]
45. Schembri, M. A., M. Givskov, and P. Klemm. 2002. An attractive surface: gram-negative bacterial biofilms. Sci. STKE 2002:RE6. [PubMed]
46. Schembri, M. A., and P. Klemm. 2001. Biofilm formation in a hydrodynamic environment by novel FimH variants and ramifications for virulence. Infect. Immun. 69:1322-1328. [PMC free article] [PubMed]
47. Serafini-Cessi, F., N. Malagolini, and D. Cavallone. 2003. Tamm-Horsfall glycoprotein: biology and clinical relevance. Am. J. Kidney Dis. 42:658-676. [PubMed]
48. Sokurenko, E. V., V. Chesnokova, R. J. Doyle, and D. L. Hasty. 1997. Diversity of the Escherichia coli type 1 fimbrial lectin. Differential binding to mannosides and uroepithelial cells. J. Biol. Chem. 272:17880-17886. [PubMed]
49. Sokurenko, E. V., V. Chesnokova, D. E. Dykhuizen, I. Ofek, X. R. Wu, K. A. Krogfelt, C. Struve, M. A. Schembri, and D. L. Hasty. 1998. Pathogenic adaptation of Escherichia coli by natural variation of the FimH adhesin. Proc. Natl. Acad. Sci. USA 95:8922-8926. [PMC free article] [PubMed]
50. Sokurenko, E. V., M. Feldgarden, E. Trintchina, S. J. Weissman, S. Avagyan, S. Chattopadhyay, J. R. Johnson, and D. E. Dykhuizen. 2004. Selection footprint in the FimH adhesin shows pathoadaptive niche differentiation in Escherichia coli. Mol. Biol. Evol. 21:1373-1383. [PubMed]
51. Steinberg, D., S. Poran, and L. Shapira. 1999. The effect of extracellular polysaccharides from Streptococcus mutans on the bactericidal activity of human neutrophils. Arch. Oral Biol. 44:437-444. [PubMed]
52. Stewart, P. S., and J. W. Costerton. 2001. Antibiotic resistance of bacteria in biofilms. Lancet 358:135-138. [PubMed]
53. Thomas, W. E., M. Forero, O. Yakovenko, L. Nilsson, P. Vicini, E. V. Sokurenko, and V. Vogel. 2006. Catch bond model derived from allostery explains force-activated bacterial adhesion. Biophys. J. 90:753-764. [PMC free article] [PubMed]
54. Thomas, W. E., L. Nilsson, M. Forero, E. V. Sokurenko, and V. Vogel. 2004. ‘Stick-and-roll’ bacterial adhesion mediated by catch-bonds. Mol. Microbiol. 53:1545. [PubMed]
55. Thomas, W. E., E. Trintchina, M. Forero, V. Vogel, and E. V. Sokurenko. 2002. Bacterial adhesion to target cells enhanced by shear force. Cell 109:913-923. [PubMed]
56. Thormann, K. M., R. M. Saville, S. Shukla, D. A. Pelletier, and A. M. Spormann. 2004. Initial phases of biofilm formation in Shewanella oneidensis MR-1. J. Bacteriol. 186:8096-8104. [PMC free article] [PubMed]
57. Weissman, S. J., S. Chattopadhyay, P. Aprikian, M. Obata-Yasuoka, Y. Yarova-Yarovaya, A. Stapleton, W. Ba-Thein, D. Dykhuizen, J. R. Johnson, and E. V. Sokurenko. 2006. Clonal analysis reveals high rate of structural mutations in fimbrial adhesins of extraintestinal pathogenic Escherichia coli. Mol. Microbiol. 59:975-988. [PMC free article] [PubMed]
58. Wood, T. K., A. F. Gonzalez Barrios, M. Herzberg, and J. Lee. 2006. Motility influences biofilm architecture in Escherichia coli. Appl. Microbiol. Biotechnol. 72:361-367. [PubMed]
59. Zhu, C., J. Lou, and R. P. McEver. 2005. Catch bonds: physical models, structural bases, biological function and rheological relevance. Biorheology 42:443-462. [PubMed]

Articles from Journal of Bacteriology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...