A polyurethane foam support (1 cm³ cube; CFH-30) was obtained from Inoac Corporation (Aichi, Japan). Polystyrene plates (PS2035-1), glass plates (FF-001), stainless steel plates (SUS430 grade; EA441WA-21), PTFE plates (J1-537-01), mica disks (V-1 grade), and square glass tubes (VitroTubes, 8100) were purchased from Hikari Co., Ltd. (Osaka, Japan), Matsunami Glass Ind., Ltd. (Osaka, Japan), ESCO Co., Ltd. (Osaka, Japan), AS ONE Corp. (Osaka, Japan), TED PELLA, Inc. (CA, United States), and Vitrocom (Mountain Lakes, NJ), respectively. A poly (mOEGMA) brush surface was prepared on glass plates as described previously (Azuma et al., 2017). The MPC polymer-coated surface for the adhesion assays was prepared as described below. A glass plate was dipped into 2% MPC polymer solution (Lipidure-CM5206, a copolymer of MPC and butyl methacrylate; NOF Corp., Tokyo, Japan) dissolved in ethanol and shaken at 70 rpm for 3 min. After shaking, the glass plate was rinsed in pure water and dried at 60°C for 3 h. The MPC polymer-coated glass tube was prepared as described below using the flow chamber system. A solution of 2% MPC polymer dissolved in ethanol was flushed through a square glass tube (50 mm in length and 1 mm in every internal dimension) at 10 cm/min for 3 min, and then pure water was flushed through at 50 cm/min for 2 min. After passing air through at 50 cm/min for 2 min, the glass tube was removed from the flow chamber system and dried at 60°C for 3 h.

AFM colloidal probes modified with silica (CP-PNPL-SIO-A), polystyrene (CP-PNPL-PS-A), and gold (CP-PNPL-Au-A) colloids 2 μm in diameter were purchased from NanoAndMore (Wetzlar, Germany). A poly (mOEGMA) brush-modified colloidal probe was prepared by grafting the poly (mOEGMA) brush onto the silica-colloidal probe as described previously (Azuma et al., 2017). An MPC polymer-coated colloidal probe was prepared by immersing the silica-colloidal probe in 0.2% MPC polymer solution and incubated for 1 h at room temperature. After incubation, the MPC polymer-coated colloidal probe was rinsed in pure water.

Microwells used in the adhesion assays with various material surfaces were constructed in our laboratory as follows: Four sheets of vinyl electrical tape (ASKUL Corporation, Tokyo, Japan) were stacked, and holes with a diameter of 6 mm were punched into the stacked tape. The punched tape was placed on a material plate as shown in Supplementary Figure S1.

Prior to use, glass was washed with piranha solution (H₂SO₄:30% H₂O₂ = 3:1) followed by rinsing with pure water. Mica was cleaved with scotch tape. The MPC polymer-coated surface was rinsed with pure water, and the other surfaces were rinsed with pure ethanol.

The static contact angles (SCAs) of air bubbles in water on each material surface were measured with a contact angle meter (CA-W, Kyowa Interface Science Co., Tokyo, Japan) at room temperature and are listed in Supplementary Table S1. Material substrates were immersed in water, and 10 μL of air bubbles were placed on the substrates.

The bacterial strains used in this study are listed in Supplementary Table S2. These bacterial strains were grown as previously described (Ishikawa et al., 2012a; Lu et al., 2013). Acinetobacter strains were grown in Luria-Bertani (LB) medium at 28°C for 8 h. An overnight culture of Yersinia enterocolitica strains grown in LB medium at 28°C was used to inoculate the LB medium at a 1:100 dilution, and the medium was incubated at 37°C for 6 h to induce yadA expression (El Tahir and Skurnik, 2001). B. henselae strains were grown for 5 days on Columbia agar supplemented with 5% sheep blood at 37°C in a humidified atmosphere with 5% CO₂. The expression of trimeric autotransporter adhesins in each strain was confirmed by Western blotting using anti-AtaA_699–1014 antiserum, anti-BadA antibodies (Riess et al., 2004), or anti-YadA antibodies (sc-22472; Santa Cruz Biotechnology, Inc., Dallas, TX, United States).

The polyurethane foam support adhesion assay was performed as previously described (Hori et al., 2015), with slight modifications. Bacterial cells were suspended in BS-N buffer (34.5 mM Na₂HPO₄, 14.7 mM KH₂PO₄, 15.5 mM K₂SO₄, pH 7.2), and the optical density of the cell suspension at 660 nm (OD₆₆₀) was adjusted to 1.0. Four pieces of the polyurethane foam support were placed into 20 mL of the cell suspension and shaken at 115 rpm at 28°C. After a 30-min incubation with shaking, the transparency of the cell suspension was observed and photographed by a digital camera.

Bacterial cells were suspended in BS-N buffer, and the OD₆₆₀ was adjusted to 0.5. The cell suspensions (50 μL each) were placed into a microwell with the materials and incubated at 28°C for 10 min. When the antiadhesive surfaces were used, the incubation time was extended to 30 min and 2 h. The cell suspensions were removed after incubation using a pipet, and the wells were washed for 10 s by shaking in BS-N buffer at 70 rpm. Cells adhering to the well were stained with 50 μL of 0.1% crystal violet solution for 15 min and washed for 10 s by shaking in BS-N buffer at 70 rpm. Finally, the stain was eluted with 200 μL of 70% ethanol, and the absorbance of the elution at 590 nm (A₅₉₀) was measured with a microplate reader (ARVO X3; PerkinElmer, Inc., MA, United States).

Bacterial cells were immobilized on a glass bottom dish (FluoroDish, FD5040-100; World Precision Ins., Sarasota, FL, United States) as previously described (Ishii et al., 2022). Single-cell force spectroscopy was performed using a NanoWizard 3 BioScience AFM system (JPK Ins., Berlin, Germany) in contact mode at room temperature in BS-N buffer. For measurement, colloidal probes with a diameter of 2 μm and spring constants of 0.06–0.09 N/m were used. The spring constants of the cantilevers were measured using the thermal noise method. The parameters used in the measurement were as follows: Z-length: 3 μm; applied force: 0.2 nN; speed: 1 μm/s; cell surface dwell time: 0.1 s; and sample rate: 5,000 Hz.

A rectangular flow chamber system, which we have previously reported (Furuichi et al., 2018), was used with slight modifications: a syringe pump (Legato 200; KD Scientific, Holliston, MA) was directly connected to a square glass tube 50 mm in length and 1 mm in every internal dimension without using a three-way stopcock valve. Tol 5 cells were suspended in BS-N buffer at an OD₆₆₀ of 0.2, and the suspension was subjected to sonication to break up the cell clumps. The glass tube with or without an MPC polymer coating was filled with the cell suspension and incubated for 10 min at room temperature. The suspension was replaced with fresh BS-N buffer by flowing slowly at 1 cm/min for 35 min, and the fluid velocity was increased stepwise. Live images of the behavior of Tol 5 cells during their adhesion and detachment to/from the inner surface of the glass tube were recorded under a digital microscope (VHX-200; Keyence, Osaka, Japan). Wall shear stress (τ) was calculated using Eq. 1, as previously described (Furuichi et al., 2018): τ = 8 μ U D (1) where μ, U, and D are the fluid viscosity, fluid velocity, and diameter of the glass tube, respectively. The equivalent diameter (De) was used for the value of D in Eq. 1, and was calculated according to Eq. 2: De = 4 A P (2) where A and P are the sectional area and wetted perimeter, respectively.

First, we compared the adhesiveness of Tol 5 and its ΔataA mutant with that of Y. enterocolitica WA-314 and B. henselae Marseille by shaking each cell suspension in the presence of a polyurethane support for 30 min. These Gram-negative bacteria have also been reported to adhere to abiotic surfaces through the peritrichate fibers of their TAAs (Muller et al., 2011), YadA and BadA, respectively. YadA is a well-studied short TAA consisting of 422 amino acids (GenBank: CBW54734.1). BadA (3,973 amino acids) (GenBank: MK993576.1) is slightly larger than AtaA, and it has also been reported to be more adhesive than other TAAs. It is worth noting that all of these TAAs show sequence and length variations among different strains and isolates (Muhlenkamp et al., 2017; Thibau et al., 2022). The expression of these TAAs was confirmed by Western blotting (see Supplementary Figure S2). In adhesion assays in shaking flasks with a polyurethane support, Tol 5 cells expressing AtaA showed overwhelmingly high adhesiveness compared with that of cells expressing other TAAs (Figure 1). Most of the Tol 5 cells adhered to the polyurethane support during shaking, and consequently, the Tol 5 cell suspension appeared clear. In contrast, the cell suspensions of Y. enterocolitica, B. henselae, and the Tol 5 ΔataA mutant remained cloudy, which indicated that many of the cells did not adhere to the polyurethane support.

Next, we measured the adhesion of bacterial cells expressing TAAs to the surface made from various materials (Table 1) using self-made microwells (Supplementary Figure S1). Cell suspensions were incubated on polystyrene (PS), glass, stainless steel, and polytetrafluoroethylene (PTFE, known as Teflon) surfaces for only 10 min. Nonadhering cells were removed by washing with fresh medium, and the instantly adhered cells on the material surface were quantified by crystal violet staining. Even in this short time, Tol 5 adhered not only to PS, glass, and stainless steel but also to PTFE, which has antiadhesive properties derived from its low surface energy (Figure 2) (Pereni et al., 2006). In contrast, the Tol 5 ΔataA mutant and Y. enterocolitica hardly adhered to any of the material surfaces. Although B. henselae showed measurable adhesiveness, the amount of adhered cells was much smaller than that of Tol 5. These results quantitatively demonstrated that Tol 5 cells exhibit markedly higher adhesiveness to various material surfaces through the action of AtaA than bacterial cells expressing other TAAs.

To investigate whether Tol 5 cells can adhere to various other antiadhesive surfaces in addition to PTFE (Table 1), we measured cell adhesion to mica, poly (oligo (ethylene glycol) methyl ether methacrylate) (poly (mOEGMA)) brush, and 2-methacryloyloxyethyl phosphorylcholine (MPC) polymer surfaces, using the self-made microwells. Mica is a phyllosilicate mineral of aluminum and potassium, and its cleaved surface is atomically flat (de Poel et al., 2014). A poly (mOEGMA) brush is a neutral hydrophilic polymer brush and exerts steric hindrance (Sin et al., 2014). An MPC polymer is a zwitterionic hydrophilic polymer that has a high free water fraction (Ishihara et al., 1998). The surfaces of these materials have been reported to have antiadhesive properties against bacterial cells (Cheng et al., 2007; Zeng et al., 2014; Kameda et al., 2019). During the incubation of bacterial cells on the antiadhesive surfaces for 10 min, Tol 5 cells quickly adhered to the surface of PTFE but not to that of mica, the poly (mOEGMA) brush, or the MPC polymer (Figure 3A). However, after incubation for 2 h, Tol 5 cells also adhered to mica and the poly (mOEGMA) brush surfaces. Nevertheless, Tol 5 cells could hardly adhere to the MPC polymer surface after the 2 h incubation. In contrast, B. henselae adhered to the surfaces of PTFE and mica but not to that of the poly (mOEGMA) brush and MPC polymer even after a 2-h incubation (Figure 3B). These results emphasized that Tol 5 cells were the only cells that adhered to the poly (mOEGMA) brush surface and showed that even sticky Tol 5 cells hardly adhered to the MPC polymer surface under these experimental conditions.

In contrast to semiquantitative adhesion assays that measure the amount of cells adhering to microwells by staining, AFM can quantify the strength of adhesion as the adhesion force in Newtons, i.e., the resistance force against peeling (Dufrene, 2015; Elbourne et al., 2019). We previously reported that Acinetobacter sp. Tol 5 exhibits an adhesion force of 2 nN to a sharp silicon nitride AFM probe, which is more than 10-fold stronger than that of the B. henselae strain (Ishii et al., 2022). To quantify the strength of adhesion of Tol 5 to various materials, including antiadhesive surfaces, compared to that of the B. henselae strain, which showed some adhesiveness to PTFE and mica, we measured the adhesion force of a single cell using AFM. To ensure that the adhesion force of a single cell was measured, a single colloidal probe made from PS, SiO₂, gold, poly (mOEGMA) brush-coated SiO₂, or MPC polymer-coated SiO₂ (Table 1) was pressed against the top of a single bacterial cell immobilized on a glass substrate beforehand through a covalent bond between the bottom side of the cell surface and a glass surface, and the force required to pull the probe away from the cell was measured (Figure 4A). Colloids of the same diameter were used for the measurements to equalize the contact area between cells and probes of different materials. This single-cell force spectroscopy demonstrated that Tol 5 exhibited a strong adhesion force over a long distance of 500–1,000 nm (Figures 4B–E). The median of the maximum adhesion forces of Tol 5 to PS, SiO₂, gold, and even the poly (mOEGMA) brush was 2.5 nN, which was much stronger than that of B. henselae (≤1 nN) (Figure 4G). The adhesion work of Tol 5, which is the energy required for detachment and is shown as the peak area of the adhesion force curve, was more than 10 times larger than that of B. henselae (Figure 4H). On the other hand, both Tol 5 and B. henselae showed only faintly weak adhesion force and small adhesion work when adhered to the MPC polymer (Figures 4F–H). These results highlighted the remarkably strong nonspecific adhesion of Tol 5 to PS, SiO₂, gold, and even the poly (mOEGMA) brush, but not to the MPC polymer, demonstrating the remarkable antiadhesive property of the MPC polymer.

To investigate the behavior of Tol 5 cells on the MPC polymer surface in the presence of shear forces, we observed Tol 5 cells on the surface by using a flow chamber system with a square glass tube (Figure 5A) (Furuichi et al., 2018). The glass tube with or without the MPC polymer coating was filled with a Tol 5 cell suspension and incubated for 10 min. Then, the cell suspension was replaced with fresh BS-N buffer by flowing slowly at 1 cm/min for rinsing, and the fluid velocity was increased stepwise, as shown in Figure 5B, while observing the inner surface of the bottom of the glass tube under a microscope. Unexpectedly, Tol 5 cells adhered to the MPC polymer-coated glass as much as to the bare (noncoated) glass under static conditions and remained adhered after rinsing at 1 cm/min (Figure 5C). When the fluid velocity was increased to 2 cm/min, a small fraction of previously adhered cell clumps started to move and slip on the surface (see Supplementary Movie S1), but many cells still resisted detachment after 10 min of flow (Figure 5C, 2 cm/min). At a high fluid velocity of 5 cm/min or more, the Tol 5 cells firmly adhered to the bare glass, whereas the cells attached to the MPC polymer were exfoliated, rolled, and washed away from the surface by shear stress (≥5.94 mN/m²) (Figure 5C, ≥5 cm/min; see Supplementary Movie S1).