We used a p-type silicon wafer [Si(100), Thickness and diameter of the wafers are 525 ± 25 μm and 100.0 ± 0.5 mm, respectively, Furuuchi Chemical Co., Japan] cut into pieces with a size of 10 × 10 mm² as the substrates. The silicon substrates were washed by ultrasonic cleaning in acetone, ethanol, and pure water, followed by a nitrogen gas blow.

Colloidal probes were prepared by attaching a washed silica bead (diameter: 18 μm, Duke Sci. Corp., CA, United States) to the end of a tipless cantilever (NP-OW, Bruker, MA, United States) using epoxy glue (Araldite, Huntsman Corp., UT, United States). The nominal spring constant of the cantilevers was 0.06 N m⁻¹. After attaching the silica bead, the probes were kept for 12 h at room temperature to fix the bead at the cantilever. Then, the colloidal probes and the silicon substrates were cleaned with a UV-O₃ cleaner (UV-300, SUMCO, Japan) for 15 min. First, the colloidal probes and substrates were coated with a wetting layer (adhesion promoter) of germanium (99.999%, Nilaco Corp., Japan) with a thickness of 5 nm by thermal evaporation under vacuum (base pressure 1.0 × 10^–5 Pa) at 350 K. Then, the germanium-coated probes and substrates were coated with gold (99.999%, Furuuchi Chemical Co., Japan) with a thickness of 100 nm at 350 K (Logeeswaran VJ et al., 2009). The deposition rates for germanium and gold were 0.01 and 0.1 nm s⁻¹, respectively.

The gold-coated probes and substrates were immersed in a phosphate-buffered saline (PBS) solution (pH 7.4 and ionic strength 167 mM) containing 0.14 mM peptide for 24 h to form covalent bonds between the thiol groups and gold. After the immersion, the probes and substrates were rinsed with pure water to remove excess adsorbed molecules. The peptide-SAMs were prepared for static water contact angle measurement, platelet adhesion test, and surface force measurement. The amino acid sequences of the peptide were summarized in Table 1.

X-ray photoelectron (XP) spectra in C1s and S2p were measured with a commercial XPS system (Theta Probe, Thermo Electron, United Kingdom). Al Kα radiation (photon energy of 1,486.6 eV and spot size of 400 µm) was used for the X-ray. All XP spectra were acquired at a take-off angle of 48° from the surface normal. The numbers of scans of the C1s and S2p spectra were 40 and 1,000, respectively.

Static water contact angles (WCA) were measured by the sessile drop method at room temperature (Model G-1-1,000, ERMA, Tokyo, Japan). 2 µl of the water droplet was placed on a peptide SAM, and the WCAs were recorded after 30 s. The measurement was carried out at five different points per substrate using three substrates for one peptide SAM and averaged (n = 15).

We employed quartz crystal microbalance with an energy dissipation system (QCM-D) (D300, Q-Sense, Sweden) to measure the amounts and viscoelasticity of fibrinogen adsorbed onto the peptide-SAMs. Firstly, gold-coated QCM sensors were cleaned with a UV-O₃ cleaner for 10 min, then immersed into a 1:1:5 solution of hydrogen peroxide (30%), ammonia (25%), and pure water heated to a temperature of 75 C for 5 min. Immediately rinse with pure water and blow of nitrogen gas. Then the gold-coated QCM sensors were cleaned with UV-O₃ cleaner for 10 min. The peptide-SAMs were prepared in the same manner for the gold-coated substrates. We used human fibrinogen (Biogenesis Ltd., United Kingdom) dissolved in PBS at a concentration of 1 mg mL⁻¹. In the QCM measurements, a measurement chamber was first filled with PBS. Then, the fibrinogen solution was injected into the chamber. After the sensor’s resonant frequency became constant, the PBS solution was injected into the chamber again for rinsing. The amount of the adsorbed fibrinogen was defined as the difference in weights before the injection and after rinsing. The conversion from resonant frequency to weight was based on the Sauerbrey equation shown as Eq. 1, Δ m = − C · Δ f n (1) where C = 17.7 ng cm⁻² Hz⁻¹, Δf is the change in the resonant frequency due to protein adsorption, and the n is the overtone number [(n = 3) in this work]. All data were collected at the n = 3 overtone. The energy dissipation factor D was measured by switching off the driving power and monitoring the amplitude decay profile. The amplitude decays as an exponentially damped sinusoidal function with a characteristic decay time (τ ₀). The decay time is related to the energy dissipation factor (D) shown as Eq. 2, D = 1 π f τ 0 (2) where f is the resonant frequency of the sensor. The viscoelasticity of the protein layer was compared with the value of Δd (change in energy dissipation)/Δf (change in resonant frequency). In this approach, a rigid layer yields a small value of Δd/Δf.

Blood was drawn from healthy volunteers and mixed with a 1/9 volume of acid citrate dextrose (ACD). Platelet-rich plasma (PRP) and platelet-poor plasma (PPP) were obtained by centrifugation of the blood at 1,500 rpm for 5 min and at 4,000 rpm for 10 min, respectively. The platelet suspension plasma containing 2 × 10⁵ cells µl⁻¹ of platelet was prepared by mixing the PRP with PPP. The platelet concentration was determined with a cell-counting hemato-cytometer (Neubauer chamber). Then, 200 µl of the plasma was placed on the peptide-SAMs and incubated for 60 min at 37 C. After the peptide-SAMs were washed with PBS solution twice, the peptide-SAMs were immersed in 1% glutaraldehyde in PBS solution overnight at 4 C to fix the adhered platelets. The fixed samples were immersed in PBS for 10 min, in a 1:1 mixture of PBS and pure water for 8 min, and in pure water for 8 min twice and dried in air overnight. Then the samples were sputter-coated using gold (JFC-1200, JEOL) prior to observation by scanning electron microscopy (VE-9800, KEYENCE, Japan). The number of adhered platelet cells was counted in five positions in the SEM images (100 × 80 μm²) using nine substrates for one peptide-SAM and averaged (n = 45).

All force curve measurements were performed with a commercial AFM system equipped with a liquid cell (MFP-3D, Oxford Instruments, United Kingdom). Spring constants of the colloidal probes were determined by monitoring the thermal fluctuations of the probes (Hutter and Bechhoefer, 1993). Velocity on approach and retraction of the probe was fixed at 200 nm s⁻¹. All force measurements were performed in PBS solution and 1 mM PB (pH 7.4 and ionic strength 2.32 mM) at room temperature. In this study, we simply defined a distance of zero as where linearity in the constant compliance region started in the force-distance curve. We measured 50 times force curves and averaged them.

Figure 1A shows the XP spectra in an S2p region. The spectra in the S2p region for all peptide-SAMs were deconvoluted into two peaks with an area ratio of 2:1 due to the spin-orbit coupling. The peaks at 162 eV and its satellite peak at 163.2 eV are assigned to the bound state of the sulfur (sulfur is covalently attached to the gold substrate) (Ishida et al., 1999). We did not observe any peaks assigned to the unbound state (usually observed at 163–164 eV), indicating that the peptide molecules are immobilized on gold substrates via a covalent bond and repeating units of the zwitterionic peptides face to the solution phase.

Figure 1B displays the XP spectra of the peptide-SAMs in the C1s region. The spectra are deconvoluted into four peaks. Peaks at about 284, 286, 287, and 289 eV can be assigned to C-(C, H), C-O, C-O next to O=C-O (carbonyl group), and carbonyl group, respectively (Pourcelle et al., 2007). Next, we compared the total peak area divided by the number of carbon atoms in the peptide molecules (Table 2). The results suggest that the peptide molecules are accumulated on the substrate and that the packing densities of DK and DR SAMs are higher than those of EK and ER SAMs. The difference may be attributed to the difference in the steric structure of the amino acid residues in the peptide molecules.

We also performed force-distance curve measurements using a probe with a sharp tip (tip radius is about 8 nm) (Supplementary Figure S1). The results showed that the thicknesses of the peptide-SAMs were around 2 nm, which corresponds to the Cα-to-Cα end-to-end distance of the EK (White et al., 2012). Taken together, the peptide molecules form a monolayer on gold substrates via covalent gold-sulfur bonds with directing the zwitterionic parts toward the solution phase.

Figure 2A shows the change in the resonant frequency and energy dissipation measured in the QCM measurements. ER and DR SAMs adsorbed fibrinogen with a significant increase in the dissipation rapidly after the injection of the protein solution, whereas we did not observe the adsorption for EK and DK SAMs without a drastic change in the energy dissipation. Figure 2B compares the amount of the adsorbed fibrinogen after rinsing, displaying the clear contrast in the degree of protein resistance, in good agreement with the report by Chen et al. (2009). In particular, the protein-resistance of EK and DK SAMs are comparable to that of the SAMs comprised of oligo (ethylene glycol)- or sulfobetaine-terminated alkanethiols, which are often widely used for the suppression of non-specific adsorption in biosensing (Hayashi et al., 2012; Tanaka et al., 2013; Sekine et al., 2015; Chang et al., 2018; Kwaria et al., 2020; Hayashi, 2021; Tanaka et al., 2021).

The Δd/Δf values also provide insight into the mechanical properties of the protein layer formed on the peptide-SAMs (Table 3). The Δd/Δf values for protein-resistant EK and DK are higher than those for protein-adsorbing ER and DR SAMs (Table 3). These results indicate that the interaction of fibrinogen with ER and DR SAMs is stronger than that with EK and DK SAMs and/or the formed protein layer is highly viscoelastic because of the denaturation of the protein molecules after adsorption.

The density of adhered platelet cells on the peptide-SAMs and the representative SEM images of the substrates are summarized in Figures 3A,B. Compared with the bare gold substrates, the adhesion and activation of platelets were suppressed on the peptide-SAMs. However, there was a distinct difference in their anti-platelet adhesion properties. Apparently, we can conclude that the platelet compatibility of EK and DK SAMs is higher compared with ER and DR SAMs.

In our platelet adhesion test, plasma proteins first accumulate on the peptide-SAMs and form a layer. Then, the protein layer interfaces the peptide-SAMs and platelet cells. The proteins’ composition, orientation, and conformation significantly affect the adhesion and activation of platelet cells (Hylton et al., 2005). There have been many works reporting a clear correlation between protein adsorption and platelet adhesion. In particular, adsorption tests of fibrinogen correlate well with platelet cells’ adhesion (Tanaka et al., 2000). Our results here are in good agreement with previous findings. The analysis of the protein composition by mass spectroscopy is underway (Hirohara et al., 2019).

Water contact angles (WCA) have often been an indicator of protein- and cell resistance for monolayer and polymer systems, although there are apparent exceptions (Alexander and Williams, 2017; Kwaria et al., 2020). We also investigated the surface wettability of peptide-SAMs by WCA on the peptide-SAMs (Table 4). The WCA showed that all peptide-SAMs are hydrophilic in terms of macroscopic wettability, indicating that strong interaction of fibrinogen with ER and DR SAMs is due not to hydrophobic interaction but to electrostatic interaction.

The main finding here is that the bioinertness of the zwitterionic peptide-SAMs cannot be explained with the macroscopic wettability of the peptide-SAMs. This has also been found for the SAMs of oligo (ethylene glycol)-terminated and several blood compatible polymers, including relatively hydrophobic poly (2-methoxyethyl acrylate) (PMEA) and hydrophilic poly (2-methacryloyloxyethyl phosphorylcholine) (PMPC) (Hayashi et al., 2007; Tanaka et al., 2013; Hayashi, 2021; Tanaka et al., 2021).

To understand the mechanism underlying the anti-biofouling property of EK and DK SAMs and the difference in the protein and cell resistance among the peptide-SAMs, we investigated the surface force induced by the peptide-SAMs. To clarify the physical origin of the surface forces, we performed surface force measurements in PBS solution, which is the solution condition used in protein adsorption and cell adhesion tests (Figure 4). The repulsive force was observed for all the peptide-SAMs, although the ranges of the repulsive force are different. As for ER and DR SAMs, the onset of the repulsion is at 4–5 nm of the separation. In contrast, the working distance of the repulsion for EK and DK SAMs is at around 8–9 nm. The repulsion’s decay lengths obtained by fitting the curves with a single exponential function are summarized in Table 5. The interaction in water is described as a sum of DLVO (electrostatic and van der Waals interaction) and non-DLVO (hydration force, steric repulsion, etc.) interactions. Among them, possible origins of the repulsive force are electrostatic double-layer force (DLVO), hydration force, and steric repulsion due to the deformation of the SAM after the physical contact of the peptide-SAMs (non-DLVO). Considering that the thicknesses of all the peptide-SAMs are about 2 nm (Supplementary Figure S1), the steric repulsion is not the origin of the difference in the repulsion observed in the force measurements.

Next, we examined the possibility of the electric double layer force for the difference in the repulsion by changing the ion concentration of the solution. We performed the measurements in phosphate buffer (PB) (1 mM) solution, whose Debye length is compared with PBS solution (Figure 5). With this solution condition, we observed the repulsion for all the peptide-SAMs, and the force curves can be fitted with the DLVO theory shown as Eq. 3, F ( D ) = − H s R 6 D 2 + 4 π R ϵ ϵ 0 ψ 2 λ D [ e x p ( − D λ D ) − e x p ( − 2 D λ D ) ] (3) where H _s , D, R, ε, ε ₀, ψ, and λ _D are the Hamaker constant for the interaction of peptide-SAMs across the water, the separation between the probe and surface, probe radius, relative dielectric constant, dielectric constant of vacuum, the surface potential of the surface and probe (peptide-SAMs), and the Debye length, respectively (Butt et al., 2005). The results of the fitting are shown in Figure 5 and Table 6.

The fitting results show that the calculated Debye lengths range between 5.4 and 6.7 nm. Compared with the decay length observed in PBS, the decay lengths of the repulsion are much longer compared with those in PBS (theoretical Debye length of PBS is 0.745 nm) and are in agreement with the theoretical value (6.32 nm). Therefore, we concluded that electrostatic double-layer force is dominant in PB solution. Unfortunately, we cannot determine the signs of the surface potentials because the same SAMs were formed on both probe and surface (Eq. (3)). By considering the strong basic character of arginine (R) (Fitch et al., 2015), we speculate that ER and DR SAMs are more positively charged than EK and DK SAMs.

The trend in the strength of the repulsion depending on the sequences in PB is different from that in PBS. Therefore, we consider that the repulsion observed for EK and DK SAMs is not electrostatic double-layer force. The decay lengths of the repulsion observed for EK and DK SAMs are far beyond the theoretical Debye length of PBS (0.745 nm) and cannot be explained by the electrostatic double-layer force. Combining the above discussions, the only possible origin for the repulsion is water-induced force.

To confirm that the origin of the repulsive force observed for EK and DK in PBS is water-induced force, we measured the interaction force in mixtures of PBS and ethanol. When water is mixed with ethanol, water and ethanol molecules are phase-separated at a low concentration of ethanol. When the concentration of ethanol increases, ethanol molecules are mixed with water molecules with disrupting three-dimensional networks of hydrogen bonding of water. At the ethanol’s molar ratio of 0.23, water and ethanol completely mix together (Matsumoto et al., 1995). As seen in Figure 6, The short-range repulsion in PBS disappeared in ethanol and PBS solution. The results indicated that water molecules near EK and DK possess the hydration barrier with a thickness of 3–4 nm (half of the working distance of the repulsion force) and prevent the approaches of protein molecules and cells, whereas ER and DR possess the hydration barrier with a thickness less than 2 nm and allowed their approaches.