Biomaterials 32 (2011) 5311e5319

Contents lists available at ScienceDirect

Biomaterials
journal homepage: www.elsevier.com/locate/biomaterials

Leading Opinion

A peptide e stainless steel reaction that yields a new bioorganic e metal state of matterq
Elisabeth M. Davis a, Dong-yang Li b, Randall T. Irvin a, *
a b

Department of Medical Microbiology and Immunology, University of Alberta, Edmonton, Alberta T6G 2H7, Canada Department of Chemical and Material Engineering, University of Alberta, Edmonton, Alberta T6G 2H7, Canada

a r t i c l e i n f o
Article history: Received 29 March 2011 Accepted 8 April 2011 Available online 7 May 2011 Keywords: Peptide Stainless steel Pilin receptor binding domain Electron work function XPS Adhesive force Corrosion

a b s t r a c t
A synthetic peptide derived from the native protein sequence of a metal binding bacterial pilus was observed to spontaneously react with stainless steel via a previously unreported type of chemical interaction to generate an altered form of stainless steel which we term bioorganic stainless steel. Bioorganic stainless steel has a signicantly increased electron work function (4.9 0.05 eV compared to 4.79 0.07 eV), decreased material adhesive force (19.4 8.8 nN compared to 56.7 10.5 nN), and is signicantly harder than regular 304 stainless steel (w40% harder). A formal or semi-formal organoemetallic covalent bond is generated between a pilin receptor binding domain and stainless steel based on XPS analysis which indicates that the electronic state of the surface is altered. Further, we establish that the peptideesteel reaction demonstrates a degree of stereospecicity as the reaction of native L-peptide, D-peptide and a retro-inverso-D-peptide yields bioorganic steel products that can be differentiated via the resulting EWF (4.867 0.008 eV, 4.651 0.008 eV, and 4.919 0.007 eV, respectively). We conclude that electron sharing between the peptide and steel surface results in the stabilization of surface electrons to generate bioorganic steel that displays altered properties relative to the initial starting material. The bioorganic steel generated from the retro-inverso-D-peptide yields a protease stable product that is harder (41% harder at a 400 mN load), and has a 50% lower corrosion rate compared with regular stainless steel (0.11 0.03 mpy and 0.22 0.04 mpy, respectively). Bioorganic steel is readily fabricated. 2011 Elsevier Ltd. All rights reserved.

1. Introduction Bacteria are amazingly efcient at binding to and growing on a vast array of surfaces, from human cells to stainless steel sinks and tubs, and surprisingly Pseudomonas aeruginosa (PA) utilizes a receptor binding domain (RBD) displayed at the tip of its type IV pili (T4P) to bind to both. T4P are powerful nanomotors that pull the bacteria along surfaces by retracting the pilus once it has bounddgenerating large forces on the RBDesurface interface. The PA RBD has an incredibly high apparent afnity for stainless steel [1,2] which cannot be readily understood given the high entropic penalty associated with a exible peptide interacting with an

q Editors Note: This paper is one of a newly instituted series of scientic articles that provide evidence-based scientic opinions on topical and important issues in biomaterials science. They have some features of an invited editorial but are based on scientic facts, and some features of a review paper, without attempting to be comprehensive. These papers have been commissioned by the Editor-in-Chief and reviewed for factual, scientic content by referees. * Corresponding author. Tel.: 1 780 492 5374; fax: 1 780 492 7521. E-mail address: Randy.Irvin@ualberta.ca (R.T. Irvin).
0142-9612/$ e see front matter 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2011.04.027

immobile surface and yet direct measurement of RBDesteel interaction with atomic force microscopy has conrmed a very strong molecular interaction [2,3]. PA is a human opportunistic pathogen that utilizes its T4P to adhere to and colonize a wide variety of biotic and abiotic surfaces. Binding is mediated by the RBD, a semi conserved 17-amino acid region that includes an intra-chain disulde loop encoded within the C-terminal region of the pilin structural protein, termed PilA, in PA T4P. The RBD has an extremely high afnity for stainless steel, exemplied by a Kiapparent of w0.2 nM for the inhibition of puried pili (thought to display 3 RBDs at the tip of the ber) and viable PA bacteria (each cell has a number of T4P located at the poles of the rod shaped cells) binding to steel, particularly when compared to the mM range of afnities previously reported for metal binding peptides [4], although recently a reasonably high afnity (116 nM) binding peptide for titanium was obtained by phage display [5]. The RBD binds directly to steel and this interaction does not appear to be mediated by hydrophobic interactions [2]. However, the RBD interaction with epithelial cells is of considerably lower afnity (w200,000 times lower) and the interaction with the minimal receptor of GalNAc-Gal appears to be mediated largely through hydrophobic interactions [6].

5312

E.M. Davis et al. / Biomaterials 32 (2011) 5311e5319 washed with dish detergent, thoroughly rinsed with distilled water, and immersed in 95% (v/v) ethanol for 15 min with gentle agitation. Coupons were then rinsed with distilled water, washed with 15 mL of acetone for 1 min, and rinsed with distilled water. They were placed in 6-well (1 coupon per well) tissue culture plates (Corning) and covered with 3 mL of sterile phosphate buffered saline (PBS)(pH 7.4) containing 10 mg/mL of K122-4(128-144)ox peptide and were incubated at room temperature (RT) for 1 h with gentle agitation. Samples were washed 6 times with distilled water and allowed to air dry. Aluminum (5 mm thick) was cut into 1 cm by 1 cm samples. Aluminum samples were cleaned, polished, and coated with K122-4(128-148)ox peptide. Aluminum was chosen for a control surface as the RBD did not appear to interact with the aluminum and hence was used as control for what the materials property of RBD deposited on a surface. For corrosion, steel coupons were annealed as described and encased in epoxy (LECO Corporation, Mississauga) prior to polishing. Once polished, samples were cleaned and coated as described. The EWF was determined using a scanning Kelvin probe [18] and three areas measuring 0.5 mm 0.5 mm were scanned per sample. Data was plotted as means in electron volts (eV) per area scanned. The electron work function (EWF) was measured using a scanning Kelvin probe (KP Techonology Ltd, UK) tted with an Au tip with a 1 mm diameter [18]. The probe was composed of three sub-systems controlled by a computer, including a digital oscillator, data acquisition, and sample translation. A three-axis microstep positioner allowed for precise position of the tip on the sample and controlled scanning steps of 0.4 mM per step. The oscillation frequency of the Kelvin probe was 173 Hz. Three separate areas of 0.5 mm by 0.5 mm were scanned on each sample to determine the electron activity, with a total of 100 reading per area scan. An AFM equipped with a silicon nitride tip was used to measure the adhesive force [3]. Fifty adhesive force measurements were obtained per sample. The adhesive force between a standard AFM silicon nitride tip (Veeco, CA) and a peptidecoated surface was measured using an atomic force microscope (AFM) (Hysitron, Minneapolis, USA). The AFM was used in contact mode. The tip was approached to the surface, allowed to made contact, and the total amount of deection of the cantilever when the tip is pulled away from the surface, which reects the adhesive force, was detected by laser beam. The related force can be quantitatively determined if the spring constant is known. The spring constant for the nitride tip was 0.06 N/m. Nanoindentation measurements were performed using a triboscope equipped with a diamond indenter. Samples were tested using loads of 50e800 mN and total displacement (in nm) of the tip into the sample was measured. A triboscope (Hysitron, Minneapolis, USA), a combination of a nanomechanical probe attachment and an AFM, was used to examine the changes in the mechanical properties of peptide-coated samples. The probe, a diamond pyramidal Vickers indenter, had a nominal radius of 150 nm with a force sensitivity of 100 nN and a displacement resolution of 0.2 nm. During nanoindentation a force-depth curve was obtained for each indentation and the total depth displacement of the tip into the surface of the sample was obtained from this curve. Five force-depth curves were obtained for each force load. Some samples were tested as to their corrosion properties. A minimum of three linear polarization and Tafel plot corrosion tests using a computerized scanning potentiostat were performed per sample to determine the polarization resistance and corrosion rate of peptide-coated samples. Electrochemical tests were performed using a computerized scanning potentiostat (Model PC4-750, Gamry). A saturated calomel electrode (SCE) and a platinum (Pt) foil were used as the counter and reference electrodes for all corrosion experiments. The electrolyte solution used was tap water and all experiments were performed at room temperature. At least three separate linear polarization corrosion tests were performed to measure the polarization resistance Rp, starting at 0.02 V below the open circuit potential (OCP) and ending 0.02 V above the OCP and using a scanning rate of 0.125 mV/s. A minimum of three Tafel plot corrosion tests were performed. A scanning rate of 1 mV/s was used and scans began from 0.25 V below the OCP and ended 0.25 V above the OCP. The corrosion rate and the polarization resistance were obtained using the SterneGeary equation. Surface electron activity was examined via XPS analysis. XPS analysis was performed using an AXIS-165 spectrometer (Kratos Analytical). The base pressure in the analytical chamber was lower than 3 108 Pa. Monochromatic Al Ka source (hn 1486.6 eV) was used at a power of 210 W. The analysis spot was 400 700 um. The resolution of the instrument is 0.55 eV for Ag 3d and 0.70 eV for Au 4f peaks. Survey scans were collected for binding energy from 1100 eV to 0 with analyzer pass-energy of 160 eV and a step of 0.35 eV. For the high-resolution spectra the passenergy was 20 eV with a step between 0.1 and 0.15 eV. No charge neutralization was required. In order to increase the surface sensitivity the samples were tilted so that the angle between the sample surface and the analyzer was 30 . No sample etching was performed prior to analysis. Data analysis was done using Casa XPS software, version 2.3.13. We examined the surface morphology samples by standard SEM and AUGER analysis methods. SEM and Auger data were collected with a eld-emission Auger microprobe JAMP-9500F (JEOL). The primary electron beam energy was 10 kV, and the probing beam current was 7 nA. The lateral resolutions for SEM and AES are 3.0

Stainless steel is polycrystalline in nature, with individual crystals being termed grains while the interfacial region located between adjoining crystals is termed the grain boundary. Grain boundaries are regions of the metal where there are signicant crystal defects and various types of dislocations. Electrons in these regions of imperfections are at higher energy levels, rendering the regions more reactive and more capable of interacting with other materials beyond the metal [7]. The grain boundary material surface properties differ from the surface properties observed within a grain [8,9]. PA binds preferentially to grain boundaries [1,10] and the force required to break the RBD-steel interaction is considerably higher (w2.2 fold) at grain boundaries than within the grains [3] suggesting that surface electron activity or the differential material composition of the grain associated with the grain boundary contributed signicantly to the interaction. Generally, bacterial binding to abiotic surfaces is thought to involve the interaction of the cells with a conditioning lm of adsorbed organic material bound to the surface. Binding is mediated through hydrophobic interactions due to energy contributions from the bulk solvent being excluded from the site of interaction [11,12]. However, the RBDesteel interaction did not require bulk solvent, and did not appear to involve a conditioning lm [2,3,13]. The surface free energy of stainless steel can be signicantly altered by regulating the grain size [3]. Nano-structured (steel with small but dened grain sizes) and of specic surface electron activities or surface free energies as determined by EWF and adhesive force measurements, was employed to demonstrate that the strength of the RBDesteel interaction is directly proportional to the surface electron activity [13]. Thus the mechanism of interaction(s) that mediate RBD binding to biotic surfaces differs substantially from that which mediates binding to stainless steel surfaces, even though the RBD is the same in both cases. While our observations concerning the RBDesteel interaction appeared to have a sound materials basis and reected the materials property of the steel surface, the molecular basis for the RBD interaction with stainless steel was puzzling from a biological point of view as most biologically relevant ligandereceptor interactions are driven by the energetics of the hydrophobic effect. Given that the afnity of the RBD for steel was extremely high (beyond what would be anticipated for a exible peptide binding to a solid surface where an entropic penalty would be encountered), that the interaction with steel did not require a conditioning lm, that hydrophobic interactions did not appear to contribute to the interaction, and that the strength of interaction was directly proportional to surface electron activity or surface free energy, we hypothesized that the RBD may form a semi-formal or formal covalent bond with the stainless steel surface to delocalize surface electrons through the RBD and thus investigated the nature of the RBDesteel interaction. We report that the PA RBD forms a previously unreported chemical interaction with steel that alters surface electron orbitals of the surface layer of stainless steel to form a new material which we term bioorganic K122-4 stainless steel (borg-K122SS).
2. Materials and methods K122-4(128-144)ox peptide was synthesized by solid phase peptide synthesis and puried by reversed-phase high-performance liquid chromatography (HPLC) as described by Wong et al. [14,15]. The disulphide bridge form of the peptide used in this study was generated by air-oxidization [16]. Biotinylation of the peptide was performed as previously described by Yu et al. [17]. Peptides were synthesized by the Peptide and Protein Chemistry Core Facility of the University of Colorado Health Sciences Center at Fitzsimons on a fee for service basis. The ability of biotinylated pili and peptide to bind to stainless steel was demonstrated by Giltner et al. [1]. Grade 304 stainless steel 2B nish plates (1 mm thick, 20 gauge steel) were cut into 1 cm 1 cm coupons and annealed at 1040  C for 1 h. The surface was polished to uniformity using sandpaper of increasing grit size (MetTech Inc, Calgary), from 120# to 1200# grit and an aqueous slurry of 0.05 mm colloidal silica. Coupons were

E.M. Davis et al. / Biomaterials 32 (2011) 5311e5319

5313

Fig. 1. Changes in surface electron reactivity, adhesive force and hardness of borg-K122SS steel compared with 304 stainless steel and aluminum with and without K122-4 (128-144)ox peptide. A) Electron work function measurements of borg-K122SS and 304 steel performed using a Kelvin probe. Three areas (1 mm 1 mm) were scanned. A total of 100 measurements were taken per area scanned and means were plotted. B) Electron work function measurements of aluminum with and without exposure to K122-4 (128-144) ox peptide. C) Adhesive force measurements on borg-K122SS and 304 stainless steel performed with an atomic force microscope (AFM) equipped with a non-conductive silicon nitride cantilever. Fifty force measurements were taken per sample. D) Adhesive force measurements on aluminum with and without exposure to K122-4 (128-144)ox peptide. E) Nanoindentation measurements of borg-K122SS and 304 steel. Five displacement measurements were taken per load. F) Nanoindentation measurements of aluminum with and without exposure to K122-4 (128-144)ox peptide.

5314

E.M. Davis et al. / Biomaterials 32 (2011) 5311e5319 and 8 nm, respectively. The sample was tilted 30 away from the primary beam toward the axis of the electron analyzer. The SEM images were used to locate positions for collecting the Auger energy spectra, line proles, and images. Auto probe tracking was in effect during the imaging to eliminate the drifting due to instabilities in power, temperature, etc. The intensity in the image and line prole distributions was calculated as (P B)/B to remove edge effects, P being the peak intensity and B being the background intensity. The images are presented in a thermal scale, the brighter area corresponding to the highest intensity. As receptor binding domain contains potential trypsin cleave sites, we examined the ability of trypsin to modify steel surfaces. Grade 304 stainless steel 2B nish plates (20 gauge, 1 mm thick) were cleaned as described previously [1] and assembled into a Schleicher and Schuell MinifoldTM System (Mandel Scientic). Synthetic biotinylated PAK(128-144)ox peptide were added at a concentration of 10 mg/mL (50 mL per well in replicates of ve) to the stainless steel manifold and incubated at RT for 1 h with gentle agitation. The manifold was washed six times with 250 mL per well of phosphate buffered saline (PBS, pH 7.4). To test whether the peptide was sensitive to trypsin, 1 mg/mL of trypsin (100 mL per well, 5 wells) was added and the manifold was incubated for 1 h at 37  C with gentle agitation. The manifold was washed six times with PBS and remaining bound peptide was assessed using streptavidin-horseradish peroxidase (HRP, Sigma) diluted 1:5000 in PBS. Following a 1 h incubation at RT, the manifold was washed with PBS and substrate buffer (0.01 M sodium citrate, pH 4.2 containing 1 M 2,20 -Azino-bis-(3-ethylbenzthiazoline-6-sulfonic acid) diammonium salt (ABTS) (Sigma) and 0.03% (v/v) hydrogen peroxide) was added (125 mL per well). The manifold was incubated for 10 min at RT with gentle agitation. One hundred mL per well of the reaction solution was transferred to a 96-well at bottom microtiter plate (Corning) and absorbance at 405 nm was determined using a FluoStar Optima plate reader. To measure the EWF of the samples following trypsin treatment, steel coupons were polished, cleaned, and coated with K122-4(128-144)ox peptide as described previously and were in 1 mg/mL trypsin for 45 min at 37  C. Samples were washed six times with distilled water and allowed to air dry before testing the EWF. Data were analyzed using the statistics program PRISM 5.0 (GraphPad Sofware) utilizing ManneWhitney analysis, as the data distribution was non-parametric in nature. Nanoindentation data were analyzed using a linear regression model to generate a best-t line through the data points. Signicant differences between data points were determined using 2-way non-parametric ANOVA analysis followed by Bonferroni posttests. A signicance cut-off of 0.05 was employed for all tests.

3. Results and discussion To elucidate the nature of the interaction between stainless steel and the T4P of P. aeruginosa, a synthetic peptide of the RBD, termed K122-4(128-144)ox, was applied to stainless steel surfaces (note that we are reporting on a single P. aeruginosa type IV pilin RBD sequence but our data indicates that all tested RBD sequences to date interact in a similar manner). Derived from the sequence of the RBD of P. aeruginosa strain K122-4 PilA protein, the peptide is N-a-acetylated, has a C-terminal amide, and contains a disulde loop (Fig. S1). This new bioorganic steel, borg-K122SS, was examined for changes in the physical and mechanical properties of its surface, including the reactivity of surface electrons, as a result of peptide binding. We speculated that any interactions between the steel surface electrons and the peptide would result in measurable changes in these surface properties. Metal electron activity is described by the electron work function (EWF). EWF measures the minimum amount of energy required to move an electron from the Fermi level to just beyond the surface of the metal. Gold, considered the standard from which all other EWFs are compared to, has an EWF of 5.1 eV [19,20]. Regions of higher electron activity, such as the grain boundary, have correspondingly lower EWFs because less energy is required for removing electrons from the regions. Conversely, surfaces with higher EWF values are more stable and inert because they possess

Fig. 2. XPS spectra analysis of elemental composition of borg-K122SS and 304 stainless steel. Overlays of borg-K122SS and 304 steel XPS spectra of A) O 1S orbitals, B) sulfur 1S orbital region with an unidentied peak associated with the borg-K122SS surface, and C) iron 3S orbital region with an unidentied peak associated with the borg-K122SS surface. XPS spectra of borg-K122SS is plotted in green while the XPS spectra of 304 stainless steel is plotted in red. (For interpretation of the references to colour in this gure legend, the reader is referred to the web version of this article.)

E.M. Davis et al. / Biomaterials 32 (2011) 5311e5319

5315

less reactive surface electrons. Borg-K122SS had a signicantly higher EWF compared to unmodied 304 steel (Fig. 1A). The presence of bound peptide increased the EWF from 4.5 to w5.0 eV (P < 0.001), suggesting that binding of the peptide stabilizes steel surface electrons. The EWF of a second metal aluminum, did not increase when peptide was added to the surface or dried on the surface, as the EWF was not signicantly different from uncoated aluminum (Fig. 1B). Aluminum is utilized as a control surface on which peptide could be deposited where the peptide would not interact and thus allows for the assessment of the materials properties of the peptide rather than a peptide coupled to a surface. Aluminum may have fairly unique properties as the type IV pilin RBD appears to be designed to interact with most solid phase surfaces. The ability of the peptide to stabilize steel but not aluminum suggests that the peptideesteel interaction occurs between the peptide and one or more specic components within the steel.

The adhesive force, a second physical property of metal surfaces, describes the ability of a metal surface to stick and interact with other surfaces and is increased at grain boundaries [21,22]. Higher adhesive forces correspond to areas where the EWF is lower [23] and the RBD has been shown to bind with higher adhesive force to grain boundaries compared with regions within grains [3]. The adhesive force of borg-K122SS and 304 steel was measured using atomic force microscopy (AFM). The AFM cantilever, possessing a silicon nitride tip, is lowered until the tip interacts with the surface. The cantilever is then raised and the amount of force necessary to break the interaction between the tip and the surface is measured (Fig. S2). Surfaces with reactive surface electrons interact more readily with the tip and greater force is necessary to break the interaction. The AFM silicon nitride tip did not interact as strongly with borg-K122SS (P < 0.001), as the adhesive force of borg-K122SS was 19.4 8.8 nN compared to an observed adhesive force of 56.7 10.5 nN for untreated 304

Fig. 3. Auger-SEM scan of distribution of elements on the surface of borg-K122SS. Elemental scan of A) nitrogen, note the presence of nanostructed material (see arrowheads) and B) sulfur, note the presence of nano-structured material (see arrowheads). C) Topographical scan of the surface of borg-K122SS. D) Analysis of the backscattered electrons obtained from the borg-K122SS surface plotted as the frequency of electrons at a given energy (in eV) and labeled as to the element generating electrons of that energy value. Note the S and N peaks (originating solely from the peptide) are low but the energies of the backscattered electrons are observed at previously reported energy levels for these elements.

5316

E.M. Davis et al. / Biomaterials 32 (2011) 5311e5319

E.M. Davis et al. / Biomaterials 32 (2011) 5311e5319

5317

steel (Fig. 1C). These results suggest that the surface of borgK122SS is less reactive than 304 steel and that the surface electrons of borg-K122SS interact less strongly with the AFM tip. The adhesive force of peptide dried on aluminum did not differ from unaltered aluminum (Fig. 1D). Another important property of surfaces is their hardness. To compare the hardness of borg-K122SS to 304 steel, an AFM equipped with a triboscope was used to obtain nanoindentation loadedisplacement plots, where total displacement of a diamond indenter tip into the surface of borg-K122SS and 304 steel was measured for loads of 50e400 mN (Fig. S3A,B). At all loads, greater displacement was observed for 304 steel compared to borg-K122SS (Fig. 1E). Even at the highest load of 400 mN, total displacement into borg-K122SS was 20 nm compared to 35 nm for 304 steel (P < 0.0001). This differential hardness was maintained in the same fashion even with a load of 800 mN, the limits for this apparatus (Fig. S3C). No marked difference in displacement was observed between aluminum with applied peptide and untreated aluminum (Fig. 1F), further suggesting that the peptide interacts with specic components in steel and that, in the presence of bound peptide, a harder material is created as a result of changes in the physical and mechanical properties of the steel when it interacts with the organic peptide. The changes in the materials properties of the steel suggested that there was a chemical interaction of the peptide with the steel surface and that the peptide was not simply binding to the surface of the steel. To conrm that the peptide was not simply bound to the surface of the steel we determined whether one could displace a limited amount of bound biotinylated-peptide (the RBD was synthesized with a N-terminal biotin and tetra-glycine linker and approximately 50% of maximal peptide loading of the surface as determined by probing with a streptavidin-horseradish peroxidase probe) with a large excess of unlabeled free peptide. Bound biotinylated K122-4(128-144)ox could not be displaced from the steel surface by exogenous peptide (Fig. S4) indicating that the bound peptide is not in dynamic equilibrium. We hypothesized that the peptide was chemically reacting with the surface electrons of steel to alter the properties of the steel surface by undergoing a previously unreported form of chemical interaction. Such an interaction that generates a new material would result in changes in the electronic state of the surface, with measurable changes in the electron orbitals of the metal. To further characterize the chemical properties of borg-K122SS, as well as to determine whether bonding was occurring and to identify which elements were involved in the interaction, X-ray photoelectron spectroscopy (XPS) analysis was used to examine the electronic state of the elements on the surface of borg-K122SS in comparison to 304 steel. Spectra analysis revealed that the iron and chromium 2p 1/2 and 3/2 orbitals (Fig. S5A,B) did not appear to play an important role in bond formation and electron stabilization as no shifts were observed in the borg-K122SS spectra when compared to 304 steel. An increase in the peak of the oxygen 1S orbital of borgK122SS was observed compared to 304 steel (Fig. 2A), with the borg-K122SS O1S peak peaking at 50 counts per second (CPS) compared to 40 CPS for 304 steel, suggesting a role for oxygen in bond formation. No signicant changes were observed in the spectra of the nitrogen 1S orbital and the carbon 1S orbital (Fig. S5C,D). The 304 steel contains negligible amounts of sulfur and

no classical sulfur electron orbitals were detected by XPS (Fig. 2B). However, the peptide contains two sulfur atoms within the disulde loop that would be visible by XPS. A signicant peak was observed near the location of unbonded sulfur 1S orbital (Fig. 2B) for borg-K122SS. This peak did not match any known shifts in sulfur peaks and could not be positively identied as sulfur based on XPS data alone. However, since sulfur must be present and was not identied anywhere else, this peak strongly suggests involvement of sulfur in this bonding event which has not been previously described. Electrons can ow through disulde bonds, making it feasible that the disulde bond could interact with and stabilize surface electrons by chemically binding with them. Further examination of the XPS spectra revealed the presence of a strong peak near the location of the iron 3S orbital (Fig. 2C) in borg-K122SS that was absent from the spectra of 304 steel. The differences between the electronic states of oxygen, iron, and sulfur on the surface of borg-K122SS and 304 steel conrm that borg-K122SS is a new material that is chemically different from 304 steel. The XPS spectra data support the involvement of several elements in the formation of borg-K122SS. To supplement the XPS data, Auger scanning electron microscopy (SEM) scans of the surface of borg-K122SS were independently performed to look at the surface distribution of the elements nitrogen, oxygen, sulfur and carbon by determining the number and source of origin of backscattered secondary electrons of the appropriate energy levels for the specic elements on the borgK122SS surface. Carbon, and oxygen were distributed across the surface in a mostly homogenous pattern, with a few areas of higher element concentration. Given the high concentration of carbon and oxygen in stainless steel, the lack of any specic distribution pattern was anticipated and no nanostructural pattern of distribution was observed for either carbon or oxygen (Fig S6A,B). Nitrogen and sulfur (which are not normally components of steel surfaces and the presence of these elements is due to presence of the peptide) were clearly seen to be distributed across the surface in discrete nanostructures (see arrowheads in Fig. 3A,B) which suggests that the peptideesteel interface was in fact imaged. These nanostructures were not discernable in a standard SEM surface scan (Fig. 3C) and one would not anticipate imaging the peptide given that this was an uncoated specimen and that the accelerating potential used was 10.0 keV (Fig. 3C). The auger-SEM energy spectra of the backscattered electrons indicated the presence of low amounts of sulfur on the surface (Fig. 3D). These results suggest that at least one of the unidentied electron orbitals observed in the XPS spectra of borg-K122SS is likely a far red-shifted-sulfur electron orbital, perhaps suggesting a conjugated electron. It is likely that electrons are delocalized throughout the disulde loop of the peptide and shared through multiple contact points with the steel surface. This hypothesis is supported by the observation that trypsin treatment of borg-K122SS releases the peptide from the steel surface (the RBD has two lysine residues, see Fig. S1) with a resulting change in the surface EWF (Fig. 4A,B). It is worth noting that a formed disulde bridge has previously been noted as a requirement for RBD functionality [14]. As protease sensitivity of bioorganic steel would severely limit potential applications, we examined the ability of synthetic D-K122-4(128-144)ox (an enantiomeric form of the native peptide, see Fig. S1) and retro-inverso

Fig. 4. Evidence that stereospecic interactions/products are differentially produced on 304 stainless steel dependent upon the stereochemistry of the synthetic peptide utilized. A) Tryspin sensitivity biotinylated K122-4(128-144)ox peptide on the surface of 304 stainless steel. B) Trypsin sensitivity of borg-K122SS as determined by measuring the EWF. C) EWF of borg-K122Ss, borg-K122SS D-amino (generated by utilizing K122-4(128-144)ox consisting of all D amino acid residues), and borg-retro-inverso-K122SS (generated by using a synthetic peptide synthesized with all D amino acid residues but in the inverted sequence or what is termed a retro-inverso peptide with a standard acetylated N-terminus and a C-terminal amide). D) EWF analysis of borg-retro-inverso-K122SS and 304 stainless steel with and without trypsin digestion. E) Nanoindentation analysis of borg-retroinverso-K122SS and 304 stainless steel with a load of 200 mN and a load of 400 mN. F) Corrosion rates of borg-retro-inverso-K122SS and 304 stainless steel.

5318

E.M. Davis et al. / Biomaterials 32 (2011) 5311e5319

D-K122-4(128-144)ox (a peptide of inverted sequence and synthesized with D-amino acid residues which will position the amino acid side chains in the correct relative positions, see Fig. S1) to alter the properties of steel as both the all D- and the retro-inverso peptide would be protease resistant. As peptides composed of D-enantiomeric amino acid residues are protease resistant, these forms of the peptide sequence would increase the potential utility of these peptides. The all D- and the retro-inverso forms of the RBD interacted with the steel surface and physical properties of the steel surface. Strikingly, the D-K122-4(128-144)ox peptide interacted with the steel but did not signicantly increase the EWF of the steel while the retro-inverso D-K122-4(128-144) signicantly increased the EWF of the steel beyond that observed for the native L-form of the peptide (Fig. 4C). The retro-inverso D-K122-4(128-144)ox was, as anticipated, resistant to trypsinization (Fig. 4D) and also significantly increased the hardness of the steel (Fig. 4E). As the EWF is a measure of the energy level of the surface electrons and the ability of those electrons to react with other materials, a high EWF generally reects good resistance to corrosion and the high EWF of the borg-retro-inverso-K122SS suggested that its corrosion rate would be low. We thus tested the corrosion rate of borg-retroinverso-K122SS. Borg-retro-inverso-K122SS is signicantly more corrosion resistant than regular stainless steel, having a corrosion rate w50% of normal steel (Fig. 4F). Our results strongly suggest that the peptideesteel reaction is a stereospecic process that results in different forms of bioorganic steel that can be differentiated based on their biophysical properties. The molecular basis for the observed stereospecicity in the products is unclear at the current time. The high strength of the RBDestainless steel interaction, which anchors the pilus to the surface and enables the progressive movement of the cell along that surface when the pilus retracts, appears reasonable given the formation of a chemical interaction with the surface. The extremely high apparent afnity of the RBD for the steel surface is also reasonable given that a bond is formed. The observation that the strength of the RBDesteel interaction varies as a function of the surface electron activity of stainless steel [13] suggests that the bond formed with the RBD requires surface electrons that can be delocalized through the peptide. The PA RBD has been demonstrated to interact with a wide range of surfaces and metals including titanium (aluminum being an interesting case where the RBD does not interact with the metal) suggesting that the nature of the surface electrons may be more important for binding than the elemental or material composition. 4. Conclusions In summary, we have characterized a new material, which we have termed borg-K122SS, which is readily and spontaneously formed from stainless steel with exposure to a synthetic peptide that is the functional adhesin of the T4P. Borg-K122SS has signicantly different physical and chemical properties compared with regular 304 stainless steel. Borg-K122SS had a higher EWF and was less adhesive than unmodied 304 steel. Borg-K122SS surfaces were signicantly harder and corroded signicantly more slowly than the parental 304 steel. XPS spectra data of borg-K122SS suggests that multiple contact points exists between the peptide and the steel, with two currently unassigned electron orbitals (one likely due to a far red-shifted-sulfur electron binding energy) which supports the formation of a chemical interaction, possibly with extensive electron de-localization, that results in chemical and physical stabilization of the surface. The observation that the RBD bonds with stainless steel to form a previously unreported material that has altered physical/chemical attributes has signicant implications for both the facile design,

modication and engineering of a common, widely used industrial material based on a natural process and our understanding of twitching motility and T4P structure and function. Acknowledgments The authors gratefully acknowledge the nancial support of the Natural Sciences and Engineering Research Council of Canada through Discovery operating grants to DYL and RTI, and through an Alexander Graham Bell Canada Graduate Scholarship to EMD. EMD was also the recipient of a University of Alberta PhD scholarship. The technical assistance and helpful suggestions of Bart Hazes, particularly in the generation the enantiomeric structure of the truncated PilA pilin protein and receptor binding domain is gratefully acknowledged. We gratefully acknowledge the technical assistance of D. Karpuzov, A. He, and S. Xu in relation to the XPS and SEM Auger analysis. We thank B. Yu, R. Chung, and X. Tang for technical assistance and helpful suggestions. The authors disclose that a patent application based on this technology has been led and assigned to Arch Biopartners Inc., and that the authors and the University of Alberta hold an equity position in Arch Biopartners Inc. Appendix. Supplementary material Supplementary data related to this article can be found online at doi:10.1016/j.biomaterials.2011.04.027. References
[1] Giltner CL, van Schaik EJ, Audette GF, Kao D, Hodges RS, Hassett DJ, et al. The Pseudomonas aeruginosa type IV pilin receptor binding domain functions as an adhesin for both biotic and abiotic surfaces. Molecular Microbiology 2006; 59:1083e96. [2] Yu B, Giltner CL, van Schaik EJ, Bautista DL, Hodges RS, Audette GF, et al. A novel biometallic interface: high afnity tip-associated binding by pilinderived protein nanotubes. Journal of Bionanoscience 2007;1:73e83. [3] Yu B, Davis EM, Hodges RS, Irvin RT, Li DY. Surface nanocrystallization of stainless steel for reduced biolm adherence. Nanotechnology 2008;19, http://dx.doi.org/10.1088/0957-4484/19/33/335101. 8335101 (p. 8). [4] Zuo R, Ornek D, Wood TK. Aluminum- and mild steel-binding peptides from phage display. Applied Microbiology and Biotechnology 2005;68:505e9. [5] Khoo X, OToole GA, Nair SA, Snyder BD, Kenan DJ, Grinstaff MW. Staphylococcus aureus resistance on titanium coated with multivalent PEGylated-peptides. Biomaterials 2010;31:9285e92. [6] Schweizer HP, Po C. Regulation of glycerol metabolism in Pseudomonas aeruginosa: characterization of the glpR repressor gene. Journal of Bacteriology 1996;178:5215e21. [7] Tao S, Li DY. Nanocrystallization effect on the surface electron work function of copper and its corrosion behaviour. Philosophical Magazine Letters 2008; 88:505e9. [8] Song S. In situ high resolution eletronc microscopy of grain-boundary migration through ledge motion in an AleMg alloy. Philosophical Magazine Letters 1990;79:3125e7. [9] Artemyev AV, Fionova LK. Grain-boundary morphology change in aluminum during heating. Fizika Metallov I Metallovedenie 1988;66:132e6. [10] Stanley PM. Factors affecting the irreversible attachment of Pseudomonas aeruginosa to stainless steel. Canadian Journal of Microbiology 1983;29: 1493e9. [11] Li X, Logan BE. Analysis of bacterial adhesion using a gradient force analysis method and colloid probe atomic force microscopy. Langmuir 2004;20: 8817e22. [12] Pereni CI, Zhao Q, Liu Y, Abel E. Surface free energy effect on bacterial retention. Colloids and Surfaces B: Biointerfaces 2006;48:143e7. [13] Yu B, Lesiuk A, Davis EM, Irvin RT, Li DY. Surface nanocrystallization for bacterial control. Langmuir 2010;26:10930e4. [14] Wong WY, Campbell AP, McInnes C, Sykes BD, Paranchych W, Irvin RT, et al. Structure-function analysis of the adherence-binding domain on the pilin of Pseudomonas aeruginosa strains PAK and KB7. Biochemistry 1995;34: 12963e72. [15] Wong WY, Irvin RT, Paranchych W, Hodges RS. Antigeneantibody interactions: elucidation of the epitope and strain-specicity of a monoclonal antibody directed against the pilin protein adherence binding domain of Pseudomonas aeruginosa strain K. Protein Sci 1992;1:1308e18.

E.M. Davis et al. / Biomaterials 32 (2011) 5311e5319 [16] Campbell AP, McInnes C, Hodges RS, Sykes BD. Comparison of NMR solution structures of the receptor binding domains of Pseudomonas aeruginosa pili strains PAO, KB7, and PAK: implications for receptor binding and synthetic vaccine design. Biochemistry 1995;34:16255e68. [17] Yu L, Lee KK, Paranchych W, Hodges RS, Irvin RT. Use of synthetic peptides to conrm that the Pseudomonas aeruginosa PAK pilus adhesin and the Candida albicans mbrial adhesin possess a homologous receptor-binding domain. Molecular Microbiology 1996;19:1107e16. [18] Baikie PJ. Estrup, low cost PC based scanning Kelvin probe. Review Scientic Instruments 1998;69:3902e7.

5319

[19] Ashcroft NW, Mernim ND. Solid state phyiscs. Philadelphia: Saunders College Publishing; 1976. [20] Bhushan B, ViGoldade A. Measurements and analysis of surface potential change during wear of single-crystal silicon (100) at ultralow loads using Kelvin probe microscopy. Applied Surface Science 2000;157:373e81. [21] Chadwick G, Smith D. Grain boundary structure and properties. London: Academic Press; 1979. [22] McLean D. Grain boundaries in metals. London: Clarendon Press; 1957. [23] Li Y, Li D. Electron work function, adhesion, and friction between 3d transition metals under light loads. Wear 2005;259:116e23.