Surface morphology, in addition to hydrophobic and electrostatic effects, can alter how proteins interact with solid surfaces. Understanding the heterogeneous dynamics of protein adsorption on surfaces with varying roughness is experimentally challenging. In this work, we use single-molecule fluorescence microscopy to study the adsorption of α-lactalbumin protein on the glass substrate covered with a self-assembled monolayer (SAM) with varying surface concentrations. Two distinct interaction mechanisms are observed: localized adsorption/desorption and continuous-time random walk (CTRW). We investigate the origin of these two populations by simultaneous single-molecule imaging of substrates with both bare glass and SAM-covered regions. SAM-covered areas of substrates are found to promote CTRW, whereas glass surfaces promote localized motion. Contact angle measurements and atomic force microscopy imaging show that increasing SAM concentration results in both increasing hydrophobicity and surface roughness. These properties lead to two opposing effects: increasing hydrophobicity promotes longer protein flights, but increasing surface roughness suppresses protein dynamics resulting in shorter residence times. Our studies suggest that controlling hydrophobicity and roughness, in addition to electrostatics, as independent parameters could provide a means to tune desirable or undesirable protein interactions with surfaces.

Protein adsorption and transport on solid surfaces are common but complicated phenomena.1,2 In 2019, the National Academies of Sciences challenged researchers to design new surface materials as part of understanding protein adsorption and dynamics in realistic complex environments.3 Achieving a better understanding of protein adsorption onto solid surfaces is fundamental to advance our knowledge in areas, such as protein sensing,4 drug delivery,5 designing anti-fouling materials,6 and protein chromatography.7 However, studying the protein–surface interactions is often challenging because multiple chemical parameters affect the nanoscale interactions at the surface.8–10 For example, in addition to chemical parameters such as hydrophobicity, morphological surface characteristics span from nano- to microscales (e.g., single-crystal faces and polymer block phase separations) and lead to various types of surface roughness.11–14 

The importance of surface roughness effects on protein adsorption has been demonstrated using atomic force microscopy (AFM),15 circular dichroism,16 quartz crystal microbalance with dissipation,13 and fluorescence17 spectroscopy. For example, surface roughness affects collagen organization15 and influences the stiffness of saturated fibronectin layers.13 Many analyses are either ensemble-averaged or static and so questions remain about underlying contributions from minor interaction mechanisms or time-dependent changes in interaction mechanisms.

Single-molecule fluorescence microscopy is a robust method that allows direct visualization of protein dynamics at the interface, uncovering mechanistic details of protein adsorption.18–21 Single-molecule methods are ideal for studying the dynamics of single molecules within complex environments22–27 and uncovering details of complicated processes, such as intracellular transport of proteins,28–33 macromolecular motion along chromatographic stationary phases,34–37 and adsorption dynamics at polymer brushes38 or self-assembled monolayers (SAMs).39 These studies highlight the potential of single-molecule spectroscopy to investigate the details of protein motion within complex systems and nanoscale inter/intramolecular interactions, including surface roughness.

In this work, we use single-molecule fluorescence microscopy to reveal the details of transport dynamics of α-lactalbumin, a well-studied model protein, at an alkanesilane SAM surface with varying surface coverages. Silica coupling via silanol groups to alkane ligands is widely used in a variety of applications, such as stationary phases for reversed-phase protein chromatography40–42 and on biosensor surfaces.43,44 Previously, Langdon et al.45 observed the presence of multiple populations for model proteins on a hydrophobic trimethylsilane surface and attributed the relative populations and their energetics to hydrophobic interactions. We expand upon that work here by showing that the relative prevalence of each population depends not only on hydrophobic interactions but also on the surface roughness and electrostatics as well as SAM surface coverage, which controls all of the chemical and physical interactions. The results of this study provide mechanistic insight into the driving forces in protein–surface interactions that may enable future improved optimization of functional materials surfaces.

Before SAM functionalization, substrates were cleaned via a multistep process. First, borosilicate microscope coverslips (No. 1; 22 × 22 mm2, VWR) were sequentially sonicated in soap water (Liquinox 2%), deionized water, methanol (ACS grade, Sigma), and acetone (ACS grade, Sigma). The coverslips were then cleaned at 80 °C for 30 min in a base piranha solution containing 4% (v/v) H2O2 (Fisher Scientific) and 13% (v/v) NH4OH. After rinsing the coverslips with DI water and drying the coverslips under a stream of nitrogen (Ultra Pure, Airgas), the coverslips were further treated with oxygen plasma (Harrick Plasma, PDC-32G) for 2 min. Octadecyltrichlorosilane (ODTS) [CH3(CH2)17SiCl3, purity > 90%, Sigma] was dissolved with toluene into varying concentrations [0.1%, 0.01%, 0.001%, and 0.0001% (w/w)] and was used as a dipping solution to deposit ODTS on the microscope coverslips for 10 min at room temperature and under ambient conditions. Reproducible ODTS coverage over all samples was achieved following the previous sample preparation procedures by Lessel et al.46 The sample preparation conditions used here promote ODTS self-polymerization and cross-linking, inducing surface roughness, which is the main interest of this study. Then the samples were placed in an ultrasonic tank at 50 °C for 10 min with toluene (99.8%, Sigma) to remove any physically adsorbed ODTS molecules. The samples were rinsed with isopropanol (HPLC Grade, Fisher) and dried under a stream of nitrogen. Microfluidic assemblies (HybriWell Chamber; Grace BioLabs) were then attached to the ODTS-covered surface with tubes (0.03 in. internal diameter; Scientific Commodities) connected at the inlet and the outlet to supply a constant solution of 100 pM labeled α-lactalbumin at 50 µl/min in 10 mM HEPES buffer (pH = 7.2, Sigma). We used fluorescent beads (540/560 nm, 0.1 µm, Invitrogen, USA) as stationary emitters on the glass slide to measure intensity profiles across the entire region of interest. Stock solutions of fluorescent beads were sonicated for 15 min and diluted into a 1:1000 ratio with DI water. The 100 μl of the diluted solution was then dropcasted onto clean glass slides and dried for 15 min before removing the excess solution with the N2 jet.

α-lactalbumin transport on ODTS-covered surface was monitored at the aqueous solution/SAM interface [Fig. 1(a)] by utilizing our custom-built microscope in the total internal reflection fluorescence (TIRF) excitation mode.19 The TIRF excitation created an exponentially decaying evanescent wave at the interface with a penetration depth of up to 100 nm into the bulk of the solution.21 Thus, only proteins at the interface were excited, detected, and tracked. The excitation light source was a 5 mW continuous-wave circularly polarized 532 nm laser (Coherent, compass 315M-100SL). The beam was expanded and focused at the edge of a 100 × NA 1.46 oil-immersion objective (alpha Plan-Apochromat, Carl Zeiss) and was then focused onto an area of 32 × 32 µm2 in the camera’s field of view on the sample slide. Emission from probe molecules was collected by the same objective and directed to the Optosplit II (Cairn Research, Faversham, UK) where emission was separated from the excitation. An electron-multiplied charge-coupled device (Andor, iXon 897) was used for detection at an integration time of 30 ms, a frame rate of 16 Hz, and an electron-multiplying gain of 300. All the collected frames were cropped from 32 ×32 to 20 ×20 µm2 central regions during post-analysis. To demonstrate that the TIRF excitation field was uniform across the analyzed region, we calculated and compared static emitters’ intensities (Fig. S1). To calculate the localization precision of the instrument, we imaged immobilized emitters on the glass substrate for 100 frames. We localized and tracked each particle for each frame using the standard deviation of the trajectory position as our localization precision for that particle. The localization precision reported (X: 12 nm, Y: 13 nm) is the average precision of each particle measured (Fig. S2).

FIG. 1.

(a) Cartoon representation of α-lactalbumin protein flowing over ODTS monolayer. Due to the TIRF design, the emission from the fluorescent label on the protein is only observable when it is adsorbed to the interface, allowing for single-molecule tracking analysis. (b) Representative α-lactalbumin trajectories on ODTS (0.01% in toluene) monolayer with two modes of motion: CTRW (blue) and localized (magenta, indicated by arrows), which were filtered using spatial trajectory filtering (inset). (c) Single-frame displacement distribution showing the presence of two modes of dynamics.

FIG. 1.

(a) Cartoon representation of α-lactalbumin protein flowing over ODTS monolayer. Due to the TIRF design, the emission from the fluorescent label on the protein is only observable when it is adsorbed to the interface, allowing for single-molecule tracking analysis. (b) Representative α-lactalbumin trajectories on ODTS (0.01% in toluene) monolayer with two modes of motion: CTRW (blue) and localized (magenta, indicated by arrows), which were filtered using spatial trajectory filtering (inset). (c) Single-frame displacement distribution showing the presence of two modes of dynamics.

Close modal

α-lactalbumin (pI = 4.2),47 with three Alexa 546 molecules per protein,48 was previously purified and acquired from the Willson Group at the University of Houston and was diluted to 10 pM in HEPES buffer (pH 7.2). As shown previously, the presence of multiple dyes on protein does not affect either adsorption dynamics49 or protein structure.34 Stock protein solution was diluted to 10 pM in HEPES buffer (pH 7.2) for all single-molecule experiments presented in this work. For salt experiments, α-lactalbumin was diluted in 10 mM HEPES buffer (pH = 7.2) containing 0.25M NaCl (GR ACS, Merck).

AFM (Park AFM NX20) operated in tapping mode was used to characterize the surface roughness of ODTS-covered surfaces. Silicon-tip on nitride lever probes (Bruker, ScanAsyst, f0 = 70 kHz) with a reflective aluminum coating was used to probe respective surfaces. AFM images were acquired with a 256 × 256 image array and a 1 Hz cantilever-tip scan rate. Samples for AFM measurements were prepared using the same protocol as for single-molecule experiments. Root mean squared surface roughness (rms) values were determined by taking surface roughness measurements using the NanoScope analysis software (version 1.5) and measuring on two separate regions of four samples for each surface concentration. Error represents the standard deviation of the sampled data for each method. In total, eight rms results per each surface concentration were averaged. To test the accuracy of the sampling method we used, we calculated rms for the smaller regions (2.5 × 2.5 μm2) and compared the values with the value for the 5 × 5 μm2 region. For one of the 0.1% ODTS sample, the calculated rms value for 5 × 5 μm2 region was 1.3 nm and rms values for four regions of 2.5 × 2.5 μm2 were 1.4, 1.3, 1.1, and 1.3 nm with the mean value 1.3 nm, which is within the standard deviation of the rms for 5 × 5 μm2 regions.

Static contact angle measurements were performed on a DSA100 Drop Shape Analyzer (Kr̈ss Instruments, Germany). 2 µl drops of DI water were cast onto three separate regions of the ODTS-covered surfaces, and the contact angles of droplets were individually evaluated through the ADVANCE software (Kr̈ss Instruments, Germany). The average and standard deviation of the contact angle values were calculated between the three individual droplets for 0.0001%, 0.001%, 0.01%, and 0.1% ODTS samples.

Positive high lateral resolution images were performed using a time-of-flight secondary ion mass spectrometry (TOF-SIMS) NCS instrument, which combines a TOF-SIMS5 instrument (ION-TOF GmbH, Münster, Germany) and an in situ scanning probe microscope (NanoScan, Switzerland) at Shared Equipment Authority from Rice University. For the TOF-SIMS chemical mapping analysis of patterned ODTS, the measurements were conducted using pulsed 60 keV Bi3++ ions (with a measured current of 0.02 pA) configurated in high lateral resolution to image a field of view of 50 × 50 μm2, with a raster of 2048 × 2048 pixels. The image raster was then binned by a factor of 64 to enhance the signal-to-noise ratio. A charge compensation with an electron flood gun was applied during the analysis. An adjustment of the charge effects was operated using a surface potential.

α-lactalbumin trajectories [Fig. 1(b)] reveal two modes of protein motion at the ODTS-covered surface. We characterize one population as a continuous-time random walk (CTRW) [Fig. 1(b), blue], while the second population undergoes localized interactions [Fig. 1(b), magenta]. The CTRW mode displays periods of flights or hopping and immobile waiting-time periods,50 whereas the localized population is treated as adsorption/desorption without long flights. Each α-lactalbumin trajectory is classified as either CTRW or localized using spatial trajectory filtering, previously used in Ref. 18. Briefly, if an α-lactalbumin molecule moves more than a threshold distance (90 nm) between two consecutive frames within a trajectory, the trajectory is classified as CTRW [Fig. 1(b), inset]. The 90 nm threshold is selected through control experiments on a bare glass surface and provides a 1% misclassification probability (see details in the supplementary material and Fig. S3).

Single-frame displacement distribution analysis allows the classification of both localized and CTRW populations [Fig. 1(c), magenta and blue, respectively]. Single-frame displacements are calculated as the distance a single protein travels between two consecutive frames. Two populations of transport mode were reported before in dye-multilayer polyelectrolyte51 in protein–polymer18 as well as in protein–SAM45 interactions. For SAMs, surface coverage is a variable parameter that could play a role in controlling these interaction modes. Thus, we tune the properties of the sample surfaces by changing the monolayer surface concentration and by lithographically templating SAMs onto glass, both of which are discussed next.

The contribution of localized protein interactions decreases as the concentration of ODTS and the corresponding surface hydrophobicity increase (Figs. 2 and 3). SAM formation depends on a variety of factors such as solvent type and polarity,52 temperature,53 humidity,54 reaction time,55 substrate roughness,56 and the concentration of monomers used for the deposition process.52,57 Rozlosnik et al. reported that as the ODTS monomer concentration increases, the overall surface coverage also increases.52 Therefore, we vary the concentration of ODTS in the immersion solution ranging between 0.1% and 0.0001% (w/w). Increasing the solution concentration results in a greater population of ODTS molecules deposited on the surface, which is confirmed by XPS (Fig. S4). The hydrophobicity also increases with ODTS concentration, confirmed by contact angle goniometry. The insets of Figs. 2(a)2(d) show the contact angles of fabricated surfaces obtained at varying concentrations of ODTS in toluene and a fixed immersion time of 10 min.

FIG. 2.

(a)–(d) Representative single-molecule α-lactalbumin trajectories at ODTS-covered interfaces with varying ODTS concentrations in the initial solutions: 0.0001%, 0.001%, 0.01% and 0.1% w/w of ODTS in toluene. Static contact angle measurements for corresponding surfaces (insets). (e)–(h) Single-frame displacement distributions for the corresponding ODTS concentrations (w/w %).

FIG. 2.

(a)–(d) Representative single-molecule α-lactalbumin trajectories at ODTS-covered interfaces with varying ODTS concentrations in the initial solutions: 0.0001%, 0.001%, 0.01% and 0.1% w/w of ODTS in toluene. Static contact angle measurements for corresponding surfaces (insets). (e)–(h) Single-frame displacement distributions for the corresponding ODTS concentrations (w/w %).

Close modal
FIG. 3.

(a) α-lactalbumin binding rates. (b) Percentage of α-lactalbumin molecules exhibiting CTRW vs localized adsorption–desorption. (c) Cumulative distributions of SRT for the corresponding ODTS concentrations (w/w %). Error bars in (a) and (b) represent the standard deviation of the mean.

FIG. 3.

(a) α-lactalbumin binding rates. (b) Percentage of α-lactalbumin molecules exhibiting CTRW vs localized adsorption–desorption. (c) Cumulative distributions of SRT for the corresponding ODTS concentrations (w/w %). Error bars in (a) and (b) represent the standard deviation of the mean.

Close modal

α-lactalbumin adsorption binding rate and surface mobility increase with increasing hydrophobicity of the ODTS surface. Varying the concentration of ODTS results in changes in protein dynamics, as illustrated by the single-molecule trajectories in Figs. 2(a)2(d) α-lactalbumin trajectories increase in number and shift toward the CTRW mode as hydrophobicity increases. Single-frame displacement histograms [Figs. 2(e)2(h)] quantify the shift toward the CTRW mode by revealing a second population with single-frame displacements larger than ∼100 nm.

Both the number of events and the fraction of CTRW events increase with hydrophobicity, which is quantified by the binding rate and localized/CTRW population ratio. The binding rates are determined by taking the average total number of events per frame per unit area. Calculated binding rates [Fig. 3(a)] suggest increasing loading capacity of α-lactalbumin with ODTS concentration. In addition, the tracking results show that the increased ODTS concentration results in a relative decrease in localized populations and an increase in CTRW populations [Fig. 3(b)]. These results agree with previous studies that suggest that proteins have a higher affinity to hydrophobic surfaces, as hydrophilic surfaces are used when protein resistance is needed.58–61 Tang et al.62 demonstrated the increased surface coverage of bovine serum albumin on alkyl modified gold substrate compared to hydrophilic carboxylate-terminated SAM by utilizing AFM imaging. It was also observed that proteins aggregate and form branched patterns. A single-molecule study of protein dynamics at a polymer surface as a function of surface hydrophobicity revealed the increased protein adsorption binding rate and mobility at a more hydrophobic surface.63 

Protein surface residence times (SRT) decrease with increasing surface hydrophobicity. SRT analysis shows the time each protein tends to spend on the surface before it desorbs. To demonstrate rare populations in single-molecule processes, we plot cumulative distributions of SRT.64Figure 3(c) demonstrates that the SRTs of adsorbed α-lactalbumin proteins decrease as the surface becomes more hydrophobic. Control experiments (Fig. S5) show that the average SRTs are not affected by changes in laser power, thus photobleaching does not play an important role in the SRT trends.

That SRT decreases with increasing ODTS coverage and hydrophobicity contradicts commonly accepted protein resistance–hydrophobicity relationships,65–67 including our own results in which Moringo et al.63 observed longer protein SRT with increased hydrophobicity of the surface. Additionally, α-lactalbumin adsorption on ODTS-covered surface induces conformational changes of the protein (Fig. S6) promoting favorable hydrophobic interactions, resulting in longer residence times on more hydrophobic surfaces.68 This discrepancy points out the complexity of surface treatments and motivates more detailed surface characterization, as discussed next.

AFM reveals that along with hydrophobicity, surface roughness increases at higher ODTS concentrations. AFM is used to uncover possible surface changes within samples prepared with different ODTS concentrations. As shown by the micrographs in Fig. 4 with calculated rms values, surface roughness increases with ODTS concentration. Previous studies show that ODTS molecules cross-link covalently, forming round aggregates that are densely packed and approximately vertically oriented with respect to the surface.54,57,69 Increased roughness with a higher density of aggregates at high ODTS concentrations is indeed observed in our samples.

FIG. 4.

AFM analyses of surface roughness on ODTS samples; all images shown are 5 × 5 μm2 area. Samples (a)–(d) are representative of the corresponding ODTS concentrations.

FIG. 4.

AFM analyses of surface roughness on ODTS samples; all images shown are 5 × 5 μm2 area. Samples (a)–(d) are representative of the corresponding ODTS concentrations.

Close modal

An increase in sample roughness, rather than hydrophobicity, explains the decrease in the SRT of α-lactalbumin. Wang et al.70 demonstrated the influence of obstacles on the motion of polymer molecules by comparing their motion on a flat surface vs pillar-patterned surfaces. In their work, the authors reported decreased mobility and increased subdiffusive motion of a polymer chain at the pillar-patterned surfaces, compared to the flat surface. In our single-molecule experiments, we do observe higher numbers of α-lactalbumin proteins interacting with ODTS surfaces with higher hydrophobicity. On the other hand, increasing surface roughness, which comes with a higher concentration of ODTS, suppresses CTRW protein motion. CTRW is composed of periods of temporal immobilization and periods of flights or surface displacements produced by brief excursions into the bulk fluid.50 Therefore, the total residence time of a protein undergoing CTRW is a summation of both time periods: immobilization(s) plus flight(s). However, increased surface roughness makes protein readsorption after the flying period less probable, which in turn decreases SRT. Additionally, to correlate CTRW protein motion on ODTS-covered surfaces with surface roughness, we perform image analyses on our single-molecule data to make protein transport-based surface maps (details in the supplementary material). By counting the number of CTRW trajectory segments in binned pixels spanning across the surface, we first make CTRW-based surface maps and demonstrate that CTRW transport distribution is uneven and increases with ODTS concentration increase (Fig. S7). Then, we calculate the rms intensity values from CTRW-based surface maps. We compare the obtained rms values from the CTRW-based surface maps with rms values from AFM analysis. The positive correlation between the two rms values indicates that the presence of CTRW on the ODTS surface is positively correlated with its surface roughness (Table S1).

Electrostatic repulsion between α-lactalbumin and ODTS-covered surfaces is another factor and acts to suppress the adsorption of proteins. The isoelectric point of α-lactalbumin is ∼4.2 leading to a net negative charge47 under our experimental conditions (pH = 7.2), while the sample surfaces are neutral because they are formed by alkyl [CH3–(CH2)17Si–] moieties. Previous studies established that increased ionic strengths suppress electrostatic forces between proteins and surfaces.71 Therefore, to investigate the role of electrostatic forces in the protein–surface interactions observed in this work, we conduct single-molecule experiments in HEPES buffer solutions with 0.1M NaCl. Figure 5 demonstrates increased α-lactalbumin SRT in the presence of salt (filled squares) compared to no salt (empty circles) conditions for 0.1% ODTS samples. The trend of increased average SRT in the presence of salt is confirmed for all other samples with varying ODTS concentrations (Fig. 5, inset). Average SRTs in the presence of salt (shown using a checkerboard pattern) are longer for all ODTS samples, revealing the presence of electrostatic forces in the system. We also observe a similar trend of increased SRT in the presence of salt for localized populations (Fig. S8).

FIG. 5.

Cumulative distributions of α-lactalbumin SRT for CTRW population at varied ODTS concentrations, at 0.1% ODTS with high ionic strength in the protein solution, and average SRT at varied ODTS concentrations with and without salt in the protein solution for CTRW population (inset).

FIG. 5.

Cumulative distributions of α-lactalbumin SRT for CTRW population at varied ODTS concentrations, at 0.1% ODTS with high ionic strength in the protein solution, and average SRT at varied ODTS concentrations with and without salt in the protein solution for CTRW population (inset).

Close modal

The existence of electrostatic forces is attributed to the presence of charged silanol groups on the exposed glass surfaces. Hozumi et al. showed72 that SiO2/Si plates covered with an octadecyltrimethoxysilane monolayer are negatively charged, regardless of functionalization by alkyl organosilane. However, the presence of SAMs on the surface reduces the surface acidity of silica due to the reduction of the density of surface silanol (Si–OH) groups. Differences in the SRT of α-lactalbumin at varied ionic strengths of the solution are explained by the shielding effect of salts. Without salt in the system, there is a repulsion between the negatively charged α-lactalbumin molecules and the ODTS-covered glass. However, the addition of NaCl decreases the Debye length for the electrostatic interaction from 3 nm for HEPES buffer to 0.9 nm for HEPES buffer with the addition of 0.1M NaCl (see the supplementary material for details), leading to increased interaction between the surface and α-lactalbumin leading to increased SRT. Observed repulsion between the surface and α-lactalbumin molecules without salt also supports the decreased binding rate with increased hydrophilicity of the sample [Fig. 3(a)]. A decrease in ODTS concentration leads to lower surface coverage, which in turn increases the density of surface silanol groups on the substrate, thereby increasing the net negative charge of the surface.72–74 Therefore, increased electrostatic repulsion, in addition to decreased hydrophobic effect, results in the decreased binding rate for the more hydrophilic surface.

To understand the role of the underlying glass substrate, we present an experiment in which we can simultaneously observe α-lactalbumin dynamics on both types of surfaces: bare glass and ODTS monolayers. We lithographically fabricate (details in the supplementary material and Fig. S9) patterned samples with alternating stripes of bare and ODTS-covered glass with widths of 5 and 10 μm, respectively. TOF-SIMS analysis of collected positive ions of the patterned surface is shown in Figs. 6(a) and 6(b). Figure 6(a) presents the distribution map of the sum of C3H2+, C4H7+, C5H7+, C5H9+, C6H7+, C6H9+, C7H7+, C7H9+, and C7H11+ ions (representative of alkyl molecules of ODTS), while Fig. 6(b) demonstrates representative ions for the glass substrate (Si+). 2D TOF-SIMS maps demonstrate the stripes with higher intensities for fragments related to ODTS [Fig. 6(a)] and glass [Fig. 6(b)]. The width dimensions of the stripes from TOF-SIMS analysis as well as from AFM surface characterization (Fig. S10) of the patterned surface are consistent with the original design of the photomask (Fig. S11) used to fabricate the pattern.

FIG. 6.

TOF-SIMS 2D maps of the patterned glass substrate with ODTS stripes with a width of 10 μm. Characteristic ions for (a) ODTS (sum of C3H2+, C4H7+, C5H7+, C5H9+, C6H7+, C6H9+, C7H7+, C7H9+, and C7H11+) and for (b) the glass substrate (Si+). Color scale represents ion counts that were normalized by dividing each pixel from the selected ion image with the same pixel location from the total ion image by the total ion intensity. Not correlated with TOF-SIMS filtered out CTRW (c) and localized (d) trajectories on the patterned surface.

FIG. 6.

TOF-SIMS 2D maps of the patterned glass substrate with ODTS stripes with a width of 10 μm. Characteristic ions for (a) ODTS (sum of C3H2+, C4H7+, C5H7+, C5H9+, C6H7+, C6H9+, C7H7+, C7H9+, and C7H11+) and for (b) the glass substrate (Si+). Color scale represents ion counts that were normalized by dividing each pixel from the selected ion image with the same pixel location from the total ion image by the total ion intensity. Not correlated with TOF-SIMS filtered out CTRW (c) and localized (d) trajectories on the patterned surface.

Close modal

Hydrophobic forces between α-lactalbumin and alkyl molecules are the driving forces for the observed CTRW motion of the proteins. Reconstructed protein trajectories from single-molecule measurements are filtered out using the same spatial filter as discussed above and classified into localized and CTRW trajectories. CTRW trajectories are only found in the ODTS-covered areas [Fig. 6(c)], whereas localized trajectories were found to be uniformly distributed [Fig. 6(d)]. Based on observed α-lactalbumin preferential adsorption on ODTS-covered areas with CTRW motion, we conclude that hydrophobic interactions between proteins and alkyl chains are the driving forces for the CTRW mode. On the other hand, the bare glass areas are responsible for the localized mode of motion.

We used single-molecule fluorescence imaging to study how protein transport is affected by the surface morphology of SAM-covered samples. Single-molecule tracking uncovered the presence of two modes of protein motion on the surface: CTRW and localized. Studying photolithographically prepared samples with alternating stripes of bare and SAM-covered glass revealed that the presence of alkyl chains of SAM causes CTRW mode of motion, whereas the bare glass substrate was responsible for the localized population. In addition, we found that the increased surface concentration of ODTS resulted not only in increased hydrophobicity but also in the increased roughness of ODTS-covered surfaces. These two effects played opposite roles in α-lactalbumin dynamics: increased hydrophobicity leads to longer protein flights, while increased surface roughness leads to a lower probability of readsorption and, therefore, shorter residence times. These findings highlight the importance of surface roughness on protein dynamics, which is of great importance for the design of next-generation materials and devices that are being exposed to biomolecules.

See the supplementary material for experimental sections describing the photolithographic sample preparation and x-ray photoelectron spectroscopy; data analysis details about single-molecule tracking algorithm and spatial trajectory filtering; and additional experimental results of the influence of laser power on the surface residence time, localization precision determination, XPS surface characterization, average surface residence times for localized populations, and AFM analyses of the patterned surface and the photomask design.

C.F.L. acknowledges support from the Welch Foundation (Grant No. C-1787) and the National Science Foundation (NSF CHE-1808382 and NSF CHE-2124983). We thank S. Link and his research group for discussions. TOF-SIMS analyses were carried out with the support provided by the National Science Foundation under Grant No. CBET-1626418. This work was conducted, in part, using resources of the Shared Equipment Authority at Rice University. We also thank Greg Wallraff (IBM) for his help with lithographic sample preparation optimization, the Willson Research Group at the University of Houston for previously providing the labeled α-lactalbumin, and an anonymous reviewer for suggesting to perform additional analysis and correlate protein mobility and surface roughness.

The authors have no conflicts to disclose.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

1.
K.
Nakanishi
,
T.
Sakiyama
, and
K.
Imamura
,
J. Biosci. Bioeng.
91
,
233
(
2001
).
2.
M.
Rabe
,
D.
Verdes
, and
S.
Seeger
,
Adv. Colloid Interface Sci.
162
,
87
(
2011
).
3.
National Academies of Sciences, Engineering, and Medicine
,
A Research Agenda for Transforming Separation Science
(
National Academies Press
,
2019
).
4.
J.-M.
Nam
,
C. S.
Thaxton
, and
C. A.
Mirkin
,
Science
301
,
1884
(
2003
).
5.
C.
Vauthier
,
C.
Dubernet
,
C.
Chauvierre
,
I.
Brigger
, and
P.
Couvreur
,
J. Controlled Release
93
,
151
(
2003
).
6.
V. B.
Damodaran
and
N. S.
Murthy
,
Biomater. Res.
20
,
18
(
2016
).
7.
K.
Hall
,
M.
Ashtari
, and
N. M.
Cann
,
J. Chem. Phys.
136
,
114705
(
2012
).
8.
S. P.
Mitra
,
J. Surf. Sci. Technol.
36
,
07
(
2020
).
9.
R.
Cagliani
,
F.
Gatto
, and
G.
Bardi
,
Materials
12
,
1991
(
2019
).
10.
S. A.
Bhakta
,
E.
Evans
,
T. E.
Benavidez
, and
C. D.
Garcia
,
Anal. Chim. Acta
872
,
7
(
2015
).
11.
A.
Dolatshahi-Pirouz
,
S.
Skeldal
,
M. B.
Hovgaard
,
T.
Jensen
,
M.
Foss
,
J.
Chevallier
, and
F.
Besenbacher
,
J. Phys. Chem. C
113
,
4406
(
2009
).
12.
K. M.
Woo
,
V. J.
Chen
, and
P. X.
Ma
,
J. Biomed. Mater. Res., Part A
67
,
531
(
2003
).
13.
M. B.
Hovgaard
,
K.
Rechendorff
,
J.
Chevallier
,
M.
Foss
, and
F.
Besenbacher
,
J. Phys. Chem. B
112
,
8241
(
2008
).
14.
M. S.
Lord
,
M.
Foss
, and
F.
Besenbacher
,
Nano Today
5
,
66
(
2010
).
15.
F. A.
Denis
,
P.
Hanarp
,
D. S.
Sutherland
,
J.
Gold
,
C.
Mustin
,
P. G.
Rouxhet
, and
Y. F.
Dufrêne
,
Langmuir
18
,
819
(
2002
).
16.
A. A.
Vertegel
,
R. W.
Siegel
, and
J. S.
Dordick
,
Langmuir
20
,
6800
(
2004
).
17.
B.
Müller
,
M.
Riedel
,
R.
Michel
,
S. M.
De Paul
,
R.
Hofer
,
D.
Heger
, and
D.
Grützmacher
,
J. Vac. Sci. Technol. B
19
,
1715
(
2001
).
18.
N. A.
Moringo
,
L. D. C.
Bishop
,
H.
Shen
,
A.
Misiura
,
N. C.
Carrejo
,
R.
Baiyasi
,
W.
Wang
,
F.
Ye
,
J. T.
Robinson
, and
C. F.
Landes
,
Proc. Natl. Acad. Sci. U. S. A.
116
,
22938
(
2019
).
19.
H.
Shen
,
L. J.
Tauzin
,
W.
Wang
,
B.
Hoener
,
B.
Shuang
,
L.
Kisley
,
A.
Hoggard
, and
C. F.
Landes
,
Anal. Chem.
88
,
9926
(
2016
).
20.
H.
Shen
,
L. J.
Tauzin
,
R.
Baiyasi
,
W.
Wang
,
N.
Moringo
,
B.
Shuang
, and
C. F.
Landes
,
Chem. Rev.
117
,
7331
(
2017
).
21.
C.
Dutta
,
L. D. C.
Bishop
,
J.
Zepeda O
,
S.
Chatterjee
,
C.
Flatebo
, and
C. F.
Landes
,
J. Phys. Chem. B
124
,
4412
(
2020
).
22.
D. A.
Higgins
,
S. C.
Park
,
K.-H.
Tran-Ba
, and
T.
Ito
,
Annu. Rev. Anal. Chem.
8
,
193
(
2015
).
23.
J. M.
Nölle
,
S.
Primpke
,
K.
Müllen
,
P.
Vana
, and
D.
Wöll
,
Polym. Chem.
7
,
4100
(
2016
).
24.
S.
Faez
,
Y.
Lahini
,
S.
Weidlich
,
R. F.
Garmann
,
K.
Wondraczek
,
M.
Zeisberger
,
M. A.
Schmidt
,
M.
Orrit
, and
V. N.
Manoharan
,
ACS Nano
9
,
12349
(
2015
).
25.
N. I.
Callaghan
,
S.-H.
Lee
,
S.
Hadipour-Lakmehsari
,
X. A.
Lee
,
M. A.
Siraj
,
A.
Driouchi
,
C. M.
Yip
,
M.
Husain
,
C. A.
Simmons
, and
A. O.
Gramolini
,
Commun. Biol.
3
,
229
(
2020
).
26.
B.
Gu
,
C. J.
Comerci
,
D. G.
McCarthy
,
S.
Saurabh
,
W. E.
Moerner
, and
J.
Wysocka
,
Mol. Cell
80
,
699
(
2020
).
27.
J.
Lu
,
H.
Mazidi
,
T.
Ding
,
O.
Zhang
, and
M. D.
Lew
,
Angew. Chem., Int. Ed.
59
,
17572
(
2020
).
28.
C. F.
Landes
,
Mol. Microbiol.
96
,
1
(
2015
).
29.
K.
Welsher
and
H.
Yang
,
Nat. Nanotechnol.
9
,
198
(
2014
).
30.
Q.
Li
,
K.-F.
Tseng
,
S. J.
King
,
W.
Qiu
, and
J.
Xu
,
J. Chem. Phys.
148
,
123318
(
2018
).
31.
J. Y.
Huang
and
C. Y.
Lin
,
J. Chem. Phys.
143
,
225101
(
2015
).
32.
J. M.
Hartley
,
T.-W.
Chu
,
E. M.
Peterson
,
R.
Zhang
,
J.
Yang
,
J.
Harris
, and
J.
Kopeček
,
ChemBioChem
16
,
1725
(
2015
).
33.
A. R.
Dun
,
G. J.
Lord
,
R. S.
Wilson
,
D. M.
Kavanagh
,
K. I.
Cialowicz
,
S.
Sugita
,
S.
Park
,
L.
Yang
,
A. M.
Smyth
,
A.
Papadopulos
,
C.
Rickman
, and
R. R.
Duncan
,
Curr. Biol.
27
,
408
(
2017
).
34.
A.
Misiura
,
H.
Shen
,
L.
Tauzin
,
C.
Dutta
,
L. D. C.
Bishop
,
N. C.
Carrejo
,
J.
Zepeda O
,
S.
Ramezani
,
N. A.
Moringo
,
A. B.
Marciel
,
P. J.
Rossky
, and
C. F.
Landes
,
Anal. Chem.
93
,
11200
(
2021
).
35.
W.
Calabrase
,
L. D. C.
Bishop
,
C.
Dutta
,
A.
Misiura
,
C. F.
Landes
, and
L.
Kisley
,
Anal. Chem.
92
,
13622
(
2020
).
36.
J. T.
Cooper
,
E. M.
Peterson
, and
J. M.
Harris
,
Anal. Chem.
85
,
9363
(
2013
).
37.
M. J.
Wirth
and
D. J.
Swinton
,
Anal. Chem.
70
,
5264
(
1998
).
38.
D. F.
Marruecos
, “
Connecting protein structure and dynamics on biomaterials with the foreign body response
,” Ph.D. thesis, University of Colorado at Boulder (
2018
).
39.
N.
Nelson
,
R.
Walder
, and
D. K.
Schwartz
,
Langmuir
28
,
12108
(
2012
).
40.
B. C.
Trammell
,
L.
Ma
,
H.
Luo
,
D.
Jin
,
M. A.
Hillmyer
, and
P. W.
Carr
,
Anal. Chem.
74
,
4634
(
2002
).
41.
L.
Ding
,
Z.
Guo
,
Y.
Xiao
,
X.
Xue
,
X.
Zhang
, and
X.
Liang
,
J. Sep. Sci.
37
,
2467
(
2014
).
42.
J. L.
Glajch
,
J. J.
Kirkland
, and
J.
Köhler
,
J. Chromatogr. A
384
,
81
(
1987
).
43.
D.
Samanta
and
A.
Sarkar
,
Chem. Soc. Rev.
40
,
2567
(
2011
).
44.
F.
Luderer
and
U.
Walschus
,
Immobilisation of DNA on Chips I
(
Springer
,
2005
), p.
37
.
45.
B. B.
Langdon
,
M.
Kastantin
, and
D. K.
Schwartz
,
Biophys. J.
102
,
2625
(
2012
).
46.
M.
Lessel
,
O.
Bäumchen
,
M.
Klos
,
H.
Hähl
,
R.
Fetzer
,
M.
Paulus
,
R.
Seemann
, and
K.
Jacobs
,
Surf. Interface Anal.
47
,
557
(
2015
).
47.
C.
Bramaud
,
P.
Aimar
, and
G.
Daufin
,
Biotechnol. Bioeng.
56
,
391
397
(
1997
).
48.
W.
Wang
,
H.
Shen
,
N. A.
Moringo
,
N. C.
Carrejo
,
F.
Ye
,
J. T.
Robinson
, and
C. F.
Landes
,
Langmuir
34
,
6697
(
2018
).
49.
N. A.
Moringo
,
H.
Shen
,
L. J.
Tauzin
,
W.
Wang
, and
C. F.
Landes
,
Langmuir
36
,
2330
(
2020
).
50.
M. J.
Skaug
,
J.
Mabry
, and
D. K.
Schwartz
,
Phys. Rev. Lett.
110
,
256101
(
2013
).
51.
L. J.
Tauzin
,
H.
Shen
,
N. A.
Moringo
,
M. H.
Roddy
,
C. A.
Bothof
,
G. W.
Griesgraber
,
A. K.
McNulty
,
J. K.
Rasmussen
, and
C. F.
Landes
,
RSC Adv.
6
,
27760
(
2016
).
52.
N.
Rozlosnik
,
M. C.
Gerstenberg
, and
N. B.
Larsen
,
Langmuir
19
,
1182
(
2003
).
53.
R.
Yamada
,
H.
Wano
, and
K.
Uosaki
,
Langmuir
16
,
5523
(
2000
).
54.
M.
Wang
,
K. M.
Liechti
,
Q.
Wang
, and
J. M.
White
,
Langmuir
21
,
1848
(
2005
).
55.
R.
Resch
,
M.
Grasserbauer
,
G.
Friedbacher
,
T.
Vallant
,
H.
Brunner
,
U.
Mayer
, and
H.
Hoffmann
,
Appl. Surf. Sci.
140
,
168
(
1999
).
56.
S.
Choi
and
J.
Chae
,
J. Micromech. Microeng.
20
,
075015
(
2010
).
57.
I.
Doudevski
and
D. K.
Schwartz
,
J. Am. Chem. Soc.
123
,
6867
(
2001
).
58.
Q.
Yu
,
Y.
Zhang
,
H.
Wang
,
J.
Brash
, and
H.
Chen
,
Acta Biomater.
7
,
1550
(
2011
).
59.
A.
Wörz
,
B.
Berchtold
,
K.
Moosmann
,
O.
Prucker
, and
J.
Rühe
,
J. Mater. Chem.
22
,
19547
(
2012
).
60.
X.
Liu
,
L.
Yuan
,
D.
Li
,
Z.
Tang
,
Y.
Wang
,
G.
Chen
,
H.
Chen
, and
J. L.
Brash
,
J. Mater. Chem. B
2
,
5718
(
2014
).
61.
I.
Banerjee
,
R. C.
Pangule
, and
R. S.
Kane
,
Adv. Mater.
23
,
690
(
2011
).
62.
Q.
Tang
,
C.-H.
Xu
,
S.-Q.
Shi
, and
L.-M.
Zhou
,
Synth. Met.
147
,
247
(
2004
).
63.
N. A.
Moringo
,
H.
Shen
,
L. J.
Tauzin
,
W.
Wang
,
L. D. C.
Bishop
, and
C. F.
Landes
,
Langmuir
33
,
10818
(
2017
).
64.
R.
Walder
,
M.
Kastantin
, and
D. K.
Schwartz
,
Analyst
137
,
2987
(
2012
).
65.
D.
Campoccia
,
L.
Montanaro
, and
C. R.
Arciola
,
Biomaterials
34
,
8533
(
2013
).
66.
M.
Aghajani
and
F.
Esmaeili
,
J. Biomater. Sci., Polym. Ed.
32
,
1770
(
2021
).
67.
F.
Wang
,
H.
Zhang
,
B.
Yu
,
S.
Wang
,
Y.
Shen
, and
H.
Cong
,
Prog. Org. Coat.
147
,
105860
(
2020
).
68.
J. S.
Weltz
,
D. K.
Schwartz
, and
J. L.
Kaar
,
ACS Nano
10
,
730
(
2016
).
69.
J. T.
Woodward
,
A.
Ulman
, and
D. K.
Schwartz
,
Langmuir
12
,
3626
(
1996
).
70.
D.
Wang
,
C.
He
,
M. P.
Stoykovich
, and
D. K.
Schwartz
,
ACS Nano
9
,
1656
(
2015
).
71.
K.
Tsumoto
,
D.
Ejima
,
A. M.
Senczuk
,
Y.
Kita
, and
T.
Arakawa
,
J. Pharm. Sci.
96
,
1677
(
2007
).
72.
A.
Hozumi
,
H.
Sugimura
,
Y.
Yokogawa
,
T.
Kameyama
, and
O.
Takai
,
Colloids Surf., A
182
,
257
(
2001
).
73.
A. J.
Hopkins
,
C. L.
McFearin
, and
G. L.
Richmond
,
J. Phys. Chem. C
115
,
11192
(
2011
).
74.
H.
Sugimura
,
Nanocrystalline Materials: Their Synthesis-Structure-Property Relationships and Applications
(
Elsevier
,
Oxford
,
2006
), p.
57
.

Supplementary Material