Surface Characterization and in situ Protein ... - ACS Publications

Subscriber access provided by UNIV OF WINNIPEG

Article

Surface characterization and in situ protein adsorption studies on carbene-modified polymers Geoffrey W Nelson, Emily Parker, Kulveer Singh, Christopher Francis Blanford, Mark Gerard Moloney, and John S. Foord Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.5b01644 • Publication Date (Web): 21 Sep 2015 Downloaded from http://pubs.acs.org on September 29, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Langmuir is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Surface characterization and in situ protein adsorption studies on carbene-modified polymers Geoffrey W. Nelson†‡, Emily M. Parker†, Kulveer Singh§†, Christopher F. Blanford§, Mark G. Moloney†* and John S. Foord† † Department of Chemistry, Chemistry Research Laboratory, University of Oxford, 12 Mansfield Road, Oxford, U.K., OX1 3TA ‡ Department of Materials, Imperial College London, Exhibition Road, London, SW7 2AZ, UK § School of Materials and Manchester Institute of Biotechnology, University of Manchester, 131 Princess Street, Manchester, U.K., M1 6GN KEYWORDS

Surface Modification, Proteins, Surface Chemistry, XPS, QCM-D

ABSTRACT Polystyrene thin films were functionalized using a facile two-step chemical protocol involving carbene insertion followed by azo-coupling, permitting the introduction of a range of chemical functional groups, including aniline, hexyl, amine, carboxyl, phenyl, phosphonate diester, and ethylene glycol.

X-ray photoelectron spectroscopy (XPS)

confirmed the success of the two-step chemical modification with a grafting density of at least 1/10th of the typical loading density (1014 - 1015) of a self-assembled monolayer (SAM). In-situ, real-time quartz crystal microbalance with dissipation (QCM-D) studies 1 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 43

show that the dynamics of binding of bovine serum albumin (BSA) are different at each modified surface. Mass, viscoelastic and kinetic data were analyzed, and compared to cheminformatic descriptors (i.e. c log P, polar surface area) typically used for drug discovery. Results show that functionalities may either resist or adsorb BSA, and uniquely influence its adsorption dynamics. It is concluded that carbene-based surface modification can usefully influence BSA binding dynamics in a manner consistent with, and more robust than, traditional systems based on self-assembled monolayer (SAM) chemistry.

1.

INTRODUCTION Surface functionalization of organic and inorganic materials using carbene-based molecules

has been known since at least the late 1960’s with a number of patents referenced by the literature.1

Early studies used carbene-based chemicals which were volatile and highly

reactive, as their synthesis or related modification strategies required high temperatures or metal catalysts.1 For these reasons, carbene insertion has not found widespread application to form functional monolayers. Instead, other chemistry has dominated research in this field, such as the past 30 years of research into thiol-based self-assembled monolayers on metal surfaces (e.g. gold).2 A new generation of stable carbene compounds has fundamentally changed the field. These particular carbenes do not suffer the disadvantages of their earlier analogues. It is now possible to directly functionalize both inorganic and organic materials in a simple, robust manner. The N-heterocyclic carbenes3, diazrines4, and diarylcarbene derivatives1,

9, 10

are

frontier molecules for this type of chemistry. In a radical departure from thiol-based SAMs, N-heterocyclic carbenes have been shown form SAMs on Au via the creation of chemically resistant Au-C bonds5, 6; however, the bulky side-groups necessary to stabilize these carbenes may be a steric limitation to their use. Photo-activated diazrines generate carbenes capable of 2 ACS Paragon Plus Environment

Page 3 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

insertion into organic molecules; however, their synthesis is considered complex.4 Diarylcarbenes are less sterically hindered, more easily synthesized, and are noted for their ability to functionalize a wide range of organic materials (e.g. polymers, diamond, cotton, etc.) via the creation of robust C-C bonds at the surface.1, 9, 10 If the stable carbene is attached to a chemical linker, then the surfaces can be further functionalized in a manner analogous to ‘click-chemistry’7 or polymer grafting8, thereby widening their potential uses. These carbene attachments are irreversible and chemically resistant (e.g. to oxidation) due to covalent C-Au or C-C bonds. The one-step carbene insertion reactions are a general means to modify a range of organic and inorganic materials with promising applications, such as: organicinorganic hybrid materials and devices; the rapid formation of biocidal surfaces; reliable and long-lasting protective coatings.1,5,6, 9, 10 Of most interest in this work is the bio-material application of diarylcarbene chemistry.9 The further functionalization of inserted diarylcarbene-based linkers by diazonium chemistry can form functional organic surfaces (i.e. bio-compatible, fluorescent, etc.) at controlled loading levels of up to 3 x 1014 molecules cm-1.1,

9, 10

The choice of terminal surface

chemistry is influenced by previous studies linking specific moieties with protein adhesion or resistance. For instance, poly(ethylene glycol) grafted onto polystyrene (PS) prevents the adsorption of fibrinogen and IgG, as well as the adhesion of Streptococcus mutans11; a similar effect has been shown on stainless steel.12

Detailed studies of the binding of

fibronectin on SAMs attached to Au surfaces has demonstrated that binding affinities vary with surface chemical functionality in the order OH > COOH > NH2 > CH3.13, 14 Tertiary amine oxide modified surfaces have recently been shown to resist nonspecific protein adsorption.15 The first real time and in situ study of protein adsorption on an exemplar organic material carbene-modified polystyrene - is undertaken here using quartz crystal microbalance with 3 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 43

dissipation (QCM-D), a recent technique developed to probe biomolecular adsorption phenomena.16 QCM-D obtains data in situ having time resolution below one second, mass sensitivity of ≈ 0.5 ng cm2, 17 and is surface sensitive to viscoelastic changes occuring within ≈ 250 nm of the interface - becoming less sensitive with the square root of the harmonic number of the frequency data.18

The extent to which the newly introduced surface

functionality influenced the behavior of a model globular protein, bovine serum albumin (BSA) was investigated on . By simultaneously monitoring frequency and dissipation at the interface of an oscillating quartz crystal, the mass and viscoelastic properties of adsorbed BSA layers was obtained, in addition to their adsorption kinetics. In addition, this work characterises the nature of the carbene-based surface modification by X-ray photoelectron spectroscopy (XPS). It is demonstrated that stable diarylcarbene molecules anchor chemical functionalities capable of affecting protein adsorption in a manner consistent with existing self-assembled monolayers on metals, inorganic materials and polymers.

2.

EXPERIMENTAL METHODS Chemicals. Reagents used in this work are described in Section S1. QCM-D studies

required the use of 0.1 M phosphate buffer (PB) made from K2PHO4●H2O, KH2PO4 (pH 7.4, both Sigma Aldrich). Supporting Substrates. Three supporting substrates were prepared: Au-coated Si squares (1 cm2), Au-coated QCM-D crystals (QSense), and spin-coated thin films of PS on either of the aforementioned substrates. Details concerning the synthesis and properties of these substrates can be found in Section S2 of the electronic supplementary information (ESI). Preparation of PS Beads and Thin Films. Two types of PS materials were studied: PS beads (Sigma Aldrich, Amberlite XAD-4, 20-40 mesh) and spin-coated thin films made from PS powder (Sigma Aldrich, MW = 211,600). PS was cross-linked using two methods: UV 4 ACS Paragon Plus Environment

Page 5 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

radiation (broadband source, 30 minutes)19 and by wet chemistry using divinylbenzene (DVB) and a thermally-activated initiator, benzoyl peroxide (BPO).20 Both cross-linking methods stabilized PS thin films towards the chemistry in Scheme 1 and enabled XPS studies, but only the DVB cross-linked films were stable for QCM-D use. All spin-coated thin films were < 100 nm thick, as confirmed by XPS and ellipsometry (not shown). Further details concerning these substrates can be found in Section S3 in the ESI. Scheme 1. Modification of polystyrene with carbene and diazo-coupling chemistry

(A) Modification of materials with carbene chemistry; (B) modification of polystyrene surfaces; (C) modification of polystyrene with bis(4-iodophenyl)diazomethane; conditions: (i) coat with a 1 w/v% solution of 1 in DCM, evaporate solvent, heat to 180 °C, wash with DCM; (ii) soak in diazonium salt solution, wash with EtOH and DCM; (iii) coat with a 1 w/v% solution of 4 in DCM, heat to 93, 120 or 180 °C, wash with DCM; (D) nitro-terminated polystyrene. Carbene Modification Chemistry. The PS materials were modified using carbene-based chemistry, furnishing functional PS surfaces that could react with diazonium salts, resulting 5 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 43

in a range of surface terminations. Scheme 1A displays the general process of ‘carbene insertion’, in which heating a diaryl diazo compound in the presence of a polymer surface results in the loss of nitrogen and the generation of a highly reactive carbene species which insert into the coated surface.1,

9, 10

The method relies upon the high level of reactivity of

carbenes with a variety of functional groups to generate new covalent bonds; this gives the approach the potential to be applicable to a wide variety of surfaces. Scheme 1B displays the development of the basic approach into a two-step method employed to prepare the library of PS materials, using diazonium coupling to achieve further chemical elaboration. Thus, the materials were first modified by inserting diaryldiazo 1 into PS to make surface 2, followed by a further diazonium coupling reaction to achieve specific chemical functionality (see surfaces 3a – 3f). Details regarding the methodology for modification of Amberlite XAD-4 beads have been reported.9 The PS films were coated with a solution of diazo 1 in DCM (dichloromethane) (approx. 0.2 ml per 1 cm2 sample, 1 w/v%) by allowing the DCM to evaporate, resulting in very pale pink colored films. The materials were heated at 180 °C until the pink color was lost, which typically took 5 minutes for each sample. After cooling, this material 2 was washed with DCM then dried under a gentle stream of nitrogen. The films were then soaked in the relevant diazonium salt solution in ethanol1 for 18 hours at ≈ 5 °C, washed with ethanol and DCM, and then dried with N2 gas. As a result, materials 3a-f (Scheme 1B) had different surface functionalities, comprising phosphonate diester (3a), amine (3b), glycol (3c), phenyl (3d), carboxyl (3e) and hexyl (3f) groups. Some PS films were also modified by bis(4iodophenyl)diazomethane (4, Scheme 1C) prepared using literature procedures and previously used on diamond10 (kindly provided by Dr Jon-Paul Griffiths (Oxford Advanced Surfaces).

6 ACS Paragon Plus Environment

Page 7 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

X-ray photoelectron spectroscopy. XPS experiments were conducted at 10−9 − 10−10 Torr, having an Al Kα source (1486.6 eV) with 10 kV anode potential and 10 mA emission current. All peak fitting was done using XPS Peak Fit (v. 4.1) software, using common protocols (e.g. Shirley backgrounds). Reported binding energies have an error of ± 0.25 eV, based on calibration to the C1s peaks. The reported elemental ratios (e.g. N1s/C1s) have ≈ 15% error and were computed using atomic sensitivity factors.21 Protein Adsorption Studies by QCM-D. QCM-D is a variation on the well-known quartz crystal microbalance (QCM) technique which is useful for protein adsorption studies.16, 22, 23 For thin and rigid overlayers, the change of oscillation frequency (f) is proportional to added mass, by the Sauerbrey equation24; the resultant mass is labelled as Sauerbrey mass (mSauerbrey).

However, this approach underestimates the mass of non-rigid layers (i.e.

hydrated proteins), as it does not account for energy loss processes at the interface (e.g. viscoelasticity).18 To overcome these limitations, frequency and energy losses (dissipation, D ) are monitored simultaneously by QCM-D, with adsorbed mass estimated from the ‘Voigtmodel’ (mVoigt), as developed by Voinova et al.18 When the signal to noise ratio was good and iterative fits reasonable, this approach was used, otherwise the Sauerbrey approach was used. The QCM-D protocols used here are similar to other protein-based QCM-D studies,25, 26 particularly that by Singh et al.17 Full details are found in Section S4 and a brief description follows. Experiments were conducted in a Q-Sense E1 module (Q-Sense, Sweden) at constant temperature (25 °C). Clean, gold-coated AT-cut quartz crystals (QSense) were mounted directly or, where necessary, spin-coated with unmodified or modified DVB-crosslinked PS thin films prior to mounting. All solutions were degassed with a rough vacuum pump for 10 minutes. Protein-free PB was flowed (0.1 mL min-1) overnight, until f and D traces stabilized. Substrate exposure to BSA occurred after switching to BSA-containing PB (0.8 g L-1). Afterwards, the QCM-D cell was rinsed with BSA-free PB, followed by Hellmanex solution 7 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 43

(2% v/v H2O, Hellma Analytics) to remove BSA contamination from the module, tubing and crystal surface. The ∆f and ∆D data was collected and fit to the Sauerbrey and Voigt models using QSoft and QTools software packages (QSense). The fixed and variable modelling parameters are found in Section S4. A one-layer Voigt model described protein adsorption onto gold, whilst on PS it was better described by the two-layer Voigt model with the shear modulus and viscosity of PS constrained to be representative of a rigid, cross-linked PS layer. Protein Adsorption Kinetics. BSA adsorption rates (dm/dt) were plotted as a function of adsorption mass (m); these plots have three characteristic regions: transport-limited, reactionlimited/intermediate , and saturation (vide infra) .27 The reaction-limited region of these plots can be evaluated qualitatively (shape)27 and quantitatively from a linear fit to the reactionlimited region, using the following expression28, 29, 30: ௗ௠ ௗ௧

= ݇௔ ܿ௕ (1 + ‫ܥ‬ଵ ݉)

(Equation 1)

where ka is the rate constant, cb is the concentration of protein in the bulk fluid, and C1 is a constant related to steric factors of the adsorbed proteins. While there exist several different kinetic interpretations of protein adsorption onto solid surfaces,31 kinetics results were interpreted using two basic adsorption models27, 31: the random sequential adsorption (RSA) model and the Langmuir model. The former assumes immobile protein adsorption with excluded space, whilst the latter assumes proteins can move after adsorption and have no excluded space.

A 50-point adjacent-average smoothing (≈ 14.5 seconds of experimental

time) was used to smooth mSauerbrey data for kinetic analysis.

3.

RESULTS AND DISCUSSION XPS Characterization of Modified Polymers. The successful functionalization of PS

surfaces using diaryl-carbene and diazonium coupling chemistry was characterized by XPS. Representative survey scans of gold and PS substrates are shown in Figure 1A. Photoelectron 8 ACS Paragon Plus Environment

Page 9 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

signals expected for PS (i.e. C1s, O1s) are seen and there is attenuation of the Au 4f peaks, further confirming its presence as an overlayer.32 I, N and P signal is absent on unmodified PS, enabling them to act as chemical markers to confirm further chemical modification. Using the carbene-insertion protocol, PS was modified with compound 4, containing two iodide groups; iodine is a useful element to follow as it is unique to compound 4 and has a large photoelectron cross-section. Compound 4 was activated at PS interfaces to create surface 5 - with different reaction temperatures and times – and washed with DCM to remove any weakly bound species. Iodine photoelectron signals appear in Figure 1B, consistent with the covalent attachment of compound 4 to PS. On modified substrates, the I3d5/2 and I3d3/2 photoelectron signal appeared at ≈ 620.1 eV and ≈ 631 eV, respectively.10 Each spin state has a profile showing iodine to be in two chemical environments. Peak i in Figure 1C is assigned to chemisorbed iodine, whilst peak ii is signal originating from physisorbed iodine; these assignments follow Wang et al.10 who attached compound 4 onto diamond substrates. From relative peak intensities, one estimates that ≈ 80% of the iodine on the surface is in the chemisorbed form, with the remaining iodine signal arising from residual physisorbed molecular dimers. The grafting densities of iodine were calculated (see Section S5). Table 1 shows their magnitude to be consistent with ≈ 1/10th of a monolayer of compound 4. The grafted molecule is relatively large and is likely to be in a closed-packed arrangement, as discussed by Wang et al.10 Notably, the I3d peak intensities, I3d5/2/C1s ratios, and grafting densities all increase with higher activation temperatures, despite ever lower reaction times. This is consistent with significant activation energy for the carbene insertion. Also, surface diffusion increases at higher temperatures, thus increasing the probability of reactions between molecules and reactive surface sites.33, 34


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 43

Figure 1. (A) XPS survey scan spectra of various supporting substrates and polystyrene overlayers: (A, i) Au; (A, ii) 2% DVB-modified polystyrene on Au; (B) survey scan spectra of UV-modified polystyrene thin films modified by bis(4-iodophenyl)diazomethane; (C) narrow scan spectra of the I3d region after bis(4-iodophenyl)diazomethane has been deposited onto UV-modified polystyrene thin films then cured 120 °C and 180 °C for 20 and 4 minutes, respectively. Physisorbed (C, i) and chemisorbed (C, ii) iodine is labelled; (D) N1s region and curve fits for XAD-4 polystyrene beads with surface terminations: R=NO2 (6) and R=NH2 (3b). Curve fits for the N1s signal originating from nitro (blue) and amine (red) are shown; (E) N1s region of DVB cross-linked polystyrene on two underlying substrates: gold (phosphonate diester, 3a) and glass slides (3b, 3f; amine and hexyl); (F) O1s region of DVB cross-linked polystyrene with two underlying substrates used: gold (2, 3a, 3e, 3f; aniline, phosphonate, carboxyl and hexyl) and glass slides (3b, 3e; amine and carboxyl). Within each sub-figure, all plots are on the same arbitrary y-axis. Table 1. XPS ratios and grafting densities for diphenyl-carbenes adsorbed at surface 5 ∆stepA (°C)

I3d5/2/C1s

Grafting Density

Time of Reaction 10

ACS Paragon Plus Environment

Page 11 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

(1013 cm −2)

(min.)

93

0.005

3.2

40

120

0.011

5.4

20

180

0.018

8.7

4

The success of the azo-coupling ultimately leading to a terminal chemical functionality was characterized by following the O1s, P2p, and particularly, N1s signals. Azo-coupling leads to ever increasing N1s/C1s ratios (see Table S1 and S2) and distinctive N1s photoelectron signal in the case of amine (3b) and amino-terminated (6) surfaces (see Scheme 1D and Figure 1D). The N1s signal at ≈ 400.5 eV and ≈ 405 eV originates from nitrogen in amine and nitro environments, respectively.10,

34

The unexpected amine peak on nitro-terminated

surfaces relates to the reduction of nitro groups by secondary electron flux produced during the XPS experiment.10, 34 Further evidence of successful azo-coupling is seen in Figure 1E where one observes negligible N1s signal for non-nitrogen containing surfaces [e.g. hexyl (3f) and phosphonate (3a)]. Phosphonate diester termination (3a) was confirmed by the presence of P2p photoelectron signal at ≈ 133.5 eV (not shown).35, 36 Oxygen-rich surfaces (cf. 3a, 3c, 3e) have high O1s/C1s ratios (see Table S2), substantial O1s signal at ≈ 532 eV (ether) and ≈ 534.5 eV (carbonyl/carboxyl)19, having binding energies reflecting their bonding character (e.g. P-O, C-O, etc.) and electronegativity (see Figure 1F). However, atmospheric sample ageing and surface charging may affect these results.19,

37

For some

functionalities, grafting densities were calculated from N1s and I 3d5/2 peak intensities (see Table 1 and Section S5).10 The grafting densities of NH2 and NO2 onto PS beads is 10.6 x 1014 cm −2 and 1.8 x 1014 cm −2, respectively (cf. Table 1). The magnitude of these values is consistent with a closed-packed monolayer. The above suggests that carbene insertion and azo-coupling chemistry successfully modified PS.


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 43

QCM-D Studies of Protein Behavior at Modified Surfaces. Typical QCM-D traces for BSA adsorption onto PS are shown in Figure 2A and 2B for f and D data, respectively. These figures show four experimental stages: first, the establishment of a stable baseline under flowing conditions in protein-free PB (2A, i); second, the decrease and increase of f and D respectively as the PB containing BSA was introduced, with saturation adsorption reached after ≈ 20 minutes (2A, ii); third, rinsing of the substrate with protein-free PB(2A, iii); and finally, cleansing of the surface with surfactant accompanied by f and D data rising above baseline values as protein and PS is stripped off the quartz crystal (2A, iv).

Figure 2. Frequency (A) and dissipation (B) QCM-D traces for a protein adsorptiondesorption experiment on unmodified DVB cross-linked polystyrene using data from the 7th harmonic; (C) Sauerbrey mass of the BSA deposition and rinsing with the start BSA deposition normalized to t = 0 seconds. The solutions used for A to C are: 0.1M PB (i, baseline; iii, rinsing); 0.08 g L–1 BSA in PB (ii, protein adsorption); 2 % v/v Hellmanex in aqueous solution (iv, cleaning).


Page 13 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Figure 3. Average Sauerbrey mass calculated with f values from the 5th, 7th, and 9th harmonics at various times during (A) adsorption or (B) rinsing phases. Sample labels are as follows: (Au) gold; (PS) polystyrene; (2) aniline; (3a) phosphonate diester; (3b) amine; (3c) ethylene glycol; (3d) phenyl; (3e) carboxyl; (3f) hexyl. The obscured data associated with 3d in B shows change of mass of 0.05, 0.25, 0.28 % after 5, 10, and 20 minutes of rinsing.

Figure 4. (A) ∆D vs. ∆f (9th harmonic) for Au, DVB-PS, and functionalized DVB-PS during BSA deposition (cf. Figure S1); (B) ∆D vs. ∆f (9th harmonic) associated with the rinsing of protein layers on various as-prepared substrates (cf. Figure S1); (C) ∆D/∆f ratios (9th harmonic) for BSA layers on the studied substrates, where the data represents the ∆D/∆f ratio at saturation (black) and the rinsing data (red) after 20 minutes of rinsing. Sample labels are


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 43

as follows: (Au) gold; (PS) polystyrene; (2) aniline; (3a) phosphonate diester; (3b) amine; (3c) ethylene glycol; (3d) phenyl; (3e) carboxyl; (3f) hexyl. The protein adsorption and rinsing frequency data was fitted to the Sauerbrey model using three different frequency harmonics, as seen in Figure 2C. As expected, mass increases until saturation as BSA is adsorbed onto the interface, and it slowly decreases during rinsing as loosely bound protein is removed. Deviations between the harmonics of the f traces can be seen, particularly during saturation and rinsing stages; they are a consequence of increasing non-rigidity in protein overlayers during the experiment. Noisy data between the dotted lines in Figure 2C is due to mechanical vibrations caused by manually switching buffer solutions within the QCM-D module. The above QCM-D responses are expected and demonstrate the as-prepared PS substrates to be stable under the experimental conditions. Application of the Sauerbrey Equation to Protein Adsorption Data. The 5th, 7th, and 9th harmonics gave similar results, suggesting the presence of fairly uniform protein layers at the interface. For single adsorption runs for each substrate, frequency dependence or drift in the measurements was reduced by applying an averaging procedure to the data: a 50-point adjacent average was applied to data from each of these harmonics around a given time point; these individual results were then averaged to represent the Sauerbrey mass at that time point. The Sauerbrey mass at the end of either the adsorption or rinse stages (‘endpoint’ mass) was calculated by first averaging the last 300 points of either stage for each harmonic; these individual results were then averaged to represent the mass at the end of the experimental stage. The difference between the Sauerbrey mass at the end of the adsorption phase and the rinsing phase is reported as a percentage change. Run-to-run reproducibility was better than 20 ng cm−2, so mass changes bigger than this reflect real changes at the interface. The results of Figure 3A show that imparted surface functionalities affect the extent of BSA adsorption. The initial modification of PS with the aniline-terminated chemical linker 14 ACS Paragon Plus Environment

Page 15 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

(2) decreased the amount of adsorbed protein. Subsequent modification with phosphonate diester (3a) and phenyl (3d) further lowered the adsorption mass. By contrast, amine (3b), glycol (3c), carboxyl (3e), and hexyl (3f) terminations increased adsorption mass. The high mass of BSA at surface 3c was unexpected, as SAMs and polymers terminated with poly(ethylenoxy) chains should resist protein adsorption.38, 39 This increased mass is likely to be due to hydration.40 The water is tightly bound and is part of the rigid ‘at surface’ layer of which changes are best observed by the Sauerbrey model. The effect of surface termination on the amount of BSA retained at the interface is shown in Figure 3B. Protein layers are considered to be bound irreversibly if mass densities change < ≈ 10 % after a rinse.23, 26 Irreversibly bound BSA layers are present on Au, PS, 3d, 3e and 3f (phenyl, carboxyl, and hexyl). In these cases, any change in mass density is a simple consequence of a flowing solution over the sample, with loosely bound proteins being forcibly removed and with the majority of bound matter retained by significant binding forces. Mass changes ≥ ≈ 10 % are due to reversibly bound protein layers23, 26, which is the case for BSA on surfaces terminated by aniline (2), glycol (3c), and phosphonate diester (3a) functionalities. Mass removal is slow at the phosphonate diester and glycol surfaces (3a and 3c). This may due to time-dependent reversibility induced by surface hydration shells.9, 39 Surprisingly, the amine (3b) functionality is associated with added mass. This result suggests the incorporation of loosely bound BSA into more strongly bound sub-layers, due to strong surface-protein affinity. The hydrogen acceptor character of amines39 and their ability to bind to the polar binding sites of BSA38, 41, 42 may facilitate the above. The magnitude of mSauerbrey in Figure 3A and 3B is comparable to that for various types of BSA monolayers: close packed (CP), RSA adsorption, or in a hydrated form (see Table S4). This comparison suggests at least a monolayer of BSA has formed at the modified PS interfaces.


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 43

Application of the Voigt Equation to Protein Adsorption Data. Proteins at solid-liquid interfaces are normally viscoelastic and better described by the Voigt model. We applied this model to the QCM-D data in a manner analogous to the Sauerbrey modelling. The results are shown in Table S3. A comparison of the mass results from each model (cf. Figure 2D and Table S3) shows that the trends in mass uptake are similar (vide supra). However, the Voigt model could not be consistently applied to all data, thus limiting its wider use here; instead the Sauerbrey data is used more extensively. The Voigt model overestimates the mass when the elastic component dominates the viscous component (i.e. 2πf [η/µ]) < ≈ 1) and underestimates mass in the opposite case.25 The former scenario is likely occurring in the case of the glycol (3c) and hexyl (3f) systems, perhaps due to the presence of extended surface hydration; for example, by hydration shells or multilayer formation.25, 43 Dissipation vs. Frequency Plots (‘∆D vs. ∆f plots’). The viscoelastic properties of protein layers are examined by plotting ∆D as a function of ∆f as done in Figure 4A-C (see Figure S1 for an alternative presentation). The ∆D vs. ∆f plots directly probe the viscoelastic changes within protein layers, essentially as a function of the adsorbed mass. The magnitude of dissipation is affected by three factors: protein structure, surface hydration, and layer thickness. The interplay of these factors during protein adsorption leads to a series of linear regions in the ∆D vs. ∆f plots, each representing a protein configuration having different viscoelastic properties.25 The sign and magnitude of the gradient and the ∆D/∆f ratio are important parameters when discussing viscoelastic changes; positive and negative values indicate the formation of more rigid or looser protein layers, respectively.44 In addition, similar values are indicative of similar overlayer structures for comparable viscoelastic regimes.44 With the above in mind, the ∆D vs. ∆f plots can be analyzed with reference to other QCM-D studies of BSA on gold and PS surfaces.41, 45, 46


Page 17 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

The differing line-shapes in Figure 4A directly track the influence of surface functionality on the viscoelasticity of adsorbed BSA.22,

45

Viscoelastic behavior in Figure 4A can be

broadly placed within four groups: Group A has linear regions with high D and includes PS and hexyl (3f); Group B has linear regions with intermediate D and includes Au, aniline (2), and phenyl (3c); Group C also has linear regions with intermediate D, but has decreasing D during later adsorption processes [i.e. phosphonate (3a), glycol (3c)]; finally, Group D is uniquely associated with BSA deposition on amine and carboxyl terminated surfaces (3b, 3e), sharing the common feature of constant D with respect to f during some, if not most, of the adsorption phase. These groups are associated with several behaviors: (i) Group A: results consistent with BSA adsorption in a single, loose conformational state, with multilayer formation being possible, as shown by linear plots with high D throughout adsorption22, 26; (ii) Group B: intermediate dissipation, with BSA layers having intermediate rigidity or partial multilayer formation22, 26; Group C: the removal of loosely bound proteins, dehydration of the layer, change of protein orientation, and/or a change in the structure of the adsorbed protein layers, as shown by loss of dissipation with added mass at the later stages of adsorption22, 46; (iv) Group D: strong binding between functional groups and the protein, forming rigid protein layers (cf. lysozyme on PAA47), as shown by curved plots having regions of constant dissipation. Feiler et al.22 note that protein layers with similar ∆D/∆f ratios have similar layer or protein structure. Therefore, we calculated the ∆D/∆f ratio after 20 minutes of adsorption from the data in Figure 4C. The results shown in Figure 4C can be grouped with decreasing ∆D/∆f as follows: (i) Group 1: 3e (carboxyl); (ii) Group 2: 3b, 3c (amine, glycol); and (iii) Group 3: 3a, 3d, and 3f (phosphonate diester, phenyl, and hexyl). The implication is that 3e is the most rigid BSA layer and 3a, 3d, and 3f are the loosest.

One concludes that surface

termination may be capable of influencing the rigidity of the BSA layer at saturation. It is


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 43

curious that phosphonate diester (3a) and phenyl-terminated surfaces (3d) have dissimilar viscoelastic behavior, but similar layer structure at saturation. In addition, the viscoelastic behavior of BSA on amine and carboxyl surfaces (3b, 3e) has Group D behavior, yet the former has looser layers at saturation. If the lineshapes of Figure 4A are considered a measure of the viscoelastic changes during an adsorption event, then surface functionality is capable of influencing both the adsorption mechanism and the final protein arrangement at the interface. This conclusion does not apply to all substrates, as exemplified by the case of PS and hexyl (3f), which share the same viscoelastic behavior and rigidity at saturation. The orientation of BSA during initial adsorption can be determined from Figure 4A by replotting the data in the form of Figure S1. Now, the gradient changes are seen more clearly, with the first such change labelled in each sub-figure. This point represents the formation of an initial monolayer of BSA. The adsorption mass of this first monolayer can be compared to those found in Table S4 for various types of BSA monolayers. It is clear that BSA has a flat orientation during initial binding to all surfaces, except for the amine-terminated surface on which it adsorbs orthogonally. By adopting the same approach, one finds that at saturation, BSA has some orthogonal orientation on all surfaces except for the phosphonate diester (3a) and phenyl (3d) terminations.

Therefore, some ∆D vs. ∆f behavior during adsorption

represents changes in protein orientation. Figure 4B shows the effect of the chemical modifications on the viscoelastic behavior of BSA during and after rinsing. Two clear viscoelastic behaviors can be identified. If there is little change in ∆D or ∆f during the rinse (e.g. carboxyl, 3e), irreversibility in the adsorption process is implied. Alternatively, when a gradual decrease in both ∆D and ∆f occurs (e.g. phosphonate diester, 3a), there exists some reversibility in the adsorption process. There are two unusual plots in Figure 4B. Compared to those surfaces promoting reversibility, rinsing of BSA at the Au interface has a similar change in ∆f, but without the


Page 19 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

same change in dissipation (i.e. more negative ∆D/∆f ); in this case, the loss of protein leads to a looser protein layer, rather than one which is more rigid. The amine terminated (3b) surface promotes increased magnitude of both ∆f and ∆D during rinsing. This cannot arise from further protein adsorption, since no protein is present in the rinsing solution, and most likely represents a change in configuration in the existing bound layers, which may be leading to the formation of a looser, more hydrated structure. Surfaces with similar rigidity after 20 minutes of rinsing can be identified from Figure 4C using averaged ∆D/∆f from the harmonic data, and are: (i) Group 1: 3c, 3e (glycol and carboxyl); (ii) Group 2: 3a (phosphonate diester) (iii) Group 3: PS, 3b, and 3d (amine and phenyl), and (iv) Group 4: 3f (hexyl). In general, the ∆D/∆f ratio has magnitude Group 1 > Group 2 > Group 3 > Group 4 with BSA layers being most rigid on Group 1 surfaces and loosest on Group 4 surfaces. The groups do not depend on reversibility, thereby suggesting different processes during rinsing can lead to similar outcomes. Protein Adsorption Kinetics. The kinetics of BSA adsorption was determined by plotting protein adsorption rate as a function of mass density (dm/dt plots, see Figure 5A-I).28, 29, 31 These dm/dt plots have three regions48, as delineated in an exemplar, Figure 5E, by dashed lines: (i) a transport-limited regime at short time-scales; (ii) a reaction-limited regime at intermediate reaction times; (iii) saturation at long reaction times where dm/dt tends to zero. Line-shapes having decreasingly negative gradients in the intermediate regime are indicative of RSA kinetics, whereas those which are linear are consistent with Langmuir-like adsorption.27 In Figure 5, we see neither type of lineshape in its pure form within the intermediate regime which suggests a mixture of immobile and mobile proteins with some, but not total, excluded space. However, the most prominent lineshape throughout Figure 5 is a convex one - as viewed from the x-axis – and so RSA kinetics will be assumed for the purpose of linearly fitting the reaction-limited regions to Equation 1. From this exercise, the


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 43

kinetic parameters ka and C1 are obtained (see Figure 5J). This analysis method is consistent both with Brusatori et al.29 who describe the adsorption of human serum albumin onto interfaces using optical waveguide light spectroscopy and Anand et al. who studied globular proteins at SAMs.28 Several points are noteworthy: ka is a direct measure of protein uptake at low coverage and gives the best correlation at short reaction times (Figure S2A); in this regime protein-protein interactions are insignificant. C1 reflects how adsorbed protein slows the subsequent rate of uptake of further protein species; if this process is simply due to blocking of the available surface area by an adsorbed molecule, the magnitude should be proportional to the surface area used (Aad). This is indeed the case in Figure S2B, which compares calculated Aad with C1 (see also Section S8). Therefore, footprint size could be calculated from adsorption mass. The calculated BSA footprint sizes (6 to 62 nm2 per molecule) compare well to literature values (16 - 70 nm2 per molecule, depending on orientation or structure).23, 49

Figure 5. dm/dt vs. mSauerbrey (9th harmonic) plots for BSA adsorption on variously modified polystyrene substrates: (A) Au, (B) polystyrene, (C) aniline 2, (D) phosphonate diester 3a, (E) amine 3b, (F) ethylene glycol 3c, (G) phenyl 3d, (H) carboxyl 3e, (I) hexyl 3f . Figure 20 ACS Paragon Plus Environment

Page 21 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

5E is the best exemplar of the three regimes of protein deposition reflected in the dm/dt vs. m data, namely: (5E, i) transport-limited , (5E, ii) reaction-limited, and (5E, iii) saturation; (J) plot of C1 versus ka with notes relating each to footprint size and protein-surface affinity, respectively. The error in both sets of data are 2-3 orders of magnitude smaller than the reported values. ‘Quadrant’ labels for Figure 5J (i-iv) are in the corners of the plot. Sample labels for Figure 5J are the same as mentioned above for Figures 5A-I. Surface functionality influences both the affinity BSA has for the modified surface and its footprint, once adsorbed. This is best seen in Figure 5J, which is sub-divided into four quadrants (i-iv), with each corresponding to a different set of protein behaviours. A surface with a large affinity for BSA could cause the protein to adopt a large footprint as it unravels or adopts a flat orientation due to strong protein-surface interactions. This scenario is represented by Quadrant iv in Figure 5J; no modification promotes this behavior. On a surface with low affinity for BSA, one might expect BSA to maintain a small, native-sized footprint, since the protein-surface interactions are not strong enough to induce major conformational change. This behavior is represented by Quadrant i and the functionalities within are known to promote some or all of the following: native state, orthogonal orientation, and multi-layers.23, 49 However, BSA can adopt a large footprint on those surfaces for which it has weak affinity, as seen in Quadrant iii. Large footprint size on surfaces with weak protein-surface interactions are more likely to originate from the protein layer itself, as it attempts to maximize protein-surface interaction. For example, changes of orientation or conformation due to strong protein-protein interactions are likely occurring at the phenyl (3d) and phosphonate diester (3a) terminated surfaces.50 The preservation of small BSA footprint on surfaces for which it has high affinity is not an obvious behavior. Yet, it is induced by all the functionalities in Quadrant ii. Various phenomena could be constraining BSA size and


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 43

vertical orientation on Quadrant ii functionalities: specific binding of BSA to the surface, clustering and aggregation due to protein-protein interactions, and/or multilayer formation. Surface Character and Cheminformatic Metrics. The surface functionalities applied to PS in this work may be characterized by various chemoinformatic metrics (e.g. molecular size, polarizability, etc.) and which in turn may be used to classify the interaction between the protein and the surface functionality, as done previously for carbene-modified PS beads.9 In that previous work, contact angle measurements determined that hydrophilicity is a feature of protein resistant surfaces.39 The range of contact angles observed here is ≈ ±8% around the median of 91.1° (see Table S5). Contact angle and mSauerbrey for substrates which do not resist BSA adsorption were compared (i.e. surfaces 2, 3b, 3c, 3e, 3f). After 5, 10 and 20 minutes of BSA adsorption, the R2 values were 0.41, 0.54, and 0.48 and the Pearson’s r values were 0.64, 0.73, 0.69. Thus, there exists a moderately strong, positive correlation between contact angle and mSauerbrey for non-resistors substrates here, which is expected. The two most hydrophilic surfaces 2 and 3a (aniline and phosphonate terminated) have low adsorption mass and the most hydrophobic sample, hexyl terminated (3f), has high adsorption mass. One notes the correlation to be only moderate and that unique protein behavior is seen at each substrate despite small variance in contact angle.

This observation suggests that differences in

wettability may not completely describe the observed protein behaviours seen here. Notably, the acid-base character of the terminal R-group did not have an effect on BSA binding. Phenyl (3d) and hexyl (3f) groups, with similar acid-base character, have very different adsorption mass, while carboxyl groups (3e) having negative charge under these experimental conditions adsorb with similar mass as a neutral surface (2). Phosphonate diester (3a) has a small positive charge and has similar binding as that for a phenyl group (3d) having a neutral charge. An amine group (3b) is a weak base (i.e. proton acceptor), but has similar adsorption mass to glycol (3c) and hexyl (3f) which have no basic character.


Page 23 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Macro-scale metrics such as contact angle and acid-base chemistry may not be able to isolate nano-scale effects at the solid-liquid interface well. Here, we use a general approach to identify such effects by ‘surface metrics’ (e.g. c log P, chain length, etc.; vide supra). These molecular metrics may be able to predict protein behavior, but with the caveat that they are not experimental surface descriptors. Their use assumes a terminal surface of comprised discrete compounds.

This type of comparison has been used to success

elsewhere9, but not explicitly in a QCM-D based protein adsorption study. These ‘surface metrics’ have been calculated (www.chemicalize.org) or experimentally derived for the as-prepared surfaces (see Table S5).

The number of surface-protein

interactions, in their various forms (i.e. hydrogen bonding, van der Waals forces, etc.) and the extent of hydration shells could be affected by modifier chain length. The c log P values (octanol/water partition coefficient) is considered a measure of surface hydrophilicity and were calculated using the method of Viswanadhan et al.51 and others.9 Polar surface area (PSA) is the fraction of the surface of a compound involving polar atoms and is a measure of surface polarity. Polar surface area correlates well with the transport properties of drugs though membranes.9,

52

Figure 6 shows the relationships between time-dependent mass

uptakes on the modified surfaces studied in this work and the aforementioned surface metrics.


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 43

Figure 6. Sauerbrey mass versus surface metrics. The linear regression lines (red) in the first two columns show that binding is proportional to c log P only for those functionalities which have a high mass of bound proteins (i.e. all except 3a and 3d); this is labelled as ‘high binding only’(HBO) in the figure. Sample labels are as follows: (Au) gold; (PS) polystyrene; (2) aniline; (3a) phosphonate diester; (3b) amine; (3c) ethylene glycol; (3d) phenyl; (3e) carboxyl; (3f) hexyl. As suggested previously, BSA binding may depend on hydrophilicity.9 This is tested here by comparing mass uptake and the c log P parameter (see Figure 6, i, iv, vii). A correlation is seen for surfaces having high adsorption mass (i.e. > 200 ng cm-2 ), but phosphonate and phenyl terminated surfaces are clear outliers. If binding occurs, then adsorption mass may depend on chain length (see Figure 6, ii, v, viii). Longer chain length could influences protein adsorption in one of three obvious ways39: increasing steric hindrance to adsorption; modifying surface hydration; influencing proteinsurface bonding interactions. The first two factors tend to increase the energy requirement for protein adsorption, whilst the third decreases this requirement for favorable proteinmodifier interactions.53, 54 In Figure 6 there is no perfect correlation, yet longer chain length tends to promote binding, with the exception of surfaces 3a and 3d. The correlation seems best at high protein coverages where both protein-protein and protein-surface interactions must be important. Using ex situ methods, Choong et al.9 observed that the polarity of the terminal functionality may affect BSA adsorption from comparisons of PSA and mass uptake. The analogous comparison here is shown in Figure 6 iii, vi and ix. The data mimics that seen by Choong et al.9 with binding increasing at low PSA, dropping at intermediate PSA, and increasing at high PSA; this observation was explained by assuming that surface functionalities can bind to either polar or non-polar ‘pockets’ within BSA. One of the 24 ACS Paragon Plus Environment

Page 25 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

reasons for the lack of perfect correlation between the aforementioned surface parameters and protein binding is likely to be that BSA binding is optimal for certain functional groups, especially amines and aromatics41,

42

and this is a phenomenon which should

override average behavior. The methodologies employed here for protein adsorption study (i.e. dm/dt plots, D/f plots, and surface metrics) have not found widespread use and the results herein show that they may provide a useful description of BSA behavior at modified interface facilitated by carbene chemistry.

4.

CONCLUSIONS Polystyrene surfaces - an exemplar organic material - has been modified by a two-step method which uses carbene and azo-coupling chemistry to produce differing but controlled surface chemistry, as confirmed by XPS analysis. This work is the first to show that such a modified organic material may influence protein adsorption in a manner consistent with traditional self-assembled monolayer and/or grafting on metals, inorganic and polymeric materials. Results of an in situ QCM-D study of this carbene-modified polystyrene enabled the extent, kinetics and mechanics of adsorption of an exemplar protein (BSA) on these functionalized interfaces to be probed. Different chemical functionalizations had a marked effect on the nature of BSA adsorption, with phosphonate diester and phenyl surface terminations imparting protein resistance, while amine and carboxylic impart specific binding.

Correlation with experimental and calculated chemoinformatic parameters

(contact angle, c log P, PSA, chain length) were partially successful in accounting for the observed adsorption trends, while contact angle was less useful. The methodologies employed here for protein adsorption study (i.e. dm/dt plots, D/f plots, and surface metrics) have not found widespread use, but were found to have value for the description and prediction of protein adsorption at a solid-liquid interface facilitated by carbene-based


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 43

modification. Despite the low surface loading (at least 1/10th grafting density of SAMs), the described carbene-based modification of organic material has a marked influence on protein adsorption, consistent with studies using traditional self-assembled monolayer or grafting chemistry. Therefore, this study further confirms carbene-based chemistry to be useful, general modification for organic materials with the promise of imparting specific chemical or biological character to surfaces.

ASSOCIATED CONTENT Supplementary Information Experimental procedures and additional figures. This material is available free of charge via the Internet at http://pubs.acs.org AUTHOR INFORMATION Corresponding Author *[email protected] Author Contributions All authors contributed to the preparation of this manuscript. Primary funding for the project was provided by CB (EP/G00434X/1), JSF and MGM.

ACKNOWLEDGMENTS We thank Dr. Sofia Pascu for the use of the spin-coater. The assistance of Dr. Robert Jacobs was invaluable for the preparation of Si-Au substrates and AFM measurements. Dr. David Staunton kindly provided access to the DLS apparatus and OMNIsize software. GWN acknowledges funding received from the June Opie Fellowship (University of Auckland, NZ), the Alexander Graham Bell Association, and Access Grants (Governments of Ontario 26 ACS Paragon Plus Environment

Page 27 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

and Canada). EMP acknowledges the EPSRC and Oxford Advanced Surfaces, Plc. for funding.

GWN acknowledges insightful conversations regarding carbene-based surface

chemistry with Professor Hugh Horton of Queen’s University. REFERENCES (1) Moloney, M. G. Functionalized Polymers by Chemical Surface Modification. J. Phys. D: Appl. Phys. 2008, 41. (2) Nuzzo, R. G.; Allara, D. L. Adsorption of Bifunctional Organic Disulfides on Gold Surfaces. J. Am. Chem. Soc. 1983, 105 (13), 4481-4483. (3) Mercs, L.; Albrecht, M. Beyond catalysis: N-heterocyclic Carbene Complexes as Components for Medicinal, Luminescent, and Functional Materials Applications. Chem. Soc. Rev. 2010, 39, 1903-1912. (4) Blencowe, A.; Hayes, W. Development and Application of Diazirines in Biological and Synthetic Macromolecular Systems. Soft Matter 2005, 1 (3), 178-205. (5) Crudden, C. M.; Horton, J. H.; Ebralidze, I. I.; Zenkina, O. V.; McLeane, A. B.; Drevniok, B.; She, Z.; Kraatz, H.-B.; Mosey, N. J.; Seki, T.; Keske, E. C.; Leake, J. D.; Rousina-Webb, A.; Wu, G. Ultra Stable Self-Assembled Monolayers of N-heterocyclic Carbenes on Gold. Nat. Chem. 2014, 6, 409-414. (6) Zhukhovitskiy, A. V.; Mavros, M. G.; Van Voorhis, T.; Johnson, J. A. Addressable Carbene Anchors for Gold Surfaces. J. Am. Chem. Soc. 2013, 135 (20), 7418-7421. (7) Such, G. K.; Johnston, A. P. R.; Liang, K.; Caruso, F. Synthesis and Functionalization of Nanoengineered Materials using Click Chemistry. Prog. Polym. Sci. 2012, 37, 985-1003.


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 43

(8) Galvin, C. J.; Genzer, J. Applications of Surface-Grafted Macromolecules Derived from Post-Polymerization Modification Reactions. Prog. Polym. Sci. 2012, 37, 871-901. (9) Choong, C.; Foord, J. S.; Griffiths, J.-P.; Parker, E. M.; Baiwen, L.; Bora, M.; Moloney, M. G. Post-Polymerisation Modification of Surface Chemical Functionality and its Effect on Protein Binding. New J. Chem. 2012, 36, 1187-1200. (10) Wang, H.; Griffiths, J.-P.; Egdell, R. G.; Moloney, M. G.; Foord, J. S. Chemical Functionalization of Diamond Surfaces by Reaction with Diaryl Carbenes. Langmuir 2008, 24, 862-868. (11) Moy, E.; Lin, F. Y. H.; Vogtle, J. W.; Policova, Z.; Neumann, A. W. Contact Angle Studies of the Surface Properties of Covalently Bonded Poly-L-lysine to Surfaces Treated by Glow-Discharge. Colloid Polym. Sci. 1994, 272, 1245-1251. (12) Caro, A.; Humblot, V.; Methivier, C.; Minier, M.; Salmain, M.; Pradier, C.-M. Grafting of Lysozyme and/or Poly(ethylene glycol) to Prevent Biofilm Growth on Stainless Steel Surfaces. J. Phys. Chem. B 2009, 113, 2101-2109. (13) Keselowsky, B. G.; Collard, D. M.; Garcia, A. J. Surface Chemistry Modulates Fibronectin Conformation and Directs Integrin Binding and Specificity to Control Cell Adhesion. J. Biomed. Mater. Res. A 2003, 66, 247–259. (14) Schmidt, D. R.; Waldeck, H.; Kao, W. J. Protein Adsorption to Biomaterials. in Biological Interactions on Materials Surfaces: Understanding and Controlling Protein, Cell and Tissue Responses. Puleo, D. A.; Bizios, R., Eds.; Springer: New York, 2009, 2-17. (15) Dilly, S. J.; Beecham, M. P.; Brown, S. P.; John M. Griffin; Clark, A. J.; Griffin, C. D.; Marshall, J.; Richard M. Napier; Taylor, P. C.; Marsh, A. Novel Tertiary Amine Oxide Surfaces That Resist Nonspecific Protein Adsorption. Langmuir 2006, 22, 8144-8150. 28 ACS Paragon Plus Environment

Page 29 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

(16) Rodahl, M.; Höök, F.; Fredriksson, C.; Keller, C. A.; Krozer, A.; Brzezinski, P.; Voinova, M.; B, K. Simultaneous Frequency and Dissipation Factor QCM Measurements of Biomolecular Adsorption and Cell Adhesion. Faraday Discuss. 1997, 229-246. (17) Singh, K.; McArdle, T.; Sullivan, P. R.; Blanford, C. F. Sources of Activity Loss in the Fuel Cell Enzyme Bilirubin Oxidase. Energy Eviron. Sci. 2013, 6, 2460-2464. (18) Voinova, M. V.; Rodahl, M.; Jonson, M.; Kasemo, B. Viscoelastic Acoustic Response of Layered Polymer Films at Fluid-Solid Interfaces: Continuum Mechanics Approach. Phys. Scripta 1999, 59, 391-396. (19) Zhang, D.; Dougal, S. M.; Yeganeh, M. S. Effects of UV Irradiation and Plasma Treatment on a Polystyrene Surface Studied by IR−Visible Sum Frequency Generation Spectroscopy. Langmuir 2000, 16, 4528-4532. (20) Pongprayoon, T.; Yanumet, N.; O'Rear, E. A.; Alvarez, W. E.; Resasco, D. E. Surface Characterization of Cotton Coated by a Thin Film of Polystyrene with and without a Crosslinking Agent. J. Colloid Interface Sci. 2005, 281, 307-315. (21) Wagner, C.D.; Davis, L.E.; Zeller, M.V.; Taylor, J.A.; Raymond, R.M.; Gale, L.H. Empirical Atomic Sensitivity Factors for Quantitative Analysis by Electron Spectroscopy for Chemical Analysis. Surf. Interface Anal. 1981, 3, 211-225. (22) Feiler, A. A.; Sahlholm, A.; Sandberg, T.; Caldwell, K. D. Adsorption and Viscoelastic Properties of Fractionated Mucin (BSM) and Bovine Serum Albumin (BSA) Studied with Quartz Crystal Microbalance (QCM-D). J. Colloid Interface Sci. 2007, 315, 475-481. (23) Höök, F.; Vörös, J.; Rodahl, M.; Kurrat, R.; Böni, P.; Ramsden, J. J.; Textor, M.; Spencer, N. D.; Tengvall, P.; Gold, J.; Kasemo, B. A Comparative Study of Protein 29 ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 43

Adsorption on Titanium Oxide Surfaces using in situ Ellipsometry, Optical Waveguide Lightmode Spectroscopy, and Quartz Cystal Microbalance/Dissipation. Colloids Surf. B. 2002, 24, 155-170. (24) Sauerbrey, G. The Use of Quartz Oscillators for Weighing Thin Layers and for Microweighing. Z. Phys. C: Part. Fields 1959, 155, 206-222. (25) Höök, F.; Kasemo, B.; Nylander, T.; Fant, C.; Sott, K.; Elwing, H. Variations in Coupled Water, Viscoelastic Properties, and Film Thickness of a Mefp-1 Protein Film during Adsorption and Cross-linking: A Quartz Crystal Microbalance with Dissipation Monitoring, Ellipsometry, and Surface Plasmon Resonance Study. Anal. Chem. 2001, 73, 5796-5804. (26) Höök, F.; Rodahl, M.; Kasemo, B.; Brzezinski, P. Structural Changes in Hemoglobin during Adsorption to Solid Surfaces: Effects of pH, Ionic Strength, and Ligand Binding. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 12271-12276. (27) Ramsden, J. Protein Adsorption Kinetics. In: Biopolymers at Interfaces. 2nd ed.; Malmsten, M., Ed.; Marcel Dekker: New York, 2003, p 199. (28) Anand, G.; Sharma, S.; Dutta, A. K.; Kumar, S. K.; Belfort, G. Conformational Transitions of Adsorbed Proteins on Surfaces of Varying Polarity. Langmuir 2010, 26, 10803-10811. (29) Brusatori, M. A.; Tie, Y.; Van Tassel, P. R. Protein Adsorption Kinetics under an Applied Electric Field: An Optical Waveguide Lightmode Spectroscopy Study. Langmuir 2003, 19, 5089-5097. (30) Van Tassel, P. R. Protein Adsorption Kinetics Under an Applied Electric Field. in: Dekker Encyclopedia of Nanoscience and Nanotechnology. 2nd ed.; Lyshevski, S.E.; Contescu, C.I.; Putyera, K., Eds.; CRC Press, 2008, pp 3569–3577. 30 ACS Paragon Plus Environment

Page 31 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

(31) Rabe, M.; Verdes, D.; Seeger, S. Understanding Protein Adsorption Phenomena at Solid Surfaces. Adv. Colloid Interface Sci. 2011, 162, 87-106. (32) Nelson, G. W.; Perry, M.; He, S.-M.; Zechel, D. L.; Horton, J. H. Characterization of Covalently Bonded Proteins on Poly(methyl methacrylate) by X-ray Photoelectron Spectroscopy. Colloids Surf. B. 2010, 78, 61-68. (33) Aglietto, M.; Alterio, R.; Bertani, R.; Galleschi, F.; Ruggeri, G. Polyolefin Functionalization by Carbene Insertion for Polymer Blends. Polymer 1989, 30, 1133-1136. (34) Olsen, D. A.; Osteraas, A. J. Difluorocarbene Modification of Polymer and Fiber Surfaces. J. Appl. Polym. Sci. 1969, 13, 1523-1535. (35) Textor, M.; Ruiz, L.; Hofer, R.; Rossi, A.; Feldman, K.; Hähner, G.; Spencer, N. D. Structural Chemistry of Self-Assembled Monolayers of Octadecylphosphoric Acid on Tantalum Oxide Surfaces. Langmuir 2000, 16, 3257-3271. (36) Puziy, A. M.; Poddubnaya, O. I.; Ziatdinov, A. M. On the Chemical Structure of Phosphorus Compounds in Phosphoric Acid-Activated Carbon. Appl. Surf. Sci. 2006, 252, 8036-8038. (37) Larrieu, J.; Held, B.; Martinez, H.; Tison, Y. Ageing of Atactic and Isotactic Polystyrene Thin Films treated by Oxygen DC Pulsed Plasma. Surf. Coat. Tech. 2005, 200, 2310-2316. (38) Chapman, R. G.; Ostuni, E.; Liang, M. N.; Meluleni, G.; Kim, E.; Yan, L.; Pier, G.; Warren, H. S.; Whitesides, G. M. Polymeric Thin Films that Resist the Adsorption of Proteins and the Adhesion of Bacteria. Langmuir 2001, 17, 1225-1233.


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 43

(39) Ostuni, E.; Chapman, R. G.; Holmlin, R. E.; Takayama, S.; Whitesides, G. M. A Survey of Structure−Property Relationships of Surfaces that Resist the Adsorption of Protein. Langmuir 2001, 17, 5605-5620. (40) Azri, A.; Giamarchi, P.; Grohens, Y.; Olier, R.; Privat, M. Polyethylene Glycol Aggregates in Water Formed through Hydrophobic Helical Structures. J. Colloid Interface Sci. 2012, 379, 14-19. (41) Silin, V.; Weetall, H.; Vanderah, D. J. SPR Studies of the Nonspecific Adsorption Kinetics of Human IgG and BSA on Gold Surfaces Modified by Self-Assembled Monolayers (SAMs). J. Colloid Interface Sci. 1997, 185, 94-103. (42) Kenawy, E. R.; Worley, S. D.; Broughton, R. The Chemistry and Applications of Antimicrobial Polymers: A State-of-the-Art Review. Biomacromolecules 2007, 8, 13591384. (43) Luan, Y.; Li, D.; Wang, Y.; Liu, X.; Brash, J. L.; Chen, H. 125I-Radiolabeling, Surface Plasmon Resonance, and Quartz Crystal Microbalance with Dissipation: Three Tools to Compare Protein Adsorption on Surfaces of Different Wettability. Langmuir 2014, 30, 1029-1035. (44) Dolatshahi-Pirouz, A.; Rechendorff, K.; Hovgaard, M. B.; Foss, M.; Chevallier, J.; Besenbacher, F. Bovine Serum Albumin Adsorption on Nano-Rough Platinum Surfaces Studied by QCM-D. Colloids Surf. B. 2008, 66, 53-59. (45) Reimhult, K.; Petersson, K.; Krozer, A. QCM-D Analysis of the Performance of Blocking Agents on Gold and Polystyrene Surfaces. Langmuir 2008, 24, 8695-8700. (46) Figueira, V. B. C.; Jones, J. P. Viscoelastic Study of the Adsorption of Bovine Serum Albumin on Gold and its Dependence on pH. J. Colloid Interface Sci. 2008, 325, 107-113. 32 ACS Paragon Plus Environment

Page 33 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

(47) Belegrinou, S.; Mannelli, I.; Lisboa, P.; Bretagnol, F.; Valsesia, A.; Ceccone, G.; Colpo, P.; Rauscher, H.; Rossi, F.O. pH-Dependent Immobilization of Proteins on Surfaces Functionalized by Plasma-Enhanced Chemical Vapor Deposition of Poly(acrylic acid)- and Poly(ethylene oxide)-like Films. Langmuir 2008, 24, 7251-7261. (48) Calonder, C.; Van Tassel, P. R. Kinetic Regimes of Protein Adsorption. Langmuir 2001, 17, 4392-4395. (49) Wertz, C. F.; Santore, M. M. Fibrinogen Adsorption on Hydrophilic and Hydrophobic Surfaces: Geometrical and Energetic Aspects of Interfacial Relaxations. Langmuir 2001, 18, 706-715. (50) Caruso, F.; Furlong, D. N.; Kingshott, P. Characterization of Ferritin Adsorption onto Gold. J. Colloid Interface Sci. 1997, 186, 129-140. (51) Viswanadhan, V. N.; Ghose, A. K.; Revankar, G. R.; Robins, R. K. Atomic Physicochemical Parameters for Three Dimensional Structure Directed Quantitative Structure-Activity Relationships. 4. Additional Parameters for Hydrophobic and Dispersive Interactions and their Application for an Automated Superposition of Certain Naturally Occurring Nucleoside Antibiotics. J. Chem. Inf. Comput. Sci. 1989, 29, 163-172. (52) Ertl, P.; Rohde, B.; Selzer, P. Fast Calculation of Molecular Polar Surface Area as a Sum of Fragment-Based Contributions and Its Application to the Prediction of Drug Transport Properties. J. Med. Chem. 2000, 43, 3714-3717. (53) Noh, H.; Vogler, E. A. Volumetric Interpretation of Protein Adsorption: Competition from Mixtures and the Vroman Effect. Biomaterials 2007, 28, 405-422.


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 43

(54) Noh, H.; Vogler, E. A. Volumetric Interpretation of Protein Adsorption: Mass and Energy Balance for Albumin Adsorption to Particulate Adsorbents with Incrementally Increasing Hydrophilicity. Biomaterials 2006, 27, 5801-5812.


Page 35 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

TABLE OF CONTENTS GRAPHIC


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

93x49mm (600 x 600 DPI)


Page 36 of 43

Page 37 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

47x12mm (600 x 600 DPI)


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

101x123mm (600 x 600 DPI)


Page 38 of 43

Page 39 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

47x12mm (600 x 600 DPI)


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

96x52mm (600 x 600 DPI)


Page 40 of 43

Page 41 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

71x61mm (600 x 600 DPI)


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 42 of 43


Page 43 of 43

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

43x22mm (600 x 600 DPI)


Surface Characterization and in situ Protein ... - ACS Publications

Recommend Documents