Fundamental Principles of the Thermodynamics and Kinetics of Protein Adsorption to Material Surfaces

Abstract

Protein adsorption is important for essentially any process that involves the contact of a protein-containing solution and a material surface, with the resulting formation of the adsorbed layer of protein determined by the thermodynamics and kinetics of the system involved. This paper presents an overview of the fundamentals of these processes. First, the hierarchical structure of proteins and the types of bonding that stabilize a protein’s native-state structure are presented. This section is then followed by a section presenting the thermodynamic driving forces that influence the way that proteins adsorb and conformationally change for three characteristically different types of surface chemistries: nonpolar (hydrophobic) surfaces, neutral hydrophilic surfaces, and charged surfaces. The final section of this paper addresses how kinetics and thermodynamics combine together to influence protein adsorption behavior, followed by concluding remarks.

Graphical Abstract

I. Introduction

The interaction of proteins with material surfaces is of fundamental importance for a broad range of applications in biomedical engineering and biotechnology [1–8]. Applications include the design of medical implant surfaces for improved biocompatibility, the understanding and control of biomineralization processes, substrates for tissue engineering and regenerative medicine to direct cellular response, and drug delivery systems. Likewise, protein-material interactions are also important for the design of surfaces for biosensors, bioassays, bioseparation, biocatalysis, and food processing as well as addressing the problem of biofouling.

In each of these applications, kinetics controls the rate that proteins arrive at the surface from solution while a combination of kinetics and thermodynamics governs how proteins adsorb and arrange on a material surface. In this paper, the fundamental principles of the thermodynamics and kinetics of these interactions are addressed and as well as how they influence the orientation and conformational changes of proteins when they adsorb onto material surfaces.

In the following sections I first provide an overview of protein structure, which is important to know in order to understand how different surfaces chemistries interact with proteins. I then provide a brief overview of the basic thermodynamic principles of energy, entropy, and free energy and how these fundamental thermodynamic driving forces influence protein adsorption behavior on three characteristic types of surface chemistry: nonpolar surfaces (i.e., hydrophobic surface chemistries), polar surfaces (i.e., neutral hydrophilic surface chemistries), and charged surfaces. This section is followed by an overview of how kinetics combines with thermodynamics to govern protein adsorption behavior and how these factors influence the structure of the final complete monolayer of surface coverage by proteins that typically occurs when material surfaces are exposed to a protein-containing solution.

II. Protein Structure

In order to understand the thermodynamics of protein adsorption to material surfaces, it is necessary to first understand the structural configuration of proteins and the types of bonding that stabilize a protein’s native-state structure.

Proteins are complex macromolecules made up of amino acids with up to four levels of hierarchical structure, which are termed primary, secondary, tertiary, and quaternary structure (Figure 1) [9]. There are 20 different types of amino acids that are coded for by DNA and the amino acids in a protein are referred to as amino acid residues. Each main-chain component of an amino acid is composed of an amine group (N-terminus) that is covalently connected to a central carbon atom (designated as Cα), which is covalently connected to a carboxyl group (C-terminus). A side chain (R-group) is connected to each Cα atom, with the structure of the R-group determining the type of amino acid and its chemical characteristic (polar, nonpolar, or +/− charge (Figure 1A).

Figure 1.

Protein structure. (A) Primary structure showing examples of four types of amino acid residues with designated three-letter and one-letter codes: nonpolar (alanine, Ala, A), positively charged (lysine, Lys, K), negatively charged (aspartic acid, Asp, D), polar (serine, Ser, S). Color scheme (online): C (gray), H (white), N (blue), O (red). (B) Secondary structures composed of an alanine sequence: α-helix (views from side and end), β-sheet. Same color code as in (A). (C) Tertiary structure (lysozyme, PDB ID 7LYZ). Color code: yellow (β-sheet), pink (α-helix), blue and white (connecting random loops). Copied from [1] with permission from the Taylor and Francis Group LLC.

The primary structure designates the specific amino acid sequence of a given type of protein with the amino acid residues joined together by non-rotatable covalent bonds (peptide bonds). This results in a single polypeptide chain that has a single N-terminus at the beginning and a single C-terminus at the end.

The secondary structure of a protein describes how the polypeptide chain wraps upon itself to form helical, β-sheet (or β-strand), and random-loop structures (Figure 1B), which are determined by the ϕ, ψ (phi, psi) dihedral angles of each amino acid. The ϕ angle represents the dihedral angle of the bond between the amino acid’s main-chain amine group and the central Cα atom while the ψ angle represents the dihedral angle of the bond between the Cα atom and the main-chain carboxyl group of the amino acid. Helical structures are designated by the number of amino acid residues that make a full 360° helical rotation: these are termed α-helix (3.7 amino acid residues per full rotation; most common type), π-helix (5 amino acid residues per full rotation), and 3₁₀ helix (3 amino acid residues per full rotation). These helices are stabilized by hydrogen bonds (H-bonds) formed between the N-H groups of the main-chain amine and the C=O group of the main-chain carboxyl unit of the amino acid residues. These H-bonds are aligned parallel to the long axis of the helix, with the R-group of each amino acid residue radiating outward from the helix, thus providing the chemical functionality of the outer surface of each helix. The β-sheet secondary structures are formed by separate segments of the polypeptide chain that are aligned in either parallel or antiparallel orientations. These structures are stabilized by H-bonds formed between N-H groups of the main-chain amino unit and the C=O group of the main-chain carboxyl unit of the adjacent chain segment, resulting in approximately planar structures. The R-group of each amino acid in a β-sheet then extend from the top and bottom of the formed sheet, thus providing the surface chemistry of the sheet structures. The third type of secondary structure is called the random loop, which describes a sequence of amino acid residues without defined organized structure. Random loops typically occur as segments joining helical and β-sheet segments or at the N- or C-termini of a polypeptide chain.

The tertiary structure of a protein describes how the secondary structures are assembled together (Figure 1C). Because the secondary structures of a protein are formed with the R-groups extending outward (except for random loops), the tertiary structure of a protein is stabilized by interactions between the R-groups of the protein’s amino acid residues. Accordingly, nonpolar R-groups typically combine together via hydrophobic interactions, with the majority of the nonpolar R-groups thus forming the non-solvent-accessible interior of the protein, with hydrophilic polar and charged amino acid residues primarily lining the outer surface of the protein to provide aqueous solubility. However, proteins also typically present some hydrophobic amino acid residues along their outer surfaces, often as part of specific ligand-binding sites. The tertiary structure of proteins is also stabilized by hydrogen bonding and +/− charged pairs of amino acid R-groups (referred to as salt bridges) along with disulfide covalent crosslinks via the side chains of adjacent cysteine amino acid residues in many cases.

The final level of structure of proteins is termed quaternary structure, which occurs in proteins such as fibrinogen, where multiple polypeptide chains combine together, with each individual polypeptide chain having its own N- and C-termini. The same types of bonding that stabilize tertiary structure typically stabilize the quaternary structure of a protein that is made up of multiple polypeptide chains.

III. Thermodynamics of Protein Adsorption to Material Surfaces

In order to understand protein adsorption behavior on different types of material surfaces, it is important to first understand how bond energy and entropy combine to determine the overall change in free energy when a protein interacts with a material surface.

According to the laws of thermodynamics [10,11], a protein will spontaneously adsorb to a material surface if the process results in a decrease in the free energy of the overall system and its surrounding environment. The change in free energy is expressed as:

$$\Delta G=\Delta H-T\Delta S$$ (1)

where ΔG is the change in free energy of the process under constant temperature and pressure, ΔH is the change in enthalpy, ΔS is the change in entropy, and T is absolute temperature. The change in enthalpy is composed of two separate terms:

$$\Delta H=\Delta E+p\Delta V$$ (2)

where ΔE is the change in bond energy, p is absolute pressure, and ΔV is the change in volume. Under aqueous solution conditions, the change in volume for protein adsorption can typically be considered to be negligible, and thus ΔH primarily represents the change in overall bond energy that occurs during protein adsorption.

Importantly, it must be understood that free energy, enthalpy, and entropy are state functions, which means that they only depend on the difference between the starting and ending states of the defined system and the surrounding environment and are not influenced by the particular path that a given process undertakes. Additionally, it is important to recognize that the changes in these thermodynamic properties don’t just represent the interactions between the protein and the surface. They also include the interactions between the protein and the surface with the surrounding aqueous solution, which, under physiological conditions, is made up of water molecules along with charged cations and anions from dissociated salts as well as all other types of molecules in solution. For simplicity in the examples below, we will consider the solution to be only composed of water molecules and the cations and anions from dissociated salts, with the influence of other molecules in solution (e.g., other types of proteins) briefly being considered in the final part of this section (e.g., Vroman effects).

The overall change in enthalpy (ΔH) during a protein adsorption process thus represents the sum of the changes in bond energy within the protein itself (i.e., due to structural changes of the protein), between the protein and the material surface, and between both the protein and the material surface with other molecules in the surrounding aqueous solution. Similarly, the overall change entropy (ΔS) represents not only the configurational changes of a protein as it adsorbs, but also the changes in entropy due to changes in the structure of both the adsorbent surface as well as the other components of the solution. In the following subsections, these interactions are generally addressed for three characteristic types of surface chemistries: nonpolar (hydrophobic), polar (neutral hydrophilic), and charged surfaces.

Readers are referred to more thorough treatments of the thermodynamics of protein adsorption that have been previously addressed in the literature. In particular, Norde, Lyklema, and coworkers have extensively covered the thermodynamics of the adsorption of proteins at solid-liquid interfaces in an excellent series of papers published in the 1970s and 1980s in which they are some of the first to determine that entropy is one of the key factors driving the spontaneous adsorption of proteins to material surfaces [12–22]. In particular, Norde, Lyklema, and coworkers presented microcalorimetry results [16] that showed spontaneous adsorption (i.e, ΔG_ads < 0) with the process being endothermic (i.e., ΔE > 0). These combined results indicated that the adsorption process must be driven by an increase in entropy (i.e., ΔS > 0) with the authors surmising that the entropic effect was primarily due to changes in hydration and ion exchange phenomena. Andrade and Hlady also published an excellent review on the topic of the thermodynamics and kinetics of protein adsorption processes in 1986 [23].

III.a. Nonpolar Surface Chemistries

Protein adsorption to nonpolar surfaces is mediated by hydrophobic interactions [1]. Because a nonpolar surface does not have functional groups that are able to form hydrogen bonds (H-bonds), the solvation layer of water molecules over this type of surface (Figure 2A) is in a much different hydrogen-bonded structural state compared a material surface functionalized with polar groups (Figure 2B), or compared to water molecules in the bulk aqueous solution away from the surface. This results in the layer water molecules over a nonpolar surface being in a higher free energy state due to the combination of both enthalpic and entropic effects compared to the water molecules in the bulk solution. Because of this phenomenon, the displacement of the surface-bound layer of water molecules from the surface into the bulk solution represents a reduction in free energy. Similarly, the displacement of water molecules surrounding nonpolar groups on the surface of a protein into the bulk aqueous solution also contributes to a decrease in free energy. These combined effects provide a strong thermodynamic driving force for the adsorption of proteins to a nonpolar surface by the reduction in the solvent-accessible surface area of both the material surface and the protein.

Figure 2.

Images from molecular dynamics simulations of water over a nonpolar surface represented by a CH₃-functionalized self-assembled-monolayer (SAM) surface (A), and a polar surface represented by an OH-functionalized SAM surface (B). Color scheme (online): C atoms (blue spheres), H atoms (white spheres), O atoms (red spheres). The hydrogen bonds of the water molecules over the SAM surfaces are designated by the red dashed lines. As shown, the network of H-bonds formed by the water molecules in the interfacial layer of water just above the CH₃-SAM surface is distinctly different than the network of H-bonds formed by the water over the OH-SAM surface, as well as the water molecules in the bulk solution away from the surface. This difference causes the water molecules at the interface of the CH₃-SAM to be in a higher free energy state compared to the water molecules over the OH-SAM surface or the water molecules in solution well above the material surface. Images produced using VMD software: Humphrey W., Dalke A., Schulten K., VMD – Visual Molecular Dynamics, J. Molec. Graphics, 14: 33–38 (1996).

Because the displacement of water from nonpolar amino acid residues from the surface of a protein provides a thermodynamic driving force for adsorption, proteins in solution will tend to initially adsorb to a nonpolar surface in orientations with the protein’s patches containing the nonpolar amino acids face down against the surface. Also, because the tertiary structure (and quaternary structure if present) of proteins is typically stabilized by hydrophobic interactions, there is a strong tendency for a protein to unfold and spread out over the surface to place its nonpolar amino acid residues against the nonpolar groups of the material surface to maximally reduce the solvent-accessible surface area of the material surface. This ‘dehydration’ effect provides a strong thermodynamic driving force to cause relatively large conformational changes of the protein, with the protein tightly adsorbed to the material surface (Figure 3). The resulting conformational changes in the protein of course also involve changes in the conformational freedom (i.e., entropic effect) and bonding energy (i.e., enthalpic effect) of the proteins themselves, which will also then contribute to the overall changes in free energy of the system.

Figure 3.

Illustration of protein adsorption to a nonpolar (hydrophobic) surface. Proteins in their native-state structure are typically structured with a layer of hydrophilic amino acids residues (AAs) and a few nonpolar AAs (‘X’) along their outer surface, thus providing for their solubility in aqueous solution. When a protein comes in contact with a nonpolar material surface, the initial adsorption will be mediated by the nonpolar AAs on the protein’s surface. The protein will then tend to unfold over the surface to spread its nonpolar core AAs to be in contact with the nonpolar material surface while maintaining its hydrophilic AAs exposed to the aqueous solution. This process is primarily thermodynamically driven by the reduction in the free energy of the system due to the displacement of surface-bound water molecules back into the surrounding aqueous solution.

As an example from my own group’s experimental studies using circular dichroism spectropolarimetry (CD) to measure the degree of adsorption-induced unfolding of human serum albumin on a hydrophobic surface compared to albumin’s solution-state structure, we measured a decrease in helical structure from 66.5% to 14.4% with an increase in β-sheet structure from 1.5% to 28.3% [24]. These changes may be considered to be due to the destabilization of helical secondary structures of the protein due to hydrophobic interactions with the material surface combined with the material surface effectively acting as a relatively flat template for segments of the polypeptide chain to conformationally conform to, thus leading to the increase in β-sheet structure.

III.b. Polar Surface Chemistries

Protein interactions with material surfaces with polar functional groups are mediated by hydrogen bond formation between the polar amino acid residues of the protein and the polar groups of the material surface. However, because these functional groups also form strong hydrogen bonds with water, adsorption of a protein to this type of surface requires the protein to displace the hydrogen-bonded layer of water. This process thus involves the breaking hydrogen bonds between water and the polar functional groups of the protein and the material surface only to then replace these bonds with hydrogen bonds between the polar amino acid residues of a protein and the material surface. While protein adsorption under this condition may seem to represent a process that is not thermodynamically favorable since it ‘simply’ represents the exchange of one set of hydrogen bonds with another, proteins do spontaneously adsorb to polar surfaces [25,26], thus indicating that there is more to this picture than initially apparent.

To explain the driving force that causes proteins to adsorb to a polar surface, we will look at the overall thermodynamics of this type of process. The interaction between a single hydrogen-bondable functional group of an amino acid residue of a protein and a hydrogen-bondable group on a material surface is illustrated in Figure 4A. As shown, for adsorption to occur, one set of hydrogen bonds must be broken (i.e., hydrogen bonds between an amino acid with water and the material surface with water), with one set of hydrogen bonds then reformed (i.e., hydrogen bonds between the amino acid with the material surface, and between the two water molecules released into the bulk solution). Thermodynamically, this effectively represents no net change in bond energy. And for the exchange of a single set of hydrogen bonds, this effectively represents approximately no net change in entropy for the system (i.e., amino acid displaced from bulk solution in exchange for a pair of water molecules released into the bulk solution), with ΔG ≈ 0. This result indicates that this process is not thermodynamically favored to spontaneously occur. However, there are additional thermodynamic factors involved beyond this individual set of interactions.

Figure 4.

Illustration of protein adsorption on a neutral hydrophilic surface covered with hydrogen-bondable functional groups (e.g., OH groups). (A) Single amino acid with a hydroxyl side group (e.g., serine) adsorbing to the material surface. For adsorption to occur, four H-bonds between water molecules and the OH functional groups of the AA and the material surface are broken and replaced by four H-bonds between the OH functional groups of the AA and the material surface and between the desorbed water molecules, with the net change in energy of ΔE ≈ 0. This process also then involves the adsorption of one molecule from solution (i.e., the amino acid) and the release of a pair of H-bonded water molecules back into solution, with the net change in entropy of ΔS ≈ 0. Thus the overall change in free energy for this process is ΔG≈0. (B) Protein with multiple hydrogen-bondable groups interacting with hydrogen-bondable groups of the surface. In this case, although any given H-bond between the protein and surface is just as likely to be replaced by H-bonds with water molecules, once the multiple sets of H-bonds are formed between the protein and the material surface, the probability of each H-bond to be replaced by water molecules at the same time is extremely low, thus representing an entropically driven adsorption process with the net change in free energy of ΔG<0.

As an approximation, consider that when a polar functional group of an amino acid residue comes in contact with a polar functional group of the material surface, that there is equal probability of them forming hydrogen bonds together vs. forming hydrogen bonds with water (i.e., probability of 0.5). Thus, this hydrogen bonding will not by itself tend to cause the amino acid to adsorb. However, a large macromolecule like a protein that is covered by many polar functional groups will not just be able to form one set of hydrogen bonds with the polar groups of a material surface but will be able to form numerous sets of hydrogen bonds with the surface. For example consider the formation of 30 individual sets of hydrogen bonds between a protein and a material surface. In this case, in order for the adsorbed protein to be desorbed from the surface, all 30 hydrogen bonds must be displaced by water at the same time. Thus, if each individual set of hydrogen bond between the protein and material surface is stable 50% of the time, the probability of all 30 hydrogen bonds to be displaced by water at the same instant is (0.50)³⁰, which results in an overall probability of 1.0 × 10⁻⁹ (Figure 4B). This obviously represents a very low probability, entropy-driven event, thus resulting in the spontaneous and effectively irreversible adsorption of proteins to material surfaces functionalized with neutral polar groups. This behavior was nicely demonstrated in a study conducted by my research group investigating the adsorption and desorption of human serum albumin to self-assembled monolayer (SAM) surfaces functionalized by hydroxyl groups [26]. After initially adsorbing the albumin to the surface, the surfaces with the adsorbed monolayer of albumin were incubated under plain buffer solution (i.e., no protein in the buffer solution) for six months, while periodically measuring the amount of albumin remaining on the material surface. The results from this study revealed no significant decrease in the amount of adsorbed albumin even at the six-month time point, thus documenting that the albumin was effectively irreversibly adsorbed.

Because the surfaces of proteins are primarily composed of hydrophilic amino acid residues, proteins will tend to adsorb to neutral hydrophilic surfaces in random orientations. Because the bonding interactions between a protein and this type of surface are mediated by hydrogen bonding, the nonpolar amino acid residues that typically stabilize a protein’s tertiary or quaternary structures will typically remain intact and stay buried within a protein’s non-solvent-accessible core. However, because the secondary, tertiary, and quaternary structures of proteins are stabilized by hydrogen bonds, adsorption to a neutral polar surface can compete with these hydrogen bonds, potentially resulting in the disruption of a protein’s structure. In fact, experimental studies that my group has conducted have shown protein adsorption to polar-group-functionalized surfaces typically results in similar trends of unfolding as with a hydrophobic material surface, although to a lesser degree—i.e., a loss in helical structure with an increase in β-sheet structure [24,27]. In this case, however, the adsorption-induced conformational changes in the protein can be considered to be due to the material surface competing with the hydrogen bonding that stabilizes helical structure while again acting as a template for the polypeptide chain of a protein to unfold and flatten out on the surface. For example, using CD to investigate the conformational changes of human serum albumin after adsorption on a SAM surface functionalized by hydroxyl groups, we observed a loss in helicity from 66.5% to 42.1% with a subsequent increase in β-sheet content from 1.5% to 13.4% [24].

Thus, although protein adsorption to a material surface presenting neutral polar functional groups can be expected to be weaker with a lesser degree of adsorption-induced unfolding compared to protein adsorption to a hydrophobic surface, adsorption still spontaneously occurs in an essentially irreversible manner.

The exception to the thermodynamically favorable adsorption of proteins to neutral hydrophilic surfaces occurs when the material surface groups result in such strong binding to water molecules that proteins are not able to displace the surface-bound layer of water [28]. The most common example of this behavior is for polyethylene glycol (PEG)-functionalize surfaces and hydrogels, which have been shown to be highly resistant to protein adsorption [29–32].

III.c. Charged Surface Chemistries

At first glance it is typical to assume protein adsorption to a charged surface to be thermodynamically driven by bond energy (i.e., ΔE) due to the electrostatic attractions between oppositely charged functional groups of the amino acid residues present on the surface of a protein and the oppositely charged functional groups of the material surface. However, this point of view overlooks the presence of the relatively high concentration of salt ions in physiological solution. Because the cations and anions from dissociated salts in physiological solution diffuse at a much faster rate than a protein, the charged functional groups of a protein and a material surface are complexed with counterions well before a protein is able to diffuse to and interact with a material surface (Figure 5). Thus, for a charged functional group of an amino acid residue of a protein to interact with an oppositely charged functional group of a material surface, the electrostatic bonds between the protein and the material surface with these counterions must be exchanged with electrostatic interactions between the protein and the material surface, with the counterions displaced into the surrounding bulk solution to interact with other ions. This interaction thus actually represents an ion-exchange process that is driven by entropic effects (i.e., ΔS) rather than bond energy effects (i.e., ΔE) [20–22]. This type of process is similar to the entropically driven adsorption of a protein to a neutral hydrophilic surface as illustrated in Figure 4, except instead of the displacement of water molecules, exchange occurs with counterions that were initially electrostatically coordinated with oppositely charged groups of the protein and the material surface.

Figure 5.

Snap-shot from a molecular simulation of two positively charged LK peptides (L = leucine, K= lysine) interacting with a carboxylate-functionalized SAM surface in physiological saline solution. Color scheme (online): C atoms (light blue spheres), H atoms (white spheres), O atoms (red spheres), N (dark blue spheres), Na⁺ ions (yellow spheres), Cl⁻ (green spheres in solution). Water molecules present in the simulation are shown as small dots so as to not block the view of the counterions and the peptides. As shown, the negatively charged carboxylate groups (COO⁻) presented by the SAM surface were very rapidly electrostatically complexed with Na⁺ ions from the physiological saline solution. At this stage of the simulation, only one of the peptides was able to adsorb to the surface by exchanging electrostatic interactions from its multiple positively charged lysine amino acid residues with initially adsorbed Na⁺ ions, while the other peptide remained in a non-adsorbed state due to the rest of the negatively charged carboxylate groups of the SAM surface still being complexed with Na⁺ counterions. Image produced using VMD software: Humphrey W., Dalke A., Schulten K., VMD – Visual Molecular Dynamics, J. Molec. Graphics, 14: 33–38 (1996).

My research group came to realize this phenomenon when we were first conducting studies to experimentally measure the adsorption free energy of a positively charged short peptide to a negatively charged surface. In our preliminary experiments, we were using a TGTG-K-GTGT peptide (T = threonine, G = glycine, K = lysine) with capped end groups to neutralize the N- and C-termini of the peptide, with the peptide thus having only one positively charged group from the middle lysine amino acid residue. Using surface plasmon resonance spectroscopy (SPR) to measure the adsorption response of this peptide over a negatively charged carboxylic acid-functionalized SAM surface, we were at first very surprised when the results showed no measurable peptide adsorption to this surface (i.e., ΔG = 0). After much discussion about why our experimental system was not working as initially expected, we came to the realization that this was the result that we should have expected based on thermodynamics. The peptide did not adsorb because the adsorption event was actually representing the separation of two electrostatically bonded pairs of interactions (i.e., the NH₃⁺ side-chain group of lysine with a Cl⁻ ion and a surface COO^— group with a Na⁺ ion) and replacing this with two other electrostatically bonded pairs of interactions (i.e., the lysine-surface NH₃⁺ -COO⁻ pair and Na⁺ ion-Cl⁻ ion pair in solution), with a net result of ΔE = 0. Also, adsorbing one molecule from solution (i.e., the peptide) but displacing an associated pair of salt ions back into solution effectively represents a condition of ΔS ≈ 0. A slight change in our peptide model remedied this problem. We replaced the original peptide design with the same peptide sequence but without the capped end groups, thus providing a peptide with two positively charged groups (N-terminus and lysine side chain) and one negatively charged group (C-terminus). The SPR studies with this peptide design then resulted in a measurable adsorption event [33,34], which we came to realize was due to an entropically driven ion-exchange process. This entropically favorable response was due to the two positively charged groups of the peptide simultaneously interacting with oppositely charged groups of the SAM surface, with the subsequent adsorption of one molecule from solution (i.e., the peptide) and the displacement of two ionically paired sets of counterions back into solution (Figure 6).

Figure 6.

Illustration of peptide adsorption to a charged surface mediated by an ion-exchange process. (A) Peptide with repeated TG amino acid sequences (T = Threonine, G = glycine) with neutral N- and C-termini (Ac = acetylation, Am = Amidization) around a central positively charged amino acid (K = lysine) over a material surface functionalized with negatively charged carboxylate groups. As indicated, the charged groups of the peptide and material surface are initially electrostatically complexed with oppositely charged counterions (sodium or chloride ions from solution). Since this process involves separating and combining the same number of electrostatic interactions while adsorbing one peptide and releasing one paired set of ions, the overall change in free energy for this process is ΔG≈0. (B) Redesigned peptide with charged N- and C-termini, thus providing two positive charges to electrostatically interact by ion-exchange with the carboxylate groups on the material surface. Although this process still involves separating and combining the same number of electrostatic interactions, it involves the adsorption of one peptide and the release of two paired set of ions, with the overall change in free energy for this process now being ΔG<0. From a statistical-mechanics point-of-view, there is equal probability of the counterions competing with the peptide and material surface in (A) while there is a relatively lower probability of counterions from solution separating both electrostatic interactions between the peptide and material surface at the same time in (B).

Because proteins often have distinct positively or negatively charged patches on their surfaces, the interaction of a protein to a charged material surface will tend to adsorb the protein with its oppositely charged area against the material surface, thus providing a means of controlling the orientation of protein to a surface in a predictable manner. This type of approach has been used to orient proteins on material surfaces such that the protein’s bioactive sites are preferentially available to interact with targeted substrates in solution [35,36]. Some of the early work demonstrating these types of influences was published by Tilton and coworkers. Tilton and coworkers [37,38] used surface force analyses and total reflection fluorescence measurements to document changes in the orientation of lysozyme on charged mica surfaces when adsorbed over a wide range of solution concentrations and pH, which revealed charge-induced changes in lysozyme adsorbing in side-on vs. end-on orientations as a function of solution conditions.

Considering the resulting structural changes of a protein when it adsorbs to a charged surface, charged groups of a material surface can interact more strongly with hydrogen-bondable groups of a protein than the native-state hydrogen bonds that serve to stabilize a protein’s secondary, tertiary, and quaternary structures, and can also compete with the salt-bridges that often stabilize a protein’s tertiary and quaternary structures. Thus, charged material surfaces can result in substantial unfolding of the adsorbed protein. For example, in CD studies conducted by my research group investigating the adsorption of what is considered to be a ‘hard’ (i.e., very stable) protein to a charged surface (lysozyme on silica glass) from low solution concentration (0.03 mg/mL), we measured a loss of in helical structure from 38% in solution to only 4% when adsorbed, with a subsequent increase in β-sheet structure from 16% in solution to 42% when adsorbed [39].

While strong protein adsorption generally occurs on charged surfaces, there is one very interesting exception to this general rule. Studies have shown that surfaces presenting zwitterionic groups (i.e., closely spaced positively and negatively charged groups with net zero charge) can be very highly resistant to protein adsorption [31,32,40,41]. The general thinking for this phenomenon is that these surfaces are able to bind to water molecules and counterions strongly enough to prevent their displacement by proteins, thus providing material surfaces that are highly resistant to protein adsorption.

IV. Kinetic Effects on the Formation of an Adsorbed Monolayer of Protein

Kinetics also plays an important concomitant role along with thermodynamics that influences protein adsorption behavior—both for relatively simple studies involving single types of protein in buffered saline solution and for the much more complex situation of protein adsorption from actual physiological solutions, such as blood plasma, which contain hundreds of different types of proteins.

For relatively simple solutions containing only one type of protein, protein adsorption can be illustrated as shown in Figure 7A, where the dynamics of the system represent a balance between the kinetics of initial protein adsorption on the surface vs. the kinetics of protein unfolding and spreading out over the surface [1]. As indicated in Figure 7B, when a protein is adsorbed from a high solution concentration, neighboring sites around an adsorbed protein will tend to fill up quickly with neighboring adsorbed proteins, with protein-protein interactions on the surface then tending to inhibit the unfolding of the adsorbed proteins. Under these conditions, each protein will occupy a small ‘footprint’ on the surface, resulting in the final adsorbed monolayer of protein exhibiting a relatively high surface concentration. In contrast to this, when protein is adsorbed from a very low solution concentration, the rate of adsorption is slowed down such that an adsorbed protein has time to undergo much more unfolding on the surface before its spreading out is inhibited by another adsorbed protein. This situation results in an adsorbed protein occupying a larger amount of surface area, with the resulting final adsorbed monolayer of protein exhibiting relative low surface concentration.

Figure 7.

(A) Illustration of the kinetic process of protein adsorption on a material surface. The initial rate of adsorption (k_a) is typically diffusion limited, thus being dependent on the protein’s solution concentration. The rate of desorption (k_d) essentially only relates to initially adsorbed protein that has not yet formed multiple interactions with the surface. After being adsorbed, proteins may unfold and spread out over unoccupied neighboring areas of the material surface, with the rate of this spreading represented by k_s. Because protein adsorption is typically an irreversible process and proteins typically do not adsorb to other proteins, adsorption will continue until the material surface is fully covered with a complete monolayer of protein. (B) Illustration of adsorbed monolayers of protein on a material surface from different solution concentrations of the protein, with the final surface concentration of the protein being dependent on the solution concentration. (C) The fact that solution concentration influences the final surface concentration of an irreversibly adsorbed monolayer of protein results in protein adsorption isotherms often exhibiting a Langmuir-like shape. However, this type of adsorption does not represent a Langmuir adsorption process.

These phenomena have been depicted in early modeling studies of protein adsorption by Van Tassel and coworkers [42–46], Talbot [47], and Lenhoff and coworkers [48,49]. Van Tassel and coworkers presented a series of papers using a modified random sequential adsorption model to theoretically represent the kinetics of irreversible protein adsorption combined with spreading to provide an explanation for differences in the amount of protein adsorption observed as a function of solution concentration. Talbot extended this theoretical approach by including a ‘memory function approach’ in which adsorbed proteins can either undergo desorption or transition to an irreversibly adsorbed state, while Lenhoff extended these 2-D analysis methods to include 3-D interactions that included electrostatic repulsion interactions between the adsorbing proteins.

As an example of this behavior from a relatively recent set of experimental studies from my own research group, the amount of fibrinogen adsorbed to SAM surfaces showed more than a 50% decrease in the final adsorbed surface concentration when the fibrinogen was adsorbed from 0.1 mg/mL vs. 10 mg/mL solution concentration [50]. This behavior is what causes protein adsorption isotherms to often exhibit a Langmuir-like shape without actually representing a Langmuir adsorption process (Figure 7C) [51]. While it can be quite tempting to fit the Langmuir model to a protein adsorption isotherm that exhibits a Langmuir shape in attempts to calculate thermodynamic properties such as an equilibrium constant or the free energy of adsorption, doing so will only provide erroneous values that have nothing to do with these thermodynamic properties.

When proteins are adsorbed from solutions containing numerous different types of proteins (e.g., blood plasma), the combination of the concentration of the protein in solution along with the size of the protein determines the diffusional flux of the protein in solution, and subsequently its rate of adsorption to a material surface [52]. Thus, when a fluid containing multiple different types of proteins comes in contact with a material surface, the relatively small proteins with high solution concentration will tend to adsorb to the surface first. Then by a combination of both kinetic and thermodynamic interactions, larger proteins that may interact more strongly with the material surface will arrive at the surface at a later timepoint and possibly displace the initially adsorbed proteins by what are referred to as Vroman effects [53,54]. These types of dynamic interactions occur until a stable adsorbed monolayer is formed on the material surface, with the resulting stoichiometry of the proteins in the adsorbed layer typically being quite different than the stoichiometric composition of proteins in the bulk solution over the surface [55].

Another important factor to recognize involving the combination of the kinetics and thermodynamics of protein adsorption is that the adsorption of proteins to material surfaces typically results in the protein molecules becoming kinetically trapped in what are referred to as metastable states. Such states represent configurational states of the protein on the surface where the protein is in a relatively deep local low free-energy ‘well’ with relatively large surrounding energy barriers that prevent the protein from transitioning to the global low free energy state for the given protein-surface system. This behavior thus typically results in proteins being adsorbed in many different conformations on the surface that are largely dependent on the orientation of the protein when it first came into contact with the surface [57].

This behavior is illustrated in the ‘protein-folding funnels’ shown in Figure 8. This figure depicts the rough free energy landscape that represents the numerous local free energy minima for protein adsorption behavior compared to the native-state structure of the protein. As illustrated in this figure, a different ‘protein folding (or refolding) funnel’ can be expected to occur for each initial orientation of an adsorbed protein, with large free-energy barriers separating the different protein conformations on a surface. While it can be expected that there will be one specific adsorbed state that would represent the global free-energy minimum of an adsorbed protein, the likelihood of an adsorbed protein actually achieving that state is small within experimentally accessible timeframes because of the large energy barriers that must be crossed between the many local low free-energy states of the system. Therefore while thermodynamics tells us the direction that the protein adsorption process should generally follow, kinetics limits the extent of the process that will occur, with a combination of both thermodynamics and kinetics thus determining the final state of the layer of adsorbed proteins for a given protein adsorption process.

Figure 8.

(A) Illustration of the protein folding funnel for a protein in aqueous solution. The horizontal axis represents the conformational phase-space and the vertical axis represents the free energy of the protein. The global free energy minimum position represents the native structure of the protein in aqueous solution. (B) Conceptual illustration of the way that protein-surface interactions shift the protein folding funnel resulting in a new low free energy state of the system. The center folding funnel represents the aqueous solution behavior, as shown in (A), with the shifted left and right folding funnels depicting how different orientations of the protein on the surface will shift the folding funnel to different free energy minima that may be separated by large free-energy barriers. Adapted from reference [56]. Copied from [57] with permission from Springer.

V. Concluding Remarks

As described above, the factors that influence the thermodynamics and kinetics of protein adsorption to material surfaces are quite complex and dependent on a number of variables such as the type of protein, surface chemistry, and solution conditions. Many other factors also influence protein adsorption, such as surface topology and solution pH, which were not addressed in this paper. Despite these complexities, the underlying thermodynamic and kinetic principles governing these interactions are relatively straightforward and can be applied to provide a general understanding of protein adsorption behavior to material surfaces. I have attempted to address these main principles in the above generalized sections by addressing some of the primary thermodynamic and kinetic factors that influence protein adsorption behavior. Interested readers are encouraged to apply these principles to better understand protein adsorption behavior for their own specific material systems and applications as well as consulting many of the excellent publications on protein adsorption behavior that have been published over the past several decades, such as the ACS series on Proteins at Interfaces I, II, and III [58–60].

Highlights.

• Fundamental principles governing thermodynamics of protein adsorption

• Fundamental principles governing kinetics of protein adsorption

• Free energy of protein adsorption on nonpolar, polar, and charged surfaces

• Influence of hydration effects on protein adsorption to materials surface

• Influence of ion-exchange on protein adsorption to charged material surfaces