The Adhesion Mechanism of Marine Mussel Foot Protein: Adsorption of L-Dopa onα- and β-Cristobalite Silica Using Density Functional Theory

Marine mussels strongly adhere to various surfaces and endure their attachment under a variety of conditions. In order to understand the basic mechanism involved, we study the adsorption of L-dopa molecule on hydrophilic geminal and terminal isolated silanols of silica (001) surface. High content of modified amino acid L-dopa is found in the glue-like material secreted by the mussels through which it sticks to various surfaces under water. To understand the adsorption behavior, we have made use of periodic Density Functional Theory (DFT) study. The L-dopa molecule adheres to silica surfaces terminated with geminal and terminal silanols via its catechol part. In both cases, the adhesion is achieved through the formation of 4 H-bonds. A binding energy of 29.48 and 31.67 kcal/mol has been estimated, after the inclusion of dispersion energy, for geminal and terminal silanols of silica, respectively. These results suggest a relatively stronger adhesion of dopa molecule for surface with terminal isolated silanols.


Introduction
Mussels are marine organisms that have the ability to attach themselves firmly to a variety of surfaces including glass, plastic, and metal oxides [1]. Their adhesion is so strong that they can even stick to nonstick materials such as Teflon [2]. Furthermore, these mussels are capable of achieving adhesion in a variety of conditions such as wet, dry, and salty environments [3][4][5]. Mussels are usually found in rocky, wind-swept, and wave-swept seacoasts. Their adhesion should be rapid and strong; otherwise they may be dislodged by the energetic incoming waves. Mussels secrete a glue-like sticky material, known as byssus, which is responsible for the strong adhesion to rocks and other surfaces in turbulent marine environment [6]. The byssus is a bundle of threadlike materials that spreads out in a radially outward direction. It consists of four parts, namely, plaque, thread, stem, and root [7]. Plaque is that part of the byssal thread which is attached with the foreign surface, whereas the stem is connected with tissues of the organism by means of the root. Stem and plaque are attached with each other through the thread. Mussel byssus is proteinaceous and does not have living cells. Byssal threads are attached to the root at the base of mussel foot where a combination of 12 retractor muscles controls the tension in them. During adhesion, the mussel foot emerges from the shell, finds suitable surface, and becomes almost motionless after the adhesion. Field tests have revealed significant information about the tenacity (ability to resist dislodgement as a result of external force) of mussels [8,9]. The tenacity of an individual California mussel (Mytilus californianus) is 300 N in lift mode and 180 N in parallel displacement mode. Since a mussel has 50-100 byssal threads at a time and all the threads contribute to load bearing in lift mode only, the per-thread tenacity of California mussel has a maximum value of approximately 6 N. More than 25 different mussel foot proteins (mfp) have been identified in byssus, out of which 5 (mfp-2 to mfp-6) are unique to plaque [10]. These 5 mfp have a high content of the usually rare modified amino acid 3,4-dihydroxy-L-phenylalanine (Ldopa). It is widely believed that it is the catechol functionality of L-dopa molecule that gets attached with external surface during the adhesion process [11][12][13][14][15]. Mian et al. demonstrated 2 Journal of Chemistry that the catechol part of dopa interacts with the surface in dry conditions [16] and displaces the preadsorbed water molecules to bind with hydrophilic silica surface [16][17][18]. By using an AFM, researchers have investigated the single molecule mechanics of mussel adhesion by measuring the force of a single dopa molecule in contact with a wet TiO 2 surface [19]. Bahri et al. performed the adsorption and surface complexation investigation of L-dopa on Rutile surface in NaCl solution. They reported that L-dopa forms two species on the surface, with their proportion strongly depending on the pH [20].
Understanding the basic principles underlying the adhesion of mussels may lead to useful applications in a variety of fields, especially in medicine, dentistry, surface coatings, biomedical engineering, and bionanotechnology. There have been efforts to design biomimetic adhesives that are efficient, nontoxic, and biocompatible. In recent years, a number of scientific studies have been dedicated to design catecholfunctionalized copolymers; for example, Messersmith et al. have created tough, nonswelling bioadhesives based on catechol-modified polymers [21]. Mussel inspired bioadhesives have recently been used in cancer drug delivery and removal of infected tumors. Gold nanorod surface has been coated with mussel inspired bioadhesives. When irradiated, the nanorods produce localized heat which destroys the cancer cells [22]. Researchers have recently reported injectable synthetic polymers inspired by mussel adhesives. These citrate-based adhesives are low-cost, nontoxic, and nonallergic [23].
In this study we will simulate the adsorption behavior of L-dopa molecule on hydroxylated, hydrophilic surfaces of silica having terminal isolated and geminal silanols. In order to understand the atomic level details of adsorption phenomenon, we will use DFT energy calculations and DFT based molecular dynamics simulation. We will consider two types of silica surfaces, namely, silica surface with geminal and terminal isolated silanols. Geminal silica is the one which is terminated with two OH ions attached with each surface silicon atom whereas terminal silanols are single OH ions attached with each surface silicon. In this study, we have chosen the (001) surface of hydroxylated --cristobalite silica to model amorphous silica because both materials have almost equal densities and refractive indices [15][16][17][18]. The present silica has a silanol density of 8.1 per nm 2 which covers the silanols density of amorphous silica with a value of 5 hydroxyls per nm 2 .

Computational Method
We performed ab initio molecular dynamics and energy calculations using version 3.2 of the SIESTA code [24]. In this study, we have used the generalized gradient approximation (GGA) as the exchange and corelation functional, by the application of revised Purdue-Burke-Enzerhof (RPBE) method [25]. Norm-conserving pseudopotentials proposed by Troullier and Martins [26] were used with standard double zeta potential (DZP) basis set. Using the Monkhorst-Pack scheme [27], Brillouin zone sampling was performed with 3 × 3 × 1 -points. The geometries were optimized separately before and the L-dopa molecule was placed 3Å above the surface to perform DFT calculations. All the calculations were carried out using the conjugate gradient (CG) method [28], and cell volume was allowed to vary. The optimization was performed until the atomic forces reached a value less than (0.04 eV/Å). A mesh cut-off [29] of 2.72 keV was used for our atomic orbitals. No plane waves were used during the calculations.
The initial configuration for optimization of bulk -cristobalite was taken from the rectangular unit cell with lattice parameters of 4.97 and 6.93Å, having the P4121 symmetry [30]. Initial geometry of both silica surfaces was constructed using the following steps: we considered a slab of bulk --cristobalite having 16 atomic layers along the -direction. The slab consisted of 3 × 3 layers in the plane normal to the surface that is the -plane, when viewed from the top. The top silica surface was obtained by cleaving the bulk system along the (001) direction. The simulation box which we considered had dimensions of = = 14.7Å and = 40Å. A vacuum gap of sufficient length was introduced along the vertical direction to remove the periodicity. During the optimization, the bottom Si atoms were kept under constraint. The number of atoms in geminal silica was 198, whereas L-dopa consisted of 25 atoms, resulting in a total of 223 atoms in our system.
Using the following relation, binding energy of the system was calculated: Throughout this text we have used a notation system in which the term ( ) represents energy of the system in the geometry of using the basis set . In (1), + ( ) represents the energy of molecule-surface complex, ( ) is the energy of the molecule (L-dopa), and ( ) denotes energy of the surface . Keep in mind that, in our definition of the binding energy, a positive sign shows attraction between the surface and the molecule. We applied the counterpoise method [31,32] to calculate the basis set superposition error using the equation: In (2), molecular and surface geometries are taken from their geometries in the molecule-surface complex. We have calculated ( ) using the molecular basis and ( ) using the basis of the surface. + ( ) and + ( ) have been calculated using the basis for molecule-surface complex.
Traditional DFT methods are generally used for the description of H-bond; however, it does not include dispersion interaction between the atoms. In this study, we determined the dispersion interaction using the method of Grimme et al. [33,34].

Optimized Geometry of L-Dopa.
The adhesive and cohesive properties of mussels have been attributed to the presence of high amount (up to 30 mol%) of L-dopa in the Table 1: Structural parameters of optimized catechol.  plaque/substrate interface. L-dopa, a usually rare catecholic amino acid, is formed as a result of posttranslational modification of tyrosine. Figure 1 shows the optimized geometry of dopa (a) and its catechol group (b). The catechol part has been labelled to indicate various bond lengths and angles. The structural parameters of dopa have been summarized in Table 1.

Adsorption Geometry of L-Dopa On Silica Surface.
As a first step, we designed the initial geometries of L-dopa and silica. These geometries were separately optimized until the forces between their atoms were less than a certain threshold value. The optimized structure of L-dopa was placed at a distance of 3Å above both surfaces that is silica with geminal and terminal isolated silanols, and the combined geometry was once again relaxed. The resultant geometries, which show the adsorption behavior of L-dopa, are shown in Figures 2 and  3. Figure 2 shows the side and top view of L-dopa adsorbed on geminal silica. For visual clarity only the top-most layer of silica has been shown as ball and stick. The siloxane bridges attached to the silanols have been shown as lines. The figure shows that the two OH groups of the catechol part of L-dopa form 4 hydrogen bonds with the surface (represented by black dashed lines). It should be noted that the L-dopa molecule acts as a donor as well as acceptor of H to form the hydrogen bonds. Figure 3 represents side view and top view of the adsorption behavior of L-dopa molecule on silica surface having terminal silanols. As expected and similar to the case of geminal surface, the two OH groups of catechol ring of the molecule form 4 hydrogen bond with the surface. Here again the OH groups act as donors as well as acceptors of H to form the bonds. The fact that catechol part of L-dopa is responsible for its strong adhesion with the silica surface is in agreement with previously reported findings [15][16][17][18]. In these studies, the catechol part of L-dopa molecule was adsorbed on silica surface under dry and wet conditions and similar behavior was observed as we have reported here.

Binding Energy and BSSE Calculations.
Using (1) and (2), respectively, the binding energy, basis set superposition error (BSSE), and dispersion energy were calculated for the dopasilica system. The result has been summarized in Table 2 for both surfaces that is silica with geminal and terminal isolated silanols. In the case of geminal silica, a binding energy of   [36]. Also in the glycine adsorption on an edingtonite silica the BSSE exceeds relatively the uncorrected binding energy value [37]. Regarding the binding energy (Δ ) calculation for ammonia (NH 3 ) on silica surface, Civalleri and Ugliengo [38] reported that the BSSE of a periodic calculation is larger. The RPBE functional which we have used in this calculation is accurate in the description of hydrogen bond [39,40] but it does not take into consideration the dispersion interaction among the atoms. Using B3LYP-D2 * for the adsorption of benzene-1,4-diol on hydroxylated silica, Rimola et al. reported that the dispersion energy term may be significant as it reaches up to 16.7 kcal/mol [41,42] in their calculations. In present calculations a method proposed by Grimme [33] has been used to estimate the dispersion energy. A binding energy value of 19.65 kcal/mol has been calculated in the case of dopa molecule adsorbed on geminal silica (001). For silica (111) surface with terminal isolated silanols, a dispersion energy of 20.22 kcal/mol has been calculated. After adding the dispersion energy, the binding energy increased to 29.48 kcal/mol and 31.67 kcal/mol for surface with geminal and terminal isolated silanols, respectively. These final values of binding energy suggest that the adsorption of L-dopa molecule is stronger for surface with terminal silanols. It has been reported [41,42] that Grimme's method generally overestimates binding energies, and to achieve a closer match Journal of Chemistry 5 with experiment a slight modification of this method is used [43].

Conclusion
Mussels are marine organisms that stick to a variety of surfaces under turbulent conditions. This strong adhesion is achieved through the secretion of a glue-like material called byssus. Mussel byssus is rich in the unusually rare Ldopa molecules. It is believed that the catecholic ring of the dopa molecule interacts with the surface to achieve adhesion. To understand the mechanism behind the phenomenon of mussel adhesion, we have used periodic Density Functional Theory (DFT). We simulated the adsorption of a single L-dopa molecule on silica surfaces having geminal and terminal silanols. The dopa molecule interacts with both surfaces through its catechol group. The catecholic ring of the molecule forms 4 hydrogen bonds and acts as a donor as well as an acceptor of hydrogen to form the bonds. The binding energy in the case of geminal silanols was estimated to be 9.83 kcal/mol which augmented to 29.48 kcal/mol when the dispersion energy term was added to it. For silica surface having terminal silanols, we calculated a binding energy of 11.5 kcal/mol which increased to 31.67 kcal/mol with the inclusion of dispersion energy. We can conclude that the catechol group indeed plays a pivotal role when dopa interacts with a silica surface. The adsorption of L-dopa molecule on both surfaces through the surface silanols is indistinguishable in terms of binding energy.