Mussel adhesion: A fundamental perspective on factors governing strong underwater adhesion

Highly saline environments are a significant challenge for technical adhesives. Nature, in particular the mussel, has used a combination of proteins and manipulation of its environment to allow adhesion to a multitude of surface types underwater. These include rocks, rope, the outer surface of other shells, painted ships, and, remarkably, Teflon.¹ The body of literature on mussel adhesion has grown in the last 30 years, with many advances, some of which are included in Fig. 1. Notably, the research community has identified the various proteins, also known as mussel foot proteins (mfps), that can be isolated from the attaching plaque^2,3 as well as probing environmental interactions,⁴ such as pH,⁵ surface chemistry,¹ secretion time,⁶ and ions present.^7,8

The redox-active catechol 3,4-dihydroxyphenolamine (DOPA) has been identified as a central functional amino acid with multiple functionalities during mussel adhesion. DOPA's importance dates back to the seminal work on the composition of insect cuticles by Pryor et al.¹⁶ where they proposed the oxidized form of DOPA to enhance reactive crosslinking, providing cohesive strength. Similar ideas were subsequently proposed for understanding mussel glue cohesion.¹⁷ In its reduced form, DOPA has also been suggested to play an important role for enhancing molecular interaction via surface coordination bond formation at the adhesive interface.¹⁸ Although the commonly accepted hypothesis recognizes DOPA, with its catechol functionality, as the one driving molecule for molecular adhesion enhancement, recent publications have highlighted substantial adhesion measured in recombinant proteins without DOPA residues.¹⁹ Similarly, barnacles can adhere well, without using any nonstandard or redox-active amino acids.^20,21 Clearly, various species of mussels and other invertebrates secreting polymer-based underwater adhesives provide inspiration for the development of new adhesives and biomaterials that can handle saline environments.^22–26 

The development of new biomaterials with potential for biomedical applications has seen a large interest in recent years. In these applications, the loading of bioactive molecules, the strong adhesion to moist areas, the enhancement of cellular adhesion, and the self-healing capabilities of biomaterials are some of the most required and desirable features.²⁷ Mussel-inspired biomaterials are becoming increasingly popular because catechol modification alters both mechanical and biological functionalities of the biomaterials. Depending on environmental conditions (pH, oxygen content, salt concentration, etc.), DOPA, dopamine, norepinephrine, gallic acid, and tannic acid assist in the covalent conjugation of catechol or gallol groups to biomedical materials. This can give biomaterials outstanding adhesive performance, oxidation resistance, antibacterial properties, coating capacity, high reactivity, chelating and coordinating abilities, bioactivity, and biological compatibility.²⁸ 

In detail, catechol functional groups have been added to various synthetic and natural polymers, such as cellulose,^29,30 chitosan,³¹ hyaluronic acid,³² and alginate,³³ to improve adhesion properties. These materials have been used in a variety of biomedical applications, including tissue engineering,³² hemostasis,³¹ wound healing,³⁴ photothermal therapy,³⁵ antibacterial surfaces,³⁶ drug delivery,³⁷ biosensors,⁸ 3D bioprinting,³⁸ etc.

Furthermore, research into the fundamental mechanism of mussel adhesion, e.g., using molecular level atomic force microscopy (AFM) force probing, showed how adhesion is controlled by the competition between different molecular interactions, also at the single molecular level.^18,39–41 In our previous work, we showed how ions compete with molecular functionalities, which can be described well using competing Langmuir isotherms.⁴² Gebbie et al. showed that interactions of surface-bound cations with the phenolic rings of aromatic amino acids such as phenylalanine (Phe), tyrosine (Tyr), and DOPA are central for strong underwater adhesion.⁴³ In their work, they showed that the Phe interaction can easily out-compete interactions by DOPA and Tyr. This suggests that the binding motives of DOPA are not essential for strong binding to a mineral surface, with further studies also demonstrating that the presence of amines is essential to strongly bind to a negatively charged mineral surface.⁴⁴ Instead DOPA, with its complex redox chemistry and electrochemical cross-linking properties,⁴⁵ may play a more prominent role to provide cohesive strength to the mussel adhesive.

In general, the controlled oxidation of catechols can be triggered either electrochemically, by adjustment to alkaline pH, the oxygen level or the addition of redox-active metal ions.^46,47 One of the features in handling DOPA is its high reactivity after oxidation, which leads to cyclization or cross-linking of molecules and the visible formation of dark precipitates in aqueous solution. This is confirmed in voltammetric studies where DOPA showed a semireversible electrochemical-chemical-electrochemical mechanism with two main degradation paths being proposed in the literature. Brun and Rosset suggest the intramolecular cyclization to DOPA-chrome, while the results of Appenroth et al. suggest an intermolecular polymerization through a Schiff-base reaction.^45,48 DOPA degradation products show reduced adhesion, which may be the result of increased cohesion. The inability to reverse this degradation process requires stable environments during DOPA integration into polymers and hydrogels to keep its adhesive properties, while at the same time allowing their use as stabilizing crosslinkers.

As already mentioned, the specific functional groups of amino acids within mussel foot proteins are clearly a factor in adhesion. While catechols in the form of DOPA are found frequently in the mfps and have been implemented in various bioinspired polymer adhesives, it appears it is not only DOPA that promotes adhesion. There are number of different ways in which the catechol functionality, DOPA, can bind to surfaces; the most commonly discussed is through bidentate hydrogen bonding, but there are also dispersion and electrostatic interactions such as $π$- $π$⁠, $π$-cation, and metal ion coordination.^7,10,49–51 In Fig. 1, these interactions are labelled in purple. Metal ion coordination additionally has an important role in cohesion as well as binding between DOPA residues on the same type of mfp via $Fe 3 +$ coordination.⁷ Cross-linking between the different types of mfp is frequently mediated by, for example, $Fe 3 +$ and $Ca 2 +$ ions, as reviewed by Lee et al.²² Furthermore, the mechanism by which metal ions are incorporated into the adhesive plaque has been elucidated; they are mixed from intracellular metal storage particles with the mfps secreted in vesicles in a microfluidiclike network.^52,53

However, many of the binding interactions are not exclusive to catechol functionality. In particular, the $π$-mediated binding, which expands the set of amino acids that should be considered in adhesion. The acid-base interactions that amine groups can facilitate at surfaces are of a similar magnitude to hydrogen bonding employed by the catechol.^39,54,55 This complex set of contributions and how they drive adhesion has slowly begun to unravel over the past 5–10 years.^19,56,57

For instance, amine groups, present as lysine in proteins, both synergistically with catechols and alone also promote adhesion.^13,56,58 The proteins can also show crowding effects related to spacing between catechol and amines that have a negative effect on adhesion.⁵⁹ Although the presence of amine can improve the adhesive properties of poly(catechol)amine, the addition of salt provokes a decrease in the expressed adhesive force.¹³ Moreover, the same principle of spacing can be applied to ions interacting with the surface, effectively participating in the adhesive process at the solid/liquid interface.⁶⁰ Ultimately, to understand the interface in detail, it is necessary to utilize a controlled model system, e.g., lipid-based, where the singular amine interaction can be evaluated.³⁹ 

Recently, Bilotto et al. utilized a thiol-lipid bilayer decorated with amine-terminated polymers to evaluate the competition of specific amine interaction and ions with bare mica.⁴²  Figure 2 shows that increasing the electrolyte concentration yields an exponential decay of adhesion with the growing population of ions at the interface. The competition of amine and ion binding to the surface could be modelled in terms of two competing Langmuir isotherms, providing thermodynamic interaction energies and kinetic parameters. In the context of adhesion in a high salinity environment, this clearly shows how ions can out-compete a specific interaction by a concentration-dependent shift of binding equilibria at the interface. In line with the strategy of the mussel, an alternative binding functionality such as a $π$-electron donor can then facilitate binding to surface bound ions. This is in line with data indicating that, in addition to DOPA and lysine (amine), a number of other amino acids have shown strong adhesion in high salt environments, including tyrosine with its single hydroxyl group substituted on the phenyl ring and Phe.^19,43

Stimulated by the research progress in mussel adhesive proteins (MAPs) and catechol-based biomimetic adhesives, the study of biological adhesives is rapidly expanding to include various other organisms, notably those that do not rely on catechol.

Table I shows a survey of organisms that have been studied for their ability to secrete polymeric adhesive materials. The degree of molecular and structural characterization, mechanical testing, overall understanding of the fundamental adhesion mechanisms, and biomimetic replication of these materials vary widely among these organisms. In mussels, adhesive proteins have been identified via rapid RNA sequencing, mass spectroscopy, chromatography, and recombinant expression; adhesion energies have been accurately quantified using sensitive force-measuring techniques, such as the surface forces apparatus (SFA) and AFM, and validated with models and simulations while also developing a multitude of catechol-based biomimetic adhesives.^92,93 On the other hand, little is known about the composition of the bioadhesives secreted by related bivalve species such as scallops and oysters, typically producing an organic-inorganic matrix containing (unidentified) proteins, polysaccharides, phospholipids, as well as calcium carbonate and silica.

When designing a new adhesive and understanding the fundamental interactions,⁴¹ the question of which functionalities to include is, therefore, more subtle than was first proposed.^2,9 We, therefore, agree with the more recently expressed ideas that the environment and the action of the mussel have a strong influence on the role a particular group plays during establishing adhesion.^5,57 The mussel, therefore, controls which proteins are extruded when and located in which areas of their byssus so that they can interact under different conditions.^6,94

The mussel further controls the pH of the local environment where it glues having a strong effect on the charge/redox state of the functional groups.⁵ When measuring the fundamental interactions of particular functional groups, in addition to pH, we can also control the oxidation state through electrochemistry.^45,58 In addition to using redox chemistry and metal ion coordination, the phenolic residues can also interact with surfaces through $π$ interactions, as discussed above and summarized by Ahn.⁵⁷ 

The final main mechanism that mussels actively control is the hydration of the surfaces to which they adhere.⁹⁵ Although how the mussel uses the hydration of the proteins themselves is also of interest. For example, the strength of adhesion and cohesion through cross-linking of mfp-1 that tends to be located to the outer regions of the plaque and byssus threads can be induced by a reduction in hydration in response to chemical and enzymatic cross-linking.⁹⁶ The impact of competition between ions and the residues of the proteins will lead to changes in local hydration; however, this is difficult to access with most techniques.

Both the adhesion mechanisms and the reversible electrochemistry of catechol derivatives have been studied extensively. However, almost always, these two aspects are considered independent of each other due to inherent methodological difficulties associated with combining the two approaches. Classically, adhesion measurements require a thick polymer or hydrogel films, while electrochemical measurements need the species to be dissolved in an electrolyte. State-of-the-art measurements can be made using the electrochemical (EC-) SFA^54,55,58 with a functionalized monolayer on the electrode surface. This allows fine control over the oxidation state of the catechol while retaining the force sensitivity of the SFA. For the data shown in Fig. 3, we prepared a mixed catechol/amine functionalized short-chain self-assembled monolayer (SAM) on a gold substrate that acted as the working electrode (see inset of Fig. 3).⁹⁷ We then measured the potential dependent adhesion forces vs an opposing mica surface.

We report two interesting aspects of these experiments:

First, we observed a clear change in adhesion strength with applied potential for this mixed catechol/amine containing a self-assembly film. However, contrary to our expectations, the adhesion was strongest at oxidizing potentials, where the less adhesive quinone form should be present at the interface. Second, we found a lower adhesion force for experiments carried out in an electrolyte solution, compared to ultra-pure water.

These preliminary observations have two important implications: First, this further supports a synergy of catechols and amines within the interfacial mfps and plaque formation steps. Amines may facilitate strong binding to the surface at higher potentials, where the surface is charged negatively due to chloride adsorption, while DOPA dominates adhesion at lower potentials. In essence, these groups complement each other at changing surface conditions. Second, it is clear and expected that ion adsorption to an interface can out-compete the specific interaction once ions are present, in line with competitive adsorption studies by Bilotto et al.⁴² 

This highlights striking similarities to how a mussel may even utilize varying environments during adhesive plaque formation. Therein, the mussel first excludes the oxidizing, high ionic-strength seawater from the interface before injecting the mfps. This may serve to remove ions from the substrate, which block the initial formation of adhesive bonds.⁵ It is, hence, necessary to further develop robust protocols to study synergies and environmental dependencies of complementing functionalities.

Furthermore, simulations such as the competing Langmuir isotherm model but also atomistic methods are, therefore, crucial to extract the most information from experiments as well as providing insight into their own right. Computational methods [e.g., molecular dynamics (MD) and density functional theory (DFT)] yield robust insights for poorly understood molecular origins of mussel adhesion and complement the experiments. For instance, a few computational studies showed that there is a competitive adsorption of catechol and water on a wet surface.^12,98,99 The role of metal ions in seawater in this competition is highly overlooked, remaining an open question. The metal ions can directly affect the adsorption process by their interactions with surface atoms, water molecules, and relevant amino acids, but they can also contribute to overall adhesion by bulk cohesion via cross-linking.^100,101 Using DFT and MD simulations, dominant molecular forces at different surfaces are differentiated by examining interaction and binding energies, the number of hydrogen bonds and their lifetimes, bond dynamics, densities, molecular geometries, and structural orientations.

These simulation approaches are indeed not limited to discovering the structural and dynamics details at microscopic level but also aid to capture macroscopic properties of mussel adhesion under different conditions (e.g., temperature, salinity, the presence of metal ions, pH value, and external forces). For example, spontaneous redox processes of DOPA or another amino acid residue result in a range of catechol derivatives for different pH values.⁴⁵ As shown in Fig. 3(c), MD simulations provide thermodynamic insights into molecular adhesion, enabling identification of repulsive and attractive interactions of functional molecules. On the other hand, accurate prediction of macroscopic physical properties is critical to design stable mussel-inspired materials for specific applications. Using MD simulations, it is possible to apply external fields such as mechanical or thermal loads on catechol-based polymers at varying cross-link concentrations and to mimic dynamic processes of a tensile test or heat transfer. As the research field moves toward exploring the multiscale complexity in mussel adhesion, the ever growing speed and power of computational methods may make these open questions accessible at the molecular scale.

In order to fully understand mussel adhesion and harness its power to the full, the main aspects that we believe require focus are as follows.

It will be essential to further understand the complex interplay between adhesion and cohesion as well as the polymer mechanics driven by the cross-linking density in the different regions of a mussel foot. The cohesion can often be tricky to access, but it is of course essential to the integrity of any adhesive. Specifically for interfacial adhesion, there are still fundamental questions regarding the balance between ion and functionality adsorption during the establishment of contact. We have discussed some of the recent progress toward understanding this nanoscale competition using microscale ensembles of molecules via a Langmuir isotherm approach. These advances in measurement technology towards the use of modeling and simulations to extract fundamental constants from ensembles rather than single molecule systems will allow a greater range of influences to be investigated, e.g., ionic strength, pH, hydroxylation, etc.

While there has historically been much focus on the bidentate binding of DOPA to surfaces, there is significant growing opinion and evidence that the presence of amines and cation-π interactions may be equally important. Essentially, at the adhesive interface, the complex redox chemistry may be overpowered by electrostatics and noncovalent interactions. This needs further clarification and research.

Clearly, the cohesion within the mussel foot proteins is driven by the local electrochemistry of DOPA. Subtle changes of pH and oxygen concentration gradients at the interface and within the bulk of the glue drive cross-linking of the quinone-form with amines, which is essential for mechanical properties as well as cohesion within the mussel foot. The nature of these cross-links should be linked to energy dissipation and mechanical aspects of the adhesive interaction, which is an interplay that should be addressed in future work.

In particular, the balance between molecular interactions of DOPA and cohesive cross-linking and how the role of DOPA evolves during plaque formation is still an open question. Important aspects are that (1) the adhesive plaque interface remains strongly reducing during its lifetime,¹⁰² (2) the adhesive plaque interface is soft with its DOPA content largely intact during its functional lifetime,¹⁰² and (3) mfp-3 and mfp-5 remain uncross-linked in plaques so long as their interface to the substratum remains intact.¹⁰³ These observations support an important role of DOPA beyond cross-linking the proteins during plaque formation. Possibly doing SFA and/or AFM studies with liquid-liquid phase separated proteins or using electrochemical probes under plaques deposited by mussels onto gold-films will better elucidate what DOPA (and thiols) are doing during the formation of adhesive interfaces. For example, Valois et al.¹⁰⁴ recently tried such an approach, and the results do not support a strong cross-linking/covalent cohesive role at the adhesive interface. As such, striving for investigating marine adhesion under conditions that are more biologically relevant appears to be essential to further clarify the adhesive vs cohesive role of DOPA in mussel foot adhesion.

Furthermore, transferring the fundamental understanding of the complex interplay between the adhesion, cohesion, and mechanical properties with the environmental conditions will be key to developing the optimal bioinspired materials. Learning from some of the many other marine species that use adhesion may also shed light on the mussel mechanism and is also an expansive area for ongoing research.

Finally, the outstanding research question that will give insight into developing antifouling applications but looks back to the mussel itself is the observation that nothing sticks to the nacre on the inside of the mussel shell, even for a while after the mussel itself has died. There are a number of factors that could be at play, for example, dissolution of calcium ions and local alkalinity at the surface. These causes can now be investigated using a combination of direct and model system measurements, such as AFM (initial data in Fig. 4) or using the EC-SFA to control the surface pH, which also relate to fundamental measurements of electric double layer repulsion and ion competition at the interface. Thus, the mussel adhesion research landscape still has many secrets to explore.

We wish to acknowledge the support of the European Research Council for funding this work (ERC Stg. 677633). J.A. and M.V. also thank the Austrian Research Promotion Agency FFG in the framework of the COMET Center of Electrochemistry and Surface Technology, CEST, through Grant No. 865864 (biomaterial synthesis, biomimetic glues, and coatings). L.M. and M.V. also acknowledge the bilateral project, No. CN 07/2020, funded by the Austrian Federal Ministry of Education, Science and Research (BMBWF), and R.S. was supported by the National Key Research and Development Program of China (No. 2019YFE0106900). B.Z. acknowledges support from the TECH4YOU - Technologies for climate change adaptation and quality of life improvement (PNRR, Missione 4, Componente 2, Investimento 1.5), funded by the Italian Minister of University and Research (MUR) in the framework of the National Recovery and Resilience Plan as a part of the Next Generation EU program. We also thank the anonymous reviewer for their insightful comments and suggestions related to the future perspective of the field.

The authors have no conflicts to disclose.

Ethics approval is not required.

Laura L. E. Mears: Conceptualization (equal); Formal analysis (equal); Project administration (equal); Supervision (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Julia Appenroth: Conceptualization (equal); Data curation (equal); Validation (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Hui Yuan: Investigation (equal); Visualization (equal). Alper T. Celebi: Conceptualization (equal); Data curation (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Pierluigi Bilotto: Conceptualization (equal); Investigation (equal); Supervision (equal); Validation (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Alexander Michael Imre: Conceptualization (equal); Investigation (equal); Methodology (equal); Validation (equal); Visualization (equal). Bruno Zappone: Conceptualization (equal); Methodology (equal); Writing – original draft (equal). Rongxin Su: Conceptualization (equal); Data curation (equal); Funding acquisition (equal); Methodology (equal); Validation (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Markus Valtiner: Conceptualization (lead); Funding acquisition (equal); Investigation (equal); Project administration (equal); Supervision (equal); Writing – original draft (equal); Writing – review & editing (equal).

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