Protein adsorption at the solid-liquid interface is an important phenomenon that often can be observed as a first step in biological processes. Despite its inherent importance, still relatively little is known about the underlying microscopic mechanisms. Here, using multivalent ions, we demonstrate the control of the interactions and the corresponding adsorption of net-negatively charged proteins (bovine serum albumin) at a solid-liquid interface. This is demonstrated by ellipsometry and corroborated by neutron reflectivity and quartz-crystal microbalance experiments. We show that the reentrant condensation observed within the rich bulk phase behavior of the system featuring a nonmonotonic dependence of the second virial coefficient on salt concentration c_{s} is reflected in an intriguing way in the protein adsorption d(c_{s}) at the interface. Our findings are successfully described and understood by a model of ion-activated patchy interactions within the framework of the classical density functional theory. In addition to the general challenge of connecting bulk and interface behavior, our work has implications for, inter alia, nucleation at interfaces.

Proteins are ubiquitous and play a critical role in many areas from living organisms to protein microchips. In humans, serum albumin has a prominent role in the foreign body response since it is the rst protein which will interact with e.g. an implant or stent. In this study, we focused on the inuence of salts (i.e., dierent cations (Y 3+ , La 3+ ) and anions (Cl − , I − )) on bovine serum albumin (BSA) in terms of its bulk behaviour, as well as its role of charges for the protein adsorption at the solid-liquid interface in order to understand and control the underlying molecular mechanisms and 1 interactions. This is part of our group's eort to gain a deep understanding of proteinprotein and protein-surface interactions in the presence of multivalent ions. In the bulk, we found not only multivalent cation-triggered phase transitions, but also a dependence on the anions. The induced attractive interactions were observed to increase from Cl − < NO − 3 < I − , resulting in iodide preventing re-entrant condensation and promoting liquidliquid phase separation in bulk. Using ellipsometry and a quartz-crystal microbalance with dissipation (QCM-D), we obtained insight into the growth of the protein adsorption layer thickness. Importantly, we found that phase transitions at the substrate can be triggered by certain interface properties, whether they exist in the bulk solution or not.Through the use of a hydrophilic, negatively charged surface (SiO 2 ), the direct binding of anions to the interface was prevented. Interestingly, this led to re-entrant adsorption even in the absence of re-entrant condensation in bulk. However, the overall amount of adsorbed protein was enhanced through stronger attractive protein-protein interactions in the presence of iodide salts. These ndings illustrate how carefully chosen surface properties and salts can directly steer the binding of anions and cations, which guide protein behaviour, thus paving the way for specic/triggered protein-protein, proteinsalt, and protein-surface interactions.

In all areas related to protein adsorption, from medicine to biotechnology to heterogeneous nucleation, the question about its dominant forces and control arises. In this study, we used ellipsometry and quartz-crystal microbalance with dissipation (QCM-D), as well as density-functional theory (DFT) to obtain insight into the mechanism behind a wetting transition of a protein solution. We established that using multivalent ions in a net negatively charged globular protein solution (BSA) can either cause simple adsorption on a negatively charged interface, or a (diverging) wetting layer when approaching liquid-liquid phase separation (LLPS) by changing protein concentration (c p) or temperature (T). We observed that the water to protein ratio in the wetting layer is substantially larger compared to simple adsorption. In the corresponding theoretical model, we treated the proteins as limited-valence (patchy) particles and identified a wetting transition for this complex system. This wetting is driven by a bulk instability introduced by metastable LLPS exposed to an ion-activated attractive substrate. Controlling and understanding protein adsorption is key to a number of phenomena in biomaterial science and medical devices such as biocompatibility, osseointegration, inflammation and contamination 1-3. One way to systematically study the underlying interaction mechanisms between proteins and solid surfaces is to alter the surface chemistry and topography e.g. through the use of alloys of different composition, self-assembled monolayers (SAMs), membrane bilayers, polymer brushes, smart biomaterials or tissue engineering 1,4-7. An interesting, and in fact efficient, alternative to modifying the surface properties would be to tune protein adsorption by exploiting suitable thermodynamic conditions, i.e. conditions that favour a certain level of adsorption driven by the underlying bulk phase behaviour. Adsorption at solid-liquid interfaces is the result of sufficiently attractive substrate-fluid and intermolecular fluid interactions. Strongly enhanced or macroscopic adsorption may in particular result in the vicinity of bulk instability regions, a phenomenon called 'wetting' that is mostly explored in the statistical physics of 'simple liquids' 8-10. Although the bulk phase behaviour of protein solutions shares intriguing similarities with that of suspensions of spherical colloids 11-14 , it is not clear a priori to what extent surface phenomena such as wetting can be transferred to solutions of proteins, in view of their significant complexity and patchy nature 15-22. Furthermore, the tailoring of adsorption beyond the monolayer would be of significant importance for the understanding of e.g. heterogeneous nucleation of crystals or for improving the biocompatibility of implants by pre adsorption, which makes this study not only important fundamentally, but also for applications. Salts provide a versatile way to manipulate the interactions. Specifically, multivalent ions can induce novel effects at interfaces, going well b...

The protein human serum albumin (HSA) is able to readily crystallize in the presence of trivalent cations, whereas this is not the case for the homologous protein in cattle, bovine serum albumin (BSA), although both have analogous functions as well as similar physicochemical properties. To understand the underlying interactions and mechanisms, we investigated their bulk phase behavior with CeCl 3 by visual inspection, optical microscopy, and small-angle X-ray scattering (SAXS). The results reveal that both proteins undergo reentrant condensation and liquid−liquid phase separation (LLPS). However, the LLPS binodal for HSA shifts toward lower protein concentrations than that for BSA, indicating a stronger intermolecular attraction in HSA solutions at the same compositions, consistent with SAXS measurements. Moreover, crystallization occurs within the condensed regime of HSA, but no crystallization was observed for BSA. Adsorption studies at a hydrophilic SiO 2 surface demonstrate that both systems show reentrant adsorption with a higher amount of adsorbed BSA, likely due to enhanced cation-mediated interactions and/or hydrogen bonds. We conclude that the higher surface hydrophobicity of HSA could explain the experimental observations. These additional hydrophobic interactions not only strengthen the attraction between the proteins but also provide directional and specific protein−protein contacts, which are favored for protein crystallization. This work further demonstrates the sensitivity and complexity of protein interactions in solution: subtle differences in molecular structure lead to a dramatic change in their phase behavior. Generalization of these findings can pave the way toward, e.g., better drug design and improve medical treatment.

No abstract