No abstract

Lactoferrin (LF) and β-lactoglobulin (BLG), strongly basic and weakly acidic bovine milk proteins, form optically clear coacervates under highly limited conditions of pH, ionic strength I, total protein concentration C(P), and BLG:LF stoichiometry. At 1:1 weight ratio, the coacervate composition has the same stoichiometry as its supernatant, which along with DLS measurements is consistent with an average structure LF(BLG2)2. In contrast to coacervation involving polyelectrolytes here, coacervates only form at I < 20 mM. The range of pH at which coacervation occurs is similarly narrow, ca. 5.7-6.2. On the other hand, suppression of coacervation is observed at high C(P), similar to the behavior of some polyelectrolyte-colloid systems. It is proposed that the structural homogeneity of complexes versus coacervates with polyelectrolytes greatly reduces the entropy of coacervation (both chain configuration and counterion loss) so that a very precise balance of repulsive and attractive forces is required for phase separation of the coacervate equilibrium state. The liquid-liquid phase transition can however be obscured by the kinetics of BLG aggregation which can compete with coacervation by depletion of BLG.

The aggregation of insulin is complicated by the coexistence of various multimers, especially in the presence of Zn(2+). Most investigations of insulin multimerization tend to overlook aggregation kinetics, while studies of insulin aggregation generally pay little attention to multimerization. A clear understanding of the starting multimer state of insulin is necessary for the elucidation of its aggregation mechanism. In this work, the native-state aggregation of insulin as either the Zn-insulin hexamer or the Zn-free dimer was studied by turbidimetry and dynamic light scattering, at low ionic strength and pH near pI. The two states were achieved by varying the Zn(2+) content of insulin at low concentrations, in accordance with size-exclusion chromatography results and literature findings (Tantipolphan, R.; Romeijn, S.; Engelsman, J. d.; Torosantucci, R.; Rasmussen, T.; Jiskoot, W. J. Pharm. Biomed. 2010, 52, 195). The much greater aggregation rate and limiting turbidity (τ(∞)) for the Zn-insulin hexamer relative to the Zn-free dimer was explained by their different aggregation mechanisms. Sequential first-order kinetic regimes and the concentration dependence of τ(∞) for the Zn-insulin hexamer indicate a nucleation and growth mechanism, as proposed by Wang and Kurganov (Wang, K.; Kurganov, B. I. Biophys. Chem. 2003, 106, 97). The pure second-order process for the Zn-free dimer suggests isodesmic aggregation, consistent with the literature. The aggregation behavior at an intermediate Zn(2+) concentration appears to be the sum of the two processes.

Lactoferrin (LF) and β-lactoglobulin (BLG) are among the protein pairs that exhibit heteroprotein coacervation, a unique and relatively unexamined type of liquid-liquid phase separation (LLPS). In prior work we found that LF and BLG undergo coacervation at highly constrained conditions of pH, ionic strength and protein stoichiometry. The molar stoichiometry in coacervate and supernatant is LF : BLG2 1 : 2 (where BLG2 represents the 38 kDa BLG dimer), suggesting that this is the primary unit of the coacervate. The precise balance of repulsive and attractive forces among these units, thought to stabilize the coacervate, is achieved only at limited conditions of pH and I. Our purpose here is to define the process by which such structural units form, and to elucidate the forces among them that lead to the long-range order found in equilibrium coacervates. We use confocal laser scanning microscopy (CLSM), small angle neutron scattering (SANS), and rheology to (1) define the uniformity of interprotein spacing within the coacervate phase, (2) verify structural unit dimensions and spacing, and (3) rationalize bulk fluid properties in terms of inter-unit forces. Electrostatic modeling is used in concert with SANS to develop a molecular model for the primary unit of the coacervate that accounts for bulk viscoelastic properties. Modeling suggests that the charge anisotropies of the two proteins stabilize the dipole-like LF(BLG2)2 primary unit, while assembly of these dipoles into higher order equilibrium structures governs the macroscopic properties of the coacervate.

The aggregation of β-lactoglobulin (BLG) near its isoelectric point was studied as a function of ionic strength and pH. We compared the behavior of native BLG with those of its two isoforms, BLG-A and BLG-B, and with that of a protein with a very similar pI, bovine serum albumin (BSA). Rates of aggregation were obtained through a highly precise and convenient pH/turbidimetric titration that measures transmittance to ±0.05 %T. A comparison of BLG and BSA suggests that the difference between pHmax (the pH of the maximum aggregation rate) and pI is systematically related to the nature of protein charge asymmetry, as further supported by the effect of localized charge density on the dramatically different aggregation rates of the two BLG isoforms. Kinetic measurements including very short time periods show well-differentiated first and second steps. BLG was analyzed by light scattering under conditions corresponding to maxima in the first and second steps. Dynamic light scattering (DLS) was used to monitor the kinetics, and static light scattering (SLS) was used to evaluate the aggregate structure fractal dimensions at different quench points. The rate of the first step is relatively symmetrical around pHmax and is attributed to the local charges within the negative domain of the free protein. In contrast, the remarkably linear pH dependence of the second step is related to the uniform reduction in global protein charge with increasing pH below pI, accompanied by an attractive force due to surface charge fluctuations.

The effect of heparin on both native and denatured protein aggregation was investigated by turbidimetry and dynamic light scattering (DLS). Turbidimetric data show that heparin is capable of inhibiting and reversing the native aggregation of bovine serum albumin (BSA), β-lactoglobulin (BLG), and Zn−insulin at a pH near pI and at low ionic strength I; however, the results vary with regard to the range of pH, I, and protein−heparin stoichiometry required to achieve these effects. The kinetics of this process were studied to determine the mechanism by which interaction with heparin could result in inhibition or reversal of native protein aggregates. For each protein, the binding of heparin to distinctive intermediate aggregates formed at different times in the aggregation process dictates the outcome of complexation. This differential binding was explained by changes in the affinity of a given protein for heparin, partly due to the effects of protein charge anisotropy as visualized by electrostatic modeling. The heparin effect can be further extended to include inhibition of denaturing protein aggregation, as seen from the kinetics of BLG aggregation under conditions of thermally induced unfolding with and without heparin.

There has been a resurgence of interest in complex coacervation, a form of liquid-liquid phase separation (LLPS) in systems of oppositely charged macroions, but very few reports describe the somewhat anomalous coacervation between acidic and basic proteins, which occurs under very narrow ranges of conditions. We sought to identify the roles of equilibrium interprotein complexes during the coacervation of β-lactoglobulin dimer (BLG2) with lactoferrin (LF) and found that this LLPS arises specifically from LF(BLG2)2. We followed the progress of complexation and coacervation as a function of r, the LF/BLG molar ratio, using turbidity to monitor the degree of coacervation and proton release and dynamic light scattering (DLS) to assess the stoichiometry and abundance of complexes. Isothermal titration calorimetry (ITC) showed that initial complex formation is endothermic, but a large exotherm related to coacervate formation obscured other regions. On the basis of turbidimetry, proton release, and DLS, we propose a speciation diagram that presents the abundance of various complexes as a function of r. Although multiple species could be simultaneously present, distinct regions could be identified corresponding to equilibria among particular protein pairs.

Surface-bound water in protein solutions has been identified with a reduction in its freezing point. We studied the presence of such nonfreezing water (NFW) in various protein-polyelectrolyte, micelle-polyelectrolyte, and protein-protein (heteroprotein) coacervates, along with appropriate concentrated solutions of macromolecules alone, finding up to 15% w/w NFW for the heteroprotein coacervate of lactoferrin (LF) and β-lactoglobulin (BLG). The level of NFW is always higher in coacervates than in the control (single macromolecule) systems, particularly for protein-containing coacervates: a coacervate of bovine serum albumin (BSA) and poly(dimethyldiallylammonium chloride) (PDADMAC) showed a ratio of NFW/protein twice that of BSA alone (0.6 vs 0.3), with a similarly high ratio for LF-BLG coacervate. These results are attributed to the maximization of water-protein contacts, structural features that reflect the mode of sample assembly, as they are not seen in a noncoacervated LF-BLG solution with identical concentrations of all species.