Role of Cosolute–Protein Interactions in the Dissociation of

Sep 21, 2015 - A descending order of cosolute binding number per mAb oligomer was found: .... Astra 4.90.07 Software (Wyatt Technology Corporation, Sa...
0 downloads 0 Views 6MB Size
Article pubs.acs.org/JPCB

Role of Cosolute−Protein Interactions in the Dissociation of Monoclonal Antibody Clusters Thomas M. Scherer* Genentech (a Member of the Roche Group), Late Stage Pharmaceutical Development, 1 DNA Way, South San Francisco, California 94080, United States S Supporting Information *

ABSTRACT: The solution thermodynamics and interactions of a reversibly self-associating monoclonal IgG1 antibody (mAb1) have been investigated as a function of cosolute type (NaCl, NaSCN, Arginine-HCl) and cosolute concentration over a wide range of protein concentrations (1−275 mg/mL) using static light scattering. Within the framework of multicomponent solution thermodynamic theory, the preferential interactions of cosolutes with mAb1 were evaluated in the concentration limit of the system. The overall interactions of cosolutes with mAb1 relative to the bulk solutions appear to be very weak, but preferential interactions alone are insufficient to account for the cosolutes’ dissociation of mAb1 clusters. As a complementary approach to understanding cosolute interactions at high concentrations, mAb1 concentration dependent light scattering was also analyzed with models of interacting hard spheres (IHS). Evaluating the cosolute concentration dependence of association constants and states as specific binding interactions to dissociate mAb1 oligomers, numerical estimates of cosolute molecules required to dissociate oligomers were obtained. A descending order of cosolute binding number per mAb oligomer was found: arginine-Cl (∼17) > NaSCN (∼12) > NaCl (∼8). Differences in the cosolute effectiveness in reducing mAb1 equilibrium oligomer formation (self-association) are in part attributable to ion binding and salt identity (Hofmeister series) effects that reduce the attractive electrostatic interactions of mAb1. However, only Arg-Cl effectively disrupts mAb1 dimer formation, likely through interactions with the hydrophobic surface features on mAb1. Results indicate that localized cosolute−protein binding interactions have an important role in modulating nonspecific protein self- or heteroassociations at high concentrations.



INTRODUCTION The important role of cosolutes in protein solution thermodynamics was identified through early investigations by pioneers1−5 of the field and remains central to understanding the biophysics of protein solutions, protein interactions, and protein stability. Proteins seldom exist outside an aqueous medium, whether as dilute solutions in vitro and in extracellular environments, or as high concentration systems found inside cells, tissues, and (increasingly) biopharmaceutical products. Inorganic salts, amino acids, carbohydrates, and other low molecular weight osmolytes often also found in the aqueous and cellular media modulate the fundamental interactions of proteins with water as well as with each other. However, while the effects of specific cosolutes are widely known and utilized (e.g., urea, guanidine−HCl, Hofmeister series salts, ammonium sulfate to name just a few), the specific roles of cosolutes and the details of their interactions with proteins are often still much less clear, and remain the subject of continuing investigations to provide a general framework for understanding their wide ranging effects.6−10 The biotechnological development of protein therapeutics has not only given the investigation of protein−cosolute interactions a new urgency, but also a set of well-defined systems [including monoclonal antibodies (mAbs)] for study. © 2015 American Chemical Society

Protein−cosolute interactions have important implications for many aspects of pharmaceutical and industrial protein purification processes and drug product stabilization. However, currently the applications of cosolutes in biotechnological processes and products are typically obtained through empirical optimization procedures as a matter of practical expediency. The investigation and new understanding of the effective benefits and causative features of a cosolute’s role in protein processing and stabilization can provide distinct technological and strategic advantages.11,12 Naturally occurring osmolytes (e.g., sucrose, glycine betaine, trehalose, urea, and guanidine) and inorganic electrolyte salts (e.g., “Hofmeister series” salts) are well-known cosolutes that modify the stability and interactions of proteins in solution.13,14 Arginine, an amino acid with numerous biotechnological applications, is now recognized as a cosolute with unique effects on the solution behavior of proteins that can be advantageous during protein purification, refolding, and storage.11 Recent investigations of arginine solution thermodyReceived: August 4, 2015 Revised: September 16, 2015 Published: September 21, 2015 13027

DOI: 10.1021/acs.jpcb.5b07568 J. Phys. Chem. B 2015, 119, 13027−13038

Article

The Journal of Physical Chemistry B

presents a richly nuanced and more complex view of protein− cosolute interactions.

namics also suggest a complex balance of preferential binding and exclusion interactions in a protein specific manner.15−18 Preferential interactions and specific binding interactions are often viewed as theories at opposite ends of the spectrum to generally describe the thermodynamics of how cosolutes interact with proteins in aqueous solutions. Light scattering permits the measurement of solution thermodynamics of macromolecule−cosolute interactions as well as the macromolecular interactions. The development of multicomponent solution theory during the 1950s demonstrated that light scattering can be used to gain insights into the solution protein−protein interactions as well as the interactions between cosolutes and proteins (but requiring careful observation of dilute solution limits).4,5,19 From the vantage point of preferential interactions, the cosolute effects on protein solution thermodynamics have been previously investigated for several instances of salt−protein interactions.20,21 Similarly, Herskovits and coauthors produced a less well-known, but still compelling, case for the specific binding role of cosolutes as the dissociative agents of protein tertiary structure and complexes.22−27 Other light scattering investigations into the role of cosolutes on protein solubility have so far produced more modest conceptual advances.13,28−30 Most, if not all, of this evaluative experimental work has been conducted in solution conditions that are highly dilute in protein and as nearly ideal as possible, out of necessity. Recent significant advances of static light scattering theory for the characterization of protein interactions at high concentrations have also been reported.31,32 These developments are based on scaled particle theory integrated effects of molecular crowding and intermolecular interactions, and have greatly facilitated static light scattering investigations of protein self-association at concentrations relevant to actual systems of interest, such as intracellular environments and biopharmaceutical products.33,34 We recently described the ionic strength and concentration dependent behavior of a monoclonal antibody (also identified as “mAb1” in numerous other citations),35−39 shown to selfassociate due in part to attractive, “patchy” electrostatic intermolecular interactions.33 Several salts or cosolutes have also previously been identified with different mitigating effects toward the attractive intermolecular interactions and rheological effects of the self-association of mAb1.40,41 In this contribution, we expand the use of SLS to investigate cosolute− protein interactions into the realm of highly nonideal solutions. The roles of NaCl, NaSCN, and arginine−hydrochloride (ArgCl) are explored to identify how cosolutes modulate protein interactions and solution thermodynamics under conditions with significant molecular crowding effects. Cosolute (from 20 to 600 mM) and protein concentration (1−275 mg/mL) dependent light scattering measurements characterize the effects of different cosolutes on mAb1 self-association in both dilute solution and at high mAb1 concentrations. mAb1 molecular weights obtained at infinite dilution are utilized to assess cosolute−protein preferential interactions relative to the bulk solution. An analysis of cosolute effects on the association constants and states (cluster sizes) over a wide range of mAb concentrations also provides more detailed insight into the specific interactions of NaCl, NaSCN, and arginine−HCl with the localized regions of protein surfaces thought to be responsible for reversible mAb1 oligomer cluster formation at high concentrations. The combined use of approaches to understand the cosolute-specific roles in mAb1 interactions



MATERIALS AND METHODS mAb1 is a humanized monoclonal antibody based on an IgG1 framework with κ-light chains. This antibody was expressed in Chinese hamster ovarian (CHO) cell lines, and purified by a series of chromatography steps, including protein A and ion exchange chromatography methods. The purified mAb1 was obtained as a concentrated solution from tangential flow filtration with added solution buffers and stabilizers at 196 mg/ mL. These stock mAb starting materials were stored at 2−8 °C until further use. Additional preparation of mAb solutions included dialysis using Spectrapore 6−8 kDa MWCO membrane (Spectrum Laboratories, CA). Approximately 50− 60 mL stock solutions were dialyzed against 1.2 L of low ionic strength buffer containing 30 mM histidine−HCl at pH 6.0 over 48 h at 2−8 °C for a minimum of 3 buffer exchanges. The dialyzed mAb solutions were filtered through 0.22 μm PVDF filters (Millipore Steriflip, Millipore Corp., MA) to remove large particulates. Typically, mAb concentrations of 140−150 mg/mL were obtained after dialysis. mAb1 antibody preparations consisted of >98% monomer as determined by size exclusion chromatrography analysis of samples diluted to 1 mg/ mL. To obtain the higher mAb concentrations, 10 mL of mAb dialyzed in 30 mM histidine−HCl, pH 6.0, was concentrated with Amicon YM30 Centriprep (Millipore Corp, MA) concentrators centrifuged at 2700 rpm (1500 G) with removal of filtrate at 3 h intervals until the desired concentration was obtained. Final mAb concentrations in the dialyzed and centrifugally concentrated preparations were determined by using gravimetric dilutions and absorptivities at 280 nm (A280) of 1.6 (mg/mL) cm−1 for mAb1 and measurement of UV absorption at 280 nm using an Agilent diode array spectrophotometer model 8453 with a 1 cm path length quartz cuvette. Extinction coefficients were determined by quantitative amino acid analysis. Deionized water from a Milli-Q Biocel purification system with 0.22 μm filter was used for all aqueous solution preparations. All buffers and reagents were analytical grade or higher purity. All buffer solutions were titrated to the correct pH and filtered using Millipore Stericup GV 0.22 μm filters. All mAb solutions used for light scattering experiments were prepared in 20 mL scintillation vials over 0.5−225 mg/mL by gravimetric dilution of known stock solution concentrations in a laminar flow hood. Sodium chloride (4.0 M), sodium thiocyanate (2.0 M), and arginine chloride (2.5 M) were prepared and filtered with 0.2 μm PVDF membranes before storage at room temperature until use. These cosolutes were introduced into the mAb solutions samples after an additional 0.1 μm filtered sample with gravimetric measurement of initial and final mass, with volumes adjusted to achieve the desired cosolute concentrations in each vial. All other sample preparation and handling procedures are described in detail elsewhere.33,41 Static Light Scattering from Solutions of a Single Species. For a single scattering species at arbitrary concentration, the generalized equation for Rayleigh scattering intensity (R(θ, w))42 reduces to the expression in eq 1. In this limiting case, which can be applied to systems of purified monodisperse molecules (e.g., many proteins), R(θ, w) is a function of the scattering angle θ and w/v concentration c2, and is directly related to the mass of the scattering species (M), the 13028

DOI: 10.1021/acs.jpcb.5b07568 J. Phys. Chem. B 2015, 119, 13027−13038

Article

The Journal of Physical Chemistry B Kc 2 R(θ = 0, c 2)

molecular size through the radius of gyration (RG), and solute interactions described by the virial coefficients (A2, A3, etc.) as shown below. Kc 2 1 = [1 + 2A 2 Mc 2 + 3A3Mc 2 2 + ...] R (θ , c ) M ⎤ ⎡ q2R G2 ⎢1 + + ...⎥ 3 ⎦ ⎣

⎧ ⎡ ex ⎪ c 2 ⎢⎡ ∂μ2 ⎤ 1 ⎨ 1 = + ⎢ ⎥ RT ⎢⎣⎣ ∂c 2 ⎦ (1 + B)2 M 2 ⎪ T , P , c3 ⎩ ⎤⎫ M 2(∂μ3 /∂c 2)T , P , c3 2 ⎥⎪ ⎬ − M3(∂μ3 /∂c3)T , P , c2 ⎥⎦⎪ ⎭

(1)

where K=

(3)

where 2 4πno(sin θ /2) 4π 2n2 ⎛⎜ dn ⎞⎟ q= 4⎝ ⎠ λo NAλo dc

B=

(∂n/∂c3)T , P , c2 (∂n/∂c 2)T , P , c3

[(∂c3/∂c 2)]T , μ1, μ3

In the limit of infinite dilution of component 2 (protein), eq 3 reduces to an expression for the measured molecular mass M2* at constant c3 (eq 4), and that within the constraints of this limit can be used to evaluate the extent of cosolute− macromolecule preferential interactions (dc3/dc2) as shown in eq 5.47

The system constant K includes the solution refractive index (n),43 the wavelength of incident light (λo), the refractive index increment of scattering solute (dn/dc), scattering vector (q) at scattering angle (θ), and Avogadro’s number (NA). In the limit of infinite dilution (c2 → 0) and the scattering vector q approaching 0, eq 1 yields the value of 1/M. In practice, light scattering measurements are made at both finite concentrations and angles, extrapolating to one or both limiting conditions to obtain M, the absolute molecular mass of the scattering species. Measurement of static light scattering intensity was conducted as a function of mAb concentration from 0.5 to 235 mg/mL, and as a function of cosolute concentrations (0−600 mM). Scattering data for each sample/vial was collected over an interval 5−10 min with data collection frequency of 12 points/ min. Astra 4.90.07 Software (Wyatt Technology Corporation, Santa Barbara, CA) was used to acquire and process the static light scattering data. Further analysis and calculations with the exported results were conducted in Microsoft Excel, Origin v7.5, and MATLAB R11b. Static Light Scattering from Multicomponent Solutions. Zimm44 and Debye provided the first fundamental understanding of Rayleigh light scattering from macromolecules in solutions as the solution refractive index variations due to concentration/composition fluctuations and density fluctuations. Subsequently, Kirkwood, Goldberg,4 and Stockmayer5 derived equations for light scattering from multicomponent solutions containing up to three components including solvent (component 1), macromolecule (2), and cosolute (3). The thermodynamic theory of multicomponent solutions has been comprehensively reviewed by Casassa and Eisenberg.19,45 The addition of a third component as cosolute (salts with counterions are considered one component) can have significant implications on the thermodynamics and evaluation of solution light scattering from polyelectrolyte molecules. Equation 3 below describes the effect of tertiary solution composition on the light scattering of macromolecular components with the ratio of the macromolecule mass M2 and cosolute mass M3 and the impact on the solution chemical potentials at constant temperature, pressure, and cosolute concentrations. The prefactor (1 + B)−2 accounts for cosolute modification of the solution refractive index and preferential interaction of cosolute components with the macromolecule.46

⎡ ⎤2 (∂n/∂c3)T , P , c2 * ⎢ [∂c3/∂c 2]T , μ3 ⎥ M2 = M2 1 + (∂n/∂c 2)T , P , c3 ⎢⎣ ⎥⎦

(4)

⎡ ⎤ ⎛ ∂c3 ⎞ ⎢ ⎛ M 2* ⎞ ⎥ (∂n/∂c 2)c3 =⎢ ⎜ − 1⎥ ⎟ ⎜ ⎟ M (∂n/∂c3)c2 ⎝ ∂c 2 ⎠T , μ , μ ⎣ ⎝ 2 ⎠c 2 → 0 ⎦ 1 3

(5)

Refractive index increment values (dn/dc3) of 0.172 mL/g were applied for the purely ionic cosolutes NaCl and NaSCN, that for for arginine−HCl was 0.201 mL/g, and a dn/dc2 value of 0.185 mL/g was used for mAb1. The expression for the Rayleigh scattering from multicomponent systems simplifies to the two-component eq 6 in the limit of weak interactions between cosolute and macromolecule.48 ⎤ ⎡ ex Kc 2 c 2 ⎛ ∂μ2 ⎞ ⎥ 1 ⎢ = 1+ ⎜ ⎟ R(θ = 0, c) M 2 ⎢⎣ RT ⎝ ∂c 2 ⎠ ⎥⎦ T ,p

(6)

In the case that cosolute effects have a negligible effect on the apparent molecular weight M2* obtained, the cosolute can be treated as part of the solvent matrix to facilitate investigation of mAb interactions at concentrations that exceed the first order terms of solution nonideality effects at high protein concentrations. At such concentrations, we apply recent developments in the theory and analysis of light scattering of high concentration solutions of interacting species of scatterers, briefly described below. Modeling of Macromolecular Solution Nonideality with Thermodynamic Expressions for Self-associating Hard Spheres. The use of interacting hard sphere models to interpret the light scattering behavior of protein solutions at high concentrations was developed by Minton, and its application to mAbs at high concentrations has been described in detail elsewhere.33,49 The association state(s) of the oligomer species (n and m) are obtained from fitting a model to the concentration dependence of light scattering with monomer mass (M1), where m-mer = reversible dimer, while n-mer = larger reversible oligomer species, along with association constants K1n and K1m for each cosolute concentration solution 13029

DOI: 10.1021/acs.jpcb.5b07568 J. Phys. Chem. B 2015, 119, 13027−13038

Article

The Journal of Physical Chemistry B

Figure 1. (a−c) Osmotic compressibility curves obtained for mAb1 concentrations from 10 to 225 mg/mL were prepared with 20, 100, 300, and 600 mM of each cosolute. Not all cosolute concentrations are shown here for clarity of illustration: (a) NaCl, (b) NaSCN, (c) arginine−HCl. The compressibility of a simple hard sphere model (R = 4.4 nm) representative of IgG1 mAb excluded volume is shown as solid black line for comparison.

Dissociation of Protein Oligomer Species or Clusters by Cosolutes. The use of cosolutes to affect the quaternary structure and subunit association states of protein complexes has been explored extensively.43,50,51 Light scattering methods were applied to the analysis of protein dissociation constants, previously interpreted in terms of cosolute properties (e.g., relative hydrophobicity, amino acid/peptide solubility), protein binding, and surface interactions. The treatment of the equilibrium dissociation constants for the reaction can be expressed as the equilibrium equation (below) between mAb species of oligomer number n in which the cosolute D directly interacts with x stoichiometry through protein binding.

condition. The effective hard spherical model also employs an adjustable parameter vexc, referred to as the effective specific exclusion volume, which reduces to the actual partial specific volume in the absence of electrostatic repulsion (i.e., near the isoionic pH of the protein). More generally, vexc exceeds the mAb partial specific volume to provide a measure of the repulsive intermolecular interactions in concentrated solution.49 In the absence of self-association, this model reproduces the scattering obtained for a uniform fluid of simple hard spheres.33 The data obtained for all angles at any single concentration were averaged to obtain an angle independent value of ⟨R(θ)⟩/ Ko since no angular dependence of scattered light intensity was observed as previously demonstrated.33 Using scripts and functions written in MATLAB (R2011b, Mathworks, Natick, MA), graciously provided by Allen Minton (National Institute of Health, Bethesda, MD),32 nonlinear least-squares modeling of the measured scattering intensity dependence upon the concentration of all scattering species (wtot) was carried out for each cosolute solution condition. To evaluate the association constants K1n and K1m for oligomer species n and m, respectively, over the cosolute range 20−600 mM, M1 was held constant with a value of 145 kDa.

MAbn + x D ⇄ n MAbDx / n

On the basis of the equilibrium reaction expression above, eq 10 was used to evaluate the association constants of mAb1 species, separately for dimer (m-mer) or higher oligomers (nmer), obtained from the best-fit interacting hard sphere models of light scattering measurements across the entire protein concentration range. 13030

DOI: 10.1021/acs.jpcb.5b07568 J. Phys. Chem. B 2015, 119, 13027−13038

Article

The Journal of Physical Chemistry B Kx =

[MAb*Dx / n]n x

[MAbn][D]

=

Kdiss (K o1n)−1 x = [D] [D]x

water were obtained from literature or measured: a (dn/dc3) of 0.172 mL/g was used for both NaCl and NaSCN, a (dn/dc3) of 0.201 mL/g was obtained for arginine−HCl, while a dn/dc2 of 0.185 mL/g for mAb1 was used. With these solution parameters known, eq 5 facilitates more thorough investigation of cosolute−protein thermodynamic interactions in dilute (mAb1) solutions. Preferential interaction terms obtained help quantify each distribution of each cosolute at the (solvent-accessible) mAb1 surface relative to the bulk cosolute concentration. The quantity dc3/dc2, obtained here using the experimentally more convenient units of g/mL, is closely related to the molal quantities found in solution thermodynamic derivations as a result of the relatively low salt concentrations employed. Thus, interpretations of the preferential interactions of cosolutes with mAb1 are derived directly from the experimentally obtained values without significant impact on the overall information gained. Figure 2 summarizes

(10)

When association or dissociation constants are obtained at several cosolute concentrations, the logarithmic relationship between association constants and cosolute concentration enables determination of cosolute binding stoichiometry required to mediate the protein−protein interactions.24,27



RESULTS AND DISCUSSION Static light scattering experiments with mAb1 previously highlighted the important contributions of electrostatic and excluded volume effects to protein interactions at high concentrations, and examined the thermodynamics of selfassociation behavior of mAb1 with both simple HS and multiple species of interacting HS models.33 The important but differential roles of monovalent salts and arginine−HCl cosolutes in modulating mAb1 intermolecular interactions, especially those found in highly concentrated solutions, were identified and qualitatively characterized by analysis of the excess chemical potential of the protein solutions.41 Here, we seek to more specifically interpret the cosolute−protein interactions under conditions of thermodynamic equilibrium, and evaluate the relative contributions of the electrostatic and hydrophobic interactions present with the use of different cosolutes. Rayleigh light scattering measurements of mAb1 were made relative to the appropriate cosolute and buffer containing solution with independent solutions placed in separate vials. This format permitted samples to be measured repeatedly and after equilibration with cosolute concentrations that ranged from 0 to 600 mM in NaCl, NaSCN, or arginine− HCl (Arg-Cl). The osmotic compressibility curves [Kc/R(θ = 0)] obtained for solutions are shown as a function of mAb1 concentration in Figure 1a−c for solutions with 0, 20, 100, 300, 600 mM added NaCl (a), NaSCN (b), and Arg-Cl (c). Figure 1a−c shows (a partial data set for clarity) representative results for experiments across cosolute concentrations, while the full data set includes additional intermediate cosolute levels between 20 and 600 mM. The full data set forms the basis of all subsequent analysis and comparison between the experiments with NaCl, NaSCN, and Arg-Cl cosolute systems. In general, the obvious nonlinearity of the Kc/R versus mAb1 concentration is typical of compressibility curves with competing effects of attractive interactions from protein selfassociation/clustering and repulsive volume exclusion (crowding) effects on intermolecular interactions, spatial distributions, and concentration fluctuations. Quantitative results from the dilute regime of light scattering experiments in Supporting Information Table 1 summarize the molar mass obtained in the limit of infinite dilution (M2), and the osmotic second virial coefficients (A2) determined from concentrations 1−10 mg/mL for each cosolute system. These results confirm that the solutions consisted essentially of monomeric mAb species, and that the self-association/cluster formation is in fact reversible. At low ionic strength the A2 values are clearly negative, which increased with each of the added cosolutes, with values ranging from −0.22 to 0.76 × 10−4 mol mL g−2 at 600 mM cosolute levels. Analysis of weightaverage molar mass values (M2*) at each cosolute concentration and M2 (experimentally obtained value in absence of cosolute) in the limit of infinite protein (mAb1) dilution (eq 5) permits interpretation of the cosolute−protein preferential interactions. Refractive index increment values for the solutes in

Figure 2. Calculated quantities of preferential interaction parameters (dc3/dc2) in the limit of protein concentration → 0 obtained from eq 5. These ratios provide the distribution of cosolute in the vicinity of the protein, relative to the cosolute concentration of the bulk solution.

the calculated preferential interaction coefficients as a function of cosolute concentration (in units of mass/vol). On the basis of the overall positive dc3/dc2 values observed, NaCl and NaSCN appear to preferentially bind to mAb1, albeit very weakly and without well-defined concentration dependence. In contrast, arginine−HCl interactions with mAb1 indicate cosolute preferential exclusion from the surface of the protein (preferentially hydrating mAb1) with increasing cosolute concentration. These observations are consistent with the expected behavior for both sodium salts52,53 and arginine− HCl.15,54 The thermodynamic preferential interaction parameters account for the net average interactions of cosolutes with the protein surfaces; however, these divergent preferential interactions appear incongruous with the observation that the three cosolutes similarly reduce A2 and the overall tendency of mAb1 to self-associate. It thus appears that preferential interactions from dilute solution thermodynamic evaluation only provide a partial understanding of each cosolute’s molecular interactions with mAb1. The weak preferential interactions of NaCl, NaSCN, and ArgCl observed, as well as the small net charge on mAb1 (+6.4 in 30 mM Hist-Cl, pH 6.0 solution),55 permit the reasonable 13031

DOI: 10.1021/acs.jpcb.5b07568 J. Phys. Chem. B 2015, 119, 13027−13038

Article

The Journal of Physical Chemistry B

Table 1. Best-Fit Parameters to Light Scattering Data (1− 200 mg/mL) Obtained from IHS Model Analysis of mAb1 Solutions (Monomer Mass (M1) = 145 kDa) in the Presence of NaCl as Cosolute NaCl (mM)

m

log Ko1m

n

log Ko1n

Vex (cm3/g)

20 20 40 75 100 200 300 600

2 2 2 2 2 2 2 2

3.7 0.0 3.4 4.2 4.3 3.8 3.2 3.8

7 6 6 6 6 5 4 4

23.8 19.1 19.7 20.4 20.3 15.0 10.5 10.0

2.15 1.98 2.00 1.90 1.94 1.95 1.82 1.58

Table 2. Best-Fit Parameters to Light Scattering Data (1− 200 mg/mL) Obtained from IHS Model Analysis of mAb1 Solutions (Monomer Mass (M1) = 145 kDa) in the Presence of NaSCN as Cosolute NaSCN (mM)

m

log Ko1m

n

log Ko1n

Vex (cm3/g)

20 40 100 100 200 300 600 600

2 2 2 2 2 2 2

0.0 0.0 0.0 0.0 4.4 3.6 2.9

8 6 4 5 4 4 4

26.1 19.2 10.8 13.6 11.2 8.5 7.0 0

2.14 1.74 1.69 1.89 1.82 1.81 2.20 1.53

Table 3. Best-Fit Parameters to Light Scattering Data (1− 200 mg/mL) Obtained from IHS Model Analysis of mAb1 Solutions (Monomer Mass (M1) = 145 kDa) in the Presence of ArgCl as Cosolute ArgCl (mM)

m

log Ko1m

n

log Ko1n

Vex (cm3/g)

20 40 75 75 100 100 150 200 200 300 600

2 2 2 2 2 2 2 2 2 2

4.4 4.4 4.0 3.8 4.7 3.5 3.6 3.6 2.7 3.2 0

7 6 5 5 5 4 4 4 3

25.0 20.6 15.7 15.3 16.6 11.2 10.7 10.4 6.7

2.28 2.09 2.00 2.00 2.14 1.97 1.98 2.08 1.88 1.74 1.69

0

multiple interacting (associating) hard spheres (IHS).32 Using this previously demonstrated approach,33,34 the mAb1 concentration dependent scattering intensity data (R/K) was fit by hard sphere models of up to 3 interacting species, including monomer and oligomers of different sizes and association constants. Figure 3a−c shows the light scattering data (experimental data is color-coded) obtained as well as the resulting best fit models (solid lines) for 20−600 mM cosolute levels. Qualitatively, the resulting fits indicate that the experimental light scattering data can be well-described by the self-association of 145 kDa monomer species with effective excluded volumes ranging from 1.5 to 2.3 cm3/g in a manner dependent on the solution cosolute/salt concentrations. Best fit parameters are summarized in data Tables 1−3. These results were obtained by fitting each solute condition independently using MatLab modules, a mAb1 molecular mass (145 kDa) at

Figure 3. Light scattering intensity data as a function of mAb1 concentration c2 (symbols) and best fit results (solid lines) obtained from interacting hard sphere (IHS) model analysis for each cosolute concentration (20−600 mM) for (a) NaCl, (b) NaSCN, and (c) arginine−HCl.

simplifying approximation of mAb1 solutions with 0−600 mM cosolutes as two-component solutions. Recognizing the cosolute as part of the background solvent matrix greatly facilitates further data treatment of light scattering data from high concentration mAb1 tertiary solutions with models of 13032

DOI: 10.1021/acs.jpcb.5b07568 J. Phys. Chem. B 2015, 119, 13027−13038

Article

The Journal of Physical Chemistry B

Figure 4. Mass fraction of mAb1 species in solution as a function of the total protein concentration calculated from IHS association constants, shown for mAb1 in the presence of Arg-Cl: (a) 30 mM buffer alone, (b) 20 mM Arg-Cl, (c) 100 mM Arg-Cl, (d) 300 mM Arg-Cl, (e) 600 mM Arg-Cl. The data illustrate how oligomer concentrations increase with protein concentration, and also show how increasing Arg-Cl concentrations reduce the presence of n-mer oligomers and finally also dimer species (only monomer throughout concentration range) in part d.

oligomer (n-mer) species of up to octamer were required to fit the concentration dependence of scattering profiles. Three different preparations of mAb1 in buffer were used to conduct all subsequent experiments with NaCl, NaSCN, and arginine−HCl cosolute systems. Stock material in buffer (30 mM histidine chloride, pH 6.0) prepared by dialysis followed by concentration may have had small differences in buffer concentration or pH, with a mild impact on the scattering profiles of the starting solutions (Supporting Information Figure S1). Initially, addition of low levels of cosolutes (20−40 mM) increased the scattering intensity from mAb1 solutions across the concentration range. With 20−40 mM NaCl and ArgCl cosolute additions the extent of this effect is quite similar; however, with the addition of low concentrations NaSCN appears to affect the scattering intensity more. As the

infinite dilution as a constrained parameter, with varying m-mer and n-mer to obtain values for association constants and effective excluded volume. Model fits obtained were evaluated on the basis of fit residuals (data not shown) and degree of fit obtained. Where two results for a single cosolute condition are reported in Tables 1−3, the two association models provided comparable results in terms of residuals and degree of fit. The interacting HS model results can be considered semiquantitative estimates of the equilibrium mAb1 oligomer or cluster species present over a wide range of concentrations and solution conditions. Low salt/cosolute concentration data sets (20 mM salts in particular) were more difficult to fit for reasons noted previously,32,33 but are evaluated here to provide a more complete analysis of the effects of ionic strength and cosolute type. At low salt concentrations, dimer (m-mer) and higher 13033

DOI: 10.1021/acs.jpcb.5b07568 J. Phys. Chem. B 2015, 119, 13027−13038

Article

The Journal of Physical Chemistry B

Figure 5. Representative analysis of mAb1 association constants as a function of cosolute concentration (Arg-Cl) within the framework of specific cosolute−protein binding interactions (a) for monomer−dimer association Ka, and (b) for monomer−oligomer association constants obtained from IHS model analysis of the concentration dependent light scattering data (1−225 mg/mL).

mass fractions. In the presence of 600 mM arginine−HCl, only monomer mAb1 species were observed across the full antibody concentration range (Figure 4e). Previously, the formation of equilibrium mAb1 clusters/ oligomers was established to be due to the presence of weak, nonspecific interactions between protein molecules that become stronger and more significant at high concentrations.33,41 To gain additional insights into of the role of cosolutes in mediating mAb1 interactions at high concentrations, the oligomer association constants were analyzed as a function of cosolute concentration using an approach first developed by Herskovits et al.25,27,50 Underlying this approach is the assumption that the effects of cosolutes on protein subunit dissociation are due to weak but specific binding interactions with the molecular surface features at the protein− protein assembly interface. Herskovits’ work demonstrated that the approximate stoichiometry of cosolute binding can be obtained from the dissociation behavior as a function of cosolute concentration. The cosolute concentration dependent IgG1 oligomer dissociation at high concentrations and equilibrium is entirely analogous, with added complexity of solution nonideality due to protein excluded volume effects. It should also be noted that, for the cosolutes of interest here, concentrations up to and including 600 mM have been shown to not affect protein denaturing or unfolding, and that monomeric IgG1 proteins retain their secondary, tertiary, and quaternary structural elements under these conditions.18 Thus, the association constants obtained from the IHS models of concentration dependent light scattering data (in Tables 1−3) can be interpreted utilizing eq 10. Using this approach we are able to more clearly identify the different specific interactions of NaCl, NaSCN, and Arg-Cl in the reduction of both dimer (mmer) and larger mAb1 oligomer (n-mer) species. Figure 5 shows log−log plots for both monomer−dimer (K°1m) and monomer−oligomer (K°1n) association constants as a function of Arg-Cl concentration as representative example (see Supporting Information Figure S2a,b for NaCl and NaSCN data analysis). Even though the analysis of oligomer dissociation has the combination of the data from all n-mer species (4−8 mers), this is justified since association constants (K°1n) are determined with reference to the monomeric state in the IHS models. The linear analysis for Arg-Cl, NaCl, and

cosolute levels were incrementally increased, the tendency of mAb1 to form higher oligomeric species was reduced (qualitatively evident in Figure 3a−c), and also in terms of association constants (Ko1n) and n-mer size (from 8 to 4 monomer units) obtained. With increasing cosolute levels above 100−200 mM, significant differences between the effects of cosolute species become apparent in Figure 3a−c. With Arg and NaSCN, the increase of cosolute levels clearly decreases the scattering intensities across the mAb1 concentration range. This manifests in the analysis of IHS models with n-mer species producing excellent fits to the scattering data with only 1 species (dimer) or no additional species required for fits of 0.3−0.6 M Arg-Cl and NaSCN solutions. At the same time NaCl, an effectively neutral ionic solute in the Hofmeister series, reduced the self-association of mAb1 to a much more limited extent when compared to NaSCN and Arg-Cl systems. As a purely ionic cosolute, NaSCN appears to be distinct from NaCl at 20−40 mM ionic strength where scattering intensities were higher, and at >100 mM where increasing levels subsequently decreased scattering intensities to a greater extent than NaCl. The addition of arginine−HCl to solutions produced consistent decreases in R/K scattering profiles and mAb1 self-association to a greater extent than NaCl. In the presence of 600 mM Arg-Cl, the scattering data and the IHS model fit results indicate mAb1 scattering is that of a single, noninteracting hard sphere species, consistent with the complete elimination of mAb1 self-association. mAb1 self-association and cluster formation in different cosolute systems can be perhaps more readily understood in terms of the amounts of each scattering species (monomer, dimer, and oligomer) present as a function of the total protein concentration. Figure 4a−e illustrates the mass fractions of monomer, dimer (m-mer), and higher oligomer (n-mer) species present with log-scale concentration in solutions of 30 mM Hist-Cl buffer (a), 20 mM (b), 100 mM (c), 300 mM (d), and 600 mM (e) levels of arginine−HCl as an example. Of note is the significant increase in mass fractions of large mAb1 oligomers at high concentrations in solutions with low Arg-Cl levels, and the generally weak concentration dependence of the formation of dimer species. As the Arg-Cl concentrations in the solutions increase to 100−600 mM, Figure 4b−d also shows the arginine dependent reduction of cluster/oligomer sizes and 13034

DOI: 10.1021/acs.jpcb.5b07568 J. Phys. Chem. B 2015, 119, 13027−13038

Article

The Journal of Physical Chemistry B

Figure 6. Total number of binding NaCl, NaSCN, and Arg-Cl cosolute moles per mol mAb1 species in solution derived from c3 concentration dependent analysis of the association constants (a) for n-mer−cosolute interactions and (b) for dimer−cosolute interactions.

Figure 7. Surface feature distribution calculated for mAb1 derived from PyMol homology models shown for side view of the full mAb and an enlarged side view of CDR only. The top two images show charge distribution and patchiness (blue = positive charge, red = negative charge) while in the bottom two images more brightly red areas are hydrophobic features and white features are hydrophilic surfaces.

13035

DOI: 10.1021/acs.jpcb.5b07568 J. Phys. Chem. B 2015, 119, 13027−13038

Article

The Journal of Physical Chemistry B

features.7,59 Of even greater interest, the differences in binding NaCl and ArgCl molecule numbers (∼8 vs 16 mol/mol) to mAb1 oligomer reveal that arginine−protein interactions impart the determinant factor for complete dissociation of mAb1 oligomer species. While the analysis shows ∼8 NaCl ions interact with negatively and positively charged patches, an additional ∼8 arginine molecules effectively bind to the oligomer’s interaction surfaces, predominantly between Fabs. Taken together, these results present a clearer understanding of the cosolute interactions with mAb1 dimer and oligomer species, as well as the types of intermolecular features between mAb1 monomers that promote the formation of such species in the first place. The observations here, as well as those from previously published experimental work and simulations, show that ionic cosolutes are generally effective in dissociating larger oligomers formed by complementary charged patches in the two Fabs, and are effective in reducing the mAb1 solution viscosities at pH 6.0.40,60,61 Furthermore, the documented stabilizing role of Arg-Cl in protein refolding intermediates11 and the dissociative effects on mAb1 dimers observed here are consistent with hydrophobic interactions (Fab and/or Fc surface features) providing a principle driving force for mAb1 assembly into dimers. Both Arg-Cl and NaSCN bind in greater numbers to mAb oligomers, further evidence that one or more hydrophobic patches also play a role in mAb1 self-association and cluster formation. Direct cosolute interactions through surface feature specific protein binding are consequently shown to mitigate the formation of reversible mAb1 oligomers through weak intermolecular interactions, rather than the limited cumulative effects of nonspecific preferential cosolute interactions.

NaSCN cosolute systems allows semiquantitative determination of mol cosolute/mol mAb1 oligomer binding, found summarized in Figure 6a,b for dimer and oligomer species, respectively. This analysis reveals that ArgCl has the greatest number of binding/interacting molecules (∼16) per mAb1 oligomer, while NaCl has the least with only ∼8 binding mol/ mol mAb oligomer. The number of bound NaSCN ions (∼12) also significantly (as judged from the nonoverlapping SE of slopes) exceeded the number of bound NaCl ions. Differences in NaSCN and ArgCl binding are present but not as clearly evident in this thermodynamic analysis or in the chemical potentials observed at high concentrations.41 This thermodynamic evaluation shows that overall the greater number of cosolute molecules binding to mAb1 oligomers species correlated positively with the cosolutes’ ability to reduce mAb1 self-association. The authors propose that the cosolute molecules interacting with a greater number (relative increase) of mAb1 do so as a consequence of more numerous interactions with a greater surface area/patch size or number of protein surface features to effectively diminish the oligomeric protein−protein interactions. However, the different cosolutes have very different binding behaviors toward the mAb1 dimer species, as illustrated in Figure 6b. Only ArgCl binds to (∼2−3 mol/mol) and dissociates the mAb1 dimer as well as the oligomeric mAb1 structures. As shown in Figure 6b (and Figure S3a), NaCl was found to have no binding interactions with mAb1 dimer species; thus, NaCl is capable of only dissociating the interactions present that form larger oligomeric mAb structures. The analysis of NaSCN binding to dimer species in Supporting Information Figure S3b also indicates that NaSCN promotes (3 mol/mol) dimer species formation (negative binding feature), but whether these are truly thermodynamic equilibrium dimer species or irreversible aggregates formed in the presence of NaSCN remains an open question. Published results on proteins in general would suggest the latter possibility could be at work in dilute solutions.56−58 There are (at least) two apparent cosolute concentration ranges over which effective cosolute−mAb1 interactions were observed, as suggested previously.41 Qualitatively, all three cosolutes appear to screen the attractive charge−charge interactions of mAb1 with little obvious differentiation at low cosolute compositions (20−200 mM). At higher cosolute concentrations (from 200 to 600 mM) the data increasingly reflect the cosolute’s unique effects on the self-association of mAb1, emphasizing that cosolute−protein interactions observed here, including direct binding, are quite weak in nature. The more quantitative thermodynamic analysis here also shows that only a small number of bound cosolute molecules (8−16 per oligomer) effectively reduced mAb1 self-association. These results also provide experimental evidence that NaCl, NaSCN, and ArgCl cosolutes interact specifically with localized regions of the mAb surface. (See Figure 7 for charge and hydrophobicity maps of mAb1 and its CDR.) Comparison of the number of bound cosolute molecules required in Figure 6a,b suggests that mAb1 interactions forming oligomers may be distinct from those that lead to dimer formation. The calculation of cosolute binding mole ratios enables further comparison between anion and cationic cosolute effects on mAb1 association. NaCl and NaSCN share a common cation, and can yield an assessment of Hofmeister salt series location effects on mAb1. While Cl− anion is near neutral in the Hofmeister series, the thiocyanate anion is considered chaotropic and capable of weak interactions with hydrophobic



CONCLUSIONS Static light scattering experiments as a function of cosolute concentration can provide detailed insights into the interactions of cosolutes with the protein clusters that result from selfassociation. In the limit of infinitely dilute protein concentrations, an assessment of the cosolute−protein interactions can be made within the framework of preferential interaction/ hydration proposed by Timasheff.20 Very weak preferential interactions appear to be present for NaCl and NaSCN in a concentration independent manner. In contrast, arginine− hydrochloride was found to be preferentially excluded from the protein surface with increasing cosolute concentration. These weak nonspecific interactions are averaged over the whole solvent-exposed surface of mAb1, and as such fail to fully explain the details of the cosolute type and concentration dependence of mAb1 self-association at high protein concentrations. Analysis of MALS data with IHS models to obtain the cosolute concentration dependent association constants of mAb1 affords an analysis of cosolute−protein interactions across the entire range of relevant protein concentrations (1−225 mg/mL). Changes in mAb1 dimer (m-mer) and higher oligomer (n-mer) formation patterns with increasing cosolute concentration verified that the cosolute interactions are indeed weak, but also indicated that the cosolute interactions with mAb1 are better described as direct binding interactions with specific features or patches at the protein surface. When evaluated in this framework, arginine− hydrochloride, NaCl, and NaSCN cosolute−protein interactions were clearly differentiated with regard to binding mAb1 oligomers and dimer species. Arg-Cl in particular was found to bind to the features of the protein−protein interface of both 13036

DOI: 10.1021/acs.jpcb.5b07568 J. Phys. Chem. B 2015, 119, 13027−13038

Article

The Journal of Physical Chemistry B

(8) Auton, M.; Bolen, D. W.; Rosgen, J. Structural Thermodynamics Of Protein Preferential Solvation: Osmolyte Solvation Of Proteins, Aminoacids, And Peptides. Proteins: Struct., Funct., Genet. 2008, 73, 802−813. (9) Auton, M.; Rosgen, J.; Sinev, M.; Holthauzen, L. M. F.; Bolen, D. W. Osmolyte Effects On Protein Stability And Solubility: A Balancing Act Between Backbone And Side-Chains. Biophys. Chem. 2011, 159, 90−99. (10) Crowe, J. H.; Carpenter, J. F.; Crowe, L. M. The Role Of Vitrification In Anhydrobiosis. Annu. Rev. Physiol. 1998, 60, 73−103. (11) Arakawa, T.; Ejima, D.; Tsumoto, K.; Obeyama, N.; Tanaka, Y.; Kita, Y.; Timasheff, S. N. Suppression of Protein Interactions By Arginine: A Proposed Mechanism Of The Arginine Effects. Biophys. Chem. 2007, 127, 1−8. (12) Liu, J.; Shire, S. J. Method of Reducing Viscosity Of High Concentration Protein Formulations. US Patent 7,666,413, 2010. (13) Curtis, R. A.; Ulrich, J.; Montaser, A.; Prausnitz, J. M.; Blanch, H. W. Protein-Protein Interactions In Concentrated Electrolyte Solutions - Hofmeister-Series Effects. Biotechnol. Bioeng. 2002, 79, 367−380. (14) Piazza, R.; Pierno, M. Protein Interactions Near Crystallization: A Microscopic Approach To The Hofmeister Series. J. Phys.: Condens. Matter 2000, 12, A443−A449. (15) Shukla, D.; Trout, B. L. Preferential Interaction Coefficients of Proteins in Aqueous Arginine Solutions and Their Molecular Origins. J. Phys. Chem. B 2011, 115, 1243−1253. (16) Ito, L.; Shiraki, K.; Matsuura, T.; Okumura, M.; Hasegawa, K.; Baba, S.; Yamaguchi, H.; Kumasaka, T. High-resolution X-ray Analysis Reveals Binding Of Arginine To Aromatic Residues Of Lysozyme Surface: Implication Of Suppression Of Protein Aggregation By Arginine. Protein Eng., Des. Sel. 2011, 24, 269−274. (17) Tsumoto, K.; Ejima, D.; Kita, Y.; Arakawa, T. Review: Why is Arginine Effective in Suppressing Aggregation? Protein Pept. Lett. 2005, 12, 613−619. (18) Lilyestrom, W. G.; Shire, S. J.; Scherer, T. M. Influence of the Cosolute Environment On Igg Solution Structure Analyzed By SmallAngle X-Ray Scattering. J. Phys. Chem. B 2012, 116, 9611−9618. (19) Eisenberg, H.: Biological Macromolecules and Polyelectrolytes in Solution; Oxford University Press: London, 1976. (20) Arakawa, T.; Timasheff, S. N. Preferential Interactions of Proteins with Salts in Concentrated-Solutions. Biochemistry 1982, 21, 6545−6552. (21) Arakawa, T.; Timasheff, S. N. Mechanism of Protein Salting in and Salting out by Divalent-Cation Salts - Balance between Hydration and Salt Binding. Biochemistry 1984, 23, 5912−5923. (22) Elbaum, D.; Herskovits, T. T. Dissociation of Human Hemoglobin By The Ureas And Amides. Osmotic Pressure And Light Scattering Studies. Biochemistry 1974, 13, 1268−1278. (23) Harrington, J. P.; Herskovits, T. T. The Effects Of Salts On The Subunit Structure And Dissociation Of Lumbricus Terrestris Hemoglobin. Biochemistry 1975, 14, 4972−4976. (24) Herskovits, T. T. Recent Aspects Of The Subunit Organization And Dissociation Of Hemocyanins. Comparative Biochemistry and Physiology. B 1988, 91, 597−611. (25) Herskovits, T. T.; Cavanagh, S. M.; San George, R. C. Lightscattering Investigations Of The Subunit Dissociation Of Human Hemoglobin A. Effects of Various Neutral Salts. Biochemistry 1977, 16, 5795−5801. (26) Herskovits, T. T.; Jacobs, R.; Nag, K. The Effects Of Salts And Ureas On The Subunit Dissociation Of Concanavalin A. Biochim. Biophys. Acta, Protein Struct. Mol. Enzymol. 1983, 742, 142−154. (27) Herskovits, T. T.; San George, R. C.; Cavanagh, S. M. Light Scattering Studies Of The Quaternary Structure And Subunit Dissociation Of Proteins: The Use Of Hydrophobic Solutes And Salts As Probes. J. Colloid Interface Sci. 1978, 63, 226−234. (28) Curtis, R. A.; Prausnitz, J. M.; Blanch, H. W. Protein-Protein And Protein-Salt Interactions In Aqueous Protein Solutions Containing Concentrated Electrolytes. Biotechnol. Bioeng. 1998, 57, 11−21.

dimer and oligomer structures, to effectively and completely reduce mAb1 self-association at moderate cosolute concentrations. Observed patterns of cosolute binding to mAb1 oligomers and dimers are consistent with electrostatic attractive interactions, caused by anisotropic charge distribution on the Fabs at pH 6.0, but also indicate the presence of hydrophobic patches that result in or promote dimer formation at higher ionic strengths. Tertiary solution components thus regulate mAb1 self-association/clustering (and likely the solution behavior of many proteins) through complex interactions that include both general cosolute−protein preferential interactions as well as important specific localized solute binding effects at protein surfaces.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcb.5b07568. A summary table of dilute solution light scattering data (0−10 mg/mL) including the apparent molar masses obtained at infinite dilution/zero scattering angle and the osmotic second virial coefficients as a function of cosolute concentration. Concentration dependent light scattering from 3 different preparations of mAb1 in 30 mM histidine-HCl buffer, as well as log−log analysis of oligomer and dimer association constants as a function of cosolute concentration range (PDF)



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The author gratefully acknowledges Dr. Allen Minton for providing insights on IHS model evaluation, as well as helpful discussion on the analysis of cosolute concentration dependent protein association constants.



REFERENCES

(1) Hofmeister, F. Zur Lehre von der Wirkung der Salze, Zweite Mitteilung. Naunyn-Schmiedeberg's Arch. Pharmacol. 1888, 24, 247− 260. (2) Scatchard, G.; Batchelder, A. C.; Brown, A. Preparation and Properties of Serum and Plasma Proteins. VI. Osmotic Equilibria in Solutions of Serum Albumin and Sodium Chloride. J. Am. Chem. Soc. 1946, 68, 2320−2329. (3) Scatchard, G.; Batchelder, A. C.; Brown, A.; Zosa, M. Preparation and Properties of Serum and Plasma Proteins. VII. Osmotic Equilibria in Concentrated Solutions of Serum Albumin. J. Am. Chem. Soc. 1946, 68, 2610−2612. (4) Kirkwood, J. G.; Goldberg, R. J. Light Scattering Arising From Composition Fluctuations In Multi-Component Systems. J. Chem. Phys. 1950, 18, 54−57. (5) Stockmayer, W. H. Light Scattering In Multi-Component Systems. J. Chem. Phys. 1950, 18, 58−61. (6) Anderson, C. F.; Courtenay, E. S.; Record, M. T. Thermodynamic Expressions Relating Different Types Of Preferential Interaction Coefficients In Solutions Containing Two Solute Components. J. Phys. Chem. B 2002, 106, 418−433. (7) Pegram, L. M.; Record, M. T. Thermodynamic Origin Of Hofmeister Ion Effects. J. Phys. Chem. B 2008, 112, 9428−9436. 13037

DOI: 10.1021/acs.jpcb.5b07568 J. Phys. Chem. B 2015, 119, 13027−13038

Article

The Journal of Physical Chemistry B

Lumbricus Terrestris Hemoglobin By The Ureas. Biochemistry 1975, 14, 4964−4971. (51) Silvers, T. R.; Myers, J. K. Osmolyte Effects on the SelfAssociation of Concanavalin A: Testing Theoretical Models. Biochemistry 2013, 52, 9367−9374. (52) Arakawa, T.; Timasheff, S. N. Preferential interactions of proteins with salts in concentrated solutions. Biochemistry 1982, 21, 6545−6552. (53) Timasheff, S. N. Preferential Interactions of Water and Cosolvents with Proteins. In Protein Solvent Interactions; Gregory, R. B., Ed.; Marcel Dekker: New York, 1995; pp 445−482. (54) Shukla, D.; Zamolo, L.; Cavallotti, C.; Trout, B. L. Understanding the Role of Arginine as an Eluent in Affinity Chromatography via Molecular Computations. J. Phys. Chem. B 2011, 115, 2645−2654. (55) Yadav, S.; Laue, T. M.; Kalonia, D. S.; Singh, S. N.; Shire, S. J. The Influence of Charge Distribution on Self-Association and Viscosity Behavior of Monoclonal Antibody Solutions. Mol. Pharmaceutics 2012, 9, 791−802. (56) Broering, J. M.; Bommarius, A. S. Evaluation of Hofmeister Effects On The Kinetic Stability Of Proteins. J. Phys. Chem. B 2005, 109, 20612−20619. (57) Rubin, J.; Linden, L.; Coco, W. M.; Bommarius, A. S.; Behrens, S. H. Salt-Induced Aggregation Of A Monoclonal Human Immunoglobulin G1. J. Pharm. Sci. 2013, 102, 377−386. (58) Tadeo, X.; Pons, M.; Millet, O. Influence Of The Hofmeister Anions On Protein Stability As Studied By Thermal Denaturation And Chemical Shift Perturbation. Biochemistry 2007, 46, 917−923. (59) Tadeo, X.; Lopez-Mendez, B.; Castano, D.; Trigueros, T.; Millet, O. Protein stabilization and the Hofmeister Effect: The Role Of Hydrophobic Solvation. Biophys. J. 2009, 97, 2595−2603. (60) Yadav, S.; Sreedhara, A.; Kanai, S.; Liu, J.; Lien, S.; Lowman, H.; Kalonia, D. S.; Shire, S. J. Establishing a Link Between Amino Acid Sequences and Self-Associating and Viscoelastic Behavior of Two Closely Related Monoclonal Antibodies. Pharm. Res. 2011, 28, 1750− 1764. (61) Chaudhri, A.; Zarraga, I. E.; Yadav, S.; Patapoff, T. W.; Shire, S. J.; Voth, G. A. The Role of Amino Acid Sequence in the SelfAssociation of Therapeutic Monoclonal Antibodies: Insights from Coarse-Grained Modeling. J. Phys. Chem. B 2013, 117, 1269−1279.

(29) Valente, J. J.; Verma, K. S.; Manning, M. C.; Wilson, W. W.; Henry, C. S. Second Virial Coefficient Studies Of Cosolvent-Induced Protein Self-Interaction. Biophys. J. 2005, 89, 4211−4218. (30) Dumetz, A. C.; Chockla, A. M.; Kaler, E. W.; Lenhoff, A. M. Comparative Effects of Salt, Organic, and Polymer Precipitants on Protein Phase Behavior and Implications for Vapor Diffusion. Cryst. Growth Des. 2009, 9, 682−691. (31) Bettelheim, F. A.; Siew, E. L. Effect of Change In Concentration Upon Lens Turbidity As Predicted By The Random Fluctuation Theory. Biophys. J. 1983, 41, 29−33. (32) Minton, A. P. Static Light Scattering from Concentrated Protein Solutions, I: General Theory for Protein Mixtures and Application to Self-Associating Proteins. Biophys. J. 2007, 93, 1321−1328. (33) Scherer, T. M.; Liu, J.; Shire, S. J.; Minton, A. P. Intermolecular Interactions Of Igg1Monoclonal Antibodies At High Concentrations Characterized By Light Scattering. J. Phys. Chem. B 2010, 114, 12948− 12957. (34) Lilyestrom, W. G.; Yadav, S.; Shire, S. J.; Scherer, T. M. Monoclonal antibody Self-Association, Cluster Formation, And Rheology At High Concentrations. J. Phys. Chem. B 2013, 117, 6373−6384. (35) Yadav, S.; Liu, J.; Shire, S. J.; Kalonia, D. S. Specific Interactions In High Concentration Antibody Solutions Resulting In High Viscosity. J. Pharm. Sci. 2010, 99, 1152. (36) Yadav, S.; Liu, J.; Shire, S. J.; Kalonia, D. S. Specific Interactions In High Concentration Antibody Solutions Resulting In High Viscosity. J. Pharm. Sci. 2010, 99, 1152−1168. (37) Yadav, S.; Scherer, T. M.; Shire, S. J.; Kalonia, D. S. Use Of Dynamic Light Scattering To Determine Second Virial Coefficient In A Semidilute Concentration Regime. Anal. Biochem. 2011, 411, 292− 296. (38) Yadav, S.; Shire, S. J.; Kalonia, D. S. Factors Affecting the Viscosity in High Concentration Solutions of Different Monoclonal Antibodies. J. Pharm. Sci. 2010, 99, 4812−4829. (39) Yearley, E. J.; Zarraga, I. E.; Shire, S. J.; Scherer, T. M.; Gokarn, Y.; Wagner, N. J.; Liu, Y. Small-angle Neutron Scattering Characterization Of Monoclonal Antibody Conformations And Interactions At High Concentrations. Biophys. J. 2013, 105, 720−731. (40) Kanai, S.; Liu, J.; Patapoff, T. W.; Shire, S. J. Shire. Reversible Self-Association Of A Concentrated Monoclonal Antibody Solution Mediated By Fab-Fab Interaction That Impacts Solution Viscosity. J. Pharm. Sci. 2008, 97, 4219−4227. (41) Scherer, T. M. Cosolute Effects On The Chemical Potential And Interactions Of An Igg1Monoclonal Antibody At High Concentrations. J. Phys. Chem. B 2013, 117, 2254−2266. (42) Wyatt, P. J. Light Scattering And The Absolute Characterization Of Macromolecules. Anal. Chim. Acta 1993, 272, 1−40. (43) Tanford, C. The Physical Chemistry of Macromolecules; John Wiley and Sons, Inc.: New York, 1961. (44) Zimm, B. H. Molecular Theory of the Scattering of Light in Fluids. J. Chem. Phys. 1945, 13, 141−145. (45) Casassa, E. F.; Eisenberg, H. Thermodynamic Analysis of Multicomponent Solutions. In Advances in Protein Chemistry; Anfinsen, C. F., Anson, M. L., Edsall, J. T., Frederic, M. R., Eds.; Academic Press: New York, 1964; Vol. 19, pp 287−395. (46) Timasheff, S. N.; Inoue, H. Preferential binding of solvent components to proteins in mixed water–organic solvent systems. Biochemistry 1968, 7, 2501−2513. (47) Vrij, A.; Overbeek, J. T. G. Scattering of light by charged colloidal particles in salt solutions. J. Colloid Sci. 1962, 17, 570−588. (48) Timasheff, S. N.; Dintzis, H. M.; Kirkwood, J. G.; Coleman, B. D. Light Scattering Investigation of Charge Fluctuations in Isoionic Serum Albumin Solutions1. J. Am. Chem. Soc. 1957, 79, 782−791. (49) Minton, A. P. The Effective Hard Particle Model Provides A Simple, Robust, And Broadly Applicable Description Of Nonideal Behavior In Concentrated Solutions Of Bovine Serum Albumin And Other Nonassociating Proteins. J. Pharm. Sci. 2007, 96, 3466−3469. (50) Herskovits, T. T.; Harrington, J. P. Solution Studies On Heme Proteins: Subunit Structure, Dissociation, And Unfolding Of 13038

DOI: 10.1021/acs.jpcb.5b07568 J. Phys. Chem. B 2015, 119, 13027−13038