Considerations for the Use of Polysorbates in Biopharmaceuticals.

PURPOSE: Polysorbates are commonly added to protein formulations and serve an important function as stabilizers. This paper reviews recent literature detailing some of the issues seen with the use of polysorbate 80 and polysorbate 20 in protein formulations. Based on this knowledge, a development strategy is proposed that leads to a control strategy for polysorbates in protein formulations. METHODS: A consortium of Biopharmaceutical scientists working in the area of protein formulations, shared experiences with polysorbates as stabilizers in their formulations. RESULTS: Based on the authors experiences and recent published literature, a recommendation is put forth for a development strategy which will lead into the appropriate control strategy for these excipients. CONCLUSIONS: An appropriate control strategy may comprise one or more elements of raw material, in-process and manufacturing controls. Additionally, understanding the role, if any, polysorbates play during stability will require knowledge of the criticality of the excipient, based upon its impact on CQAs due to variations in concentration and degradation level.

Free fatty acid particles in protein formulations, part 1: microspectroscopic identification.

We report, for the first time, the identification of fatty acid particles in formulations containing the surfactant polysorbate 20. These fatty acid particles were observed in multiple mAb formulations during their expected shelf life under recommended storage conditions. The fatty acid particles were granular or sand-like in morphology and were several microns in size. They could be identified by distinct IR bands, with additional confirmation from energy-dispersive X-ray spectroscopy analysis. The particles were readily distinguishable from protein particles by these methods. In addition, particles containing a mixture of protein and fatty acids were also identified, suggesting that the particulation pathways for the two particle types may not be distinct. The techniques and observations described will be useful for the correct identification of proteinaceous versus nonproteinaceous particles in pharmaceutical products.

Free fatty acid particles in protein formulations, part 2: contribution of polysorbate raw material.

Polysorbate 20 (PS20) is a nonionic surfactant frequently used to stabilize protein biopharmaceuticals. During the development of mAb formulations containing PS20, small clouds of particles were observed in solutions stored in vials. The degree of particle formation was dependent on PS20 concentration. The particles were characterized by reversed-phase HPLC after dissolution and labeling with the fluorescent dye 1-pyrenyldiazomethane. The analysis showed that the particles consisted of free fatty acids (FFAs), with the distribution of types consistent with those found in the PS20 raw material. Protein solutions formulated with polysorbate 80, a chemically similar nonionic surfactant, showed a substantial delay in particle formation over time compared with PS20. Multiple lots of polysorbates were evaluated for FFA levels, each exhibiting differences based on polysorbate type and lot. Polysorbates purchased in more recent years show a greater distribution and quantity of FFA and also a greater propensity to form particles. This work shows that the quality control of polysorbate raw materials could play an important role in biopharmaceutical product quality.

Optimized UV detection of high-concentration antibody formulations using high-throughput SE-HPLC.

High-concentration antibody solutions (>100 mg/mL) present significant challenges for formulation and process development, including formulation attributes such as increased solution viscosity, and the propensity for self-association. An additional challenge comes from the adaptation of analytical methods designed for low-concentration formulations to the high-concentration regime. The oligomeric state is a good example: it is a quality attribute monitored during pharmaceutical development and is one that can be affected by dilution; a typical first step in the analysis of high-concentration solutions. The objective of this work was to develop a size-exclusion HPLC (SE-HPLC) method that would allow the injection of high-concentration antibody formulations without the need for dilution prior to injection and their analysis in a high-throughput manner that does not create a bottleneck for the execution of complex formulation development studies. It was found that changing the UV detection wavelength from 215 to 235 nm simplified sample preparation by allowing for an approximately fivefold increase in injection load while maintaining the signal within the linear range of detection. In addition, the chromatographic peak properties (i.e., peak symmetry, resolution, and sensitivity) were determined to be consistent when compared with analytical methods developed for formulations with lower antibody concentrations.

Antibody solubility behavior in monovalent salt solutions reveals specific anion effects at low ionic strength.

Protein solubility was measured using the crystalline precipitate of a recombinant therapeutic antibody, in monovalent salt solutions containing KF, KCl, and KSCN (up to ∼ 0.7 M) at different pH conditions. For all three anions, the antibody solubility demonstrated complex behavior, both monotonic and nonmonotonic, with dependence on pH and salt concentration. At pH 7.1, close to the isoelectric point (pI) of 7.2, a typical salting-in behavior was observed with the salting-in constants of 12.7, 8.0, and 2.8 M for KSCN, KCl, and KF, respectively, suggesting that the anions follow the order of SCN(-) > Cl(-) > F(-) for increasing antibody solubility. Nonmonotonic behavior, as described by an initial solubility decrease followed by a solubility increase with ionic strength, was observed at pH 5.3, far below its pI. The effectiveness of the anion for reducing the solubility followed the order of SCN(-) > Cl(-) > F(-) . After the solubility reached the minimum, the anion's effectiveness for raising the antibody solubility was in agreement with that at pH 7.1. The mechanisms for the above phenomena are discussed based upon specific binding of the anions to the antibody surface. The mechanistic view of anion binding and its charge neutralization effect at pH 5.3 was supported by the results from the effective charge and zeta-potential measurements.

Effective charge measurements reveal selective and preferential accumulation of anions, but not cations, at the protein surface in dilute salt solutions.

Specific-ion effects are ubiquitous in nature; however, their underlying mechanisms remain elusive. Although Hofmeister-ion effects on proteins are observed at higher (>0.3 M) salt concentrations, in dilute (<0.1 M) salt solutions nonspecific electrostatic screening is considered to be dominant. Here, using effective charge (Q*) measurements of hen-egg white lysozyme (HEWL) as a direct and differential measure of ion-association, we experimentally show that anions selectively and preferentially accumulate at the protein surface even at low (<100 mM) salt concentrations. At a given ion normality (50 mN), the HEWL Q* was dependent on anion, but not cation (Li(+), Na(+), K(+), Rb(+), Cs(+), GdnH(+), and Ca(2+)), identity. The Q* decreased in the order F(-) > Cl(-) > Br(-) > NO(3)(-) ∼ I(-) > SCN(-) > ClO(4)(-) ≫ SO(4)(2-), demonstrating progressively greater binding of the monovalent anions to HEWL and also show that the SO(4)(2-) anion, despite being strongly hydrated, interacts directly with the HEWL surface. Under our experimental conditions, we observe a remarkable asymmetry between anions and cations in their interactions with the HEWL surface.

Diffusion and sedimentation interaction parameters for measuring the second virial coefficient and their utility as predictors of protein aggregation.

The concentration-dependence of the diffusion and sedimentation coefficients (k(D) and k(s), respectively) of a protein can be used to determine the second virial coefficient (B2), a parameter valuable in predicting protein-protein interactions. Accurate measurement of B2 under physiologically and pharmaceutically relevant conditions, however, requires independent measurement of k(D) and k(s) via orthogonal techniques. We demonstrate this by utilizing sedimentation velocity (SV) and dynamic light scattering (DLS) to analyze solutions of hen-egg white lysozyme (HEWL) and a monoclonal antibody (mAb1) in different salt solutions. The accuracy of the SV-DLS method was established by comparing measured and literature B2 values for HEWL. In contrast to the assumptions necessary for determining k(D) and k(s) via SV alone, k(D) and ks were of comparable magnitudes, and solution conditions were noted for both HEWL and mAb1 under which 1), k(D) and k(s) assumed opposite signs; and 2), k(D) ≥k(s). Further, we demonstrate the utility of k(D) and k(s) as qualitative predictors of protein aggregation through agitation and accelerated stability studies. Aggregation of mAb1 correlated well with B2, k(D), and k(s), thus establishing the potential for k(D) to serve as a high-throughput predictor of protein aggregation.

Ion-specific modulation of protein interactions: anion-induced, reversible oligomerization of a fusion protein.

Ions can significantly modulate the solution interactions of proteins. We aim to demonstrate that the salt-dependent reversible heptamerization of a fusion protein called peptibody A or PbA is governed by anion-specific interactions with key arginyl and lysyl residues on its peptide arms. Peptibody A, an E. coli expressed, basic (pI = 8.8), homodimer (65.2 kDa), consisted of an IgG1-Fc with two, C-terminal peptide arms linked via penta-glycine linkers. Each peptide arm was composed of two, tandem, active sequences (SEYQGLPPQGWK) separated by a spacer (GSGSATGGSGGGASSGSGSATG). PbA was monomeric in 10 mM acetate, pH 5.0 but exhibited reversible self-association upon salt addition. The sedimentation coefficient (s(w)) and hydrodynamic diameter (D(H)) versus PbA concentration isotherms in the presence of 140 mM NaCl (A5N) displayed sharp increases in s(w) and D(H), reaching plateau values of 9 s and 16 nm by 10 mg/mL PbA. The D(H) and sedimentation equilibrium data in the plateau region (>12 mg/mL) indicated the oligomeric ensemble to be monodisperse (PdI = 0.05) with a z-average molecular weight (M(z)) of 433 kDa (stoichiometry = 7). There was no evidence of reversible self-association for an IgG1-Fc molecule in A5N by itself or in a mixture containing fluorescently labeled IgG1-Fc and PbA, indicative of PbA self-assembly being mediated through its peptide arms. Self-association increased with pH, NaCl concentration, and anion size (I(-) > Br(-) > Cl(-) > F(-)) but could be inhibited using soluble Trp-, Phe-, and Leu-amide salts (Trp > Phe > Leu). We propose that in the presence of salt (i) anion binding renders PbA self-association competent by neutralizing the peptidyl arginyl and lysyl amines, (ii) self-association occurs via aromatic and hydrophobic interactions between the ..xxCTRWPWMC..xxxCTRWPWMCxx.. motifs, and (iii) at >10 mg/mL, PbA predominantly exists as heptameric clusters.

Effect of ions on agitation- and temperature-induced aggregation reactions of antibodies.

PURPOSE: The impact of ions on protein aggregation remains poorly understood. We explored the role of ionic strength and ion identity on the temperature- and agitation-induced aggregation of antibodies. METHODS: Stability studies were used to determine the influence of monovalent Hofmeister anions and cations on aggregation propensity of three IgG(2) mAbs. The C(H)2 domain melting temperature (T (m1)) and reduced valence (z*) of the mAbs were measured. RESULTS: Agitation led to increased solution turbidity, consistent with the formation of insoluble aggregates, while soluble aggregates were formed during high temperature storage. The degree of aggregation increased with anion size (F(-) < Cl(-) < Br(-) < I(-) < SCN(-) ~ ClO(4) (-)) and correlated with a decrease in T (m1) and z*. The aggregation propensity induced by the anions increased with the chaotropic nature of anion. The cation identity (Li(+), Na(+), K(+), Rb(+), or Cs(+)) had no effect on T (m1), z* or aggregation upon agitation. CONCLUSIONS: The results indicate that anion binding mediates aggregation by lowering mAb conformational stability and reduced valence. Our observations support an agitation-induced particulation model in which anions enhance the partitioning and unfolding of mAbs at the air/water interface. Aggregation predominantly occurs at this interface; refreshing of the surface during agitation releases the insoluble aggregates into bulk solution.

Anion binding mediated precipitation of a peptibody.

PURPOSE: Understand the underlying mechanism governing the salt-induced precipitation of a basic (pI = 8.8) protein, Peptibody A (PbA), in acidic solutions. METHODS: The rate, extent, and reversibility of PbA precipitation was monitored over 4-weeks as a function of pH (3.7-5.0), salt concentration (0-400 mM), and ion identity using a series of monovalent, Hofmeister anions (F(-), Cl(-), Br(-), I(-), ClO(4) (-), SCN(-)) and cations (Li+, Na+, K+, Rb+, Cs+). The effects of salt on conformational stability and reduced valence were determined using Fourier-transform infrared spectroscopy, circular dichroism, and capillary electrophoresis/analytical ultracentrifugation. RESULTS: PbA precipitation occurred upon salt addition and could be modulated with solution pH, salt identity & concentration. The precipitation was sensitive to anions, but not cations, and increased with anion size. A reverse Hofmeister effect (SCN(-) approximately ClO(4) (-)>I(-)>Cl(-)>Br(-)>F(-)) was observed with "salting-in" anions being the more effective precipitants. An increase in the precipitation rate below pH 4.3 indicated that protonation of aspartyl and glutamyl side-chains was also important for precipitation. The reversibility of precipitation was excellent (100%) at 4 degrees C but decreased upon storage at 25 degrees C and 37 degrees C; the loss in reversibility correlated with an increase in intermolecular beta-sheet content of the precipitate. CONCLUSION: Salts, employed as buffering, tonicifying, and viscosity modifying agents, may adversely affect the solubility of basic proteins formulated under acidic conditions.

Structural insights for designed alanine-rich helices: comparing NMR helicity measures and conformational ensembles from molecular dynamics simulation.

The temperature dependence of helical propensities for the peptides Ac-ZGG-(KAAAA)(3)X-NH(2) (Z = Y or G, X = A, K, and D-Arg) were studied both experimentally and by MD simulations. Good agreement is observed in both the absolute helical propensities as well as relative helical content along the sequence; the global minimum on the calculated free energy landscape corresponds to a single alpha-helical conformation running from K4 to A18 with some terminal fraying, particularly at the C-terminus. Energy component analysis shows that the single helix state has favorable intramolecular electrostatic energy due to hydrogen bonds, and that less-favorable two-helix globular states have favorable solvation energy. The central lysine residues do not appear to increase helicity; however, both experimental and simulation studies show increasing helicity in the series X = Ala --> Lys --> D-Arg. This C-capping preference was also experimentally confirmed in Ac-(KAAAA)(3)X-GY-NH(2) and (KAAAA)(3)X-GY-NH(2) sequences. The roles of the C-capping groups, and of lysines throughout the sequence, in the MD-derived ensembles are analyzed in detail.

Self-buffering antibody formulations.

Monoclonal antibodies (mAbs) often require the development of high-concentration formulations. In such cases, and when it is desirable to formulate a mAb around pH 5.0, we explored a novel approach of controlling the formulation pH by harnessing the ability of mAbs to "self-buffer." Buffer capacities of four representative IgG(2) molecules (designated mAb1 through mAb4) were measured in the pH 4-6 range. The buffer capacity results indicated that the mAbs possessed a significant amount of buffer capacity, which increased linearly with concentration. By 60-80 mg/mL, the mAb buffer capacities surpassed that of 10 mM acetate, which is commonly employed in formulations for buffering in the pH 4-6 range. Accelerated high temperature stability studies (50 degrees C over 3 weeks) conducted with a representative antibody in a self-buffered formulation (50 mg/mL mAb1 in 5.25% sorbitol, pH 5.0) and with solutions formulated using conventional buffers (50 mg/mL mAb1 in 5.25% sorbitol, 25 or 50 mM acetate, glutamate or succinate, also at pH 5.0) indicated that mAb1 was most resistant to the formation of soluble aggregates in the self-buffered formulation. Increased soluble aggregate levels were observed in all the conventionally buffered (acetate, glutamate, and succinate) formulations, which further increased with increasing buffer strength. The long-term stability of the self-buffered liquid mAb1 formulation (60 mg/mL in 5% sorbitol, 0.01% polysorbate 20, pH 5.2) was comparable to the conventionally buffered (60 mg/mL in 10 mM acetate or glutamate, 5.25% sorbitol, 0.01% polysorbate 20, pH 5.2) formulations. No significant change in pH was observed after 12 months of storage at 37 and 4 degrees C for the self-buffered formulation. The 60 mg/mL self-buffered formulation of mAb1 was also observed to be stable to freeze-thaw cycling (five cycles, -20 degrees C --> room temperature). Self-buffered formulations may be a better alternative for the development of high-concentration antibody and protein dosage forms.

Minimization and optimization of designed beta-hairpin folds.

Minimized beta hairpins have provided additional data on the geometric preferences of Trp interactions in TW-loop-WT motifs. This motif imparts significant fold stability to peptides as short as 8 residues. High-resolution NMR structures of a 16- (KKWTWNPATGKWTWQE, DeltaG(U)(298) >or= +7 kJ/mol) and 12-residue (KTWNPATGKWTE, DeltaG(U)(298) = +5.05 kJ/mol) hairpin reveal a common turn geometry and edge-to-face (EtF) packing motif and a cation-pi interaction between Lys(1) and the Trp residue nearest the C-terminus. The magnitude of a CD exciton couplet (due to the two Trp residues) and the chemical shifts of a Trp Hepsilon3 site (shifted upfield by 2.4 ppm due to the EtF stacking geometry) provided near-identical measures of folding. CD melts of representative peptides with the -TW-loop-WT- motif provided the thermodynamic parameters for folding, which reflect enthalpically driven folding at laboratory temperatures with a small DeltaC(p) for unfolding (+420 J K(-)(1)/mol). In the case of Asx-Pro-Xaa-Thr-Gly-Xaa loops, mutations established that the two most important residues in this class of direction-reversing loops are Asx and Gly: mutation to alanine is destabilizing by about 6 and 2 kJ/mol, respectively. All indicators of structuring are retained in a minimized 8-residue construct (Ac-WNPATGKW-NH(2)) with the fold stability reduced to DeltaG(U)(278) = -0.7 kJ/mol. NMR and CD comparisons indicate that -TWXNGKWT- (X = S, I) sequences also form the same hairpin-stabilizing W/W interaction.

Chemical shifts provide fold populations and register of beta hairpins and beta sheets.

A detailed analysis of peptide backbone amide (H(N)) and H alpha chemical shifts reveals a consistent pattern for beta hairpins and three-stranded beta sheets. The H alpha's at non-hydrogen-bonded strand positions are inwardly directed and shifted downfield approximately 1 ppm due largely to an anisotropy contribution from the cross-strand amide function. The secondary structure associated H alpha shift deviations for the H-bonded strand positions are also positive but much smaller (0.1-0.3 ppm) and the turn residues display negative H alpha chemical shift deviations (CSDs). The pattern of (H(N)) shift deviations is an even better indicator of both hairpin formation and register, with the cross-strand H-bonded sites shifted downfield (also by approximately 1 ppm) and with diagnostic values for the first turn residue and the first strand position following the turn. These empirical observations, initially made for [2:2]/[2:4]-type-I' and -II' hairpins, are rationalized and can be extended to the analysis of other turns, hairpin classes ([3:5], [4:4]/[4:6]), and three-stranded peptide beta-sheet models. The H alpha's at non-hydrogen-bonded sites and (H(N))'s in the intervening H-bonded sites provide the largest and most dependable measures of hairpin structuring and can be used for melting studies; however the intrinsic temperature dependence of (H(N)) shifts deviations needs to reflect the extent of solvent sequestration in the folded state. Several observations made in the course of this study provide insights into beta-sheet folding mechanisms: (1) The magnitude of the (H(N)) shifts suggests that the cross-strand H-bonds in peptide hairpins are as short as those in protein beta sheets. (2) Even L-Pro-Gly turns, which are frequently used in unfolded controls for hairpin peptides, can support hairpin populations in aqueous fluoroalcohol media. (3) The good correlation between hairpin population estimates from cross-strand H-bonded (H(N)) shift deviations, H alpha shift deviations, and structuring shifts at the turn locus implies that hairpin folding transitions approximate two-state behavior.

Hairpin folding rates reflect mutations within and remote from the turn region.

Hairpins play a central role in numerous protein folding and misfolding scenarios. Prior studies of hairpin folding, many conducted with analogs of a sequence from the B1 domain of protein G, suggest that faster folding can be achieved only by optimizing the turn propensity of the reversing loop. Based on studies using dynamic NMR, the native GB1 sequence is a slow folding hairpin (k(F)(278)=1.5 x 10(4)/s). GB1 hairpin analogs spanning a wide range of thermodynamic stabilities (DeltaG(U)(298)=-3.09 to+3.25 kJ/mol) were examined. Fold-stabilizing changes in the reversing loop can act either by accelerating folding or retarding unfolding; we present examples of both types. The introduction of an attractive side-chain/side-chain Coulombic interaction at the chain termini further stabilizes this hairpin. The 1.9-fold increase in folding rate constant observed for this change at the chain termini implies that this Coulombic interaction contributes before or at the transition state. This observation is difficult to rationalize by "zipper" folding pathways that require native turn formation as the sole nucleating event; it also suggests that Coulombic interactions should be considered in the design of systems intended to probe the protein folding speed limit.

Studies of helix fraying and solvation using 13C' isotopomers.

Both NMR and IR studies of carbonyl (13C') isotopomers of designed helices can provide residue-level details regarding the fractional occurrence and melting behavior of helical phi/psi angles along the sequence of helical peptides, details that cannot be obtained from CD or 1H-NMR studies. We have studied a classic series of helical models, Ac-YGG-(KAXAA)3K-NH2 (X=A,V), in both aqueous and helix-favoring media containing fluoroalcohol cosolvents, including a solvent system allowing the observation of cold denaturation. These studies confirmed the strong N-capping associated with this sequence and revealed more extensive C-terminal fraying than that calculated using current helicity prediction algorithms. In the X=A series, the central residues are somewhat resistant to thermal melting; it instead occurs predominantly at the frayable C terminus. For the X=V series under cold-denaturing conditions, the temperature of maximal helicity is not uniform along the sequence and both solvated and nonsolvated helical alanine sites (13C=O stretches at 1592 cm(-1) and 1615 cm(-1), respectively) are apparent. Correlation between the two spectroscopies employed yielded the intriguing observation that the valine side chain is able to desolvate the i - 4 amide in short monomeric helices. In addition, we report further measurements of the temperature dependence of alanine statistical coil chemical shifts, the temperature dependence of the 13C chemical shift of urea (employed as chemical shift reference), and a useful formula for converting 13C' shifts into fractional helicities.

Hairpin folding dynamics: the cold-denatured state is predisposed for rapid refolding.

Cold denaturation is a general phenomenon in globular proteins, and the associated cold-denatured states of proteins have important fundamental and practical significance. Here, we have characterized the cold-denatured state of a beta-hairpin forming peptide, MrH3a, in 8% hexafluoro-2-propanol (HFIP) and the dynamics of its refolding following a laser-induced T-jump. Beta-hairpins constitute an important class of protein structural elements, yet their folding mechanisms are not fully understood. Characterization of MrH3a using NMR, CD, and IR spectroscopies reveals residual structure in the cold-denatured state, in contrast with the highly disordered heat-denatured state. The residual structure in the cold-denatured state comprises relatively compact and solvent protected conformations. Furthermore, we find a substantial acceleration in the rate of folding from the cold-denatured state compared to that of the heat-denatured state. In addition, the cold-denatured state is not populated in 20% HFIP; folding occurs only from the fully unfolded state and is significantly slower. We interpret the acceleration of the folding rate of MrH3a in 8% HFIP as a direct consequence of the collapsed conformations of the cold-denatured state. Finally, there may be some reduction of the loop search cost when starting from the cold-denatured state, since this state may have some of the stabilizing cross-strand interactions already formed.

The mechanism of beta-hairpin formation.

Beta-hairpins constitute an important class of connecting protein secondary structures. Several groups have postulated that such structures form early in the folding process and serve to nucleate the formation of extended beta-sheet structures. Despite the importance of beta-hairpins in protein folding, little is known about the mechanism of formation of these structures. While it is well established that there is a complex interplay between the stability of a beta-hairpin and loop conformational propensity, loop length, and the formation of stabilizing cross-strand interactions (H-bonds and hydrophobic interactions), the influence of these factors on the folding rate is poorly understood. Peptide models provide a simple framework for exploring the molecular details of the formation of beta-hairpin structures. We have explored the fundamental processes of folding in two linear peptides that form beta-hairpin structures, having a stabilizing hydrophobic cluster connected by loops of differing lengths. This approach allows us to evaluate existing models of the mechanism of beta-hairpin formation. We find a substantial acceleration of the folding rate when the connecting loop is made shorter (i.e., the hydrophobic cluster is moved closer to the turn). Analysis of the folding kinetics of these two peptides reveals that this acceleration is a direct consequence of the reduced entropic cost of the smaller loop search.

Enhanced hairpin stability through loop design: the case of the protein G B1 domain hairpin.

A mutational study of the peptide corresponding to the second hairpin of the protein G B1 domain (GB1p) provided a series of mutants with significantly increased fold stability. Mutations focused on improvement of the direction-reversing loop and the addition of favorable Coulombic interactions at the sequence termini. The loop optimization was based on a database search for residues that occur with the greatest probability in similar hairpin loops in proteins. This search suggested replacing the native DDATKT sequence with NPATGK, which resulted in a 4.5 kJ/mol stabilization of the hairpin fold. The introduction of positively charged lysines at the N-terminus provided an additional 2.4 kJ/mol of stabilization, affording a GB1p mutant that is 86 +/- 3% folded at 25 degrees C with a melting temperature of 60 +/- 2 degrees C. The trpzip version of this peptide, in which three of the hydrophobic core residues were mutated to tryptophan, yielded a sequence that melted at 85 degrees C. Throughout, fold populations and melting temperatures were derived from the mutation and temperature dependence of proton chemical shifts and were corroborated by circular dichroism (CD) melts. The study also suggests that the wild-type GB1p sequence is significantly less stable than reported in some other studies: only 30% folded in water at 25 degrees C.

Possible locally driven folding pathways of TC5b, a 20-residue protein.

A novel computational procedure for modeling possible locally driven folding pathways by stepwise elongations of the peptide chain was successfully applied to TC5b, a 20-residue miniprotein. Systematic exploration of the possible locally driven pathways showed that the Trp-cage structure of TC5b could be obtained by stepwise elongation starting from the noncentral local nucleation centers preexisting in the unfolded state of TC5b. The probable locally driven folding pathway starts with folding of alpha-helical fragment 4-9, followed by formation of the proper three-dimensional structure of fragment 4-12, and then 4-18. Accordingly, the Trp-cage-forming interactions emerge successively, first Trp(6)-Pro(12), then Trp(6)-Pro(18), and then Trp(6)-Tyr(3). The Trp-cage-like structures of TC5b found in this study by independent energy calculations are in excellent agreement with the NMR experimental data. The same procedure rationalizes the incomplete Trp-cage formation observed for two analogs of TC5b. Generally, the success of this novel approach is encouraging and provides some justification for the use of computational simulations of locally driven protein folding.