Skip to main content
NCBI home page
Protein Science : A Publication of the Protein Society logoLink to Protein Science : A Publication of the Protein Society
. 2015 Apr 11;24(5):803–811. doi: 10.1002/pro.2651

Nonspecific shielding of unfavorable electrostatic intramolecular interactions in the erythropoietin native-state increase conformational stability and limit non-native aggregation

Douglas D Banks 1,*
PMCID: PMC4420528  PMID: 25628168

Abstract

Previous equilibrium and kinetic folding studies of the glycoprotein erythropoietin indicate that sodium chloride increases the conformational stability of this therapeutically important cytokine, ostensibly by stabilizing the native-state [Banks DD, (2011) The Effect of Glycosylation on the Folding Kinetics of Erythropoietin. J Mol Biol 412:536–550]. The focus of the current report is to determine the underlying cause of the salt dependent increase in erythropoietin conformational stability and to understand if it has any impact on aggregation, an instability that remains a challenge to the biotech industry in maintaining the efficacy and shelf-life of protein therapeutics. Isothermal urea denaturation experiments conducted at numerous temperatures in the absence and presence of sodium chloride indicated that salt stabilizes erythropoietin primarily by increasing the difference in enthalpy between the native and unfolded sates. This result, and the finding that the salt induced increases in erythropoietin melting temperatures were independent of the identity of the salt cation and anion, indicates that salt likely increases the conformational stability of erythropoietin at neutral pH by nonspecific shielding of unfavorable electrostatic interaction(s) in the native-state. The addition of salt (even low concentrations of the strong chaotrope salt guanidinium hydrochloride) also exponentially decreased the initial rate of soluble erythropoietin non-native aggregation at 37 °C storage.

Keywords: non-native protein aggregation, conformational stability, colloidal stability, melting temperature, kinetics, glycosylation, electrostatics

Introduction

Aggregation remains a primary concern of the biotechnology industry in maintaining the shelf-life and biological activity of protein therapeutics. To meet this challenge, the intrinsic aggregation rates of protein biologics are minimized through selection and engineering processes while extrinsic factors such as the solution pH, ionic strength, and the addition of various cosolutes are optimized during formulation development to further limit protein self-association. Formulations that most effectively limit aggregation, necessarily address its underlying cause. For instance, if there are strong intermolecular attractive forces between monomers a successful formulation strategy may be to increase the protein colloidal stability. Examples of this approach include numerous monoclonal antibodies that are often formulated at low ionic strengths and at pH values far from their isoelectric point (pI) where they tend to be less conformationally stable, yet exhibit greater colloidal stability due to stronger electrostatic repulsion.1–3

Conversely, if the native-state ensemble is in dynamic equilibrium with a partially unfolded intermediate state(s) of lower colloidal stability (as is often the case for proteins stored at elevated temperatures or under other stressful conditions), aggregation may be more effectively limited by either increasing the colloidal stability of the intermediate or by reducing its concentration by shifting the equilibrium to favor the native-state, i.e. increasing conformational stability. A prime example of this second strategy is the Escherichia coli (E. coli) expressed nonglycosylated recombinant human granulocyte colony stimulating factor (rhGCSF). When this four-helical bundle cytokine is stored at physiological temperature and pH nearly the entire helical core becomes more solvent accessible then when stored at refrigerated temperatures and irreversible aggregation/fibrillation and total precipitation follow within days.4–6 Reducing the concentration of this structurally expanded state (and unfolded-state) by addition of the preferentially excluded cosolute sucrose increases rhGCSF conformational stability and significantly reduces the rate of rhGCSF non-native aggregation.4,6,7

The focus of the present study is another therapeutically important member of the hematopoietic growth factor family responsible for the production and maturation of red blood cells, recombinant human erythropoietin (rhEPO).8,9 Despite a lack of sequence homology, this heavily glycosylated cytokine shares a high degree of structural similarity with rhGCSF (their four-helical bundles overlay with a root mean square deviation of only 1.6 Å10), folds by the same three-state on-pathway mechanism,11,12 and yet is about 2 kcal mol−1 less conformationally stable near physiological conditions.7,12 The E. coli expressed nonglycosylated rhEPO is even more susceptible to aggregation/precipitation than rhGCSF at neutral pH, despite having a higher pI and greater net positive surface charge.13–15 For this reason, a previous report that compared the kinetic folding mechanisms and equilibrium stabilities of the Chinese hamster ovary (CHO) cell expressed fully glycosylated and E. coli derived nonglycosylated rhEPOs were performed at low ionic strength solvent conditions [10 mM sodium phosphate (NaPi), pH 6.9] to enhance colloidal stability.16 This study together with a previous folding study of the fully glycosylated rhEPO formulated in 20 mM NaPi (pH 6.9) and 140 mM sodium chloride (NaCl)12 revealed that the conformational stability of the glycosylated rhEPO is greater at higher ionic strength solvent conditions. The goals of the current investigation are two-fold, first to determine the cause of the salt dependent increase in glycosylated rhEPO conformational stability, and second to understand if it has any bearing on aggregation propensity. A brief discussion comparing the relative importance of colloidal versus conformational stability, in the context of glycosylation and ionic strength, in limiting aggregation for rhEPO and rhGCSF is also presented.

Results and Discussion

Previous equilibrium and kinetic folding studies revealed that the CHO cell derived fully glycosylated rhEPO folds by a sequential three-state on-pathway mechanism and is about 1.8 times more conformationally stable in the presence of 20 mM NaPi (pH 6.9), 140 mM NaCl than when formulated in 10 mM NaPi (pH 6.9) alone.12,16 For comparison, the urea concentration dependence of the natural log of the folding and unfolding rate constants in both formulations are overlaid in Figure 1. In the absence of denaturant, the folding phases to the on-pathway kinetic intermediate (UI) and from the intermediate to the native-state (IN) are roughly equivalent in both solvent conditions, while the two unfolding phases [from the native-state to the intermediate (NI) and from the intermediate to the unfolded-state (IU)] are significantly slower at higher ionic strength (Fig. 1). Since the folding and unfolding rate constants are proportional to the height of the energy barriers from the ground to transition states, the simplest interpretation of these data is that the increase in ionic strength stabilizes the kinetic intermediate and native ground states as depicted in Figure S1A, Supporting Information. Along these lines of interpretation, the greater denaturant concentration dependence of the NI unfolding phase, or kinetic m‡-value, for the higher ionic strength formulation suggests that salt induces a greater degree of burial of surface area in the native-state.17

Figure 1.

Figure 1

Open in a new tab

The natural log of the urea concentration dependence of the folding and unfolding rate constants for rhEPO formulated in 10 mM NaPi (pH 6.9) (blue) and in 20 mM NaPi (pH 6.9), 140 mM NaCl (red). Kinetic responses initiated from manual mixing and monitored by far-UV CD and intrinsic tryptophan fluorescence are shown as squares and circles, respectively; stopped-flow fluorescence data are shown as triangles. Solid lines represent the global fits to a three state on-pathway mechanism described in detail elsewhere.16 The urea concentration dependence of the microscopic rate constants derived from this fit are shown as dashed lines. For clarity a fast off-pathway unfolding phase and data from sequential mixing unfolding experiments with rate constants >1 are not shown. Data used to construct the chevron plot are from Banks et al.12,16

It is worth noting; however, that since absolute free energies cannot be determined, the reaction coordinate in Figure S1A, Supporting Information may be equivalently drawn by equating the energies of the native-state (Fig. S1B, Supporting Information), signifying that increased conformational stability arises from destabilization of the unfolded-state. Although this mechanism cannot be ruled out by the folding data alone, it would require that the unfolded, kinetic intermediate and transition-state species be destabilized by the same amount. This seems possible for the transition-state between the unfolded and intermediate species (TS1) based on its relatively low Tanford β-value (βT) of 0.2, but much less likely for the kinetic intermediate and the transition-state from the intermediate to the native-state (TS2) that have βT-values of 0.7 and 0.74, respectively.12,16 The βT-value is a qualitative measure derived from the kinetic m‡-values and range from 0 to 1, signifying the degree of unfolded and native-like compactness, respectively.17–19 In other words, it seems doubtful that the same favorable electrostatic interaction(s) in the unfolded-state (that are destabilized by the addition of salt) are also present in the intermediate and TS2 ensembles that have roughly 70–75% native-like solvent accessibility.

Mechanism of salt induced increase in native-state stability

Working from a model of salt induced increase in native-state stability, the following three mechanisms are proposed: (1) specific ion binding to the native-state, (2) salt nonspecifically shielding energetically unfavorable electrostatic intramolecular interaction(s) in the native-state, and (3) salt enhancing the hydrophobic interactions of the three N-linked glycans with the surface of the rhEPO native-state. This last mechanism is included since increased binding affinities of free N-glycans to nonglycosylated E. coli derived rhEPO with addition of molar concentrations of NaCl and ammonium sulfate have been observed.20 Note that the general Hofmeister preferential hydration effect was not considered as a possible mechanism of salt induced increase in conformational stability of the protein moiety of rhEPO (as opposed to the protein plus glycans) since NaCl is not a particularly strong kosmotrope and it seems to have a large stabilizing effect even at relatively low concentrations (<0.1M). Moreover, addition of strong kosmotropes would be expected to destabilize the species with the greater amount of exposed hydrophobic surface area, i.e. the unfolded-state, and result in faster folding rates, which was not observed (Fig. 1).

To test the validity of either of these mechanisms, the rhEPO stability was monitored as a function of salt concentration and by systematically changing both the cation and anion of the salt. To screen numerous salts at multiple concentrations, the rhEPO melting temperature (Tm) was used to assess conformational stability. Compared with chemical unfolding studies, thermal denaturation experiments monitored by spectroscopy are easy to perform and require less protein. A caveat to using such an approach; however, is that a change in melting temperature may not necessarily correspond to a similar change in conformational stability at the lower 25 °C temperature used for the previous equilibrium and kinetic folding studies.21 This possibility, and the past finding that the rhEPO free energy of unfolding (ΔG) derived from thermal and chemical denaturation studies do not agree,22 was explored by collecting isothermal equilibrium urea unfolding data as a function of temperature in both low and high ionic strength solvent conditions as detailed in the Supporting Information. On the basis of these results, salt was found to primarily increase the ΔG at all temperatures (i.e., increase the enthalpy of unfolding;23,24 Fig. 2 and Table1), signifying that changes in melting temperature (ΔTm) will correlate with changes in ΔG (ΔΔG) at lower temperatures and therefore be a meaningful measure of the salt induced change in conformational stability.

Figure 2.

Figure 2

Open in a new tab

Representative isothermal urea equilibrium unfolding data collected as a function of temperature for rhEPO formulated in 10 mM NaPi (pH 6.9) (A) and in 20 mM NaPi (pH 6.9), 400 mM NaCl (B). For clarity only the far UV-CD data collected at 222 nm are shown with the two-state global fits (solid lines) derived from fitting all wavelengths (224–220 nm); data and fits are shown normalized to Fapp. The ΔG and m-values derived from the 224–220 nm global fits are shown in (C) and the inset, respectively; error bars represent the standard deviation of the fitted parameters and the solid lines are the fits to the Gibbs-Helmholtz equation [Eq. (S2)] as described in the Supporting Information. Lines in the inset of (C) are drawn to guide the eye.

Table 1.

Thermodynamic Unfolding Parameters of rhEPOa

Parameter Low ionic strength High ionic strength
Hs (kcal mol−1) 5.76 (0.04) 8.23 (0.04)
ΔCp (kcal mol−1 K−1) 2.0 (0.1) 2.4 (0.1)
Ts (K) 283.6 (0.4) 283.3 (0.7)
Open in a new tab
a

Results are from isothermal urea induced equilibrium unfolding experiments conducted as a function of temperature and fit to Eq. (S2) of the Supporting Information; Error bars represent the standard deviations of the fitted parameters. Low ionic strength: 10 mM NaPi (pH 6.9). High ionic strength: 20 mM NaPi (pH 6.9), 400 mM NaCl. Ts is the temperature of maximum conformational stability, ΔHs is the change in enthalpy at this temperature (where ΔG = ΔH) and ΔCp is the change in heat capacity of unfolding.

The effect of ionic strength on the rhEPO Tm is shown in Figure 3. Focusing initially on NaCl, the Tm increases sharply up to an ionic strength of about 0.1M NaCl and then gradually becomes less dependent on the ionic strength at higher values. This trend is consistent with either saturating a specific weak ion binding site or the nonspecific screening of unfavorable electrostatic interaction(s), but not with a general Hofmeister effect where a linear dependence on ionic strength would be expected.25,26 Changing the cation to potassium had little effect, suggesting that sodium does not bind specifically to the rhEPO surface. Likewise, changing the anion to phosphate, or both cation and anion to potassium phosphate (KPi) resulted in nearly identical increases in Tm up to an ionic strength of ∼0.1M. Beyond this ionic strength the curvature of both phosphate salts was less than that of the chloride salts. This is likely because PO42− is a strong Hofmeister kosmotrope and beyond ionic strengths of roughly 0.1M begins to stabilize rhEPO by preferential exclusion. These findings are most consistent with nonspecific screening of unfavorable electrostatic interaction(s) on the surface of the protein moiety of the native-state, than for instance between the negatively charged sialic acids at the termini of the N-linked glycans since their removal does not alter conformational stability.13,27 Consistent with this mechanism, the surface of the rhEPO native state has a number of basic side-chains within 5–10 Å of each other that may be the regions responsible for destabilizing the native-state at neutral pH and low ionic strength solvent conditions (Fig. 4).

Figure 3.

Figure 3

Open in a new tab

Representative thermal denaturation experiments of rhEPO at increasing concentrations of NaCl (A); data are normalized to the Fapp and solid lines are the fits to the two-state thermal unfolding model. The Tm values derived from these fits as well from unfolding curves collected in the presence of increasing concentrations of KCl, NaPi, KPi, and GdnHCl (Fig. S3 of the Supporting Information) are shown in (B); error bars represent the standard deviation of the fitted parameters and lines are drawn between the data points to guide the eye. Conditions: 10 mM NaPi (pH 6.9).

Figure 4.

Figure 4

Open in a new tab

Electrostatic solvent accessible surface potential of the E. coli expressed nonglycosylated rhEPO; positive and negative charge at pH 7.0 are shown as blue and red, respectively. Potential areas of unfavorable electrostatic interactions are shown between R139, K140, and R143 (A); R150 and K152 (B); K52 and R53 (C); numbers are the distances (Å) between the side-chains.

If the native-state is destabilized by unfavorable electrostatic intramolecular interactions, then even strong chaotropic salts such as guanidinium hydrochloride (GdnHCl) would be expected to stabilize rhEPO at low concentrations. As shown in Figure 3, increasing the concentration of GdnHCl up to about 75 mM had the same effect on the rhEPO Tm as NaCl. At concentrations greater than 0.1M GdnHCl, the incremental increase in Tm with further addition of this strong denaturant becomes less than that of NaCl due the preferential interaction of this chaotrope with the rhEPO unfolded-state. Still, it is not until ∼0.7M GdnHCl that the stabilization of the unfolded-state begins to outweigh the contribution of this salts effect on the native-state stabilization. Similar observation have been made with other proteins where conformational stability assessed by urea and GdnHCl induced equilibrium unfolding did not yield similar results by virtue of the two denaturants having different effects on protein electrostatic interactions.28 This finding also rules out the possibility of salt increasing the hydrophobic interaction between the attached glycans and surface of the native-state, since GdnHCl is weakly hydrated and proposed to disrupt hydrophobic interactions.29,30

The effect of salt on rhEPO non-native aggregation

According to the extended Lumry-Eyring model of non-native aggregation, aggregation precedes from an aggregation competent partial unfolded intermediate species and can be prevented by shifting the equilibrium to favor the native state.31,32 To understand whether the salt induced increase in rhEPO conformational stability effect its rate of aggregation, rhEPO was stored for one month at 37 °C; a temperature considered stressful compared to the refrigerated temperatures (2–8 °C) that protein biologics are typically stored at, in the presence of increasing concentrations of NaCl and at a single GdnHCl concentration of 0.1M. At this storage temperature the rhEPO equilibrium m-value is only about half its value at refrigerated conditions [Fig. 2(C), inset], consistent with a more open and solvent accessible native-state.21 In agreement with the extended Lumry-Eyring model, the presence of increasing concentrations of NaCl progressively lowered the initial rate of soluble rhEPO aggregate formation, up until a concentration of ∼0.4M [Fig. 5(C)]. Even the presence of 0.1M GdnHCl limited aggregation to nearly the same extent as 0.1M NaCl, consistent with the nonspecific shielding of unfavorable electrostatic intramolecular interaction(s) model of increasing conformational stability [Fig. 5(B)].

Figure 5.

Figure 5

Open in a new tab

Representative chromatograms of rhEPO following 28 days storage at 37 °C in the presence of increasing NaCl concentrations and at a single GdnHCl concentration of 0.1M (A). The kinetics of soluble high molecular weight (HMW) aggregate formation for rhEPO stored in the same conditions is shown in (B). Error bars represent the standard deviations of triplicate independent experiments. The initial rate of rhEPO aggregation decreases exponentially (red symbols) with increasing concentrations of NaCl (C) and correlate well with the rhEPO melting temperature (inset); the initial rate of aggregation in the 0.1M GdnHCl formulation is plotted for comparison (green symbol). Conditions: 3 mg/mL rhEPO, 20 mM NaPi (pH 6.9).

Glycosylation and the effect of colloidal stability on non-native aggregation

Increasing the ionic strength likely only slows the aggregation rate of the fully glycosylated CHO derived rhEPO. The E. coli expressed and nonglycosylated rhEPO undergoes irreversible aggregation/precipitation within hours under physiological conditions similar to rhGCSF.13,14 This finding is perhaps not altogether surprising since both of these hematopoietin growth factors appear susceptible to the same temperature dependent opening of the native state, as evidenced by the strong temperature dependence of their equilibrium m-values.6 Unlike rhGCSF; however, nonglycosylated rhEPO carries a net positive surface charge at neutral pH (its measured pI is 9.2 as compared to 6.1 for rhGCSF7,33) and would be expected to be more susceptible to irreversible aggregation/precipitation as the solvent ionic strength is increased because of the shielding of repulsive positive surface charges between monomers.* In fact, only low ionic strength solution conditions (10 mM NaPi, pH 6.8) were found to be appropriate for initial NMR structural analysis of the E. coli derived rhEPO.14 For this reason, final structural determination of the nonglycosylated rhEPO ultimately had to be performed with a triple lysine mutant (replacing the three Asn N-linked glycosylation sites with Lys) that compared with the wild-type sequence had no appreciable impact on conformational stability, but did further increase the pI, colloidal stability, and suppress aggregation.10,14,16

In this sense, the increase in colloidal stability afforded by the presence of glycosylation has a much larger impact than salt in minimizing aggregation, rendering all aggregate fully soluble, as the total integrated area of the SEC chromatograms in Figure 5(A) remain constant at all time-points. This finding is also consistent with other proteins susceptible to non-native aggregation/fibrillation such as the amyloidogenic hen egg white lysozyme,34 chicken cystatin,35 and human prion protein fragment 175–19536 where glycosylation was found to improve solubility and suppress intermolecular cross-β-structure formation/fibrillation. Interestingly, attachment of a small polyethylene glycol (PEG) moiety to the N-terminus of rhGCSF was shown by Rajan et al. to have the similar effect of rendering all rhGCSF aggregate soluble and limiting its overall aggregation rate while having little effect on conformational stability.37 Perhaps the best strategy for minimizing aggregation for this important family of growth factors is to first enhance the colloidal stability of the expanded state populated at elevated temperatures by glycosylation/PEGylation and then to decrease its concentration by either destabilizing it with addition of preferentially excluded cosolutes, as in the case of rhGCSF, or by stabilizing the native-state, as shown for rhEPO by increasing the solvent ionic strength.

Materials and Methods

Materials

Pharmaceutical quality fully glycosylated rhEPO was expressed in CHO cells and purified as previously described.38,39 Protein concentrations were determined by absorbance at 280 nm using a theoretical extinction coefficient of 22,710 M−1 cm−1 determined by the methods of Pace et al.40 Ultra-pure urea and GdnHCl were from MP Biomedicals (Solon, OH) and were prepared fresh for each experiment; concentrations were determined by refractive index.41 High purity sodium and potassium phosphate (monobasic monohydrate and dibasic dihydrate) and all other reagents were purchased from Sigma-Aldrich (St. Louis, MO).

Thermal unfolding experiments

Thermal denaturation of 10 µM rhEPO samples were performed on a Jasco J-815 CD spectrophotometer (Jasco, Tokyo, Japan). Samples were heated from 20 to 90 °C in a 0.2 cm cuvette while monitoring the CD signal at 222 nm using a bandwidth of 2 nm. Because of a shift in the thermal unfolding curve at heating rates greater than 60 °C/hr. for rhEPO formulated at high ionic strength (Fig. S2, Supporting Information), a heating rate of 45 °C/h was used for all experiments. Under these conditions, thermal unfolding was greater than 95% reversible as assessed by allowing a thermally unfolded sample cool back to 20 °C and monitoring the far UV–CD spectrum [Fig. S2(C), Supporting Information), and repeating the thermal unfolding experiment. Thermal unfolding curves were fit to an equilibrium two-state thermal unfolding model given elsewhere42 to derive the Tm and the change in the enthalpy of unfolding at this temperature (ΔHm). Unfolding curves collected at different NaCl concentrations were normalized to the fraction of unfolded protein (Fapp) according to Eq. 1.

graphic file with name pro0024-0803-m1.jpg 1

Here Yi is the observed signal at a given temperature and YF and YU are the extrapolated values of the folded and unfolded signal at the same temperature.

Aggregation assay

Multiple NaCl formulations ranging in concentration from 0 to 0.6 M were prepared at a final rhEPO concentration of 3 mg/mL buffered by 20 mM NaPi (pH 6.9) by addition from a 1M NaCl, 20 mM NaPi (pH 6.9) stock solution. A single 0.1M GdnHCl formulation was prepared in a similar manner from a 1M GdnHCl, 20 mM NaPi (pH 6.9) stock solution. All formulations were prepared in triplicate and filter sterilized in a laminar flow hood after which 0.2 mL aliquots of each were dispensed into sterile 0.5 mL screw-cap microcentrifuge tubes. All formulations were then stored for 28 days at 37 °C. At given time-points during storage a complete set of all formulations were analyzed by size exclusion chromatography.

Size-exclusion chromatography

Size exclusion chromatography (SEC) was performed on an Agilent 1100 series quaternary pump liquid chromatography system (Agilent, Palo Alto, CA) equipped with a single TosoHaas TSK-gel SW2000xl column. Protein was eluted isocratically at a flow rate of 0.5 mL/min with mobile phase consisting of 100 mM sodium phosphate (pH 6.9) and 150 mM sodium chloride. Absorbance was monitored at 280, 230, and 215 nm.

Electrostatic surface potential

The electrostatic solvent accessible surface potential of rhEPO (excluding the N-linked glycans) was generated by first replacing the three lysine residues at positions 24, 38, and 83 of the triple lysine mutant used for solution structure determination (PDB 1BUY)10 with asparagine residues to restore the wild-type sequence. Continuum electrostatic calculations were performed with the pyMOL 1.6 (Schrödinger LLC) adaptive Poisson-Boltzmann software (APBS)43 plugin44 after first optimizing hydrogen bonding, assigning atomic charge and radius parameters using the PDB2PQR web service (http://agave.wustl.edu/pdb2pqr/)45 that utilizes PROPKA46,47 to determine the protonation state at a particular pH. The electrostatic surface potential was calculated at 37 °C, pH 7.0 with an ionic strength of 0.1M monovalent salt and rendered with pyMOL 1.6 (Schrödinger LLC).

Glossary

ΔG

the difference in free energy between the native and unfolded states

Fapp

fraction of unfolded protein

far-UV CD

far-ultra violet circular dichroism spectroscopy

GdnHCl

guanidinium hydrochloride

HPLC

high pressure liquid chromatography

KCl

potassium chloride

KPi

potassium phosphate

NaCl

sodium chloride

NaPi

sodium phosphate

rhEPO

recombinant human erythropoietin

rhGCSF

recombinant human granulocyte colony-stimulating factor

SEC

size exclusion chromatography

Tm

melting temperature

Footnotes

*

Note that this may not be the case for PNGase F treated rhEPO, since this enzyme deamidates N-linked asparagine residues to aspartic acid during the deglycosylation reaction, which would change the net charge of the protein. This may also lower the conformational stability of rhEPO since there are glutamic acid residues adjacent to two of these N-linked sites at positions 23 and 37.

Supporting Information

Additional Supporting Information may be found in the online version of this article.

Supporting Information

pro0024-0803-sd1.pdf (901.1KB, pdf)

References

  1. Wang Y, Latypov RF, Lomakin A, Meyer JA, Kerwin BA, Vunnum S, Benedek GB. Quantitative evaluation of colloidal stability of antibody solutions using PEG-induced liquid–liquid phase separation. Mol Pharm. 2014;11:1391–1402. doi: 10.1021/mp400521b. [DOI] [PubMed] [Google Scholar]
  2. Arzenšek D, Kuzman D, Podgornik R. Colloidal interactions between monoclonal antibodies in aqueous solutions. J Colloid Interface Sci. 2012;384:207–216. doi: 10.1016/j.jcis.2012.06.055. [DOI] [PubMed] [Google Scholar]
  3. Banks DD, Latypov RF, Ketchem RR, Woodard J, Scavezze JL, Siska CC, Razinkov VI. Native-state solubility and transfer free energy as predictive tools for selecting excipients to include in protein formulation development studies. J Pharm Sci. 2012;101:2720–2732. doi: 10.1002/jps.23219. [DOI] [PubMed] [Google Scholar]
  4. Krishnan S, Chi EY, Webb JN, Chang BS, Shan D, Goldenberg M, Manning MC, Randolph TW, Carpenter JF. Aggregation of granulocyte colony stimulating factor under physiological conditions: characterization and thermodynamic inhibition. Biochemistry. 2002;41:6422–6431. doi: 10.1021/bi012006m. [DOI] [PubMed] [Google Scholar]
  5. Raso SW, Abel J, Barnes JM, Maloney KM, Pipes G, Treuheit MJ, King J, Brems DN. Aggregation of granulocyte-colony stimulating factor in vitro involves a conformationally altered monomeric state. Protein Sci. 2005;14:2246–2257. doi: 10.1110/ps.051489405. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Banks DD, Zhang J, Siska CC. Relationship between native-state solubility and non-native aggregation of recombinant human granulocyte colony stimulating factor: practical implications for protein therapeutic development. Mol Pharm. 2014;11:3431–3442. doi: 10.1021/mp500165j. [DOI] [PubMed] [Google Scholar]
  7. Chi EY, Krishnan S, Kendrick BS, Chang BS, Carpenter JF, Randolph TW. Roles of conformational stability and colloidal stability in the aggregation of recombinant human granulocyte colony-stimulating factor. Protein Sci. 2003;12:903–913. doi: 10.1110/ps.0235703. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Graber SE, Krantz SB. Erythropoietin and the control of red cell production. Ann Rev Med. 1978;29:51–66. doi: 10.1146/annurev.me.29.020178.000411. [DOI] [PubMed] [Google Scholar]
  9. Goldwasser E, Kung CKH. Progress in the purification of erythropoietin. Ann NY Acad Sci. 1968;149:49–53. doi: 10.1111/j.1749-6632.1968.tb15135.x. [DOI] [PubMed] [Google Scholar]
  10. Cheetham JC, Smith DM, Aoki KH, Stevenson JL, Hoeffel TJ, Syed RS, Egrie J, Harvey TS. NMR structure of human erythropoietin and a comparison with its receptor bound conformation. Nature Struct Mol Biol. 1998;5:861–866. doi: 10.1038/2302. [DOI] [PubMed] [Google Scholar]
  11. Brems DN. The kinetics of G-CSF folding. Protein Sci. 2002;11:2504–2511. doi: 10.1110/ps.0206202. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Banks DD, Scavezze JL, Siska CC. Kinetic folding mechanism of erythropoietin. Biophys J. 2009;96:4221–4230. doi: 10.1016/j.bpj.2009.02.049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Narhi LO, Arakawa T, Aoki KH, Elmore R, Rohde MF, Boone T, Strickland TW. The effect of carbohydrate on the structure and stability of erythropoietin. J Biol Chem. 1991;266:23022–23026. [PubMed] [Google Scholar]
  14. Narhi LO, Arakawa T, Aoki K, Wen J, Elliott S, Boone T, Cheetham J. Asn to Lys mutations at three sites which are N-glycosylated in the mammalian protein decrease the aggregation of Escherichia coli-derived erythropoietin. Prot Eng. 2001;14:135–140. doi: 10.1093/protein/14.2.135. [DOI] [PubMed] [Google Scholar]
  15. Tsuda E, Kawanishi G, Ueda M, Masuda S, Sasaki R. The role of carbohydrate in recombinant human erythropoietin. Eur J Biochem. 1990;188:405–411. doi: 10.1111/j.1432-1033.1990.tb15417.x. [DOI] [PubMed] [Google Scholar]
  16. Banks DD. The effect of glycosylation on the folding kinetics of erythropoietin. J Mol Biol. 2011;412:536–550. doi: 10.1016/j.jmb.2011.07.061. [DOI] [PubMed] [Google Scholar]
  17. Bachmann A, Kiefhaber T. Kinetic mechanisms in protein folding. In: Buchner J, Kiefhaber T, editors. Protein folding handbook. Vol. 1. Weinheim: Wiley-VCH Verlag GmbH; 2005. pp. 379–410. [Google Scholar]
  18. Roder H, Colón W. Kinetic role of early intermediates in protein folding. Curr Opin Struct Biol. 1997;7:15–28. doi: 10.1016/s0959-440x(97)80004-8. [DOI] [PubMed] [Google Scholar]
  19. Capaldi AP, Shastry MCR, Kleanthous C, Roder H, Radford SE. Ultrarapid mixing experiments reveal that Im7 folds via an on-pathway intermediate. Nature Struct Mol Biol. 2001;8:68–72. doi: 10.1038/83074. [DOI] [PubMed] [Google Scholar]
  20. Toyoda T, Arakawa T, Yamaguchi H. N-glycans stabilize human erythropoietin through hydrophobic interactions with the hydrophobic protein surface: studies by surface plasmon resonance analysis. J Biochem. 2002;131:511–515. doi: 10.1093/oxfordjournals.jbchem.a003128. [DOI] [PubMed] [Google Scholar]
  21. Myers JK, Nick Pace C, Martin Scholtz J. Denaturant m values and heat capacity changes: relation to changes in accessible surface areas of protein unfolding. Protein Sci. 1995;4:2138–2148. doi: 10.1002/pro.5560041020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Lah J, Prislan I, Kržan B, Salobir M, Francky A, Vesnaver G. Erythropoietin unfolding: thermodynamics and its correlation with structural features. Biochemistry. 2005;44:13883–13892. doi: 10.1021/bi0512952. [DOI] [PubMed] [Google Scholar]
  23. Deutschman WA, Dahlquist FW. Thermodynamic basis for the increased thermostability of CheY from the hyperthermophile Thermotoga maritima. Biochemisry. 2001;40:13107–13113. doi: 10.1021/bi010665t. [DOI] [PubMed] [Google Scholar]
  24. Razvi A, Scholtz JM. Lessons in stability from thermophilic proteins. Protein Sci. 2006;15:1569–1578. doi: 10.1110/ps.062130306. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Baldwin RL. How Hofmeister ion interactions affect protein stability. Biophys J. 1996;71:2056–2063. doi: 10.1016/S0006-3495(96)79404-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Batchelor JD, Olteanu A, Tripathy A, Pielak GJ. Impact of protein denaturants and stabilizers on water structure. J Am Chem Soc. 2004;126:1958–1961. doi: 10.1021/ja039335h. [DOI] [PubMed] [Google Scholar]
  27. Toyoda T, Itai T, Arakawa T, Aoki KH, Yamaguchi H. Stabilization of human recombinant erythropoietin through interactions with the highly branched N-glycans. J Biochem. 2000;128:731–737. doi: 10.1093/oxfordjournals.jbchem.a022809. [DOI] [PubMed] [Google Scholar]
  28. Monera OD, Kay CM, Hodges RS. Protein denaturation with guanidine hydrochloride or urea provides a different estimate of stability depending on the contributions of electrostatic interactions. Protein Sci. 1994;3:1984–1991. doi: 10.1002/pro.5560031110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Huerta-Viga A, Woutersen S. Protein denaturation with guanidinium: a 2D-IR study. J Phys Chem Lett. 2013;4:3397–3401. doi: 10.1021/jz401754b. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Mason PE, Neilson GW, Dempsey CE, Barnes AC, Cruickshank JM. The hydration structure of guanidinium and thiocyanate ions: Implications for protein stability in aqueous solution. Proc Natl Acad Sci U S A. 2003;100:4557–4561. doi: 10.1073/pnas.0735920100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Lumry R, Eyring H. Conformation changes of proteins. J Phys Chem. 1954;58:110–120. [Google Scholar]
  32. Roberts CJ. Kinetics of irreversible protein aggregation: analysis of extended Lumry−Eyring models and implications for predicting protein shelf life. J Phys Chem B. 2003;107:1194–1207. [Google Scholar]
  33. Davis JM, Arakawa T, Strickland TW, Yphantis DA. Characterization of recombinant human erythropoietin produced in Chinese hamster ovary cells. Biochemistry. 1987;26:2633–2638. doi: 10.1021/bi00383a034. [DOI] [PubMed] [Google Scholar]
  34. Song Y, Azakami H, Hamasu M, Kato A. In vivo glycosylation suppresses the aggregation of amyloidogenic hen egg white lysozymes expressed in yeast. FEBS Lett. 2001;491:63–66. doi: 10.1016/s0014-5793(01)02151-2. [DOI] [PubMed] [Google Scholar]
  35. He J, Song Y, Ueyama N, Saito A, Azakami H, Kato A. Prevention of amyloid fibril formation of amyloidogenic chicken cystatin by site-specific glycosylation in yeast. Protein Sci. 2006;15:213–222. doi: 10.1110/ps.051753306. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Bosques CJ, Imperiali B. The interplay of glycosylation and disulfide formation influences fibrillization in a prion protein fragment. Proc Natl Acad Sci U S A. 2003;100:7593–7598. doi: 10.1073/pnas.1232504100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Rajan RS, Li T, Aras M, Sloey C, Sutherland W, Arai H, Briddell R, Kinstler O, Lueras AMK, Zhang Y, Yeghnazar H, Treuheit M, Brems DN. Modulation of protein aggregation by polyethylene glyol conjugation: GCSF as a case study. Protein Sci. 2006;15:1063–1075. doi: 10.1110/ps.052004006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Lin FK, Suggs S, Lin CH, Browne JK, Smalling R, Egrie JC, Chen KK, Fox GM, Martin F, Stabinsky Z. Cloning and expression of the human erythropoietin gene. Proc Natl Acad Sci U S A. 1985;82:7580–7584. doi: 10.1073/pnas.82.22.7580. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Takeuchi M, Inoue N, Strickland TW, Kubota M, Wada M, Shimizu R, Hoshi S, Kozutsumi H, Takasaki S, Kobata A. Relationship between sugar chain structure and biological activity of recombinant human erythropoietin produced in Chinese hamster ovary cells. Proc Natl Acad Sci U S A. 1989;86:7819–7822. doi: 10.1073/pnas.86.20.7819. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Pace CN, Vajdos F, Fee L, Grimsley G, Gray T. How to measure and predict the molar absorption coefficient of a protein. Protein Sci. 1995;4:2411–2423. doi: 10.1002/pro.5560041120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Pace CN. Determination and analysis of urea and guanidine hydrochloride denaturation curves. Methods in Enzymology. 1986;131:266–280. doi: 10.1016/0076-6879(86)31045-0. Edited by C.H.W. Hirs and Serge N. Timasheff Academic Press Inc. Orlando, FL. Vol. ) Academic Press. [DOI] [PubMed] [Google Scholar]
  42. Pace CN, Hebert EJ, Shaw KL, Schell D, Both V, Krajcikova D, Sevcik J, Wilson KS, Dauter Z, Hartley RW, GR Grimsley. Conformational stability and thermodynamics of folding of ribonucleases Sa, Sa2 and Sa3. J Mol Biol. 1998;279:271–286. doi: 10.1006/jmbi.1998.1760. [DOI] [PubMed] [Google Scholar]
  43. Baker NA, Sept D, Joseph S, Holst MJ, McCammon JA. Electrostatics of nanosystems: application to microtubules and the ribosome. Proc Natl Acad Sci U S A. 2001;98:10037–10041. doi: 10.1073/pnas.181342398. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Lerner MG. 2006. CHA. APBS plugin for PyMOL. Ann Arbor: University of Michigan.
  45. Dolinsky TJ, Nielsen JE, McCammon JA, Baker NA. PDB2PQR: an automated pipeline for the setup of Poisson–Boltzmann electrostatics calculations. Nucleic Acids Res. 2004;32:W665–W667. doi: 10.1093/nar/gkh381. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Olsson MHM, Søndergaard CR, Rostkowski M, Jensen JH. PROPKA3: consistent treatment of internal and surface residues in empirical pKa predictions. J Chem Theory Comput. 2011;7:525–537. doi: 10.1021/ct100578z. [DOI] [PubMed] [Google Scholar]
  47. Søndergaard CR, Olsson MHM, Rostkowski M, Jensen JH. Improved treatment of ligands and coupling effects in empirical calculation and rationalization of pKa values. J Chem Theory Comput. 2011;7:2284–2295. doi: 10.1021/ct200133y. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting Information

pro0024-0803-sd1.pdf (901.1KB, pdf)

Articles from Protein Science : A Publication of the Protein Society are provided here courtesy of The Protein Society

RESOURCES