Search

In the previous paper we described the effect of several different solvents on the structure of antibodies and demonstrated that 0.1 M glycine, pH 2.9, 7 M urea, pH 4.0, and 6 M guanidine-HCl, pH 4.0, unfold the antibodies to different degrees. Antibodies can be refolded from all of these solvents by dialysis. Polyclonal antibodies (pAbs) are a mixture of antibodies which recognize and bind different epitopes on the same antigen, with the strength of the antigen-antibody binding varying with each subpopulation. When rabbit antisera to the extracellular domain of Her2 receptor (sHer2), derived from Chinese hamster ovary cells, was applied to an antigen column, bound pAbs were recovered with a step-wise elution of 0.1 M glycine, pH 2.9 (44% of the total recovered pAb), 7 M urea, pH 4.0 (29%), and 6 M guanidine-HCl, pH 4.0 (27%), with baseline resolution between them. Fluorescence spectra of the pAbs confirmed that the 0. 1 M glycine pH 2.9 sample had near-native structure, the pAbs in 7 M urea, pH 4.0, were partially unfolded, and the pAbs in the 6 M guanidine-HCl, pH 4.0, were totally unfolded. The glycine- or urea-eluted sample was refolded by dialysis into PBS, while the guanidine-HCl-eluted sample was first dialyzed into the 7 M urea pH 4.0 buffer and then into PBS. The refolded material from glycine or urea had native-like spectra, while the spectrum of the protein refolded from 6 M guanidine-HCl was slightly perturbed. All three of these subpopulations of pAbs formed antigen-antibody complexes which could be isolated by gel-filtration chromatography, precipitated sHer2 during immunoprecipitation, and recognized sHer2 in Western blots. The guanidine-HCl-eluted material was most sensitive for Western blotting. Identical results were obtained with pAbs applied either in the batch mode or to the top of the column, indicating that antibody aggregation which may occur when applied from the top of the column is not responsible for the distribution of pAbs into different subpopulations. These results indicate that the sequential use of these three increasingly chaotropic solvents to elute antibodies results in both increased recovery of antibodies and fractionation of pAbs into subpopulations with potentially different antigen binding characteristics., (Copyright 1997 Academic Press.)

Antibodies are routinely purified by acid/salt elution from antigen affinity columns. The antibodies recovered with this procedure are active, but the recovery of protein is often low. We investigated the effect of acid and other denaturing or chaotropic solvents on the conformation of monoclonal antibodies (mAbs) made against the extracellular region of Her2 receptor (sHer2) derived from Chinese hamster ovary cells. The mAb remain almost completely folded in the 0.1 M glycine, pH 2.9, commonly used for elution, with the beta-sheet secondary structure intact, and only very small changes detected in the environment of the tryptophans. In 7 M urea, 50 mM NaAc pH 4.0, the antibody was partially unfolded, with the Trp environment further perturbed and some of the beta-sheet structure converted to disordered structure. In 6 M guanidine HCl, 50 mM NaAc, pH 4.0, the antibody is completely unfolded, with no secondary or tertiary structure present. The antibodies exposed to glycine or urea were refolded by dialysis into phosphate-buffered saline (PBS), while the guanidine HCl-denatured antibodies were refolded by dialysis into 7 M urea, pH 4.0, followed by dialysis into PBS. The refolded antibodies were capable of forming antigen-antibody complexes which could be isolated by gel filtration chromatography. Two different mAbs were subjected to immunoaffinity chromatography on sHer2-Sepharose. mAb86 was eluted by 0.1 M Gly, pH 2.9, while mAb52 was eluted with the 7 M urea, 50 mM NaAc, pH 4.0. The isolated antibodies were refolded by dialysis into PBS, analyzed for their ability to recognize native sHer2 by immunoprecipitation, and denatured sHer2 by Western blot analysis. Both preparations recognized the native protein, but precipitated slightly different forms of sHer2, indicating that they might recognize different epitopes. The mAb52 is a more sensitive reagent for Western blot analysis. Thus, this procedure can be used to recover antibodies which would not be recovered with glycine as the only eluate. It is also possible that the antibodies can be fractionated by the different eluants into populations which can be used for different applications., (Copyright 1997 Academic Press.)

At acidic pH many proteins exist in a partially unfolded form, called the "A" state. This is defined as a flexible, expanded structure with well-defined, usually native-like secondary structure, but no unique tertiary structure, and showing no cooperativity during thermal-induced denaturation. Granulocyte-colony stimulating factor (G-CSF), a four-helix bundle cytokine, maintains both thermal stability and tertiary structure at pH 2.0. We therefore examined the conformation and thermal unfolding of G-CSF at pH 2.0, 4.0 and 7.0 using circular dichroism (CD) and Fourier transform infrared spectroscopy (FTIR). The secondary structure of the molecule remains highly helical as the pH is lowered from 7.0 to 2.0. The tertiary structure of the protein is slightly different at each pH value, but even at pH 2.0 G-CSF maintains a regular three-dimensional structure. The structure is hydrodynamically compact at these different pH values, with no increase in Stoke's radius even at pH 2.0. The thermal-induced denaturation of G-CSF was determined by monitoring changes in the CD or FTIR spectra. At pH 2.0 the temperature at which thermal-induced denaturation begins is higher than it is at pH 4.0 or 7.0, the thermal unfolding transition remains cooperative and some alpha-helical structure persists even at 86 degrees C. At pH 4.0 and 7.0, secondary and tertiary structures disappear simultaneously during thermal denaturation, whereas at pH 2.0 small changes in the far-UV CD region begin to occur first, followed by the simultaneous cooperative loss of tertiary structure and much of the remaining secondary structure. The structure of G-CSF at pH 2.0 is thus revealed as compact, with a unique, three-dimensional structure, highly helical secondary structure, and most importantly, a cooperative thermal unfolding transition. G-CSF at acid pH thus does not adopt the "A" state.

The solution structure and stability of rat and human ciliary neurotrophic factor (CNTF) were examined by circular dichroism (CD), Fourier transform infrared (FTIR) and fluorescence spectroscopy and sedimentation equilibrium analyses. The secondary structure of both proteins, as assessed by CD and FTIR, consists primarily of alpha-helix, consistent with CNTF being a member of the four-helical bundle family of cytokines and neurokines, with rat CNTF containing slightly less helix (about 10% less) and slightly more disordered structure. The environment of the tyrosine and tryptophan residues, assessed by intrinsic fluorescence emission spectroscopy, appears to be the same in both proteins. Binding of anilinonaphthalene sulfonate is identical for both proteins, indicating that these two proteins have similar surface hydrophobicities in the native state. The thermal stability of the human CNTF is significantly less than that of the rat CNTF, yet their stabilities to guanidine HCl-induced denaturation are equivalent. This apparent discrepancy in stability between the two proteins may be explained by solubility differences upon thermal unfolding. Although the human protein precipitates as it is denatured by heat, the rat protein does not. It thus appears that the unfolded state of human CNTF is less soluble and more prone to aggregation than that of the rat protein upon heating, although their conformational stability is similar. Both proteins remain largely folded at pH 3.0. Sedimentation equilibrium analysis demonstrates that both rat and human CNTF exist primarily as monomers; however, significant dimer formation occurs as the protein concentrations are increased to greater than 3 mg/mL, particularly in the presence of ammonium sulfate.

Insulin-like growth factor-I (IGF-I) has three disulfide bonds and refolding of the fully reduced molecule generates varying ratios of correctly (PII) and incorrectly (PI) folded forms via several intermediates. All of the intermediates have the disulfide bond between Cys18 and 61 formed, indicating that formation of this disulfide is the first step in refolding. In order to further understand the refolding pathway, two intermediate froms, PIII with the additional disulfide Cys(6/47) formed and PIIIa with Cys(6/48) formed, were isolated. The oxidation of the remaining Cys48 and 52 in PIII and Cys47 and 52 in PIIIa would lead to PI and PII, respectively; however, air oxidation of these resulted in a rapid reshuffling into other intermediates as well as folding into the fully oxidized forms, and this occurred whether refolding was started with PIII or PIIIa. When oxidation occurred in the presence of an excess of oxidized glutathione, the predominant species generated were various glutathione adducts regardless of the initial intermediate form, indicating that formation of the last disulfide bond is not a favorable process relative to disulfide exchange when excess disulfides from oxidized glutathione are present. Interestingly, if 80 microM copper sulfate, an oxidant, is added to the refolding buffer, PIII resulted in formation of the PI form alone, whereas PIIIa resulted in the PII form alone. It was concluded from these results that the intermediate forms of IGF-1 can rapidly reshuffle between different disulfide structures, and that formation of the last disulfide bond is not as favorable a process as the conversion to other intermediates. The oxidation to form the last disulfide bond in PIII or PIIIa is accelerated and hence the interconversion to other intermediates is kinetically minimized only in the presence of copper sulfate. It appears, therefore, that the two intermediate forms, PIII and PIIIa, are the precursors of the corresponding fully oxidized forms, but their conversions are not energetically a favorable process.

Erythropoietin (EPO) is a glycoprotein hormone which belongs to the four-helical-bundle cytokine family and regulates the level of circulating red blood cells. The EPO receptor (EPOR) belongs to the cytokine-receptor family of proteins. While many of the downstream events following receptor/ligand interaction have been defined, both ligand-induced receptor dimerization and conformational changes induced by binding have been implicated as the initial step in signal transduction. In a recent paper [Philo et al. (1996), Biochemistry 38, 1681-1691] we described the formation of both 1:1 and 2:1 EPOR/EPO complexes. In this paper, we examine changes in protein conformation and stability resulting from the formation of both 1:1 and 2:1 complexes of the soluble extracellular domain of EPOR and the recombinant EPO derived from either Chinese hamster ovary cells or from Escherichia coli cells. Occupation of the first binding site results in a slight conformational change that is apparent in both the far- and near-UV circular dichroism spectra. Formation of the 2:1 complex results in an even greater change in conformation which involves the local environment of one or more aromatic amino acids, accompanied perhaps by a small increase in helical content of the complex. This change in local conformation could occur in the EPO molecule, in the EPOR, in both EPOR molecules due to dimerization, or in all molecules in the trimer. The 1:1 complex exhibits increased stability to thermal-induced denaturation relative to the individual protein component; indeed, the E. coli-derived (nonglycosylated) EPO stays folded in the complex at temperatures where the EPO alone would have unfolded and precipitated. Glycosylation of the receptor increases the reversibility of thermal denaturation, but does not affect the temperature at which this unfolding reaction occurs.

Soluble Escherichia coli-derived recombinant human stem cell factor (rhSCF) forms a non-covalently associated dimer. We have determined a dimer association constant (Ka) of 2-4 x 10(8) M-1, using sedimentation equilibrium and size exclusion chromatography. SCF has been shown previously to be present at concentrations of approximately 3.3 ng/ml in human serum. Based on the dimerization Ka, greater than 90% of the circulating SCF would be in the monomeric form. When 125I-rhSCF was added to human serum and the serum analyzed by size exclusion chromatography, 72-49% of rhSCF was monomer when the total SCF concentration was in the range of 10-100 ng/ml, consistent with the Ka determination. Three SCF variants, SCF(F63C), SCF (V49L,F63L), and SCF(A165C), were recombinantly expressed in Escherichia coli, purified, and characterized. The dimer Ka values, biophysical properties, and biological activities of these variants were studied. Dimerization-defective variants SCF(F63C)S-CH2CONH2 and SCF(V49L,F63L) showed substantially reduced mitogenic activity, while the activity of the Cys165-Cys165 disulfide-linked SCF(A165C) dimer was 10-fold higher than that of wild type rhSCF. The results suggest a correlation between dimerization affinity and biological activity, consistent with a model in which SCF dimerization mediates dimerization of its receptor, Kit, and subsequent signal transduction.

The degradation products of recombinant human interleukin-1 receptor antagonist (rhIL-1ra) formed during storage at 30 degrees C in aqueous solution were characterized. Cationic exchange chromatography of the stored sample showed two major, new peaks eluting before (P1) and after (L2) the native protein, which were interconvertible. Size-exclusion chromatography and electrophoresis documented that both the P1 and L2 fractions were irreversible dimers, formed by noncovalent interactions. A competition assay with interleukin-1 indicated that on a per monomer basis the P1 and L2 dimers retained about two-thirds of the activity of the native monomer. Infrared and far-UV circular dichroism spectroscopies showed that only minor alterations in secondary structure arose upon the formation of the P1 dimer. However, alteration in the near-UV circular dichroism spectrum suggested the presence of disulfide bonds in the P1 dimer, which are absent in the native protein. Mass spectroscopy and tryptic mapping, before and after carboxymethylation, demonstrated that the P1 dimer contained an intramolecular disulfide bond between Cys-66 and Cys-69. Although conversion of native protein to the P1 dimer was irreversible in buffer alone, the native monomer could be regained by denaturing the P1 dimer with guanidine hydrochloride and renaturing it by dialysis, suggesting that the intramolecular disulfide bond does not interfere with refolding. Analysis of the time course of P1 formation during storage at 30 degrees C indicated that the process followed first-order, and not second-order, kinetics, suggesting that the rate-limiting step was not dimerization. It is proposed that a conformational change in the monomer is the rate-limiting step in the formation of the P1 dimer degradation product. Sucrose stabilized the native monomer against this process. This result can be explained by the general stabilization mechanism for this additive, which is due to its preferential exclusion from the protein surface.

The unfolding and refolding of alpha-helical proteins has been extensively studied, demonstrating formation of intermediate structures which retain the native-like alpha-helix but lack the tertiary structure. Studies on the folding of proteins consisting primarily of beta-sheet are interesting since, unlike the alpha-helix, the beta-sheet requires the formation of peptide hydrogen bonds between two or more polypeptide segments which may be far apart in the linear sequence. Here we have studied the unfolding of the beta-sheet-containing protein tumor necrosis factor-alpha (TNF-alpha). This protein exists as a symmetric trimer in solution. Murine TNF-alpha begins to melt at 60 degrees C and unfolds to a soluble structure with a transition midpoint of 66 degrees C. This reaction is irreversible. This unfolded form contains a considerable amount of (approximately 30%) alpha-helix, as determined by circular dichroism. Human TNF-alpha begins to melt at 60 degrees C and precipitates concurrently with unfolding, such that there is no soluble protein present by 70 degrees C. The secondary and tertiary structures of murine TNF-alpha unfold simultaneously, suggesting that unfolding from the native to the unfolded state occurs cooperatively. The thermal-induced denaturation is very insensitive to protein concentration, indicating that trimer to monomer conversion, if it occurs, is not rate-limiting. Trifluoroethanol induces alpha-helix in both human and murine TNF-alpha, further demonstrating the propensity of TNF-alpha to form alpha-helix. The different behavior of human versus murine TNF-alpha upon thermal unfolding is due to differences in the solubility of the unfolded protein, the murine form being more soluble. These results indicate that TNF-alpha can form alpha-helix when the long range interactions conferred by the native structure are removed during unfolding.

Acid-induced unfolding of proteins often results in an intermediate structure, called the molten globule structure or "A" state, which retains at least partial secondary structure but lacks a rigid tertiary structure. Acid-induced unfolding has been studied extensively for alpha-helical proteins, while few studies have been done on proteins containing only beta-strands. Tumor necrosis factor-alpha (TNF-alpha) is a trimer in which the individual subunits consist of antiparallel beta-sheet, organized into a jellyroll beta-sandwich. We have found previously [Narhi et al. (1996) Biochemistry 35, 11447-11453] that thermal denaturation of TNF-alpha results in an aggregate which contains a substantial amount of alpha-helix and that the addition of trifluoroethanol induces alpha-helix in both murine and human TNF-alpha. Here we show that acid also can induce alpha-helix in these proteins. At acidic pH (below 4), both human and murine TNF-alpha convert to a monomeric form, as determined by sedimentation and diffusion constants obtained from sedimentation velocity experiments. The sedimentation coefficient indicated that this monomer was only slightly expanded relative to the native state. Near-UV circular dichroic (CD) analysis showed a loss of tertiary structure. These structural features coincide with the notion that the acid-induced structure of TNF-alpha is a molten globule. What is unique in this protein is that TNF-alpha acquires alpha-helical structure, which is not present in the native structure as determined by both CD and Fourier transform infrared spectroscopy. Even more surprising is that TNF-alpha at pH 3.3 undergoes a very gradual noncooperative change in secondary structure upon heating, which results in an increase in alpha-helical content. At pH 2.2 in the absence of salt, TNF-alpha shows considerable alpha-helix, although heating does not change the spectrum. At pH 2.2, physiological salt decreases the amount of alpha-helix at ambient temperature, and upon heating, we see the noncooperative increase in alpha-helix as observed at pH 3.3 with low salt. The addition of salt at low pH induces reassociation but to a range of oligomers rather than a unique trimer structure. This acid-induced formation of an alpha-helical monomer of TNF-alpha may be related to its known interaction with lipid bilayers.

We report here the expression of recombinant human prokallikrein and kallikrein in engineered Chinese hamster ovary cells transfected with a human genomic gene encoding preprokallikrein. At high expression levels, recombinant prokallikrein, an inactive proenzyme form, is predominantly secreted into the culture medium. Upon chromatographic separations, the inactive prokallikrein as well as the mature kallikrein after thermolysin activation of the proenzyme can be prepared to apparent purity. Both prokallikrein and kallikrein can be further separated into two distinct high- and low-molecular-weight isoforms. Kallikrein preparations are fully active in standard kallikrein activity assays such as esterase activity and kinin release from kininogen. Both kallikrein and prokallikrein display multiple molecular forms with differences in both molecular sizes and charges. The structural differences in high- and low-molecular-weight kallikreins or prokallikreins were found to be due to glycosylation, with the high-molecular-weight species glycosylated at three Asn-linked sites and the low-molecular-weight species at two of the three Asn-linked sites. The multiply charged kallikrein isoforms are derived from different numbers of sialic acids attached at the detected Asn-linked carbohydrates. In comparison with kallikrein, prokallikrein appears to show a significant decrease in the magnitude of near uv-circular dichroism bands, suggesting a change in local conformation. This conformational change correlates with the loss of activity in proenzyme due to the presence of propeptide.

The effect of oxidation of the methionine residues of Escherichia coli-derived recombinant human stem cell factor (huSCF) to methionine sulfoxide on the structure and activity of SCF was examined. Oxidation was performed using hydrogen peroxide under acidic conditions (pH 5.0). The kinetics of oxidation of the individual methionine residues was determined by quantitation of oxidized and unoxidized methionine-containing peptides, using RP-HPLC of Asp-N endoproteinase digests. The initial oxidation rates for Met159, Met-1, Met27, Met36, and Met48 were 0.11 min-1, 0.098 min-1, 0.033 min-1, 0.0063 min-1, and 0.00035 min-1, respectively, when SCF was incubated in 0.5% H2O2 at room temperature. Although oxidation of these methionines does not affect the secondary structure of SCF, the oxidation of Met36 and Met48 affects the local structure as indicated by CD and fluorescence spectroscopy. The 295-nm Trp peak in the near-UV CD is decreased upon oxidation of Met36, and lost completely following the oxidation of Met48, indicating that the Trp44 environment is becoming significantly less rigid than it is in native SCF. Consistent with this result, the fluorescence spectra revealed that Trp44 becomes more solvent exposed as the methionines are oxidized, with the hydrophobicity of the Trp44 environment decreasing significantly. The oxidations of Met36 and Met48 decrease biological activity by 40% and 60%, respectively, while increasing the dissociation rate constant of SCF dimer by two- and threefold. These results imply that the oxidation of Met36 and Met48 affects SCF dimerization and tertiary structure, and decreases biological activity.

Oxidative folding of recombinant human stem cell factor (rhSCF) produced in Escherichia coli was investigated in vitro. Folding of denatured and reduced rhSCF involves at least five intermediate forms, I-1 to I-5, detectable by their differences in hydrophobicity using reverse-phase high performance liquid chromatography. Both I-1 and I-2 contain a native-like disulfide bond, Cys4-Cys89 and Cys43-Cys138, respectively, and I-3 forms a mispaired disulfide, Cys43-Cys89. These forms appear to reach steady state equilibrium and are important folding intermediates. I-1 was found to be the prominent intermediate that directly folds into native rhSCF (N); and the thermodynamically less stable I-2 favors rearrangment into I-1. I-3 may serve as an intermediate for disulfide rearrangment between I-1 and I-2. I-4 and I-5, which are disulfide-linked dimers, are in equilibrium with reduced rhSCF and other intermediates and may not play an important role in rhSCF folding. Both trifluoroacetic acid-trapped I-1 and I-2, after isolation by high performance liquid chromatography, proceed with the remaining oxidative folding process after reconstitution. Iodoacetate-trapped I-1 and I-2 contain low alpha-helical content and some tertiary structure, while I-3 and reduced rhSCF have little ordered structure. Gel filtration/light-scattering experiments indicate that reduced rhSCF and iodoacetate-trapped I-1, I-2, and I-3 exist as dimeric forms, indicating that rhSCF dimerization precedes formation of disulfide bonds. I-1, I-2, I-3, and the C43,138A analog lacking Cys43-Cys138 bond are not biologically active or exhibit significantly lower activity. The two disulfide bonds in rhSCF seem to be essential for the molecule to maintain an active conformation required for its receptor binding and biological activities.

Distinct from the noncovalently linked recombinant human stem call factor (rhSCF) dimer, we report here the isolation and identification of an SDS-nondissociable dimer produced during folding/oxidation of rhSCF. Experimental evidence using various cleavage strategies and analyses shows that the isolated dimer is composed of two rhSCF monomers covalently linked by four disulfide bonds. The cysteines are paired as in the noncovalently associated dimer except that all pairings are intermolecular rather than intramolecular. Other structural models, involving intertwining of intramolecular disulfide loops, are ruled out. The molecule behaves similarly to the noncovalently associated dimer during ion-exchange or gel permeation chromatography. However, the disulfide-linked dimer exhibits increased hydrophobicity in reverse-phase columns and in the native state does not undergo spontaneous dimer dissociation-association as seen for the noncovalent dimer. Spectroscopic analyses indicate that the disulfide-linked and noncovalently associated rhSCF dimers have grossly similar secondary and tertiary structures. In vitro, the disulfide-linked dimer exhibits approximately 3-fold higher biological activity in supporting growth of a hematopoietic cell line and stimulating hematopoietic cell colony formation from enriched human CD34+ cells. The molecule binds to the rhSCF receptor, Kit, with an efficiency only half that of the noncovalently associated dimer. Formation of intermolecular disulfides in the disulfide-linked dimer with retention of biological activity has implications for the three-dimensional structure of noncovalently held dimer and disulfide-linked dimer.

Although there is considerable evidence that signaling by the erythropoietin (EPO) receptor is initiated when it is dimerized by binding EPO, it has been previously reported that the soluble extracellular domains of the EPO receptor (sEPOR) are not dimerized in the presence of EPO and are able to form only 1:1 complexes with EPO. We have now shown unambiguously by light scattering, sedimentation equilibrium, and titration calorimetry that two molecules of sEPOR can bind to a single EPO monomer but that the binding of the second sEPOR is approximately 1000-fold weaker than that of the first. Because this second binding interaction is quite weak (Kd of approximately 1 microM), the 2:1 sEPOR.EPO complexes are easily dissociated during chromatography (forming the 1:1 complexes reported previously) and cannot be isolated in pure form. Global analysis of the sedimentation equilibrium data has enabled us to determine the binding constants and is consistent with a model in which EPO has two independent binding sites for sEPOR but cannot exclude anticooperative or sequential binding models. The influence of glycosylation of EPO and/or sEPOR on the binding affinities has also been investigated. Titration calorimetry is consistent with the sedimentation data and shows that the weaker binding site has a more negative delta H. The relation of these results to the binding of EPO to membrane-bound receptors and to the phenomenon of apparent high-affinity and low-affinity classes of receptors is discussed.

High-affinity binding of basic fibroblast growth factor (bFGF) to the tyrosine kinase receptor requires cell-surface heparan sulfate proteoglycan or exogenous addition of heparin. The crystal structure of bFGF shows Arg40 and 45 on the surface opposite to the heparin-binding region, suggesting that these charged residues may be involved in the receptor binding. Therefore, these amino acids were mutated to aspartic acid separately or simultaneously, and also a simultaneous mutation to glutamic acid was introduced. These mutants displayed a mitogenic activity decreased greater than tenfold compared to the wild-type protein. Addition of heparin had no effect on the activity, while these mutants showed heparin-binding characteristics resembling those of the native sequence protein. The mutants exhibited decreased stability compared to the native sequence protein. Gradual changes in conformation were observed by circular dichroic and infrared spectroscopy. Heparin chromatography also showed the presence of denatured form for these mutants. However, in the presence of multivalent anions such as citrate, sucrose octasulfate, and heparin, the conformation of the mutants resembled that of the wild-type protein, as revealed by X-ray crystallography and circular dichroism spectra of the mutant with a Arg40-->Asp substitution.

Soluble CD14 (sCD14) is a 55-kDa serum protein that binds lipopolysaccharide (LPS) and mediates LPS-dependent responses in a variety of cells. Using recombinant sCD14 expressed in Chinese hamster ovary (CHO) cells, we examined the structural characteristics of sCD14 and sCD14.LPS complexes. The circular dichroism and fluorescence spectra of the sCD14 indicate that it contains substantial beta-sheet (40%) and a well-defined tertiary structure with the tryptophan residues located in environments with different degrees of hydrophobicity and solvent exposure. The spectra of the sCD14.LPS complex are identical within experimental error to the uncomplexed sCD14. Changes in surface accessibility upon LPS binding were examined using limited proteolysis with endoproteinase Asp-N. This analysis revealed that aspartic acid residues at amino acids 57, 59, and 65 are susceptible to cleavage by Asp-N, while the same residues are protected from proteolytic cleavage in the sCD14.LPS complex. These results suggest that a region including amino acids 57 to 64 is involved in LPS binding by sCD14.

CD14 is a 55-kDa glycoprotein which binds lipopolysaccharide (LPS) and enables LPS-dependent responses in a variety of cells. Recent limited proteolysis studies have implicated a region in CD14 between amino acids 57 and 64 as being involved in LPS interaction. To specifically assess the importance of this region with respect to LPS binding, we constructed a mutant sCD14 (sCD14 delta 57-64) lacking amino acids 57-64. sCD14 delta 57-64 was isolated from the serum-free conditioned medium of this cell line, and, in all assays, the purified protein failed to recognize LPS or enable LPS-dependent responses in cells. We also demonstrated that the region between amino acids 57 and 64 is required for binding of a neutralizing CD14 mAb, MEM-18. Native polyacrylamide gel electrophoresis assays were used to demonstrate that MEM-18 and LPS compete for the same binding site on CD14. These data strongly suggest that the region spanning amino acids 57-64 binds LPS and that formation of sCD14.LPS complex is required in order for sCD14-mediated responses to occur.

Comparative analyses of both glycosylated and nonglycosylated neu differentiation factor (NDF) isoforms revealed significant similarities and differences of their overall structures and functions. Biophysical analyses confirmed that all NDF isoforms are monomeric, but have an extended ellipsoidal shape in solution. All full-length NDFs are similar in secondary and tertiary structures and they contain no alpha-helix but are abundant in beta-strand structures. A small NDF fragment containing only the epidermal growth factor domain is also rich in beta-strand structures, but exhibits tertiary structure different from the long NDF forms. Monoclonal antibodies that selectively recognize epidermal growth factor domains of human NDF-alpha and -beta can specifically bind the respective NDF-alpha and -beta isoforms independent of NDF origins. Western blot analysis and quantitative binding assays further identify that an NDF preparation produced naturally from Rat1-EJ cells contains both alpha and beta isoforms in a 3 to 2 ratio. In receptor-binding competition experiments, human and rat NDF-beta isoforms have higher affinity than NDF-alpha isoforms. NDF-beta isoforms can dramatically enhance the stimulation of DNA synthesis for transfected NIH3T3 cells that overexpress HER-3 and HER-4 receptors, while NDF-alpha isoforms can only stimulate proliferation of HER-4-transfected cells with lower activity. Taken together, NDF-alpha and -beta isoforms share similar gross protein conformations but are biologically distinct.

Two different cDNAs have been isolated, coding for two forms of granulocyte colony-stimulating factor (G-CSF): one for a polypeptide of 174 amino acids and the other for a polypeptide of 177 amino acids. In this paper, we have expressed these two forms in Chinese hamster ovary cells and characterized the purified proteins for activity and conformation. In vitro mitogenic assay showed a 50-fold lower activity for the 177 form than for the 174 form. In vitro receptor binding assay showed that binding of the 177 form to the purified extracellular domain of G-CSF receptor was also diminished, while the 174 form complexed with the receptor. Circular dichroic spectra showed that both forms are similar in the secondary structure, but are slightly different in the tertiary structure. Infrared spectra also showed a slight difference between the two forms. Both techniques also demonstrated differences in stability; i.e., the 174 form is more stable than the 177 form during storage or against heat denaturation.

LIF is a multi-functional cytokine that elicits effects on a broad range of cell types. In this report, we present the high level expression of human LIF (hLIF) from a chemically synthesized gene template in Escherichia coli where it comprises up to 25% of the cellular protein. The recombinant hLIF, after purification and folding, was examined using CD, FTIR spectroscopy and light scattering. CD and FTIR spectra showed that the hLIF is an alpha-helical protein and has a distinct tertiary structure. The IFTR spectrum resembles that of other four helical bundle proteins including G-CSF and IL-6. Light scattering analysis indicated that it is a monomeric protein, distinguishing it from M-CSF and interferon gamma, which also belong to the class of four helical bundle proteins but are dimeric. Recombinant hLIF was assayed for its activity on the murine leukemic cell line, M-1 as well as on human leukemic cell line, ML-1. It inhibited the growth of M-1 cells and differentiated them towards macrophages. However, it did not have any differentiation inducing effect on human leukemic cell lines alone or in combination with other cytokines.

Recombinant human keratinocyte growth factor (rhKGF) is a fairly unstable protein, posing a challenging problem for long-term storage. During storage, the protein unfolds at relatively low temperatures and the unfolded proteins aggregate rapidly, leading to the formation of large visible precipitates. Thermal unfolding of rhKGF displays a similar pattern, i.e., unfolding is followed immediately by aggregation as the temperature is increased. As the unfolding and aggregation (precipitation) of rhKGF limit the storage life of the protein, a search for stabilizers to suppress rhKGF unfolding and aggregation has been done by examining the effects of excipients on thermal melting temperature and on the rate of protein aggregation during storage. Sulfated polysaccharides and citrate are found to be effective in increasing the melting temperature of rhKGF or preventing its aggregation. In particular, 0.5% (w/v) heparin and high molecular weight dextran sulfate, and 0.5 M citrate are highly effective, decreasing the rates of rhKGF aggregation by about 50-fold. Other negatively charged small ions, such as phosphate, also have moderate stabilizing effects on rhKGF. A mechanistic study of the aggregation pathway of rhKGF has led to a better understanding of the stabilizing effects of these molecules. Molecules which enhance rhKGF conformational stability are capable of effectively suppressing rhKGF aggregation.

Three neurotrophic factors, brain-derived neurotrophic factor (BDNF), neurotrophin-3 (NT-3), and nerve growth factor (NGF) form noncovalent homodimers in solution. Since they are highly homologous proteins, it seemed probable that two monomers of these proteins might associate together to form a heterodimer. This was tested by denaturing the two different proteins together in 6 M guanidine HCl and refolding them in phosphate-buffered saline. When the refolded mixture of BDNF and NT-3 was subjected to Mono S cation exchange chromatography, a new peak was observed eluting between NT-3 and BDNF, which accounted for about 30% of the protein used. This new protein species migrated as a single band upon native gel electrophoresis with mobility between that of the NT-3 homodimer and the BDNF homodimer, indicating that a complex had been formed. Sedimentation equilibrium data show that the dissociation constant of this heterodimer is < 3 x 10(-10) M. The heterodimer was stable upon incubation at 37 degrees C in phosphate-buffered saline over 11 days. Having determined that the heterodimer is highly stable, it was subjected to various biological assays. Autophosphorylation assay using TrkB receptor showed that the heterodimer is indistinguishable from the BDNF or NT-3 homodimer in the ability to induce phosphorylation of the receptor. It was also indistinguishable from the homodimers in the neurotrophic activity using chick dorsal root ganglion explant. In the sympathetic neuron survival assay, the heterodimer behaved more similarly to NT-3, whereas in the dopamine uptake assay, it was intermediate between the two homodimers. In addition, the heterodimer was shown to be retrogradely transported in the dorsal root ganglion neurons. A heterodimer between NGF and BDNF is formed but much less effectively than the NT-3.BDNF heterodimer, and it is not stable even at 4 degrees C. These results indicate that BDNF and NT-3 have an intersubunit contact surface for dimerization resembling each other's but different from the contact surface of NGF.

Native platelet-derived growth factor-B (PDGF-B) forms a covalent dimer through interchain disulfide bonds. In a previous study, an analog of PDGF-B was produced by replacing cysteine 43 and 52, which are involved in the interchain disulfide bonds, with serine. It was revealed that this analog protein has the dimeric molecule weight at pH 4 to 7, forming a non-covalent dimer in solution, and its mitogenic activity is similar to the native covalent dimer. However, the analog protein was more labile to pepsin digestion and low pH treatment, indicating that the interchain disulfides contribute to the stability of the protein. It is interesting to see if the conformation of the protein is affected by elimination of the interchain disulfide bonds, and if the interchain disulfides play any role in the stability of the protein. Circular dichroism and Fourier transform infrared spectroscopic analyses of the analog showed that it has a conformation similar to the wild type at pH 7.5, but is unfolded at pH 2.5, while the native PDGF-B disulfide-linked dimer shows an apparently unaltered conformation at pH 2.5. The analog is also less stable to sodium dodecylsulfate and guanidine HCl-induced denaturation at neutral pH. These results indicate that the non-covalent interactions are sufficient for proper folding and dimer formation at neutral pH, but that the interchain disulfide bonds greatly stabilize the native conformation of PDGF-B.

Class I major histocompatibility complex (MHC) proteins present peptide antigens to T cells during the immune response against viruses. Peptides are loaded into newly synthesized class I heterodimers in the endoplasmic reticulum such that most or all cell surface class I molecules contain peptides derived from endogenous or foreign proteins. We previously reported the assembly of empty heterodimers of the murine class I MHC molecule H-2Kd, from denatured heavy and light chains from which endogenous peptides had been removed [Fahnestock et al. (1992) Science 258, 1658-1662]. Here we measure thermal stability profiles of empty versus peptide-filled molecules and compare the effects of human versus murine light chains on the overall stability of the Kd heterodimer. The majority of empty heterodimers are stable at 37 degrees C regardless of the species of light chain, indicating that our previous report of the unexpectedly high thermal stability was an intrinsic property of the Kd molecule and not due to use of a murine/human chimeric protein. Binding constants are derived for a series of peptides interacting with empty Kd heterodimers. The dissociation constants of four known Kd-restricted peptides range from 2.3 x 10(-7) to 3.4 x 10(-8) M. Using a series of 24 analog peptides, the effects of length and peptide composition on binding affinity of one Kd-restricted peptide are explored, and the results are interpreted with reference to the known three-dimensional structures of class I MHC protein/peptide complexes.

The gene for human stem cell factor (SCF) encodes a leader sequence followed by 248 amino acids (Martin et al., 1990, Cell 63, 203). Of these 248 amino acids, the first 189 correspond to an extracellular domain and the remainder correspond to a hydrophobic transmembrane domain plus a cytoplasmic domain. A naturally occurring soluble form, released by proteolytic cleavage after amino acid 165, has been described. An alternatively spliced mRNA, lacking the codons for exon 6, has also been described. Since the amino acids encoded by exon 6 include the proteolytic cleavage site, the form expressed from the alternatively spliced mRNA tends to remain membrane-bound. In the present study, we have begun to explore structure/function relationships within the extracellular domain of SCF. Forms beginning at amino acid 1 (after the leader sequence) and ranging from 127 to 189 at the C-terminus have been recombinantly expressed in Escherichia coli and purified. In addition, forms missing the amino acids encoded by exon 6, forms missing up to 10 amino acids from the N-terminus, and forms with disulfide bond alterations have been expressed and purified. The forms have been characterized structurally, as well as functionally, in quantitative cell proliferation and receptor-binding assays. The results indicate that amino acids 1-141 comprise a structural and functional core and allow conclusions about the necessity of each of the two disulfide bonds for structure and function.

Recombinant human brain-derived neurotrophic factor in acid undergoes a slow loss of tertiary structure as monitored by both near-UV circular dichroism and fluorescence, and appears to retain some secondary structure, as monitored by far-UV circular dichroism and Fourier transform infrared spectroscopy. This loss of tertiary structure parallels a decrease in the weight average molecular weight, from dimer to monomer, when examined using light scattering. Increasing the temperature accelerates this slow reaction. This process may be described most simply as N2 in equilibrium with 2D where N and D are the native and denatured forms of the protein, respectively. However, the acid denaturation strongly depends on the protein concentration, with higher concentration resulting in a lower rate and extent of denaturation. This suggests that the more complicated mechanism N2 in equilibrium with 2N in equilibrium with 2D more accurately describes the denaturation, where the dissociation into a native monomer is the rate-limiting step, and the conversion of N to D occurs relatively rapidly. Size-exclusion chromatography (at neutral pH) at several points during denaturation further demonstrated that the amount of tertiary structure remaining paralleled the dimer concentration and also that the monomer form was long-lived, remaining as monomer during the course of the chromatography. Size-exclusion chromatography and sedimentation velocity determination indicated that the acid-denatured form is a compact molecule. On the basis of the above data, the acid-denatured form may be considered to be a monomeric compact intermediate A state with no tertiary structure but considerable secondary structure.(ABSTRACT TRUNCATED AT 250 WORDS)

We have studied the pathway and kinetics of refolding of recombinant human brain-derived neurotrophic factor (BDNF), which is a very tightly-associated dimer in its native state. When BDNF unfolded in 6 M guanidine hydrochloride is diluted 20-fold into phosphate-buffered saline, a partially folded intermediate is rapidly formed (< 1 min). Circular dichroism and fluorescence spectroscopy show that this intermediate has extensive secondary structure, but no well-defined tertiary structure. Size-exclusion chromatography with light scattering detection shows that it is compact and monomeric, and therefore corresponds to what is often called a "collapsed form" or "molten globule". This collapsed form disappears with a half-time of approximately 30 min, simultaneously with the appearance of native dimers, without accumulation of monomeric species with a native tertiary structure. Remarkably, the monomer-dimer association constant of the collapsed form is approximately 10(10) weaker than the native structure, and it has a low tendency to form large aggregates. Given the very large hydrophobic surface present at the dimer interface of nerve growth factor (and presumably in BDNF), these results indicate that these hydrophobic groups are not exposed in the collapsed form, and that it is therefore quite dissimilar from the native structure. A significant conformational change in the collapsed form is necessary to re-expose these hydrophobic groups to form the dimer interface, making this the rate-limiting step in reaching the native conformation.

Oh-eda et al. have shown instability of granulocyte-colony stimulating factor (G-CSF) upon storage above pH 7.0 [J. Biol. Chem. (1990) 265, 11,432-11,435]. To clarify the mechanism of this instability, the accessibility of a free cysteinyl residue at position 17 for disulfide exchange reaction was examined using a sulfhydryl reagent. The results show that the cysteine is partially solvent-exposed in both glycosylated and nonglycosylated forms, suggesting that the exposure of the cysteine plays a critical role in the instability of the protein. This is supported by the facts that at low pH where the cysteine is protonated, both proteins have much greater stability and that a Cys17-->Ser analog is extremely stable at neutral pH and 37 degrees C. It was observed that the rate of sulfhydryl titration is slower for the glycosylated form than for the nonglycosylated form, suggesting that the cysteine residue is less solvent-exposed for the former protein or that the pKa is somewhat more basic. In either case, the carbohydrate appears to affect the reactivity of the sulfhydryl group through steric hindrance or alteration in local conformation. Both the glycosylated and nonglycosylated proteins showed essentially identical conformation as determined by circular dichroism, fluorescence, and infrared spectroscopy. Unfolding of these two proteins, induced either by guanidine hydrochloride or by pH, showed an identical course, indicating comparable conformational stability. Contribution of conformational changes to the observed instability at higher pH is unlikely, since little difference in fluorescence spectrum occurs between pH 6.0 and 8.0.(ABSTRACT TRUNCATED AT 250 WORDS)

Recombinant human brain-derived neurotrophic factor (BDNF) is now under extensive investigation because of its potential clinical applications. Radioactively labeled proteins are usually required to study receptor binding and pharmacokinetic properties of proteins. This study was undertaken to see if iodination affects the biological and conformational properties of a recombinant BDNF. BDNF was iodinated using a stoichiometric amount of nonradioactive cold NaI to minimize multiple iodinations. Of the four tyrosines present in BDNF--Tyr-52, Tyr-54, Tyr-63, and Tyr-86--only Tyr-63 and Tyr-86 were iodinated under the experimental conditions used. Iodination of Tyr-63 resulted in modification without alteration of the biological activity, whereas iodination of Tyr-86 resulted in a molecule with highly compromised biological activity. Similar inactivation was observed if both Tyr-63 and Tyr-86 were iodinated. These modified proteins exhibited conformation and dimerization apparently identical to those of the native protein, as demonstrated by analytical ultracentrifugation, gel filtration, light scattering, and circular dichroism. From these results, we concluded that Tyr-52 and Tyr-54 are not accessible to the reagent and are probably buried in the hydrophobic core, whereas Tyr-63 and Tyr-86 are exposed on the surface of the molecule; of the two exposed residues, only Tyr-86 contributes to the biological activity.

Several amino groups of recombinant human erythropoietin are selectively cross-linked by specific cross-linkers including disuccinimidyl suberate or dithiobis(succinimidyl propionate). Intramolecular cross-linkings are obtained without significant change of the protein conformation using appropriate concentrations (0.2 mM) of the cross-linkers, which possess an 11-12-A length of a spacer between two reacting groups. Intramolecularly cross-linked peptides obtained suggest that several amino groups in erythropoietin (EPO) are positioned at a distance of near 12 A in the solution state. These interfacing amino groups include Lys 20-Lys 154, Lys 45-Lys 140, Lys 52-Lys 154, Lys 52-Lys 140, and Ala 1-Lys 116. A comparison of the cross-linking results between nonglycosylated EPO and glycosylated EPO suggests that both proteins retain high similarity regarding protein conformation. These results fit a structural model similar to that of human growth hormone, in which four alpha-helical bundles and a long stretch of beta-sheet structure are involved in the active protein.

We have used recombinant DNA methods to produce two forms of bovine acidic fibroblast growth factor (aFGF), one with alanine substituted for the cysteine at position 47 and the other with the Ala47 change plus the substitution of glycine for the naturally occurring histidine at position 93. Both forms were expressed at high levels in Escherichia coli and purified to near homogeneity by solubilization of the inclusion bodies containing the aFGF, ion exchange chromatography, refolding of the protein and hydrophobic interaction chromatography. Circular dichroic and infrared spectra suggested that the proteins are similar in secondary and tertiary structures and contain little or no alpha-helical conformations. Hydrophobic interaction chromatography showed that aFGF C47A/H93G is slightly more hydrophobic than the aFGF C47A form, suggesting that residue 93 is exposed to the solvent. Half-maximal activity in an in vitro bioassay system was reached at a 10- to 20-fold lower dose for the aFGF C47A/H93G form than for the aFGF C47A form, suggesting that alteration of this residue has an effect on the region responsible for receptor binding. Addition of 50 micrograms/ml heparin enhanced the in vitro activity of the aFGFs, reducing the half-maximal dose to approximately 100 pg/ml for both forms, comparable to that observed previously for basic FGF with or without heparin in this assay system.

The structural properties of human brain-derived neurotrophic factor (BDNF) and neurotrophin-3 (NT-3) were studied using sedimentation equilibrium and circular dichroism (CD), fluorescence and Fourier-transform infrared spectroscopies, and compared with those of human nerve growth factor (NGF). Both the far UV CD and infrared spectra indicate that these three proteins have similar, but not identical, secondary structures which contain primarily beta-sheet and irregular structures. NGF appears to contain the most beta-sheet while NT-3 contains a small fraction of alpha-helix. The near UV CD spectra appear to indicate that the three proteins contain disulfide bonds in similar environments, suggesting a resemblance in tertiary structure. The fluorescent tryptophans found in the molecules are relatively solvent exposed, while Trp102 found only in NT-3 is possibly quenched. The fluorescent Trp(s) in NGF are significantly quenched relative to those in the other two neurotrophic factors. Both NT-3 and BDNF have very hydrophilic surfaces at neutral pH, as indicated by a low binding affinity to a hydrophobic probe, anilinonaphthalenesulfonate. Sedimentation equilibrium showed that BDNF, NT-3, and NGF exist as strongly associated dimers in phosphate-buffered saline, pH 7.1. Fits of the observed fringe displacements to various association models suggested that the BDNF, NT-3, and NGF samples contain, in addition to the principal dimeric species, some oligomers, and that NT-3 contains a small fraction of incompetent monomer.

Folding of recombinant human insulin-like growth factor-I (IGF-I) results in two distinct species as resolved by reversed-phase high-performance liquid chromatography (RP-HPLC). The earlier eluting peak (PI) has a nonnative disulfide structure, while the later eluting peak (PII) assumes the native disulfide structure. This folding problem causes a lower yield and requires expensive RP-HPLC separation. In contrast, IGF-II folds mainly into a single form with all three disulfide bonds correctly formed. Sequence comparison of the two molecules revealed that IGF-I has arginine at residues 55 and 56, while IGF-II has alanine and leucine, respectively, at these positions. Two analogs of IGF-I, IGF-I (Ala55/Leu56) and IGF-I (Leu56), behave similarly to IGF-II upon refolding and RP-HPLC; that is, a single peak eluted from the RP-HPLC column. However, when the peaks isolated by RP-HPLC were subjected to hydrophobic interaction chromatography, circular dichroism, and peptide mapping, they were found to be a mixture of PI and PII. It was then concluded that factors other than just these two residues contribute to correct folding of IGF-II and that the PI and PII of the above two IGF-I mutants assume different conformation at neutral pH but similar conformation under the RP-HPLC condition.

Can one protein sequence encode two structures? Oxidative folding of human insulin-like growth factor 1 (IGF-1), a globular protein of 70 residues, is shown to yield two products of similar thermodynamic stability. This observation is of particular interest in light of the recent demonstration that two of the three disulfide bonds in native IGF-1 rearrange in the presence of dithiothreitol [Hober, S., et al. (1992) Biochemistry 31, 1749-1756]. Kinetics of the IGF-1 folding pathway were monitored by high-performance liquid chromatography (rp-HPLC). Disulfide-pairing schemes of intermediates and products were established by peptide mapping. Two disulfide isomers were obtained as products: one with native insulin-like pairing [6-48; 18-61; 47-52] (designated native IGF-1; 60% yield) and the other with alternative pairing [6-47; 18-61; 48-52] (designated IGF-swap; 40% yield). The predominant early intermediate contains the single disulfide 18-61, which is shared in common by the two products. Relative yields of native IGF-1 and IGF-swap are independent of protein concentration under dilute conditions. In the absence of an added thiol reagent, each isomer is stable indefinitely at neutral pH; in the presence of an added thiol reagent, the two isomers interconvert with an Arrhenius activation barrier of 12 kcal/mol. Interconversion does not require complete reduction and yields the same ratio of products as initial folding, demonstrating thermodynamic control. Spectroscopic studies using circular dichroism (CD), infrared spectroscopy (FTIR), two-dimensional 1H-NMR (2D-NMR), and photochemical dynamic nuclear polarization (photo-CIDNP) suggest that IGF-1 and IGF-swap adopt similar secondary structures but distinct tertiary folds. Implications of these observations for understanding the topology of protein-folding pathways are discussed.

Insulin and insulin-related proteins contain three motif-specific disulfide bonds. Here we examine the role of these disulfide bonds in the folding and function of one family member, human insulin-like growth factor 1 (IGF-1). Analogues containing pariwise Cys-->Ser or Cys-->Ala substitutions were expressed in Escherichia coli, purified, and analyzed with respect to receptor-binding, solution structure, and thermodynamic stability. An analogue lacking all three disulfide bonds (designated des-Cys-IGF-1) is inactive and unfolded. Introduction of the [18-61] disulfide bond, previously shown to occur in an early intermediate in oxidative refolding [Miller, J. A., Owers-Narhi, L., Hua, Q. X., Rosenfeld, R., Arakawa, T., Rohde, M., Prestrelski, S., Lauren, S., S. Stoney, K. S., Tsai, L., & Weiss, M. A. (1993) Biochemistry (preceding paper in this issue)], results in a compact partially folded state with low but significant biological activity. Additional but incomplete structural organization and biological activity are observed following introduction of either the [6-48] or the [47-52] disulfide bonds. Native function, structure, and stability require the presence of all three disulfide bonds. These analogues provide genetic models of IGF-1 protein-folding intermediates. Their characterization suggests that bifurcation of the IGF-1 folding pathway reflects alternative late steps in the folding of a molten-globule intermediate.

In order to probe the role of the individual tryptophans of granulocyte-colony stimulating factor (G-CSF) in pH and guanidine HCl-induced fluorescence changes, site-directed mutagenesis was used to generate mutants replacing Trp118,Trp58, or both with phenylalanine. Neither Trp to Phe mutation affected the folding or activity of the recombinant G-CSF, and the material expressed in yeast behaved identically to that expressed in Escherichia coli. All of the G-CSF species responded to pH and guanidine HCl in qualitatively the same manner. Trp58 has a fluorescence maximum at 350 nm and is quenched to a greater extent by the addition of guanidine HCl, indicating that it is fully solvent-exposed. Trp118 has a fluorescence maximum at 344 nm, and is less solvent-accessible than Trp58. The analog in which both tryptophans have been replaced with phenylalanine shows only tyrosine fluorescence, with a peak at 304 nm which decreases with increasing pH. The intensity of the tyrosine fluorescence in this analog is much greater than that of the native sequence protein or single tryptophan mutants, indicating that energy transfer is taking place from tyrosine to tryptophan in these molecules. Below neutral pH the tyrosine fluorescence is much greater in the [Phe58]G-CSF than in the [Phe118]G-CSF, indicating that Trp58 might be a more efficient recipient of energy transfer from the tyrosine(s).

A new technique, temperature gradient gel electrophoresis (TGGE), was applied to the study of heat-induced protein denaturation. The gels used contained 30 mM Borax + 75 mM boric acid, pH 8.4, and various concentrations of urea. When this technique was applied to bovine serum albumin, the protein showed discontinuous bands upon melting, indicating that the thermal transition is irreversible. The apparent melting temperature, calculated based on the relative intensity of two bands in the transition region, was 58 degrees C in 2 M urea. When the thermal denaturation of bovine serum albumin was analyzed spectroscopically, the transition was again irreversible, with an apparent transition temperature of 57 degrees C, consistent with the TGGE results. Recombinant stem cell factor, recombinant granulocyte macrophage-colony stimulating factor, and catalase were also analyzed by TGGE, indicating that the technique can be used to analyze denaturation of monomeric and multimeric proteins.

The conformational properties of several platelet-derived growth factors (PDGFs) were characterized by circular dichroism (CD), Fourier transform infrared spectroscopy (FTIR), gel filtration and sedimentation equilibrium. Three different forms of disulfide linked dimer, PDGF-AA, PDGF-AB, and PDGF-BB, showed similar far UV CD spectra with evidence for slight beta-structure, but little evidence of other regular secondary structures. These spectra were, however, different from the far UV CD spectra of the glutathione adducts of PDGF-A and B, suggesting that the latter two proteins adopt different conformations in the absence of intra- or inter-molecular disulfide bonds. FTIR studies confirmed this by showing that the glutathione adducts of the PDGF-B protein have a significantly lower amount of regular secondary structures than PDGF-BB. Additionally, the increased bandwidths of the amide I components of the FTIR spectrum of the glutathione adduct indicates a more flexible structure relative to the dimeric form. Sedimentation equilibrium analysis showed that PDGF-BB is primarily a dimer and that the glutathione form is primarily a monomer. Thus, it was concluded that the glutathione derivative has little affinity to form non-covalent dimers in neutral solution.

The folding and oxidation of recombinant human granulocyte colony-stimulating factor solubilized from Escherichia coli inclusion bodies was investigated. During the folding process, two intermediates, I1 and I2, were detected by kinetic studies using high performance liquid chromatography. I1 exists transiently and disappears quickly with the concomitant formation of I2. In contrast, I2 requires a longer time to fold into the final oxidized form, N. CuSO4 catalysis increases the folding rate of I2 from I1, while CuSO4 and elevated temperature (37 degrees C) have a dramatic effect on the folding rate of N from I2. These observations suggest the following sequential oxidative folding pathway. [sequence: see text] Peptide map analysis of the iodoacetate-labeled intermediates revealed that I1 represents the fully reduced granulocyte colony-stimulating factor containing 5 free cysteines; I2 is the partially oxidized species containing a single Cys36-Cys42 disulfide bond; and N, the final folded form, has two disulfide bonds. The physicochemical properties and biological activities of I1, I2, N, and several Cys----Ser analogs made by site-directed mutagenesis were further investigated. In guanidine hydrochloride-induced denaturation studies, the disulfide-reduced intermediates and the analogs missing either of the disulfide bonds are conformationally less stable than those of the wild type molecule or the analog with the free Cys at position 17 changed to Ser. Recombinant human granulocyte colony stimulating factor lacking either disulfide bond or both has overall secondary and tertiary structures different from those of the wild type molecule and exhibits lower biological activity. These studies show that disulfide bond formation is crucial for maintaining the molecule in a properly folded and biologically active form.

The stabilities of trypsin and soybean trypsin inhibitor in sodium dodecylsulfate (SDS) were examined by SDS-polyacrylamide gel electrophoresis (PAGE). Both samples contained several bands, all of which migrated to positions corresponding to the appropriate molecular weight or less, even when the samples were unheated, suggesting that both the trypsin and trypsin inhibitor are susceptible to SDS-induced denaturation. When they were mixed together prior to addition of SDS-PAGE sample buffer (1% SDS), a new smearing band appeared which corresponded to a molecular weight of around 46,000, suggesting that these proteins form a stable complex in SDS. This was confirmed by electroblotting and sequence analysis, which indicated that this band contains both the trypsin and inhibitor sequences. At a fixed concentration of the inhibitor, increasing concentrations of the trypsin resulted in an increase in the intensity of the complex band. When the mixture was heated for 10 min in 1% SDS, the complex band disappeared in a temperature-dependent manner. The melting temperature determined under the experimental conditions used was about 35 degrees C. Similar results were obtained with Bowman-Birk trypsin inhibitor, except that the complex with the above inhibitor had a higher melting temperature, around 41 degrees C, suggesting that the Bowman-Birk inhibitor/trypsin complex is more stable than the soybean inhibitor/trypsin complex.

Unfolding of a fungal alpha-amylase in aqueous sodium dodecylsulfate (SDS) solution was examined by SDS-polyacrylamide gel electrophoresis (PAGE). When the alpha-amylase was incubated with 1% SDS at room temperature and subjected to SDS-PAGE, it showed a much higher mobility than expected from the molecular weight. Circular dichroic and gel filtration analyses indicated that the protein is apparently in the native conformation upon incubation with 1% SDS. When the protein was heated in the presence of 1% SDS at 90 degrees C for 10 min, it had a lower mobility in SDS-PAGE and showed characteristics of an unfolded protein by circular dichroism and gel filtration. The melting temperatures of the protein were determined in the absence and presence of SDS by incubating it for 10 min at various temperatures. The melting temperatures were 70, 55, and 49 degrees C in the presence of 0, 1, and 2% SDS, respectively. The observed small shift of the melting temperatures by SDS suggests that the destabilizing action of SDS on the alpha-amylase is weak. However, the unfolding in SDS is not reversible process, since prolonged incubation of the protein with 1% SDS at 50 degrees C gradually increased the amount of unfolded protein. This indicates that the SDS-induced unfolding of the alpha-amylase is a slow process.

The prt gene, encoding a protease (Prt) from Streptomyces lividans TK24, was cloned and sequenced. An S. lividans host with plasmid-borne prt secreted 200 micrograms/ml of a 22-kDa Prt into the culture medium. Prt is classified as a metalloprotease since its activity is significantly inhibited by 1,10-phenanthroline or EDTA. The region upstream from prt codes for an incomplete open reading frame (ORF) oriented opposite to prt. This ORF has a strong similarity to a gene family (lysR) whose members regulate the transcription of structural genes required for either biosynthesis or degradation.

To further elucidate the role of the disulfide bonds in determining the protein folding of recombinant human epidermal growth factor (r-HuEGF) we studied the structure of reduced and oxidized r-HuEGF using circular dichroism (CD). The far UV CD spectrum of reduced r-HuEGF in 10 mM sodium phosphate pH 3.0 is very different from that of the oxidized molecule. The spectrum of the reduced molecule consists of a plateau from 225 to 200 nm, consistent with the presence of alpha-helix, beta-sheet, and unordered structure. The addition of the alpha-helix inducer trifluoroethanol to the reduced molecule resulted in an enhancement of alpha-helix, at the apparent expense of beta-sheet, while the oxidized molecule was unaffected by the presence of this reagent. Secondary structure predictions based on the amino acid sequence of EGF correlate most closely with the structure of the reduced molecule. From these results, it appears that the r-HuEGF has a more regular secondary structure in the absence of the disulfide bonds than in their presence. This suggests that the folding of EGF occurs by destroying the regular secondary structure that was present in the reduced state, and that the structure of the native molecule is dictated largely by disulfide bonding.

Circular dichroism (CD) and Fourier transform infrared spectroscopic studies have shown that the secondary structure of transforming growth factor alpha (TGF-alpha) is very similar to that of epidermal growth factor (EGF). The infrared spectra revealed a minor difference between the two proteins, in particular in the beta-sheet structure. A large difference was observed with CD between the two proteins in the apparent conformation each adopts when the disulfide bonds are reduced. Reduced TGF-alpha showed a distinct alpha-helical conformation only at a high trifluoroethanol concentration, whereas reduced EGF assumed an alpha-helical conformation in the absence of trifluoroethanol. This indicates that these two proteins adopt different secondary structures in the absence of disulfide bonds, although they assume similar folding structures in their presence. These data suggest that the disulfide bonds to a large degree dictate the conformation of these two proteins. Additionally, differences in the dynamic behavior between EGF and TGF-alpha were also observed. Infrared experiments showed that the hydrogen-deuterium exchange rate is much higher for TGF-alpha than for EGF, indicating that TGF-alpha is a more flexible molecule. The rate of reduction of the disulfide bonds by dithiothreitol was also faster for TGF-alpha. Therefore, it can be concluded that although EGF and TGF-alpha have a similar overall conformation, TGF-alpha is a more flexible molecule than EGF.

Erythropoietin is a glycoprotein hormone that stimulates the maturation of late erythroid progenitor cells. It has three N-linked and one O-linked carbohydrates which play an important role in the biosynthesis and biological activities of the protein. To determine the role the carbohydrate might have in maintaining the conformational stability of the protein, the protein expressed in mammalian cells (fully glycosylated), the asialo mammalian-expressed protein, and the protein expressed in Escherichia coli (no carbohydrate) were compared for their stability to guanidine HCl, pH, and temperature. Circular dichroism was used to follow protein unfolding. Both the intact and asialo mammalian-expressed proteins unfolded with a cooperative transition in guanidine HCl, with a midpoint at 1.75 M guanidine HCl. The E. coli-expressed material unfolded with a midpoint of 1.2 M guanidine HCl, and a delta G of unfolding which was 1.4 kcal/mol less than that of the two glycosylated molecules. The E. coli-derived protein was also significantly less stable to pH-induced conformational changes, showing a cooperative transition in 35% glycerol with a midpoint at pH 4.4, while both the intact and asialo mammalian-expressed molecules had a transition midpoint of pH 3.75 in the absence of glycerol, and approximately pH 3 in the presence of 35% glycerol. The E. coli-expressed molecule unfolded and precipitated upon heating to 44 degrees C, while the asialo and intact mammalian-expressed proteins remained soluble, with a Tm of 56 degrees C. From these experiments, the carbohydrate appears to play a critical role in stabilizing the erythropoietin molecule to denaturing conditions, and this increased stability does not depend on the presence of sialic acid.

We have recently described the identification, isolation, and characterization of a factor, termed stem cell factor (SCF), which acts on primitive hematopoietic progenitors of the marrow. A soluble form of the factor was isolated from the conditioned medium of a rat cell line (Zsebo, K. M., Wypych, J., McNiece, I. K., Lu, H. S., Smith, K. A., Karkare, S. B., Sachdev, R. K., Yuschenkoff, V. N., Birkett, N. C., Williams, L. R., Satyagal, V. N., Tung, W., Bosselman, R. A., Mendiaz, E. A., and Langley, K. E. (1990) Cell 63, 195-201) and rat and human cDNAs have been cloned (Martin, F. H., Suggs, S. V., Langley, K. E., Lu, H. S., Ting, J., Okino, K. H., Morris, C. F., McNiece, I. K., Jacobsen, F. W., Mendiaz, E. A., Birkett, N. C., Smith, K. A., Johnson, M. J., Parker, V. P., Flores, J. C., Patel, A. C., Fisher, E. F., Erjavec, H. O., Herrera, C. J., Wypych, J., Sachdev, R. K., Pope, J. A., Leslie, I., Wen, D., Lin, C.-H., Cupples, R. L., and Zsebo, K. M. (1990) Cell 63, 203-211). The cDNAs encode amino acids C-terminal to those found in the isolated natural form, including a putative transmembrane domain. This paper describes the structural characterization of soluble forms of recombinant human SCF purified from Escherichia coli (unglycosylated) and from Chinese hamster ovary (CHO) cells (glycosylated). Fluorescence emission spectra indicate that the single Trp residue is present in a hydrophobic environment. Circular dichroism and infrared spectroscopy indicate considerable secondary structure, including both alpha-helix and beta-sheet. Molecular weight determinations by sedimentation equilibrium show that the molecules are dimeric (noncovalently associated), and gel filtration analyses are consistent with this conclusion. The CHO cell-derived SCF is about 30% carbohydrate by weight, with both N-linked and O-linked sugar. The presence or absence of the carbohydrate does not influence the results of the various structural analyses.

Subtilisin (Sbt) and Streptomyces subtilisin inhibitor (SSI) were analyzed either alone or together using sodium dodecylsulfate (SDS)-polyacrylamide gel electrophoresis (PAGE). With all ratios of Sbt to SSI tested, the proteins formed a stoichiometric complex, and migrated abnormally at the top of the gel. Electroblotting and amino acid sequence analysis of the complex band showed both Sbt and SSI present at approximately equal molar ratios. When excess Sbt was present, it migrated as a free but still folded form slightly above the band corresponding to the complex. When excess SSI was present, it migrated as several species with molecular weights smaller than the intact form; in fact, the sequences of some of these species indicated that they lacked different amounts of N-terminal and possibly C-terminal residues.

Fluorescence and circular dichroism were used to follow the pH-dependent conformational changes of granulocyte colony stimulating factor (G-CSF). Tryptophan fluorescence of the spectra monitored at 344 nm, or after deconvolution of the emission spectra, at 345 nm, showed a decrease in intensity on going from pH 7 to 4, with a midtransition pH of 5.8. On the other hand, tyrosine fluorescence measured either by the ratio of intensity at 308 nm to that at 344 nm, or by the fluorescence intensity at 303 nm after deconvolution of the spectra, increased in intensity as the pH was changed from 6 to 2.5, with a midtransition pH of 4.5. Near UV circular dichroic spectra also showed changes between pH 7.5 and 4.5, which correlated with the transition monitored by the tryptophan fluorescence. The guanidine hydrochloride-induced conformational changes of G-CSF at five pH values from 2.5 to 7.5 were also studied. Circular dichroic and fluorescence spectra revealed minor conformational changes by the addition of 1 or 2 M guanidine HCl at all pH values examined, while the major conformational transition occurred between 2 and 4 M guanidine hydrochloride. The secondary structure of the protein was most stable between pH 3.3 and 4.5. The guanidine HCl-induced denaturation of G-CSF involved more than a two-state transition, with detectable intermediate(s) present, and the structure of the intermediate(s) appeared to depend on the pH used. These results are consistent with the pH dependence of the structure described above, and demonstrate the complex conformational properties of G-CSF.

Solvents play a critical role in hydrophobic interaction chromatography (HIC), since the separation of proteins by HIC is based on the hydrophobicity of the proteins presented to the solvents. This review first describes the solvent properties which determine the effect of cosolvents on the binding and elution of proteins in HIC; i.e., the protein solvent interactions and the surface tension of water/cosolvent mixture. Second are presented the various cosolvents which have been tested as facilitating binding or elution of the proteins. Last, some examples of solvent manipulation which resolved complex mixtures of proteins by HIC are reviewed.