Cation-dependent stability of subtilisin.

Subtilisin BPN' contains two cation binding sites. One specifically binds calcium (site A), and the other can bind both divalent and monovalvent metals (site B). By binding at specific sites in the tertiary structure of subtilisin, cations contribute their binding energy to the stability of the native state and increase the activation energy of unfolding. Deconvoluting the influence of binding sites A and B on the inactivation rate of subtilisin is complicated, however. This paper examines the stabilizing effects of cation binding at site B by using a mutant of subtilisin BPN' which lacks calcium site A. Using this mutant, we show that calcium binding at site B has relatively little effect on stability in the presence of moderate concentrations of monovalent cations. At [NaCl] =100 mM, site B is >or=98% occupied with sodium, and therefore its net occupancy with a cation varies little as subtilisin is titrated with calcium. Exchanging sodium for calcium results in a 5-fold decrease in the rate of inactivation. In contrast, because of the high selectivity of site A for calcium, its occupancy changes dramatically as calcium concentration is varied, and consequently the inactivation rate of subtilisin decreases approximately 200-fold as site A becomes saturated with calcium, irrespective of the concentration of monovalent cations.

Stabilizing mutations and calcium-dependent stability of subtilisin.

Stability is a property of subtilisin which has proven particularly amenable to enhancement via random mutagenesis and screening, yet the effects of most stabilizing mutations are not understood in structural and energetic detail. This paper seeks to explain the longstanding observation that stabilizing mutations are usually calcium-dependent in their stabilizing effect, irrespective of their proximity to the calcium binding sites. Stabilizing mutations in subtilisin fall into one of three classes. The largest class of mutations stabilize only in the presence of excess calcium. A smaller number of mutations stabilize independently of [calcium], and a few mutations stabilize only in the presence of chelating agents, such as EDTA. This study compares the effects of mutations from each class when introduced into subtilisin BPN' and two calcium-free versions of subtilisin. The calcium-dependent effects of mutations can be explained by considering subtilisin to be in conformational equilibrium between two structurally similar but energetically distinct states: N and N*. The equilibrium from the N* to the N state can be altered either by calcium binding to site A or by mutation. Mutations which stabilize only in the presence of calcium stabilize the N state relative to N*. Mutations which stabilize only in the presence of chelants stabilize the N* state relative to N. As a byproduct of this analysis, we have developed a hyperstable variant of subtilisin whose inactivation at high temperature in the presence of EDTA is 10(5) times slower than wild-type subtilisin.

Stability and global fold of the mouse prohormone convertase 1 pro-domain.

We have purified the mouse prohormone convertase 1 (PC1) pro-domain expressed in Escherichia coli cells and demonstrated, using a number of biophysical methods, that this domain is an independent folding unit with a T(m) of 39 degrees C at a protein concentration of 20 microM and pH 7.0. This differs significantly from similar pro-domains in bacteria and human furin, which are unfolded at 25 degrees C and require the catalytic domain in order to be structured [Bryan et al. (1995) Biochemistry 34, 10310-10318; Bhattacharjya et al. (2000) J. Biomol. NMR 16, 275-276]. Using heteronuclear NMR spectroscopy, we have determined the backbone (1)H, (13)C, and (15)N assignments for the pro-domain of PC1. On the basis of (1)H/(13)C chemical shift indices, NOE analysis, and hydrogen exchange measurements, the pro-domain is shown to consist of a four-stranded beta-sheet and two alpha-helices. The results presented here show that both the bacterial pro-domain in complex with subtilisin and the uncomplexed mouse PC1 pro-domain have very similar overall folds despite a lack of sequence homology. The structural data help to explain the location of the secondary processing sites in the pro-domains of the PC family, and a consensus sequence for binding to the catalytic domain is proposed.

Structure and dynamics of an acid-denatured protein G mutant.

NMR studies of protein denatured states provide insights into potential initiation sites for folding that may be too transient to be observed kinetically. We have characterized the structure and dynamics of the acid-denatured state of protein G by using a F30H mutant of G(B1) which is on the margin of stability. At 5 degrees C, F30H-G(B1) is greater than 95% folded at pH 7.0 and is greater than 95% unfolded at pH 4.0. This range of stability is useful because the denatured state can be examined under relatively mild conditions which are optimal for folding G(B1). We have assigned almost all backbone (15)N, H(N), and H(alpha) resonances in the acid-denatured state. Chemical shift, coupling constant, and NOE data indicate that the denatured state has considerably more residual structure when studied under these mild conditions than in the presence of chemical denaturants. The acid-denatured state populates nativelike conformations with both alpha-helical and beta-hairpin characteristics. To our knowledge, this is the first example of a denatured state with NOE and coupling constant evidence for beta-hairpin character. A number of non-native turn structures are also detected, particularly in the region corresponding to the beta1-beta2 hairpin of the folded state. Steady-state ¿(1)H-(15)N¿ NOE results demonstrate restricted backbone flexibility in more structured regions of the denatured protein. Overall, our studies suggest that regions of the helix, the beta3-beta4 hairpin, and the beta1-beta2 turn may serve as potential initiation sites for folding of G(B). Furthermore, residual structure in acid-denatured F30H-G(B1) is more extensive than in peptide fragments corresponding to the beta1-beta2, alpha-helix, and beta3-beta4 regions, suggesting additional medium-to-long-range interactions in the full-length polypeptide chain.

Protein engineering of subtilisin.

The serine protease subtilisin is an important industrial enzyme as well as a model for understanding the enormous rate enhancements affected by enzymes. For these reasons along with the timely cloning of the gene, ease of expression and purification and availability of atomic resolution structures, subtilisin became a model system for protein engineering studies in the 1980s. Fifteen years later, mutations in well over 50% of the 275 amino acids of subtilisin have been reported in the scientific literature. Most subtilisin engineering has involved catalytic amino acids, substrate binding regions and stabilizing mutations. Stability has been the property of subtilisin which has been most amenable to enhancement, yet perhaps least understood. This review will give a brief overview of the subtilisin engineering field, critically review what has been learned about subtilisin stability from protein engineering experiments and conclude with some speculation about the prospects for future subtilisin engineering.

Rapid folding of calcium-free subtilisin by a stabilized pro-domain mutant.

In vitro folding of mature subtilisin is extremely slow. The isolated pro-domain greatly accelerates in vitro folding of subtilisin in a bimolecular reaction whose product is a tight complex between folded subtilisin and folded pro-domain. In our studies of subtilisin, we are trying to answer two basic questions: why does subtilisin fold slowly without the pro-domain and what does the pro-domain do to accelerate the folding rate? To address these general questions, we are trying to characterize all the rate constants governing individual steps in the bimolecular folding reaction of pro-domain with subtilisin. Here, we report the results of a series of in vitro folding experiments using an engineered pro-domain mutant which is independently stable (proR9) and two calcium-free subtilisin mutants. The bimolecular folding reaction of subtilisin and proR9 occurs in two steps: an initial binding of proR9 to unfolded subtilisin, followed by isomerization of the initial complex into the native complex. The central findings are as follows. First, the independently stable proR9 folds subtilisin much faster than the predominantly unfolded wild-type pro-domain. Second, at micromolar concentrations of proR9, the subtilisin folding reaction becomes limited by the rate at which prolines in the unfolded state can isomerize to their native conformation. The simpliest mechanism which closely describes the data includes two denatured forms of subtilisin, which form the initial complex with proR9 at the same rate but which isomerize to the fully folded complex at much different rates. In this model, 77% of the subtilisin isomerizes to the native form slowly and the remaining 23% isomerizes more rapidly (1.5 s-1). The slow-folding population may be unfolded subtilisin with the trans form of proline 168, which must isomerize to the cis form during refolding. Third, in the absence of proline isomerization, the rate of subtilisin folding is rapid and at [proR9] 3 s-1. The implications of these results concerning why subtilisin folds slowly without the pro-domain are discussed.

Stabilizing the subtilisin BPN' pro-domain by phage display selection: how restrictive is the amino acid code for maximum protein stability?

We have devised a procedure using monovalent phage display to select for stable mutants in the pro-domain of the serine protease, subtilisin BPN'. In complex with subtilisin, the pro-domain assumes a compact structure with a four-stranded antiparallel beta-sheet and two three-turn alpha-helices. When isolated, however, the pro-domain is 97% unfolded. These experiments use combinatorial mutagenesis to select for stabilizing amino acid combinations at a particular structural locus and determine how many combinations are close to the maximum protein stability. The selection for stability is based on the fact that the independent stability of the pro-domain is very low and that binding to subtilisin is thermodynamically linked to folding. Two libraries of mutant pro-domains were constructed and analyzed to determine how many combinations of amino acids at a particular structural locus result in the maximum stability. A library comprises all combinations of four amino acids at a structural locus. Previous studies using combinatorial genetics have shown that many different combinations of amino acids can be accommodated in a selected locus without destroying function. The present results indicate that the number of sequence combinations at a structural locus, which are close to the maximum stability, is small. The most striking example is a selection at an interior locus of the pro-domain. After two rounds of phagemid selection, one amino acid combination is found in 40% of sequenced mutants. The most frequently selected mutant has a deltaG(unfolding) = 4 kcal/mol at 25 degrees C, an increase of 6 kcal/mol relative to the naturally occurring sequence. Some implications of these results on the amount of sequence information needed to specify a unique tertiary fold are discussed. Apart from possible implications on the folding code, the phage display selection described here should be useful in optimizing the stability of other proteins, which can be displayed on the phage surface.

Engineering the independent folding of the subtilisin BPN' pro-domain: correlation of pro-domain stability with the rate of subtilisin folding.

The 77-amino acid pro-domain greatly accelerates the in vitro folding of subtilisin in a bimolecular reaction whose product is a tight complex between folded subtilisin and folded pro-domain. In this complex the pro-domain has a compact structure with a four-stranded antiparallel beta-sheet and two three-turn alpha-helixes. When isolated from subtilisin, however, the pro-domain is 97% unfolded even under optimal folding conditions. The instability of the isolated pro-domain suggests that there may be a thermodynamic linkage between the stability of the pro-domain and its ability to facilitate subtilisin folding. On the basis of the X-ray crystal structure of the pro-domain subtilisin complex, we have designed stabilizing mutations in three areas of the pro-domain: alpha-helix 23-32 (E32Q), beta-strands 35-51 (Q40L), and alpha-helix 53-61 (K57E). These amino acid positions were selected because they do not contact subtilisin in the complex and because they appear to be in regions of the structure which are not well packed in the wild type pro-domain. Since none of the mutations directly contact subtilisin, their effects on the folding of subtilisin are linked to whether or not they stabilize a conformation of the pro-domain which promotes subtilisin folding. By sequentially introducing the three stabilizing mutations, the equilibrium for folding the pro-domain was shifted from 97% unfolded to 65% folded. By measuring the ability of these mutants to fold subtilisin, we are able to establish a correlation between the stability of the pro-domain and its ability to accelerate subtilisin folding. As the pro-domain is stabilized, the folding reaction becomes faster and distinctly biphasic. A detailed mechanism was determined for the double mutant, Q40L-K57E, which is 50% folded: P + Su if (30 800 M-1 s-1, 0.04 s-1) PSI if (0.07 s-1, <0.005 s-1) PS. PSI is an intermediate complex which accumulates in the course of the reaction, and PS is the fully folded complex. The more stable the pro-domain, the faster the folding reaction up to the point at which the isomerization of the intermediate into the fully folded complex becomes the rate-limiting step in the folding process.

Engineering the independent folding of the subtilisin BPN' prodomain: analysis of two-state folding versus protein stability.

In complex with subtilisin BPN', the 77 amino acid prodomain folds into a stable compact structure comprising a four-stranded antiparallel beta-sheet and two three-turn alpha-helices. When isolated from subtilisin, the prodomain is 97% unfolded even under optimal folding conditions. Traditionally, to study stable proteins, denaturing cosolvents or temperatures are used to shift the equilibrium from folded to unfolded. Here we manipulate the folding equilibrium of the unstable prodomain by introducing stabilizing mutations generated by design. By sequentially introducing three stabilizing mutations into the prodomain we are able to shift the equilibrium for independent folding from 97% unfolded to 65% folded. Spectroscopic and thermodynamic analysis of the folding reaction was carried out to assess the effect of stability on two-state behavior and the denatured state. The denatured states of single and combination mutants are not discernably different in spite of a range of DeltaGunfolding from -2.1 to 0.4 kcal/mol. Conclusions about the nature of the denatured state of the prodomain are based on CD spectral data and calorimetric data. Two state folding is observed for a combination mutant of marginal stability (DeltaG = 0). Evidence for its two-state folding is based on the observed additivity of individual mutations to the overall DeltaGunfolding and the conformity of DeltaGunfolding vs T to two-state assumptions as embodied in the Gibbs-Helmholz equation. We believe our success in stabilizing the two-state folding reaction of the prodomain originates from the selection of mutations with improved ability to fold subtilisin rather than selection for increase in secondary structure content. The fact that a small number of mutations can stabilize the independent folding of the prodomain implies that most of the folding information already exists in the wild-type amino acid sequence in spite of the fact that the unfolded state predominates.

Generation of soluble and active subtilisin and alpha-chymotrypsin in organic solvents via hydrophobic ion pairing.

With very low concentrations of anionic detergents, such as sodium dodecyl sulfate (SDS) and Aerosol OT (AOT), it is possible to solubilize proteases in organic solvents, while retaining enzymatic activity. For example, the SDS-subtilisin BPN' complex catalyzes transesterification of Ac-Phe-OMe in ethanol with a kcat/Km of 36 M-1 s-1 for mutant M1 and 39 M-1 s-1 for the wild type. By comparison, M1 suspended in ethanol is approximately 1000-fold less active, with a kcat/Km of 0.03 M-1 s-1. Similarly, AOT complexes of alpha-chymotrypsin were found to be approximately 1000 times more active (kcat/Km = 100-350 M-1 s-1) than the suspended enzyme.

Prodomain mutations at the subtilisin interface: correlation of binding energy and the rate of catalyzed folding.

The in vivo folding of subtilisin is dependent on a 77 amino acid prosequence, which is eventually cleaved from the N-terminus of subtilisin to create the 275 amino acid mature form of the enzyme. The recent determination of the structure of a complex of the prodomain and a calcium-free subtilisin mutant has suggested how the prodomain may catalyze subtilisin folding [Bryan, P., Wang, L., Hoskins, J., Ruvinov, S., Strausberg, S., Alexander, P., Almog, O., Gilliland, G., & Gallagher, T. (1995) Biochemistry 34, 10310-10318]. In the complex, the prodomain packs against the two parallel surface helices of subtilisin (residues 104-116 and residues 133-144) and supplies caps to the N-termini of the two helices. The binding site is contained almost entirely in the linear sequence 100-144 of subtilisin. The C-terminus of the prodomain (residues 72-77) extends out from its central part to bind like a substrate in subtilisin's active site cleft. The simplest model of catalyzed folding is one in which the observed binding interaction in the complex accelerates folding by stabilizing an intermediate which includes the 45 amino acid alpha beta alpha substructure in subtilisin. According to our hypothesis, amino acids 100-144 would have a native-like fold in the intermediate which the prodomain stabilizes. Guided by the structure of the bimolecular complex of subtilisin and its prodomain, we have constructed mutations in the C-terminal region of the prodomain. Analysis of five mutants reveals a general correlation between the ability of the prodomain to bind to native subtilisin and its ability to accelerate subtilisin folding.(ABSTRACT TRUNCATED AT 250 WORDS)

Directed evolution of a subtilisin with calcium-independent stability.

Extracellular proteases of the subtilisin-class depend upon calcium for stability. Calcium binding stabilizes these proteins in natural extracellular environments, but is an Achilles' heel in industrial environments which contain high concentrations of metal chelators. Here we direct the evolution of calcium-independent stability in subtilisin BPN'. By deleting the calcium binding loop from subtilisin, we initially destabilize the protein but create the potential to use new structural solutions for stabilization. Analysis of the structure and stability of the loop-deleted prototype followed by directed mutagenesis and selection for increased stability resulted in a subtilisin mutant with native-like proteolytic activity but 1000-times greater stability in strongly chelating conditions.

Large increases in general stability for subtilisin BPN' through incremental changes in the free energy of unfolding.

Six individual amino acid substitutions at separate positions in the tertiary structure of subtilisin BPN' (EC 3.4.21.14) were found to increase the stability of this enzyme, as judged by differential scanning calorimetry and decreased rates of thermal inactivation. These stabilizing changes, N218S, G169A, Y217K, M50F, Q206C, and N76D, were discovered through the use of five different investigative approaches: (1) random mutagenesis; (2) design of buried hydrophobic side groups; (3) design of electrostatic interactions at Ca2+ binding sites; (4) sequence homology consensus; and (5) serendipity. Individually, the six amino acid substitutions increase the delta G of unfolding between 0.3 and 1.3 kcal/mol at 58.5 degrees C. The combination of these six individual stabilizing mutations together into one subtilisin BPN' molecule was found to result in approximately independent and additive increases in the delta G of unfolding to give a net increase of 3.8 kcal/mol (58.5 degrees C). Thermodynamic stability was also shown to be related to resistance to irreversible inactivation, which included elevated temperatures (65 degrees C) or extreme alkalinity (pH 12.0). Under these denaturing conditions, the rate of inactivation of the combination variant is approximately 300 times slower than that of the wild-type subtilisin BPN'. A comparison of the 1.8-A-resolution crystal structures of mutant and wild-type enzymes revealed only independent and localized structural changes around the site of the amino acid side group substitutions.(ABSTRACT TRUNCATED AT 250 WORDS)

The engineering of binding affinity at metal ion binding sites for the stabilization of proteins: subtilisin as a test case.

A weak Ca2+ binding site in the bacterial serine protease subtilisin BPN' (EC 3.4.21.14) was chosen as a model to explore the feasibility of stabilizing a protein by increasing the binding affinity at a metal ion binding site. The existence of this weak Ca2+ binding site was first discovered through a study of the rate of thermal inactivation of wild-type subtilisin BPN' at 65 degrees C as a function of the free [Ca2+]. Increasing the [Ca2+] in the range 0.10-100 mM caused a 100-fold decrease in the rate of thermal inactivation. The data were found to closely fit a theoretical titration curve for a single Ca2+ specific binding site with an apparent log Ka = 1.49. A series of refined X-ray crystal structures (R less than or equal to 0.15, 1.7 A) of subtilisin in the presence of 0.0, 25.0, and 40.0 mM CaCl2 has allowed a detailed structural characterization of this Ca2+ binding site. Negatively charged side chains were introduced in the vicinity of the bound Ca2+ by changing Pro 172 and Gly 131 to Asp residues through site-directed and random mutagenesis techniques, respectively. These changes were found to increase the affinity of the Ca2+ binding site by 3.4- and 2-fold, respectively, when compared with the wild-type protein (ionic strength = 0.10). X-ray studies of these new variants of subtilisin revealed the carboxylate side chains to be 6.8 and 13.2 A, respectively, from the bound Ca2+. These distances and the degree of enhanced binding are consistent with simple electrostatic theory.(ABSTRACT TRUNCATED AT 250 WORDS)

Engineering thermostability in subtilisin BPN' by in vitro mutagenesis.

A procedure has been developed for the isolation and identification of mutants of the bacterial serine protease, subtilisin, which exhibit enhanced thermostability. The cloned subtilisin BPN' gene from Bacillus amyloliquefaciens was treated with a variety of chemical mutagens to introduce random mutations in the coding sequence. Strains containing the cloned, mutagenized subtilisin gene which produced subtilisin with enhanced thermostability were selected by a simple plate assay procedure, which screens for esterase activity on nitrocellulose filters after preincubation at elevated temperatures. The identification and characterization of eight different stabilizing mutations are described. Several mutants containing various combinations of these stabilizing mutations were constructed by oligonucleotide-directed mutagenesis. Combining independent, stabilizing mutations in the same subtilisin molecule has resulted in an approximate multiplicative decrease in the rate of thermal inactivation. In this way, a variant of subtilisin has been constructed which is about 12-fold more stable than wild-type subtilisin, with no radical changes in the tertiary protein structure but rather minor, independent alterations in amino acid sequence. The ultimate goal in these studies is to be able to accurately predict where stabilizing changes can be made in a protein.

Protein engineering of subtilisin BPN': enhanced stabilization through the introduction of two cysteines to form a disulfide bond.

Introduction of a disulfide bond by site-directed mutagenesis was found to enhance the stability of subtilisin BPN' (EC 3.4.21.14) under a variety of conditions. The location of the new disulfide bond was selected with the aid of a computer program, which scored various sites according to the amount of distortion that an introduced disulfide linkage would create in a 1.3-A X-ray model of native subtilisin BPN'. Of the several amino acid pairs identified by this program as suitable candidates, Thr-22 and Ser-87 were selected by using the additional requirement that the individual cysteine substitutions occur at positions that exhibit some degree of variability in related subtilisin amino acid sequences. A subtilisin variant containing cysteine residues at positions 22 and 87 was created by site-directed mutagenesis and was shown to have an activity essentially equivalent to that of the wild-type enzyme. Differential scanning calorimetry experiments demonstrated the variant protein to have a melting temperature 3.1 degrees C higher than that of the wild-type protein and 5.8 degrees C higher than that of the reduced form (-SH HS-) of the variant protein. Kinetic experiments performed under a variety of conditions, including 8 M urea, showed that the Cys-22/Cys-87 disulfide variant undergoes thermal inactivation at half the rate of that of the wild-type enzyme. The increased thermal stability of this disulfide variant is consistent with a decrease in entropy for the unfolded state relative to the unfolded state that contains no cross-link, as would be predicted from the statistical thermodynamics of polymers.

Protein engineering.

The techniques of protein engineering are proving to be a revolutionary experimental tool for understanding protein structure-function relationships. Even at this early stage, proteins of improved characteristics for specific industrial and therapeutic uses have already been produced. Tailoring enzymatic properties for non-physiological substrate conditions, altering pH optima, changing substrate specificity, and improving stability have already been demonstrated to be feasible. Nevertheless, the ability to make useful proteins which radically differ from a natural structure or designing altogether new structures exceeds present understanding.

Proteases of enhanced stability: characterization of a thermostable variant of subtilisin.

A procedure has been developed for the isolation and identification of mutants in the bacterial serine protease subtilisin that exhibit enhanced thermal stability. The cloned subtilisin BPN' gene from Bacillus amyloliquefaciens was treated with bisulfite, a chemical mutagen that deaminates cytosine to uracil in single-stranded DNA. Strains containing the cloned, mutagenized subtilisin gene which produced subtilisin with enhanced thermal stability were selected by a simple plate assay procedure which screens for esterase activity on nitrocellulose filters after preincubation at elevated temperatures. One thermostable subtilisin variant, designated 7150, has been fully characterized and found to differ from wild-type subtilisin by a single substitution of Ser for Asn at position 218. The 7150 enzyme was found to undergo thermal inactivation at one-fourth the rate of the wild-type enzyme when incubated at elevated temperatures. Moreover, the mid-point in the thermally induced transition from the folded to unfolded state was found to be 2.4-3.9 degrees C higher for 7150 as determined by differential scanning calorimetry under a variety of conditions. The refined, 1.8-A crystal structures of the wild-type and 7150 subtilisin have been compared in detail, leading to the conclusion that slight improvements in hydrogen bond parameters in the vicinity of position 218 result in the enhanced thermal stability of 7150.

Enhancer sequences responsible for DNase I hypersensitivity in polyomavirus chromatin.

DNase I preferentially cleaves polyomavirus minichromosomes at two sites in the enhancer, each of which comprises the sequence AAGCAPuPuAAG flanked by short inverted repeats. A tandem duplication of this sequence generates an additional hypersensitive locus. Mutations which alter either the AAGCAPuPuAAG or flanking repeats diminish hypersensitivity. This region must determine the chromatin conformation recognized by DNase I.