15 Subtilisin: X-Ray Structure

Prediction of protein domain boundaries from inverse covariances

It has been known even since relatively few structures had been solved that longer protein chains often contain multiple domains, which may fold separately and play the role of reusable functional modules found in many contexts. In many structural biology tasks, in particular structure prediction, it is of great use to be able to identify domains within the structure and analyze these regions separately. However, when using sequence data alone this task has proven exceptionally difficult, with relatively little improvement over the naive method of choosing boundaries based on size distributions of observed domains. The recent significant improvement in contact prediction provides a new source of information for domain prediction. We test several methods for using this information including a kernel smoothing-based approach and methods based on building alpha-carbon models and compare performance with a length-based predictor, a homology search method and four published sequence-based predictors: DOMCUT, DomPRO, DLP-SVM, and SCOOBY-DOmain. We show that the kernel-smoothing method is significantly better than the other ab initio predictors when both single-domain and multidomain targets are considered and is not significantly different to the homology-based method. Considering only multidomain targets the kernel-smoothing method outperforms all of the published methods except DLP-SVM. The kernel smoothing method therefore represents a potentially useful improvement to ab initio domain prediction.

Stabilization of substilisin E in organic solvents by site-directed mutagenesis

Subtilisin E was rationally engineered to improve its stability in polar organic solvents such as dimethylformamide (DMF). A charged surface residue, Asp248, was substituted by three amino acids of increasing hydrophobicity, Asn, Ala, and Leu; all three variants were stabilized with respect to wild type in 80% DMF. This stabilization was only observed in the presence of high concentrations of the organic solvent: no stability enhancements were observed in 40% DMF. In contrast, the mutation Asn218 --> Ser alters internal hydrogen bonding interactions and stabilizes subtilisin E in both 40% and 80% DMF. This study provides additional evidence that substitution of surface-charged residues is a generally useful mechanism for stabilizing enzymes in organic media and that the stabilizing effects of such substitutions are unique to highly altered solvent environments. The effects of the single amino acid substitutions on free energies of stabilization are additive in the Asp248 --> Asn + Asn218 --> Ser combination variant, yielding an enzyme that is 3.4 times more stable than wild type in 80% DMF.

Enzyme engineering for nonaqueous solvents: random mutagenesis to enhance activity of subtilisin E in polar organic media

Enzyme activity is often dramatically reduced in polar organic solvents, even under conditions where the folded structures are stable. We have utilized random mutagenesis by polymerase chain reaction (PCR) techniques combined with screening for enhanced activity in the presence of dimethylformamide (DMF) to probe mechanisms by which improved enzymes for chemical synthesis in polar organic media might be obtained. Two amino acid substitutions which enhance subtilisin E activity in the presence of DMF, Q103R and D60N, were identified by screening on agar plates containing DMF and casein. The two substitutions are located near the substrate binding pocket or in the active site, and their effects on the catalytic efficiency kcat/KM for the hydrolysis of a peptide substrate are additive. The effects of D60N are apparent only in the presence of DMF, highlighting the importance of screening in the organic solvent. Protein engineering is an effective approach to enhancing enzyme activity in organic media: the triple mutant D60N + Q103R + N218S is 38 times more active than wild-type subtilisin E in 85% DMF. An evolutionary approach consisting of multiple steps of random mutagenesis and screening in continually higher concentrations of organic solvent should result in enzymes that are substantially more active in organic media.

Nucleotide sequence of the gene for aqualysin I (a thermophilic alkaline serine protease) of Thermus aquaticus YT-1 and characteristics of the deduced primary structure of the enzyme

Aqualysin I is an alkaline serine protease which is secreted into the culture medium by Thermus aquaticus YT-1, an extreme thermophile [Matsuzawa, H., Hamaoki, M. & Ohta, T. (1983) Agric. Biol. Chem. 47, 25-28]. The gene encoding aqualysin I was cloned into Escherichia coli using synthetic oligodeoxyribonucleotides as hybridization probes. The nucleotide sequence of the cloned DNA was determined. The primary structure of aqualysin I, deduced from the nucleotide sequence, agreed with the NH2-terminal sequence previously reported and the determined amino acid sequences, including the COOH-terminal sequence, of the tryptic peptides derived from aqualysin I. Aqualysin I comprised 281 amino acid residues and its molecular mass was determined to be 28,350. On alignment of the whole amino acid sequence, aqualysin I showed high sequence homology with the subtilisin-type serine proteases, and 43% identity with proteinase K, 37-39% with subtilisins and 34% with thermitase. Extremely high sequence identity was observed in the regions containing the active-site residues, corresponding to Asp32, His64 and Ser221 of subtilisin BPN'. The nucleotide sequence of the cloned DNA (1105 nucleotides) revealed that it contains the entire gene encoding aqualysin I and one open reading frame without a translational stop codon. Therefore, aqualysin I was considered to be produced as a large precursor, which contains a NH2-terminal portion, the protease and a COOH-terminal portion. The G + C content of the coding region for aqualysin I was 64.6%, which is lower than those of other Thermus genes (68-74%). The codon usage in the aqualysin I gene was rather random in comparison with that in other Thermus genes.

The high-resolution X-ray crystal structure of the complex formed between subtilisin Carlsberg and eglin c, an elastase inhibitor from the leech Hirudo medicinalis. Structural analysis, subtilisin structure and interface geometry

Triclinic crystals of the complex formed by eglin with subtilisin Carlsberg were analyzed by X-ray diffraction. The crystal and molecular structure of this complex was determined with data that extended to 0.12-nm resolution by a combination of Patterson search methods and isomorphous replacement techniques. Its structure was refined to a crystallographic R value of 0.178 (1.0-0.12 nm) using an energy-restraint least-squares procedure. The complete subtilisin molecule could be traced without ambiguity in the refined electron density. The eglin component, from which an amino-terminal segment is cleaved off, is only defined from Lys8I (i.e. the lysine residue 8 of the inhibitor) onwards. Per unit cell, 436 fixed solvent molecules and 2 calcium ions were located. In spite of 84 amino acid replacements and one deletion, subtilisin Carlsberg exhibits a very similar polypeptide fold to subtilisin BPN'. The root-mean-square deviations of all alpha-carbon atoms (excluding those at the deletion site) from models of subtilisin BPN' [Alden, R. A., Birktoft, J. J., Kraut, J., Robertus, J. D. & Wright, C. S. (1971) Biochem. Biophys. Res. Commun. 45, 337-344] and subtilisin Novo [Drenth, J., Hol, W. G. J., Jansonius, J. N. & Kockoek, R. (1972) Eur. J. Biochem. 25, 177-181] are 0.077 nm and 0.103 nm. Most of these deviations result from global shifts rather than changes of the local geometry. The single-residue deletion at position 56 affects only the surrounding conformation. Two sites of high electron density and close distances to surrounding oxygen ligands have been found in the Carlsberg enzyme which are probably occupied by calcium ions. Eglin consists of a twisted four-stranded beta-sheet flanked by an alpha-helix and by an exposed proteinase binding loop on opposite sides. Around the reactive site, Leu45I-Asp46I, this loop is mainly stabilized by electrostatic/hydrogen bond interactions with the side chains of two arginine residues which project from the hydrophobic core [Bode, W., Papamokos, E., Musil, D., Seemüller, W. & Fritz, H. (1986) EMBO J. 5, 813-818]. The reactive site loop conformation resembles that found in other 'small' proteinase inhibitors. The scissile peptide bond is not cleaved but its carbonyl group is slightly distorted from planar geometry. Most of the intermolecular contacts are contributed by the nine residues of the reactive-site loop Gly40I-Arg48I.(ABSTRACT TRUNCATED AT 400 WORDS)

Solid-phase active site inhibitors of proteinases

Phenylalanine chloromethyl ketone covalently attached to porous glass beads was synthesized to serve as a solid-phase active site directed inhibitor of chymotrypsin-like proteolytic enzymes. The solid-phase reagent inhibited 20 nmol of bovine chymotrypsin per gram of glass and covalently bound 30 nmol of protein per gram of glass. Sepharose-bound lysine chloromethyl ketones were synthesized to serve as inhibitors of trypsin-like enzymes. Sepharose-MethionylLysyl chloromethyl ketone inactivated and bound about 6.8 nmol of enzyme per ml of settled gel. In a preliminary experiment, a cyanogen bromide cleavage of the methionine residues showed that it should be possible to release all peptides but the peptide containing the active-site histidine. The immobilized trypsin was also reduced, carboxymethylated and digested with chymotrypsin. The potential of the solid-phase approach is in the isolation of a specific serine proteinase and in the sequence determination of residues surrounding the active-site histidine.