The determinants of alpha-amylase pH-activity profiles.

The glycosyl hydrolases present a large family of enzymes that are of great significance for industry. Consequently, there is considerable interest in engineering the enzymes in this family for optimal performance under a range of very diverse conditions. Until recently, tailoring glycosyl hydrolases for specific industrial processes mainly involved stability engineering, but lately there has also been considerable interest in engineering their pH-activity profiles. We mutated four neutral residues (N190, F290, N326 and Q360) in the chimeric Bacillus Ba2 alpha-amylase to both charged and neutral amino acids. The results show that the pH-activity profile of the Ba2 alpha-amylase can be changed by inserting charged residues close to the active site. The changes in the pH-activity profile for these neutral --> charged mutations do not, however, correlate with the predictions from calculations of the p K(a) values of the active site residues. More surprisingly, the neutral --> neutral mutations change the pH-activity profile as much as the neutral --> charged mutations. From these results, it is concluded that factors other than electrostatics, presumably the dynamic aspects of the active site, are important for the shape of the pH-activity profiles of the alpha-amylases.

In vitro selection of enzymatically active lipase variants from phage libraries using a mechanism-based inhibitor.

The 'detergent lipase' Lipolase, from Thermomyces lanuginosa was subjected to a combinatorial protein engineering/phage display approach with the aim of identifying new enzyme variants with improved characteristics in the presence of detergents. First it was demonstrated that wild-type Lipolase could be produced in Escherichia coli retaining full activity and be displayed as an active enzyme fused to coat protein 3 on E. coli phage M13. A phagemid library designed to result in approximately two to three mutations per lipase gene was then constructed. Nine amino acids located in two regions close to the active site were targeted for randomization. Selections using a mechanism-based biotinylated inhibitor showed that phages displaying Lipolase could be specifically enriched from a population of control phages. Selections on a library phage stock in the presence of inhibitor and a commercial powder detergent resulted in a step-wise increase in the proportion of active clones. Analysis of 84 active clones revealed that they all expressed lipase activity, but with lower activities than that of a wild-type Lipolase-producing clone. In six of the seven most active clones a wild-type serine at position 83 had been replaced by threonine, a substitution known to alter the substrate chain length preference of Lipolase variants. Furthermore, the selection had enriched enzyme variants with a high degree of conservatism in one of the variegated regions, suggesting that this region is important for enzymatic activity and that the designed selection procedure was relevant. The selected variants contained primarily basic amino acid residues within the other variegated region. Taken together, the described results show that selection protocols based on enzymatic activity can be designed for this enzyme class which should be of importance for future protein engineering attempts.

Efficient gene targeted random mutagenesis in genetically stable Escherichia coli strains.

We describe a method to generate in vivo collections of mutants orders of magnitude larger than previously possible. The method favors accumulation of mutations in the target gene, rather than in the host chromosome. This is achieved by propagating the target gene on a plasmid, in Escherichia coli cells, within the region preferentially replicated by DNA polymerase I (Pol I), which replicates only a minor fraction of the chromosome. Mutagenesis is enhanced by a conjunction of a Pol I variant that has a low replication fidelity and the absence of the mutHLS system that corrects replication errors. The method was tested with two reporter genes, encoding lactose repressor or lipase. The proportion of mutants in the collection was estimated to reach 1% after one cycle of growth and 10% upon prolonged cell cultivation, resulting in collections of 10(12)-10(13) mutants per liter of cell culture. The extended cultivation did not affect growth properties of the cells. We suggest that our method is well suited for generating protein variants too rare to be present in the collections established by methods used previously and for isolating the genes that encode such variants by submitting the cells of the collections to appropriate selection protocols.

Design and synthesis of triglyceride analogue biotinylated suicide inhibitors for directed molecular evolution of lipolytic enzymes.

The design, synthesis, and inhibition properties of two new triglyceride analogue biotinylated suicide inhibitors (2) and (3) for directed molecular evolution of lipolytic enzymes by phage-display is described.

Structural analysis of a chimeric bacterial alpha-amylase. High-resolution analysis of native and ligand complexes.

Several chimeric alpha-amylases genes were constructed by an in vivo recombination technique from the Bacillus amyloliquefaciens and Bacillus licheniformis genes. One of the fusion amylases (hereafter BA2), consisting of residues 1-300 from B. amyloliquefaciens and 301-483 from B. licheniformis, has been extensively studied by X-ray crystallography at resolutions between 2.2 and 1.7 A. The 3-dimensional structure of the native enzyme was solved by multiple isomorphous replacement, and refined at a resolution of 1.7 A. It consists of 483 amino acids, organized similarly to the known B. lichiniformis alpha-amylase structure [Machius et al. (1995) J. Mol. Biol. 246, 545-559], but features 4 bound calcium ions. Two of these form part of a linear cluster of three ions, the central ion being attributed to sodium. This cluster lies at the junction of the A and B domains with one calcium of the cluster structurally equivalent to the major Ca(2+) binding site of fungal alpha-amylases. The third calcium ion is found at the interface of the A and C domains. BA2 contains a fourth calcium site, not observed in the B. licheniformis alpha-amylase structure. It is found on the C domain where it bridges the two beta-sheets. Three acid residues (Glu261, Asp328, and Asp231) form an active site similar to that seen in other amylases. In the presence of TRIS buffer, a single molecule of TRIS occupies the -1 subsite of the enzyme where it is coordinated by the three active-center carboxylates. Kinetic data reveal that BA2 displays properties intermediate to those of its parents. Data for crystals soaked in maltooligosaccharides reveal the presence of a maltotriose binding site on the N-terminal face of the (beta/alpha)(8) barrel of the molecule, not previously described for any alpha-amylase structure, the biological function of which is unclear. Data for a complex soaked with the tetrasaccharide inhibitor acarbose, at 1.9 A, reveal a decasaccharide moiety, spanning the -7 to +3 subsites of the enzyme. The unambiguous presence of three unsaturated rings in the (2)H(3) half-chair/(2)E envelope conformation, adjacent to three 6-deoxypyranose units, clearly demonstrates synthesis of this acarbose-derived decasaccharide by a two-step transglycosylation mechanism.

Conversion of the maltogenic alpha-amylase Novamyl into a CGTase.

Novamyl is a thermostable five-domain maltogenic alpha-amylase that shows sequence and structural homology with the cyclodextrin glycosyltransferases (CGTases). Comparing X-ray crystal structures of Novamyl and CGTases, two major differences in the active site cleft were observed: Novamyl contains a loop insertion consisting of five residues (residues 191-195) and the location of an aromatic residue known to be essential to obtain an efficient cyclization reaction. To convert Novamyl into a cyclodextrin (CD)-producing enzyme, the loop was deleted and two substitutions, F188L and T189Y, were introduced. Unlike the parent Novamyl, the obtained variant is able to produce beta-CD and showed an overall conversion of starch to CD of 9%, compared with CGTases which are able to convert up to 40%. The lower conversion compared with the CGTase is probably due to additional differences in the active site cleft and in the starch-binding E domain. A variant with only the five-residue loop deleted was not able to form beta-CD.

A novel biotinylated suicide inhibitor for directed molecular evolution of lipolytic enzymes.

A bifunctional activity label (8) for directed molecular evolution of lipolytic enzymes has been designed and synthesized. The structure is composed of a 4-nitrophenyl activated phosphonate, that is, a suicide substrate of lipases/esterases, connected to a biotin moiety through a spacer containing a disulfide bridge. The phosphonate (3) was prepared by Michaelis-Arbuzov reaction of trimethylsilyl-protected 11-bromoundecanol (2) with triethyl phosphite. The deprotected omega-hydroxyalkylphosphonate (4) was transformed into an active N-hydroxysuccinimide carbonate (5) followed by 4-nitrophenyl activation of the phosphonate using standard procedures. The biotinylated phosphonate inhibitor (8) was then synthesised by coupling the phosphonate inhibitor (6) to the epsilon-amino-caproic acid and cystamine containing biotinyl spacer (7). The function of all relevant groups of the final activity label (8) (biotin-label, cleavable disulfide bridge, phosphonate-inhibitor) have been successfully tested with the commercial lipase Lipolase (Novo Nordisk). Hence, a tool for directed molecular evolution of lipolytic enzymes has been developed.

Display of active subtilisin 309 on phage: analysis of parameters influencing the selection of subtilisin variants with changed substrate specificity from libraries using phosphonylating inhibitors.

Many attempts have been made to endow enzymes with new catalytic activities. One general strategy involves the creation of random combinatorial libraries of mutants associated with an efficient screening or selection scheme. Phage display has been shown to greatly facilitate the selection of polypeptides with desired properties by establishing a close link between the polypeptide and the gene that encodes it. Selection of phage displayed enzymes for new catalytic activities remains a challenge. The aim of this study was to display the serine protease subtilisin 309 (savinase) from Bacillus lentus on the surface of filamentous fd phage and to develop selection schemes that allow the extraction of subtilisin variants with a changed substrate specificity from libraries. Subtilisins are produced as secreted preproenzyme that mature in active enzyme autocatalytically. They have a broad substrate specificity but exhibit a significant preference for hydrophobic residues and very limited reactivity toward charged residues at the P4 site in the substrate. Here, we show that savinase can be functionally displayed on phage in the presence of the proteic inhibitor CI2. The free enzyme is released from its complex with CI2 upon addition of the anionic detergent LAS. The phage-enzyme can be panned on streptavidin beads after labelling by reaction with (biotin-N-epsilon-aminocaproyl-cystamine-N'-glutaryl)-l-Ala-l-Ala-l-P ro-Phe(P)-diphenyl ester. Reactions of libraries, in which residues 104 and 107 forming part of the S4 pocket have been randomised, with (biotin-N-epsilon-aminocaproyl-cystamine-N'-glutaryl)-alpha-l-Lys-l-A la-l-Pro-Phe(P)-diphenylester allowed us to select enzymes with increased specific activity for a substrate containing a lysine in P4. Parameters influencing the selection as for instance the efficiency of maturation of mutant enzymes in libraries have been investigated.

Protein engineering of bacterial alpha-amylases.

alpha-Amylases constitute a very diverse family of glycosyl hydrolases that cleave alpha1-->4 linkages in amylose and related polymers. Recent structural and mutagenic studies of archeael, mammalian and bacterial alpha-amylases have resulted in a wealth of information on the catalytic mechanism and on the structural features of this enzyme class. Because of their high thermo-stability, the Bacillus alpha-amylases have found widespread use in industrial processes, and much attention has been devoted to optimising these enzymes for the very harsh conditions encountered there. Stability has been a major area of focus in this respect, and several remarkably stable bacterial alpha-amylases have been produced by bioengineering techniques. Protein engineering studies of pH-activity profiles and of substrate specificities have also been initiated, although without much success. In the coming years it is likely, however, that the focus of alpha-amylase engineering will shift from engineering stability to these new areas.

Electrostatics in the active site of an alpha-amylase.

The importance of electrostatics in catalysis has been emphasized in the literature for a large number of enzymes. We examined this hypothesis for the Bacillus licheniformis alpha-amylase by constructing site-directed mutants that were predicted to change the pKa values of the catalytic residues and thus change the pH-activity profile of the enzyme. To change the pKa of the catalytic residues in the active site, we constructed mutations that altered the hydrogen bonding network, mutations that changed the solvent accessibility, and mutations that altered the net charge of the molecule. The results show that changing the hydrogen bonding network near an active site residue or changing the solvent accessibility of an active site residue will very likely result in an enzyme with drastically reduced activity. The differences in the pH-activity profiles for these mutants were modest. pH-activity profiles of mutants which change the net charge on the molecule were significantly different from the wild-type pH-activity profile. The differences were, however, difficult to correlate with the electrostatic field changes calculated. In several cases we observed that pH-activity profiles shifted in the opposite direction compared to the shift predicted from electrostatic calculations. This strongly suggests that electrostatic effects cannot be solely responsible for the pH-activity profile of the B. licheniformis alpha-amylase.

DNA shuffling of subgenomic sequences of subtilisin.

DNA family shuffling of 26 protease genes was used to create a library of chimeric proteases that was screened for four distinct enzymatic properties. Multiple clones were identified that were significantly improved over any of the parental enzymes for each individual property. Family shuffling, also known as molecular breeding, efficiently created all of the combinations of parental properties, producing a great diversity of property combinations in the progeny enzymes. Thus, molecular breeding, like classical breeding, is a powerful tool for recombining existing diversity to tailor biological systems for multiple functional parameters.

X-ray structure of Novamyl, the five-domain "maltogenic" alpha-amylase from Bacillus stearothermophilus: maltose and acarbose complexes at 1.7A resolution.

The three-dimensional structure of the Bacillus stearothermophilus "maltogenic" alpha-amylase, Novamyl, has been determined by X-ray crystallography at a resolution of 1.7 A. Unlike conventional alpha-amylases from glycoside hydrolase family 13, Novamyl exhibits the five-domain structure more usually associated with cyclodextrin glycosyltransferase. Complexes of the enzyme with both maltose and the inhibitor acarbose have been characterized. In the maltose complex, two molecules of maltose are found in the -1 to -2 and +2 to +3 subsites of the active site, with two more on the C and E domains. The C-domain maltose occupies a position identical to one previously observed in the Bacillus circulans CGTase structure [Lawson, C. L., et al. (1994) J. Mol. Biol. 236, 590-600], suggesting that the C-domain plays a genuine biological role in saccharide binding. In the acarbose-maltose complex, the tetrasaccharide inhibitor acarbose is found as an extended hexasaccharide species, bound in the -3 to +3 subsites. The transition state mimicking pseudosaccharide is bound in the -1 subsite of the enzyme in a 2H3 half-chair conformation, as expected. The active site of Novamyl lies in an open gully, fully consistent with its ability to perform internal cleavage via an endo as opposed to an exo activity.

Oligosaccharide specificity of a family 7 endoglucanase: insertion of potential sugar-binding subsites.

Family 7 of the glycosyl hydrolases contains both endoglucanases and cellobiohydrolases. In addition to their different catalytic activities on crystalline substrates, the cellobiohydrolases differ from the endoglucanases in their activity on longer soluble substrates, indicative of a greater number of subsites on the enzyme. A double mutant (S37W, P39W) of the Humicola insolens endoglucanase I (EG I) has been constructed in order to mimic aspects of the subsite structure of the corresponding family 7 cellobiohydrolase, cellobiohydrolase-I (CBH I). The 3-D crystal structure of the double mutant has been solved and refined to a crystallographic R-factor of 0.17 at a resolution of 2.2 A (1 A = 0.1 nm). The two mutant tryptophans are clearly visible in the electron density and are in the same orientation as those found in the substrate binding groove of CBH I. In addition to the substitutions, the C-terminal amino acids (399QELQ), disordered in the native enzyme structure, are clearly visible and there are a small number of minor loop movements associated with differences in crystal packing. Kinetic determinations show that the S37W, P39W mutant EG I has almost identical activity, compared to native EG I, on small soluble cellodextrins. On phosphoric acid swollen cellulose there is a small (30%), but significant, decrease in the apparent KM indicating that the double mutant may indeed exhibit stronger binding to longer polymeric substrates.

An interface point-mutation variant of triosephosphate isomerase is compactly folded and monomeric at low protein concentrations.

Wild-type trypanosomal triosephosphate isomerase (wtTIM) is a very tight dimer. The interface residue His-47 of wtTIM has been mutated into an asparagine. Ultracentrifugation data show that this variant (H47N) only dimerises at protein concentrations above 3 mg/ml. H47N has been characterised at a protein concentration where it is predominantly a monomer. Circular dichroism measurements in the near-UV and far-UV show that this monomer is a compactly folded protein with secondary structure similar as in wtTIM. The thermal stability of the monomeric H47N is decreased compared to wtTIM: temperature gradient gel electrophoresis (TGGE) measurements give Tm-values of 41 degrees C for wtTIM, whereas the Tm-value for the monomeric form of H47N is approximately 7 degrees C lower.

Identification of a Bacillus subtilis secretion mutant using a beta-galactosidase screening procedure.

High-level synthesis of exportable beta-galactosidase (LacZ) fusion proteins in Bacillus subtilis results in a lethal phenotype, and has been suggested as a tool for the selection of secretion mutants. We tested a plasmid-based, inducible lacZ fusion gene system for this purpose, but frequent mutations in cis, which reduced expression of the fusion gene, forced abandonment of the induction-selection strategy. Instead, after modification of the indicator plasmid, a screening procedure for increased basal LacZ activity levels was adopted. This led to the identification of a conditional B. subtilis secretion mutant after nitrosoguanidine mutagenesis. At 42 degrees C, but not at 30 degrees C, this mutant displayed extreme growth retardation when the LacZ fusion protein was produced, and and was also defective in the secretion of subtilisin Carlsberg. The processing kinetics and secretion of a subtilisin Carlsberg-alkaline phosphatase fusion derivative were found to be defective specifically at the non-permissive temperature. The secretion defect was not linked to the secA/div locus.

Three new crystal structures of point mutation variants of monoTIM: conformational flexibility of loop-1, loop-4 and loop-8.

BACKGROUND: Wild-type triosephosphate isomerase (TIM) is a very stable dimeric enzyme. This dimer can be converted into a stable monomeric protein (monoTIM) by replacing the 15-residue interface loop (loop-3) by a shorter, 8-residue, loop. The crystal structure of monoTIM shows that two active-site loops (loop-1 and loop-4), which are at the dimer interface in wild-type TIM, have acquired rather different structural properties. Nevertheless, monoTIM has residual catalytic activity. RESULTS: Three new structures of variants of monoTIM are presented, a double-point mutant crystallized in the presence and absence of bound inhibitor, and a single-point mutant in the presence of a different inhibitor. These new structures show large structural variability for the active-site loops, loop-1, loop-4 and loop-8. In the structures with inhibitor bound, the catalytic lysine (Lys13 in loop-1) and the catalytic histidine (His95 in loop-4) adopt conformations similar to those observed in wild-type TIM, but very different from the monoTIM structure. CONCLUSIONS: The residual catalytic activity of monoTIM can now be rationalized. In the presence of substrate analogues the active-site loops, loop-1, loop-4 and loop-8, as well as the catalytic residues, adopt conformations similar to those seen in the wild-type protein. These loops lack conformational flexibility in wild-type TIM. The data suggest that the rigidity of these loops in wild-type TIM, resulting from subunit-subunit contacts at the dimer interface, is important for optimal catalysis.

Modular mutagenesis of a TIM-barrel enzyme: the crystal structure of a chimeric E. coli TIM having the eighth beta alpha-unit replaced by the equivalent unit of chicken TIM.

The crystal structure of a hybrid Escherichia coli triosephosphate isomerase (TIM) has been determined at 2.8 A resolution. The hybrid TIM (ETIM8CHI) was constructed by replacing the eighth beta alpha-unit of E. coli TIM with the equivalent unit of chicken TIM. This replacement involves 10 sequence changes. One of the changes concerns the mutation of a buried alanine (Ala232 in strand 8) into a phenylalanine. The ETIM8CHI structure shows that the A232F sequence change can be incorporated by a side-chain rotation of Phe224 (in helix 7). No cavities or strained dihedrals are observed in ETIM8CHI in the region near position 232, which is in agreement with the observation that ETIM8CHI and E.coli TIM have similar stabilities. The largest CA (C-alpha atom) movements, approximately 3 A, are seen for the C-terminal end of helix 8 (associated with the outward rotation of Phe224) and for the residues in the loop after helix 1 (associated with sequence changes in helix 8). From the structure it is not clear why the kcat of ETIM8CHI is 10 times lower than in wild type E.coli TIM.

Comparison of the structures and the crystal contacts of trypanosomal triosephosphate isomerase in four different crystal forms.

Triosephosphate isomerase (TIM) is a dimeric enzyme consisting of 2 identical subunits. Trypanosomal TIM can be crystallized in 4 different spacegroups: P2(1)2(1)2(1), C2(big cell), C2(small cell), and P1. The P1 crystal form only grows in the presence of 1.4 M DMSO; there are 2 DMSO binding sites per subunit. The structures have been refined at a resolution of 1.83 A, 2.10 A, 2.13 A, and 1.80 A, respectively. In the 4 different spacegroups the TIM subunit can be observed in the context of 7 different crystallographic environments. In the C2 cells, the dimer 2-fold axis coincides with a crystallographic 2-fold axis. The similarities and differences of the 7 subunits are discussed. In 6 subunits the flexible loop (loop 6) is open, whereas in the P2(1)2(1)2(1) cell, the flexible loop of subunit 2 is in an almost closed conformation. The crystal contacts in the 4 different crystal forms are predominantly generated by polar residues in loops. A statistical analysis of the residues involved in crystal contacts shows that, in particular, serines are frequently involved in these interactions; 19% of the exposed serines are involved in crystal contacts.

The crystal structure of human CskSH3: structural diversity near the RT-Src and n-Src loop.

SH3 domains are modules occurring in diverse proteins, ranging from cytoskeletal proteins to signaling proteins, such as tyrosine kinases. The crystal structure of the SH3 domain of Csk (c-Src specific tyrosine kinase) has been refined at a resolution of 2.5 A, with an R-factor of 22.4%. The structure is very similar to the FynSH3 crystal structure. When comparing CskSH3 and FynSH3 it is seen that the structural and charge differences of the RT-Src loop and the n-Src loop, near the conserved Trp47, correlate with different binding properties of these SH3 domains. The structure comparison suggests that those glycines and acid residues which are very well conserved in the SH3 sequences are important for the stability of the SH3 fold.

Design, creation, and characterization of a stable, monomeric triosephosphate isomerase.

Protein engineering on trypanosomal triosephosphate isomerase (TIM) converted this oligomeric enzyme into a stable, monomeric protein that is enzymatically active. Wild-type TIM consists of two identical subunits that form a very tight dimer involving interactions of 32 residues of each subunit. By replacing 15 residues of the major interface loop by another 8-residue fragment, a variant was constructed that is a stable and monomeric protein with TIM activity. The length, sequence, and conformation of the designed fragment were suggested by extensive modeling.