On the activity loss of hydrolases in organic solvents: I. Rapid loss of activity of a variety of enzymes and formulations in a range of organic solvents

Dehydrated enzyme powders have been used extensively as suspensions in organic solvents to catalyze synthetic reactions. Prolonged enzyme activity is necessary to make such applications commercially successful. However, it has recently become evident that the stability and thus activity of many enzymes is compromised in organic solvents. Herein we explore the stability of various hydrolases (i.e., lipases from Mucor meihei and Candida rugosa, -chymotrypsin, subtilisin Carlsberg, and pig-liver esterase) and various formulations (lyophilized powder, cross-linked enzyme crystals, poly(ethylene glycol)-enzyme conjugates) in different organic solvents. The results show a roughly exponential activity decrease for all enzymes and formulations studied after exposure to organic solvents. Inactivation was observed independent of the enzyme, formulation details, and the solvent. In addition, no relationship was found between the magnitude of inactivation and the value of initial activity. Thus, quite active formulations lost their activity as quickly as less active formulations. The estimated half-times (t_1/2) for all enzymes and preparations ranged from 1.8 h for subtilisin C. co-lyophilized with methyl-β-cyclodextrin to 61.6 h for the most stable poly(ethylene glycol)--chymotrypsin preparation. The data here presented indicates that the inactivation is likely not related to changes in enzyme structure and dynamics.     

Hydrogen/deuterium exchange study of subtilisin Carlsberg during prolonged exposure to organic solvents

It has been previously reported that prolonged exposure of an enzyme to organic solvents leads to substantial decrease of activity. This effect was found to be unrelated to the catalysts' structure or their possible aggregation in organic solvents, and up to the present day the cause for activity loss remains unclear. In the present work, the structural dynamics of the serine protease subtilisin Carlsberg (SC) have been investigated during prolonged exposure to two organic solvents by following hydrogen/deuterium (H/D) exchange of mobile protons. The enzyme, after lyophilization, was incubated in organic solvents at controlled deuteriated water activity for different times and the H/D exchange was allowed to take place. The amount of deuterium exchanged was evaluated by (2)H NMR, which in turn gave us a picture of the changing dynamics of our model enzyme during incubation and under different experimental conditions. Our results show that the flexibility of SC decreases during prolonged storage in 1,4-dioxane (Diox) and acetonitrile (ACN) as indicated by the observed 3- to 10-fold decrease in the apparent rate constants of exchange (k) of fast exchangeable protons (FEP) and slow exchangeable protons (SEP) in the protein. Our study also shows that SC is more flexible in ACN than in Diox (k 3-20 times higher in ACN for the FEP and SEP), suggesting that enzyme dynamics are affected by solvent physicochemical properties. Additionally, the enzyme dynamics are also affected by the method of preparation: decreased flexibility (k decreases 3- to 10-fold for FEP and SEP) is observed when the enzyme is chemically modified with poly ethylene glycol (PEGylated) or colyophilized with crown ethers. A possible relationship between activity, enantioselectivity (E), and structural dynamics is discussed, demonstrating that direct correlations, as have been attempted in the past, are hampered by the multi-variable nature and complexity of the system.     

Molecular dynamic study of subtilisin Carlsberg in aqueous and nonaqueous solvents

Using molecular dynamics simulations, we have obtained an important insight into the structural and dynamical changes exerted by a nonaqueous solvent on the serine protease subtilisin Carlsberg. Our findings show that the structural properties of the subtilisin–acetonitrile (MeCN) system were sensitive to the amount of water present at the protein surface. A decrease or lack of water promoted the enzyme–MeCN interaction, which increased structural changes of the enzyme primarily at the surface loops. This effect caused variations on the secondary and tertiary structure of the protein and induced the opening of a pathway for the solvent to the protein core. Also, disturbance of the oxyanion hole was observed due to changes in the orientation in the Asn-155 side chain. The disruption of the oxyanion hole and the changes of the tertiary structure should affect the optimal activity of the enzyme.     

Effect of crown ethers on structure, stability, activity, and enantioselectivity of subtilisin Carlsberg in organic solvents.

Colyophilization or codrying of subtilisin Carlsberg with the crown ethers 18-crown-6, 15-crown-5, and 12-crown-4 substantially improved enzyme activity in THF, acetonitrile, and 1,4-dioxane in the transesterification reactions of N-acetyl-L-phenylalanine ethylester and 1-propanol and that of (+/-)-1-phenylethanol and vinylbutyrate. The acceleration of the initial rate, V(0), ranged from less than 10-fold to more than 100-fold. All crown ethers activated subtilisin substantially, which excludes a specific macrocyclic effect from being responsible. The secondary structure of subtilisin was studied by Fourier-transform infrared (FTIR) spectroscopy. 18-Crown-6 and 15-crown-5 led to a more nativelike structure of subtilisin in the organic solvents employed when compared with that of the dehydrated enzyme obtained from buffer alone. However, the high level of activation with 12-crown-4 where this effect was not observed excluded overall structural preservation from being the primary cause of the observed enzyme activation. The conformational mobility of subtilisin was investigated by performing thermal denaturation experiments in 1,4-dioxane. Although only a small effect of temperature on subtilisin structure was observed for the samples prepared with or without 12-crown-4, both 18-crown-6 and 15-crown-5 caused the enzyme to denature at quite low temperatures (38 degrees C and 56 degrees C, respectively). No relationship between this property and V(0) was evident, but increased conformational mobility of the protein decreased its storage stability. The possibility of a "molecular imprinting" effect was also tested by removing 18-crown-6 from the subtilisin-18-crown-6 colyophilizate by washing. V(0) was only halved as a result of this procedure, an effect insignificant compared with the ca. 80-fold rate enhancement observed prior to washing in THF. This suggests that molecular imprinting is likely the primary cause of subtilisin activation by crown ethers, as recently suggested.     

Catalytic activity and conformation of chemically modified subtilisin Carlsberg in organic media

Subtilisin Carlsberg, an alkaline protease from Bacillus licheniformis, was modified with polyoxyethylene (PEG) or aerosol-OT (AOT), and the solubility, conformation, and catalytic activity of the modified subtilisins in some organic media were compared under the same conditions. The solubility of modified subtilisins depended on the solubility of the modifier. On the other hand, the conformational changes depended on the solubility, rather than the property, of the modifier. When the modified subtilisin was dissolved in water-miscible polar solvents such as dimethylsulfoxide, acetonitrile, and tetrahydrofuran, significant conformational changes occurred. When modified subtilisin was dissolved in water-immiscible organic solvents, such as isooctane and benzene, the solvent did not induce significant conformational changes. The catalytic activity in the transesterification reaction of the N-acetyl-L-phenylalanine ethylester of the modified subtilisin in organic solvents was higher than that of native subtilisin. The high activity of modified subtilisin was thought to be due to a homogeneous reaction by the dissolved enzymes.     

Solvent effects on the catalytic activity of subtilisin suspended in organic solvents

We studied a model transesterification reaction catalyzed by subtilisin Carlsberg suspended in toluene, n-hexane, diisopropyl ether, and mixtures of these solvents. To account for solvent effects due to differences in water partitioning between the enzyme and the bulk solvents, we measured water sorption isotherms for the enzyme in each solvent. We measured catalytic activity as a function of enzyme hydration and obtained bell-shaped curves with maxima at the same enzyme hydration in all the solvents. However, the activity maxima were different in all the media, being the lowest in toluene. Differences in the partitioning of substrates and product between the bulk solvent phase and the enzyme active site were accounted for but could not explain the lower catalytic activity observed in toluene. The fact that toluene is very similar to one of the substrates suggested the possibility of competitive inhibition by this solvent. We derived a model allowing for differences in solvation of the substrates, by using thermodynamic activities instead of concentrations, as well as for competitive inhibition by toluene. The model fit the experimental data well, confirming that toluene had a direct adverse effect on the catalytic activity of the enzyme. (c) 1996 John Wiley & Sons, Inc.     

Random mutagenesis to enhance the activity of subtilisin in organic solvents: Characterization of Q103R subtilisin E

Q103R subtilisin E was isolated following random mutagenesis and screening for improved activity in the presence of dimethylformamide (DMF). Our goal is to identify the mechanism(s) by which amino acid substitutions can enhance enzyme activity in polar organic solvents. A quantitative framework for comparing substrate binding and catalytic activities of mutant and wild-type enzymes in the presence and absence of DMF is outlined. Kinetic experiments performed at high salt concentration (1M KCl) reveal that the mechanism behind the Q103R variant's enhanced activity toward succinyl-Ala-Ala-Pro-Phe-p-nitroanilide is both electrostatic and nonelectrostatic in origin. Favorable electrostatic interactions between the negatively charged succinyl group of the substrate and the positive charge on Arg 103 are responsible for tighter substrate binding. This conclusion is supported by kinetic experiments performed on the related substrate Ala-Ala-Pro-Phe-p-nitroanilide and the hydrolysis kinetics of the Q103E, Q103K, and Q103S variants constructed by site-directed mutagenesis. These results highlight the importance of the choice of the substrate used to screen for improvements in catalytic activity.     

Enzyme activation in organic solvents: co-lyophilization of subtilisin Carlsberg with methyl-beta-cyclodextrin renders an enzyme catalyst more active than the cross-linked enzyme crystals

In this study we explored the efficiency of the additive methyl-beta-cyclodextrin (M beta CD) to enhance the activity and enantioselectivity of the serine protease subtilisin Carlsberg in organic solvents. These two parameters, measured for different transesterification reactions and in several solvents, are compared with results obtained by using two additional preparations of the same enzyme: lyophilized powder and cross-linked enzyme crystals (CLEC). The results suggest that co-lyophilization of subtilisin with M beta CD preserves the enzyme's active site tertiary structure rendering a highly active and enantioselective catalyst.     

Engineering subtilisin E for enhanced stability and activity in polar organic solvents

We examined the effect of a novel disulfide bond engineered in subtilisin E from Bacillus subtilis based on the structure of a thermophilic subtilisin-type serine protease aqualysin I. Four sites (Ser163/Ser194, Lys170/Ser194, Lys170/Glu195, and Pro172/Glu195) in subtilisin E were chosen as candidates for Cys substitutions by site-directed mutagenesis. The Cys170/Cys195 mutant subtilisin formed a disulfide bond in B. subtilis, and showed a 5-10-fold increase in specific activity for an authentic peptide substrate for subtilisin, N-succinyl-L-Ala-L-Ala-L-Pro-L-Phe-p-nitroanilide, compared with the single-Cys mutants. However, the disulfide mutant had a 50% decrease in catalytic efficiency due to a smaller k(cat) and was thermolabile relative to the wild-type enzyme, whereas it was greatly stabilized relative to its reduced form. These results suggest that an electrostatic interaction between Lys170 and Glu195 is important for catalysis and stability in subtilisin E. Interestingly, the disulfide mutant was found to be more stable in polar organic solvents, such as dimethylformamide and ethanol, than the wild-type enzyme, even under reducing conditions; this is probably due to the substitution of uncharged Cys by charged surface residues (Lys170 and Glu195). Further, the amino-terminal engineered disulfide bond (Gly61Cys/Ser98Cys) and the mutation Ile31Leu were introduced to enhance the stability and catalytic activity. A prominent 3-4-fold increase in the catalytic efficiency occurred in the quintet mutant enzyme over the range of dimethylformamide concentration (up to 40%).     

Computer-assisted modeling of subtilisin enantioselectivity in organic solvents

In order to rationalize our discovery of a marked dependence of subtilisin's enantioselectivity on the organic solvent used as the reaction medium, we empolyed the X-ray crystal structure of the enzyme and the means of interactive computer modeling to construct the structures of the reactive enzyme-substrate complexes. For subtilisin-catalyzed transesterifications between vinyl butyrate and S and R enantiomers of chiral secondary alcohols XCH(OH)Y, the computer simulation data clearly explain a higher reactivity of the former enantiomer on the basis of severe steric hindrances experienced by the latter enantiomer in the active site of subtilisin. The models of binding derived by computer modeling also successfully predicted changes in subtilisin enantioselectivity as a function of the sizes of the X and Y substituents in the nucleophile and upon addition of certain inhibitors. (c) 1992 John Wiley & Sons, Inc.     

The applicability of subtilisin Carlsberg in peptide synthesis.

The synthesis of peptide bonds catalysed by subtilisin Carlsberg was studied in different hydrophilic organic solvents with variable H2O concentration. Z-Val-Trp-OMe and Z-Ala-Phe-OMe were used as acyl donors, and a series of amino acid derivatives, di- and tripeptides of the general structure Xaa-Gly, Gly-Xaa, Gly-Gly-Xaa (Xaa represents all natural L-amino acids except cysteine) and other peptides were used as nucleophiles. A comparative study of the enzymatic synthesis in aqueous DMF (50%, v/v) and acetonitrile containing 10% (v/v) of H2O demonstrated that the yields of peptide products were higher in most cases when acetonitrile with low H2O concentration was used. The acylation of weak nucleophiles was improved in organic solvents with very low H2O concentration (2%). The reactions in anhydrous Bu(t)-OH proceeded with substantially lower velocity. Generally, the restricted nucleophile specificity of the enzyme for glycine and hydrophilic amino acid residues in P1' position, as well as numerous side reactions, limit the utilization of subtilisin in peptide synthesis, especially in the case of the segment condensations. Contrary to the published data, we have proved that proline derivatives were not acylated in any media with the help of subtilisin Carlsberg. Effective ester hydrolysis of a protected nonapeptide corresponding to the N-terminal sequence of dicarba-eel-calcitonin catalysed by subtilisin was achieved.     

Correlation between catalytic activity and secondary structure of subtilisin dissolved in organic solvents

Fourier-transform infrared (FTIR) spectroscopy has been used to quantify the alpha-helix and beta-sheet contents of subtilisin Carlsberg dissolved in several nonaqueous, as well as aqueous, solvents. Independently, the catalytic activity of the enzyme has been measured in the same solvents. While our previous FTIR studies revealed no connection between the secondary structure and enzymatic activity for subtilisin suspended in various organic solvents, a very different situation is observed herein for the dissolved enzyme. Specifically, if either the alpha-helix or beta-sheet content in a given solvent is higher or lower than in water, no appreciable enzymatic catalysis is observed. Conversely, when the secondary structure of subtilisin dissolved in a given nonaqueous solvent is similar to that in water, so is the enzymatic activity. (c) 1997 John Wiley & Sons, Inc. Biotechnol Bioeng 56: 485-491, 1997.     

Enzyme engineering for nonaqueous solvents. II. Additive effects of mutations on the stability and activity of subtilisin E in polar organic media.

Single amino acid substitutions increase the activity and stability of subtilisin E in mixtures of organic solvents and water, and the effects of these mutations are additive. A variant of subtilisin E that exhibits higher activity in mixtures of dimethylformamide (DMF) and water (Q103R) was created by random mutagenesis combined with screening for improved activity (K. Chen and F. H. Arnold, in preparation). Another mutation, N218S, known to improve both the activity and stability of subtilisin BPN', also improves the activity and stability of subtilisin E in the presence of DMF. The effects of the two substitutions on transition-state stabilization are additive. Furthermore, the Q103R mutation that improves activity has no deleterious effect on subtilisin stability. The double mutant Q103R+N218S is 10 times more active than the wild-type enzyme in 20% (v/v) DMF and twice as stable in 40% DMF. Although the effects of single mutations can be impressive, a practical strategy for engineering enzymes that function in nonaqueous solvents will most likely require multiple changes in the amino acid sequence. These results demonstrate the excellent potential for engineering nonaqueous-solvent-compatible enzymes.     

A study of subtilisin types Novo and Carlsberg by circular polarization of fluorescence.

The circular polarization of the luminescence of a chromophore, in addition to its circular dichroism and optical rotatory dispersion, is a manifestation of its asymmetry. In the study of proteins, the circular polarization of luminescence yields more specific information than circular dichroism or optical rotatory dispersion since nonfluorescent chromophores do not contribute, and the spectra of the tyrosine and the tryptophan residues are much better resolved in emission than in absorption. The circular polarization of the fluorescence of the tyrosine and tryptophan residues in derivatives of subtilisin Carlsberg and subtilisin Novo were indeed resolved in this study. The tyrosine residues in the Carlsberg protein, and both tyrosine and tryptophan residues in the Novo protein, were found to be heterogeneous with respect to their optical activity and emission spectra. Changes in the environment of the emitting tyrosine residues in both proteins and in the tryptophan residues in the Novo protein were found on changing the pH from 5.0 to 8.3. The pH dependence of the enzymatic activity of these proteins may thus be due, at least in part, to conformational changes in the molecules. Fluorescence circular polarization also revealed that covalently bound inhibitors at the active site of subtilisin Novo affect the environment of the emitting aromatic side chains, presumably via changes in conformation.     

Kinetic analysis of the mechanism for subtilisin in essentially anhydrous organic solvents

Steady-state kinetic analysis has been used to confirm the catalytic mechanism of lyophilized subtilisin suspended in a variety of organic solvents. Specifically, this article demonstrates that partial reactions can occur between subtilisin and ester substrates in organic solvents. Partitioning of common intermediates between competing acceptors at a constant ratio of products has also been described. The decomposition of a common intermediate formed from different substrates at the same rate is also further evidence of an acyl-enzyme mechanism for subtilisin suspended in anhydrous solvents. Partitioning of a common intermediate to give two products at a constant total rate, and saturation kinetics at varying substrate concentrations, complete a kinetic investigation of the enzyme mechanism. All the data generated support the formation of a stable acyl enzyme during the transesterification reaction catalzyed by subtilisin in the solvents used.     

Effects of charged water-soluble polymers on the stability and activity of yeast alcohol dehydrogenase and subtilisin Carlsberg.

Remarkable increases in enzyme catalytic stability resulting from addition of charged water-soluble polymers have recently been reported, suggesting that use of these polymers may be an attractive general strategy for enzyme stabilization. To test the proposed hypothesis that coulombic forces between water-soluble polymers and enzymes are primarily responsible for enzyme stabilization, we examined the catalytic stability and activity of two enzymes in the presence of polymers differing in net charge. All polymers tested increased enzyme lifetimes, regardless of their net charge, suggesting that stabilization of these enzymes by water-soluble polymers is not solely dependent on simple electrostatic interactions between the polymers and enzymes.     

On enzymatic activity in organic solvents as a function of enzyme history

Catalytic activities of alpha-chymotrypsin and subtilisin Carlsberg in various hydrous organic solvents were measured as a function of how the enzyme suspension had been prepared. In one method, lyophilized enzyme was directly suspended in the solvent containing 1% water. In another, the enzyme was precipitated from its aqueous solution by a 100-fold dilution with an anhydrous solvent. In most cases, the reaction rate in a given nonaqueous enzymatic system strongly (up to an order of magnitude) depended on the mode of enzyme preparation. The magnitude of this dependence was markedly affected by the nature of the solvent and enzyme. A mechanistic hypothesis proposed to explain the observed dependencies was verified in additional experiments in which the water contents and enzyme history were further varied.     

Glutathione oxidase activity of selenocystamine: a mechanistic study.

Selenocystamine (RSe-SeR) was shown to catalyze the oxygen-mediated oxidation of excess GSH to glutathione disulfide, at neutral pH and ambient PO2. This glutathione oxidase activity required the heterolytic reduction of the diselenide bond, which produced two equivalents of the selenolate derivative selenocysteamine (RSe-), via the transient formation of a selenenylsulfide intermediate (RSe-SG). Formation of RSe- was the only reaction observed in anaerobic conditions. At ambient PO2, the kinetics and stoichiometry of GSSG production as well as that of GSH and oxygen consumptions demonstrated that RSe- performed a three-step reduction of oxygen to water. The first step was a one-electron transfer from RSe- to dioxygen, yielding superoxide and a putative selenyl radical RSe., which decayed very rapidly to RSe-SeR. In the second step, RSe- reduced superoxide to hydrogen peroxide through a much faster one-electron transfer, also associated with the decay of RSe. to RSe-SeR. The third step was a two-electron transfer from RSe- to hydrogen peroxide, again much faster than oxygen reduction, which resulted in the production of RSe-SG, presumably via a selenenic acid intermediate (RSeOH) which was trapped by excess GSH. This third step was studied on exogenous hydroperoxide in anaerobic conditions, and it could be eliminated from the glutathione oxidase cycle in the presence of excess catalase. The role of RSe- as a one- and two-electron reductant was confirmed by competitive carboxymethylation with iodoacetate. RSe- was able to rapidly reduce ferric cytochrome c to its ferrous derivative. The overall rate of catalytic glutathione oxidation was GSH concentration dependent and oxygen concentration independent. Excess glutathione reductase and NADPH increased the catalytic oxidation of GSH, probably by switching the rate-limiting step from selenylsulfide to diselenide cleavage. When GSH was substituted for dithiothreitol, it was shown to reduce RSe-SeR to RSe- in a fast and quantitative reaction, and selenocystamine behaved as a dithiothreitol oxidase, whose catalytic cycle was dependent on oxygen concentration. The oxidase cycle of glutathione was inhibited by mercaptosuccinate, while that of dithiothreitol was not affected. When mercaptosuccinate was substituted for GSH, a stable selenenylsulfide was formed. These observations suggest that electrostatic interactions affect the reductive cleavage of diselenide and selenenylsulfide linkages. This study illustrates the ease of one-electron transfers from RSe- to a variety of reducible substrates. Such free radical mechanisms may explain much of the cytotoxicity of alkylselenols, and they demonstrate that selenocystamine is a poor catalytic model of the enzyme glutathione peroxidase.