Dr Frank Bordusa, Max-Planck Society, Research Unit ‘Enzymology of Protein Folding’, Weinbergweg 22, D-06120 Halle/Saale, Germany. Fax: + 49 3419736998, Tel.: + 49 3419736918, E-mail: firstname.lastname@example.org
The function of acyl-4-guanidinophenyl esters as substrate mimetics for the serine protease α-chymotrypsin was investigated by protein-ligand docking, hydrolysis, and acyl transfer experiments. On the basis of protein-ligand docking studies, the binding and hydrolysis properties of these artificial substrates were estimated. The predictions of the rational approach were confirmed by steady-state hydrolysis studies on 4-guanidinophenyl esters derived from coded amino acids (which α-chymotrypsin is not specific for), noncoded amino acids, and even simple carboxylic acid moieties. Enzymatic peptide syntheses qualify these esters as suitable acyl donors for the coupling of acyl components far from the natural enzyme specificity, thus considerably expanding the synthetic utility of α-chymotrypsin.
Enzymatic peptide bond formation using proteases is considered to be an attractive alternative to chemical solution and solid-phase peptide synthesis [1–3]. The advantages of an enzymatic synthesis are the mild reaction conditions, the freedom from racemization and the need for side-chain protection, the possibility of using immobilized enzyme technology with catalyst recovery, and the scope for industrial scale-up. However, there are some disadvantages. Beyond dipeptides, there is the permanent risk of undesired proteolytic side reactions. No new case can be treated as routine. Furthermore, the specificities of the available proteases prevent the synthesis of any proteinogenic sequences. Finally, noncoded amino acids are not usually accepted as substrates.
To overcome these limitations, medium [4–6], protein [7–10], and substrate engineering [4,11] were found to be useful. In this context, the application of substrate mimetics represents one of the most powerful strategies that was firstly reported for trypsin-catalysed peptide syntheses [12–14]. This concept is based on the binding site specific 4-guanidinophenyl ester (OGp) functionality to mediate acceptance of nonspecific amino acid moieties in the specificity-determining S1 position of the enzyme (notation according to ). Recently, we found that other Arg-specific proteases, e.g. thrombin and clostripain, act in the same manner . Docking studies indicate binding of the OGp moiety in the active site mimicking the specific Arg side chain of natural substrates. Further investigations revealed that the formation of the acyl enzyme results from a reversed binding of these artificial substrates leading to an acyl enzyme intermediate which bears the acyl residue at the S′ subsite region . Suggested by similar S′ specificity data found for the deacylation of normal and reversed-type substrates, the acyl residue of this intermediate should rearrange to the S subsite of the enzyme during catalysis (Scheme 1).
On the basis of these new mechanistic insights, the concept of substrate mimetic-mediated catalysis was postulated to be of general validity, which could be confirmed by the use of Staphylococcus aureus V8 protease as a biocatalyst introducing a novel type of substrate mimetics . In this case, the new mimetic moiety was developed in an empirical way on the basis of the natural S1 specificity of the enzyme because of its unknown structure. For proteases of known 3D structure, rational approaches for the design of specifically binding mimetics should alternatively be applicable. Moreover, on the basis of knowledge on substrate mimetics recognition and the mechanism of catalysis, this approach may allow an estimation of whether the calculated binding leads to a hydrolysis or not. In this paper, computer-assisted protein-ligand docking studies were used to predict the function of the OGp moiety as an artificial recognition site for α-chymotrypsin. The validity of the results of the rational approach was proved by hydrolysis studies and enzyme-catalysed peptide syntheses employing acyl-OGps derived from acyl moieties comprising coded amino acids, which α-chymotrypsin is not specific for, as well as noncoded amino acids and simple carboxylic acid derivatives.
Materials and methods
Tos-Lys-CH2Cl-treated bovine α-chymotrypsin (EC 184.108.40.206) was purchased from Fluka (Switzerland). It was used without further purification. Amino acid derivatives, peptides, 4-aminophenol, N,N′-dicyclohexylcarbodiimide, 4-dimethylaminopyridine, benzyl chloroformate, S-methylisothiourea, and p-toluenesulfonic acid were products of Bachem (Switzerland), Fluka and Merck (Germany), respectively. All reagents were of the highest commercial purity. Solvents were purified and dried by the usual methods.
Boc-amino acid-OGp and Bz-amino acid-OGp were prepared according to a previously described procedure . Bz-l-Phe-methyl ester (OMe) was synthesized by benzoylation of H-l-Phe-OMe. The identity and purity of all final products were checked by analytical HPLC analysis at a wavelength of 220 nm, NMR, thermospray mass spectroscopy, and elemental analysis. Satisfactory analytical data were found in all cases (± 0.4% for C, H, N).
The docking studies were based on the crystal structure of α-chymotrypsin (RCSB Protein Data Bank, 1cho) . All solvent molecules and ions were removed from the protein, polar hydrogens were added, and template charges were assigned employing the quanta96 program package . The ligand molecule Boc-l-Ala-OGp was treated in the same way. Following a previously described protocol [17,21], the program package autodock, version 2.4 , was employed for the protein-ligand docking experiments, which successfully reproduces the crystal structures of nonflexible macromolecules with small flexible ligands [23,24]. Within each of the 20 independent docking runs, 20 million protein-ligand arrangements were generated in a 35 × 35 × 35 Å cubic grid centred around the catalytic Ser195 residue applying a grid spacing of 0.35 Å. Free rotation was allowed around all relevant bonds of the ligand. The resulting protein-ligand complexes were clustered, employing an rmsd tolerance of 1 Å. According to the trypsin-model  and the general ideas on the catalytic mechanism , those lowest-energy complexes between chymotrypsin and the ligands corresponding to a productive orientation, i.e. showing hydrogen bonds between Gly193 (oxanion hole) and Ser195 (active residue) of the enzyme and the carbonyl carbon and oxygen of the scissile ester bond, were the basis of all discussions.
Hydrolysis reactions were performed at 25 °C using an assay mixture containing 0.025 m Mops buffer, pH 7.6, 0.1 m NaCl, 5 mm CaCl2, and 20% methanol as cosolvent. The substrate concentrations were between 0.006 and 4.0 mm and the enzyme concentrations between 9.8 × 10−7 and 1.6 × 10−5m. The active enzyme concentration was determined by active site titration using 4-nitrophenyl-4′-guanidino-benzoate as titrant . After thermal equilibration of the assay mixtures the reactions were initiated by addition of 5 µL of enzyme stock solution. The rate of reaction was analysed by reversed phase HPLC determining the disappearance of the substrate esters for at least 10 different concentrations. For this purpose, aliquots were withdrawn at defined time intervals and diluted with stop solution containing 50% methanol and 1% trifluoroacetic acid. The kinetic parameters were calculated by iterative nonlinear curve fitting of the untransformed data according to the Michaelis–Menten formalism using the software sigmaplot 1.01 (Jandel, USA). In cases where saturation of the enzyme by the substrate could not be reached, the specificity constants were calculated from the linear slope of the initial Michaelis–Menten curve in relation to the active enzyme concentration. All values reported are the average of at least two independent experiments.
The enzymatic reactions were performed at 25 °C in a total volume of 50 µL containing 0.2 m Hepes buffer, pH 8.0, 0.2 m NaCl, and 10 mm CaCl2. Stock solutions of acyl donor esters (4 mm) were prepared in water containing 20% methanol as cosolvent (10% dimethylformamide in the cases of Bz-l-Phe-OMe, Bz-d-Phe-OGp, and Bz-OGp). Acyl acceptors (stock solution 40 mm) were dissolved in 0.4 m Hepes buffer, pH 8.0, 0.4 m NaCl, and 20 mm CaCl2. The concentrations of acyl acceptors were calculated as free, Nα-unprotonated species [HN]0 according to the Henderson–Hasselbalch equation [HN]0 = [N]0/(1 + 10pK–pH) . The pK values of the α-amino group of the nucleophiles were determined by inflection point titration on a Video Titrator VIT 90 (Radiometer, Denmark). After thermal equilibration of the assay mixture the reactions were initiated by addition of 2.5 µL of enzyme stock solution leading to an active enzyme concentration of 37 µm for substrate mimetics and 0.1 µm for Bz-l-Phe-OMe. A reaction time of 5 min usually led to a complete ester consumption. A control experiment without enzyme was performed for each reaction to determine the extent of nonenzymatic ester hydrolysis which was generally less than 5%. Based on the same control experiments, nonenzymatic aminolysis of the acyl donor esters could be ruled out. For HPLC analysis aliquots (40 µL) were withdrawn and diluted with 100 µL of stop solution containing 50% methanol and 1% trifluoroacetic acid. The data reported are the average of at least three independent experiments.
Samples were analysed by analytical reversed phase HPLC using RP C18 columns (Vydac 218TP54, 10 µm, 300 Å, 25 × 0.4 cm and Grom Capcell SG 120, 5 µm, 5 Å, 25 × 0.4 cm), a C8 column (Grom Capcell SG 300, 5 µm, 300 Å, 25 × 0.4 cm), and a C4 column (Kromasil 5 µm, 300 Å, 125 × 4.6 mm). Detection was at 254 nm. The reaction rates and product yields were calculated from peak areas of the substrate esters and the hydrolysis and aminolysis products, respectively. p-Toluenesulfonic acid was used as internal standard.
To evaluate whether acyl-OGps bind to α-chymotrypsin, Boc-l-Ala-OGp was selected as a model ligand and docked towards the enzyme. Figure 1 shows the arrangement of the ligand in the active site of α-chymotrypsin in the lowest energy complex (a), in comparison to that found with trypsin (b) . The ligand interacts with α-chymotrypsin in a specific manner. Besides the leaving group-mediated binding of Boc-l-Ala-OGp to several positions within the S′ subsite region of the enzyme, by far the lowest interaction energies were calculated for complexes with the OGp moiety binding at S1. Corresponding to the natural specificity of α-chymotrypsin for aromatic amino acid residues, hydrophobic contacts between the phenyl moiety of the ester group and the amino acid residues Cys191 and Val213 of the enzyme predominate. Moreover, the guanidino functionality favours this binding mode by formation of additional hydrogen bonds to the serine residues Ser189, Ser190, and Ser217 which are located at the bottom of the S1 binding pocket. Mediated by the specific binding of the leaving group, the whole substrate realizes an arrangement where the carbonyl carbon of the scissile ester bond is, nevertheless, close to the hydroxy group of the reactive Ser195 (about 2.7 Å). In the case of trypsin, this distance was found to be only slightly shorter (about 2.5 Å) . Furthermore, the carbonyl oxygen of the substrate ester points towards Gly193 which is important for the formation of the oxanion transition state (oxanion hole). Obviously, the enzyme–substrate arrangements calculated for trypsin and α-chymotrypsin show a high degree of similarity despite the considerable differences in their natural specificities. Since this arrangement leads to a specific cleavage of the substrate in the case of trypsin, α-chymotrypsin should be capable of hydrolysing acyl-OGps too.
To evaluate the predictions provided by the docking studies, steady-state-hydrolysis kinetic studies were performed using several Boc-Xaa-OGp esters (Xaa =l-Ala, l-Gly, l-Glu, l-Pro, d-Ala, d-Leu, and d-Phe) and Bz-OGp. Plots of the initial rates of hydrolysis for selected derivatives are shown in Fig. 2. As indicated by the straight line found for the hydrolysis of Boc-l-Glu-OGp, a saturation of the enzyme by the substrate could not be observed in all cases. For substrates reaching the maximum rate of hydrolysis, the plots generally show the typical Michaelis–Menten kinetics without any hint at substrate inhibition or activation. The kinetic data are summarized in Table 1. They demonstrate that all substrates with l-amino acids and even those with d-amino acids are specifically cleaved by α-chymotrypsin, as predicted by the computational docking studies. Analysing the second order rate constants kcat/Km, the highest specificities were found for Boc-l-Ala-OGp, Boc-d-Phe-OGp, and Boc-d-Leu-OGp. The substrates derived from Gly, l-Glu, and d-Ala show a decrease in specificity by a factor of about 10. This also holds for the nonamino acid-derived Bz-OGp analogue. For Boc-l-Pro-OGp, the lower kcat/Km-value three orders of magnitude indicates the lowest specificity of all substrates under investigation. Interestingly, as already found for trypsin-catalysed reactions, α-chymotrypsin also catalyses the hydrolysis of Boc-d-Ala-OGp by more than one order of magnitude less specifically than the hydrolysis of the l-enantiomer. As indicated by the individual kinetic parameters, this effect can directly be attributed to the decrease of the corresponding kcat values, whereas the Km values are quite similar for the two enantiomers.
Table 1. Steady-state kinetic parameters for the α-chymotrypsin-catalysed hydrolysis of OGps. Conditions were 0.025 m Mops buffer, pH 7.6, 0.1 m NaCl, 5 mm CaCl2, 25 °C, 20% MeOH. All errors are less than 15%.
Calculated as linear slope of the initial Michaelis–Menten curve in relation to the active enzyme concentration.
The basic properties of acyl-OGps for α-chymotrypsin-catalysed peptide syntheses were characterized by analytical acyl transfer reactions. For this purpose, the substrate esters listed in Table 1 were used in form of the corresponding Nα-Bz-substituted analogues as acyl donors in coupling reactions with amino acid and dipeptide amides, respectively. The normal-type substrate Bz-l-Phe-OMe served as a reference. To control for spontaneous hydrolysis and aminolysis of the acyl donor esters, the reactions were analysed without enzyme. On the basis of these experiments, nonenzymatic aminolysis could be ruled out and the extent of spontaneous hydrolysis was found to be less than 5%. The yields obtained for the enzymatic coupling reactions are summarized in Table 2 for esters derived from coded amino acids, which α-chymotrypsin is not specific for, and in Table 3 for esters derived from noncoded amino acid or nonamino acid moieties. Generally, the results show that each OGp is able to serve as an acyl donor for α-chymotrypsin-catalysed peptide synthesis independently of its acyl moiety, as was predicted by the docking and hydrolysis kinetic studies. On analysing the efficiency of the catalysis for substrate mimetics with coded amino acids (Table 2), yields similar to those found for the normal-type acyl donor Bz-l-Phe-OMe could be obtained in most cases. Remarkably, this fact also holds for the ester derived from the bulky imino acid proline, where even higher yields were obtained. Only in a few reactions, using Bz-l-Glu-OGp as acyl donor, did somewhat lower product yields result. The data listed in Table 3 show that the OGp functionality mediates coupling not only of nonspecific coded acyl moieties, but also of d-amino acid or simple carboxylic acid derivatives. Interestingly, the comparison of the enantiomers of Boc-l/d-Ala-OGp shows about 10–40% higher product yields for the reactions with the d-isomer. A comparable influence of the configuration on the efficiency of peptide bond formation was already found in trypsin-catalysed reactions . The data observed for Bz-d-Phe-OGp show, however, that this finding cannot be generalized. Finally, the high yields found for reactions with the non amino-acid-derived benzoic acid ester (Bz-OGp) indicate that OGps of simple carboxylic acids can also act as efficient acyl donor components for α-chymotrypsin-catalysed coupling.
Table 2. Yields in percentage of α-chymotrypsin-catalysed peptide syntheses using OGps bearing unspecific, coded, amino acid derivatives. Conditions were 0.2 m Hepes buffer, pH 8.0, 0.2 m NaCl, 0.02 m CaCl2, 25 °C, 10% MeOH; [acyl donor] 2 mm, [acyl acceptor] 20 mm, [enzyme] 37 µm. All errors are less than 5%.
Table 3. Yields in percentage of α-chymotrypsin-catalysed peptide syntheses using OGps bearing noncoded amino acid and carboxylic acid derivatives. Conditions were 0.2 m Hepes buffer, pH 8.0, 0.2 m NaCl, 0.02 m CaCl2, 25 °C, 10% MeOH; [acyl donor] 2 mm, [acyl acceptor] 20 mm, [enzyme] 37 µm. All errors are less than 5%.
Up to now, leaving groups of protease substrates which interact in the manner of substrate mimetics could only be designed in an empirical way on the basis of the natural S1 specificity of the enzyme. The application of rational approaches has mainly been prevented by the lack of knowledge on substrate binding and the mechanism of catalysis of this type of artificial substrate. Recently, we suggested an extended kinetic model for the hydrolysis of such reversed-type substrates by trypsin . The protein-ligand docking approach used in these studies confirmed the experimental findings and, furthermore, provided a detailed picture of the binding behaviour of these substrates. The results of the present study illustrate that it is possible on the basis of this approach to estimate a priori whether a substrate mimetic becomes hydrolysed by the protease. It can be shown whether the conditions for the catalytic mechanism and the binding behaviour of substrate mimetics, namely the orientation of the carbonyl oxygen to Gly193 (oxanion hole), the distance between the carbonyl carbon of the scissile ester bond and the active Ser195, and the reversed binding of the acyl moiety, are fulfilled . Thus the protein-ligand docking approach can serve as a useful tool to predict acceptance of a given substrate mimetic by the enzyme.
For the design of substrate mimetics with higher specificities towards α-chymotrypsin, the analysis of the kinetic constants found for the hydrolysis of acyl-OGps is meaningful. As revealed by the kinetic parameters, α-chymotrypsin generally catalyses the hydrolysis of these substrate mimetics less specifically than the hydrolysis of normal-type substrates. For the most specific Boc-Xaa-OGp esters, specificity constants kcat/Km were obtained, which are about two orders of magnitude lower than those found, for instance, for Ac-Tyr-OEt . This effect is mainly based on the lower kcat values, whereas the values for Km are quite similar for the two types of substrate. Since the deacylation is the rate-limiting step of the hydrolysis of normal-type esters , kcat corresponds to k3 and Km describes the binding (KS) and the rate of acylation k2, respectively (see Scheme 1). On the basis of this assumption, the Km values found for Boc-Xaa-OGp esters indicate similar efficiencies of binding and (or) similar rates of acylation. Accordingly, the decrease in the kcat values might suggest lower rates of deacylation for acyl enzymes derived from OGps, which was already found for the analogous trypsin-catalysed hydrolyses . Corresponding to the trypsin-model, the decrease in kcat might be caused by the additional rearrangement step KR which results from the reverse binding of the substrate mimetics. Since the rate of deacylation is not usually affected by the leaving group, this finding also indicates that an increase of the specificity of the substrate mimetics can only be reached by an improvement of binding and an increase of the acylation rate. Studies on the applicability of the docking approach in optimization or design of novel types of substrate mimetics leading to more specific binding and higher acylation rates are presently under investigation. From a synthetic point of view, the increase of the substrate specificity is generally advantageous since the increase in specificity usually decreases the risk of proteolytic side reactions as well as the amount of enzyme required for catalysing the synthesis reaction.
As substrate mimetics are obviously meaningful in protease-mediated peptide syntheses, the predictions of the computational approach were also examined by acyl transfer reactions. The results of these studies show that the OGp functionality enables α-chymotrypsin to catalyse the peptide bond formation between nonspecific acyl residues and numerous amino acid and peptide derivatives. This behaviour also holds for esters derived from coded amino acids (which α-chymotrypsin is not specific for), noncoded amino acids, and even simple carboxylic acid moieties. The product yields obtained demonstrate the high synthetic utility of these types of acyl donors despite the relatively low specificity of chymotrypsin for their hydrolysis. With the exception of Boc-d-Phe-OGp, the detailed analysis of the relationship between specificity and efficiency of synthesis reveals a slight increase of the yields with a decrease in substrate specificity. As already discussed in our previous paper  for the analogous trypsin-catalysed reactions, this remarkable effect could be based on the reverse binding of substrate mimetics to the active site of the protease which seems to inhibit the hydrolysis, but favours the aminolysis of the acyl enzyme intermediate. Consequently, α-chymotrypsin can be used as an efficient biocatalyst for coupling a wide spectrum of different acyl components in the form of their OGps far away from its natural specificity which is unfeasible with common acyl donors. Furthermore, due to the acceptance of OGps by trypsin and proteases with trypsin-like specificity, this type of substrate mimetics becomes likewise universal for numerous proteases.
This work has been supported by the Deutsche Forschungsgemeinschaft (Innovationskolleg ‘Chemisches Signal und biologische Antwort’) and Fonds der Chemischen Industrie (Liebig Scholarship, F. B.). We would like to thank Prof. Dr H.-D. Jakubke for inspiring discussions and Mrs R. Schaaf and Mrs. D. Haines for skillful technical assistance. The authors thank Fluka AG and ASTA Medica AG for special chemicals.