Shiao, D. D. F., Lumry, R. & Rajender, S. (1972) Eur. J. Biochem. 47. Zaslavsky, B. Yu., Mestechkina, N. M. & Rogozhin, S. V. (1983). 48. Mozhaev, V. V., &kinis ...
Eur. J. Biochem. 173,147-154(1988) 0FEBS 1988
Protein stabilization via hydrophilization Covalent modification of trypsin and a-chymotrypsin
',
Vadim V. MOZHAEV Virginius A. SIKSNIS', Nikolay S. MELIK-NUBAROV', Nida Z. GALKANTAITE', Gervydas J. DENIS2, Eugenius P. BUTKUS ', Boris Yu. ZASLAVSKY ', Nataliya M. MESTECHKINA' and Karel MARTINEK
' Chemistry Department, Moscow State University All-Union Research Institute of Applied Enzymology, Vilnius Chemistry Department, Vilnius State University Institute of Elementorganic Compounds, USSR Academy of Sciences, Moscow Institute of Organic Chemistry and Biochemistry, Czechoslovak Academy of Sciences, Prague (Received July 3/November 18,1987) - EJB 87 0766
This paper experimentally verifies the idea presented earlier that the contact of nonpolar clusters located on the surface of protein molecules with water destabilizes proteins. It is demonstrated that protein stabilization can be achieved by artificial hydrophilization of the surface area of protein globules by chemical modification. Two experimental systems are studied for the verification of the hydrophilization approach. 1. The surface tyrosine residues of trypsin are transformed to aminotyrosines using a two-step modification procedure: nitration by tetranitromethane followed by reduction with sodium dithionite. The modified enzyme is much more stable against irreversible thermoinactivation: the stabilizing effect increases with the number of aminotyrosine residues in trypsin and the modified enzyme can become even 100 times more stable than the native one. 2. a-Chymotrypsin is covalently modified by treatment with anhydrides or chloroanhydrides of aromatic carboxylic acids. As a result, different numbers of additional carboxylic groups (up to five depending on the structure of the modifying reagent) are introduced into each Lys residue modified. Acylation of all available amino groups of cr-chymotrypsin by cyclic anhydrides of pyromellitic and mellitic acids results in a substantial hydrophilization of the protein as estimated by partitioning in an aqueous Ficoll-400/Dextran-70 biphasic system. These modified enzyme preparations are extremely stable against irreversible thermal inactivation at elevated temperatures (65 -98 "C); their thermostability is practically equal to the stability of proteolytic enzymes from extremely thermophilic bacteria, the most stable proteinases known to date.
Applications of enzymatic catalysis in biotechnology, fine organic synthesis, analysis, medicine, and other areas are often hindered because many enzymes, if isolated from their natural environment in vivo, become unstable and rapidly inactivate (denature); for review, see [l, 21. For these reasons the problem of enzyme stabilization has received considerable attention in recent years [3 - 81. The analysis of the structure-stability relationships in the protein leads us to the conclusion [S] that the structure of proteins is not optimal with regard to their stability. There are still some reverses for stabilization which are used by nature for the production of extremely stable proteins, such as, for example, enzymes from thermophilic microorganisms ;for review, see [S, 91. Reserves of such a kind may be useful for artificial stabilization of enzymes. We shall discuss here, from this viewpoint, how to improve the protein structure in order to make enzymes more stable. According to X-ray crystallographic data, about one half of the surface area of proteins is occupied by nonpolar amino Corre2pondence to K. Martinek, Ustav OrganickC Chemie a Biochemie, Ceskoslovenska Akademie Vtid, Flemingovo nimEsti 2, CS-166-10 Praha, Czechoslovakia The authors wish to dedicate this paper to the memory of their teacher and friend. Professor Iliva V. Berezin. Enzymes. cr-Chymotrypsin (EC 3.4.21.2), trypsin (EC 3.4.21 -4)
acids [lo- 121. Nonpolar residues are organized very often as hydrophobic surface clusters [13, 141. These play an important role in vivo, since they enable proteins to bind via hydrophobic interactions to other proteins (forming multienzyme complexes), to lipids in biological membranes, to polysaccharides in cell walls, to substrates and effectors during enzymatic catalysis, i.e. to function in the optimal manner [15, 161. However, the contact of nonpolar residues with water is thermodynamically disadvantageous [13,17,18] and is harmful for protein stability in vitro [6, 7, 161. Hence, a reduction of the nonpolar surface area should stabilize proteins. This is in fact observed when comparing the in vitro structure and stability of mesophilic and thermophilic proteins [19-211 and of enzymes from wild and mutant strains [16, 221; for review, see [8]. Undoubtedly, the same effect (protein stabilization by hydrophilization of the nonpolar surface area) may also be achieved artificially. Stellwagen suggested recently [16] the use of mutagenesis for hydrophilization. However, at present, the techniques of site-directed mutagenesis are still complicated. Therefore, other (more simple) methods of hydrophilization of proteins must be also tried. Covalent modification of proteins is one of the first lines of approach to be taken. The experimental procedure of hydrophilization involves several important problems. The main one is the criterion of hydrophobicity (hydrophilicity) of the modifying reagents.
148 hydrophilic modifier hydrophobic surface cluster
-0
chemical modification
r
or
hydrophilk modifier
0-
Fig. 1 . Schematic presentation ofthe approach to hydrophilization of nvnpolar surface areas m protein globule by chemical modification
We suggest using the n values of functional groups and atoms introduced by Hansch [23, 241 for this purpose; see also [8] for details. Another principal problem is to carry out the modification so that it may affect the functional groups located on the surface of a protein only but not the groups in their interior. This selectivity can be achieved by the choice of experimental conditions. The surface functional groups are modified, for example, by using a small molar excess of the modifying reagent over the protein. By contrast, 100- 1000-fold molar excess of the modifier should be used to modify the internal amino acid residues [25]. The modification can be performed in one of the two ways shown in Fig. 1. First, the hydrophobic amino acid residues are directly modified by the hydrophilic reagent. Obviously, the most hydrophobic are the amino acids with either aliphatic or aromatic side chains [13]. It is not easy to modify aliphatic amino acids under mild (non-denaturing) conditions [25]. There are some reactions, however, which bring about the hydrophilization of Tyr (see below) or Trp [25]. Second, any (not necessarily hydrophobic) residue located near a nonpolar surface cluster, can be modified by the hydrophilic reagent to obtain shielding of the cluster from the aqueous medium (Fig. 1). In this study we have realized both hydrophilization approaches shown in Fig. 1. First, a two-step introduction [26, 271 of an amino group into the tyrosine residues of trypsin was performed:
This reaction was chosen for hydrophilization since the hydrophilic increment of the amino group introduced into the aromatic ring of tyrosine is equal [24] to about -5 kJ/mol (- 1.2 kcal/mol). Second, for the acylation of amino groups of a-chymotrypsin benzoic acid chloroanhydride and a series of aromatic carboxylic acid anhydrides were used:
group, especially in its ionized form, is highly negative: about - 22 kJ/mol (- 5 kcal/mol) for COO- group in the aromatic ring [24]. No doubt, amino groups in the surface layer of the protein globule are first modified by acylation. X-ray crystallographic data have shown [28] that almost all amino groups are located on the surface of the a-chymotrypsin globule. More than half of the residues adjacent to lysines are hydrophobic amino acids [28]. We assume therefore that the hydrophilic fragments of the bulky aromatic carboxylic acids (used as modifying reagents) shield at least some of the surface hydrophobic residues, thus stabilizing the enzyme.
EXPERIMENTAL PROCEDURES Catalytic activity assay The catalytic activity of trypsin and a-chymotrypsin from bovine pancreas (Sigma) was evaluated by determining the initial rate of hydrolysis of 5 mM N-benzoyl-L-arginine ethyl ester (Sigma) or 5 mM N-acetyl-L-tyrosine ethyl ester (Sigma), respectively, in 0.1 M KCl, pH 8.0,25"C, using a Radiometer pH-stat. Enzyme thermoinactivation
A volume of 5- 10 ml of approximately 0.1 1 M enzyme solution in aqueous buffer (3 mM KH2P04, 0.1 M KC1, pH 8.0) was incubated with stirring in a stoppered thermostatted cell at 56 f 0.5"C for trypsin; for a-chymotrypsin the temperature was varied over the ranges 40 - 54°C and 60 97°C for the native and modified enzyme, respectively. At intervals aliquots were taken, cooled quickly to room temperature in ice water and the residual activity of trypsin was assayed as described above. Autolysis and aggregation are negligible under the experimental conditions as we have shown earlier [29]. Nitration of tyrosine residues of trypsin Nitration was performed using a slightly modified method of Kenner et al. [30]. To 1-2 ml 0.4 mM trypsin in buffer (50 mM Tris/HCI, 50 mM CaC12, pH 8.0, with or without 5 mM benzamidine), aliquots of 0.1 -2 M tetranitromethane in 96% ethanol were added (a 10-200 molar excess over trypsin). After incubation for 1-2 h at 20°C with stirring the reaction mixture was fractionated by gel filtration on Sephadex G-25 with 1 mM HCl as eluent.
(E)
As a result of such a modification several new carboxylic groups (up to five) are introduced into each protein group modified. This brings about an appreciable hydrophilization of the enzyme since the hydrophilic increment of the COOH
Reduction qf nitrotyrosine residues in trypsin to aminotrypsin [26, 271 0.5 ml of 6 mM aqueous solution of sodium dithionite (Serva) was added to 3-5 ml of 20 1M nitrated trypsin in
149 50 mM Tris/HCl (pH 8.0). After 3-5 min of incubation at 25 "C, the solution was fractionated by gel filtration on Sephadex G-25 with 1 mM HC1 as eluent. Special attention should be paid to the quality of sodium dithionite. Even in the best preparations commercially available the concentration of sodium dithionite does not exceed 85 - 90%. Impurities, such as sodium sulfite and thiosulfate, cleave the S-S bonds of some enzymes [31].In trypsin they can split the most labile S-S bond thus destabilizing the enzyme significantly. Therefore, preparations with a maximal content of dithionite were used to avoid undesirable side reactions. The concentration of dithionite in a commercial preparation was determined by titration with Cu2+ ions [32].
ated by gel filtration on Sephadex G-50 with 0.1 M KCl as eluent. Determination of degree of acylation
The degree of modification (the number of acylated amino groups) was determined by spectrophotometric titration of free (unmodified) amino groups of the protein with trinitrobenzenesulfonic acid (Serva) using the Hitachi Perkin-Elmer model, as described in [34]. In all thermoinactivation experiments, we used only the enzyme preparations whch showed maximal degree of acylation (no free amino groups were determined by spectrophotometric titration).
Determination of nitrotyrosines
The pH value of the enzyme solution after nitration and gel filtration was adjusted to 9.5; then the absorbance at 428 nm was measured and compared with native trypsin. The number of nitrotyrosines in the trypsin molecule was calculated taking the molar absorption coefficient of nitrotyrosine at 428 nm equal to 4200 M-' cm-' (cf. [33]). Determination of aminotyrosines
The number of aminotyrosines in the trypsin molecule was determined by two methods. a) The pH value of the enzyme solution after reduction and gel filtration was adjusted to 8.0; then, the absorbance at 428 nm was determined and compared with the absorbance of nitrated trypsin at 428 nm. The number of aminotyrosines (nTyr-NH,)in trypsin was calculated taking the molar absorption coefficient of nitrotyrosine at 428 nm equal to 4200 M-' cm- [33] and the degree of dissociation of nitrotyrosine equal to 0.9 at pH 8.0. b) Spectrophotometric titration with picrylsulfonic acid (Sigma) [34] was carried out at pH 6.0, where aliphatic amino groups of the protein are protonated and are thus unreactive towards the titrant. They do not therefore interfere [34] with the determination of deprotonated reactive aromatic amino groups of aminotyrosines (pK, = 4.75). The molar absorption coefficient of the reaction product is 13000 M-' cm-' at 420 nm. The results obtained by the two methods coincide well; moreover, the number of aminotyrosines determined spectrophotometrically coincides with the results [33] of amino acid analysis of the modified enzyme.
'
Modtfication of a-chymotrypsin
The acylation of the enzyme was effected by benzoyl chloride (Sigma) and a series of cyclic anhydrides of aromatic carboxylic acids : o-phthalic acid (Sigma), trimellitic acid (Aldrich), pyromellitic acid (Sigma) and mellitic acid synthesized by us. Modification of a-chymotrypsin with benzoyl chloride (A) was described elsewhere [35]. Modification of the enzyme with cyclic anhydrides of phthalic (B), trimellitic (C), pyromellitic (D) and mellitic (E) acids was carried out according to a slightly modified procedure described in [36]. A solution (1 ml) of the anhydride (2 mM) in dimethylsulfoxide was added dropwise at 4°C to 10 nil 40 pM a-chymotrypsin in 0.1 M KH2P04 buffer (pH 8.0) with stirring. The pH of the reaction mixture was kept constant by addition of a concentrated alkali solution. The reaction proceeded for 2 h, the solution was then fraction-
Partitioning of a-chymotrypsinpreparations in the aqueous Ficoll-400/Dextran-70biphasic system
The biphasic system, 16.9% (w/w) Ficoll400 (Pharmacia) and 13.9% (w/w) Dextran 70 (lot 310670, Minmedprom, USSR), was prepared as described previously [37]. The ionic composition of the system was 0.15 M NaCl in 0.01 M sodium phosphate buffer (pH 7.4). The difference in the relative hydrophobicities of the phases was estimated [37] by partitioning sodium salts of 2,4-dinitrophenylated amino acids with aliphatic side chains. Partitioning experiments were carried out as described previously 1371. The enzyme solution (0.1 -2 mg/ml) was added to the biphasic system and the phases were allowed to settle for 21 -24 h at 25"C, the aliquots from both phases were then withdrawn and diluted with water or buffered solution. The concentration of the enzyme was determined by the enzyme activity assay (see above). The partition coefficient of enzyme preparations ( P ) in the biphasic system is defined as a ratio of the enzyme concentrations in the Ficoll-rich and the dextran-rich phases. The values of P were measured for each a-chymotrypsin preparation over approximately tenfold concentration ranges and were found to be independent of the protein concentration.
RESULTS AND DISCUSSION Nitration and amination of tyrosine residues in trypsin
Trypsin has a relatively large number of tyrosine residues on the surface of its molecule [38]. Spectral investigations have shown [39] that four tyrosines are entirely exposed to the solvent and can undergo chemical modification more readily than the other tyrosine residues in trypsin. Kenner et al. [30] showed that the treatment of trypsin with a moderate (10 200-fold) excess of tetranitromethane specifically modifies these external four tyrosines; a much greater excess of the reagent must be used to make the internal tyrosines react. Tryptophan and other amino acid residues are marginally [30, 401 affected by tetranitromethane under these conditions. The next step, i.e. the reduction of nitrotyrosines, also proceeds specifically, and no other residues in the covalent structure of trypsin [30] or other enzymes [27] are affected as a result of treatment with sodium dithionite. After the two-step modification of tyrosines in trypsin, the enzyme activity dropped to 50%. However, the addition of trypsin inhibitor (benzamidine) before the modification procedure increased the residual activity of modified enzyme up to 70 - 80% (by eliminating the autolysis).
150 Thermoinactivation of modifiedpreparations of trypsin Fig. 2 shows that kinetic curves of thermoinactivation of both the native and modified trypsin preparations are biphasic. The reason could be the inhomogeneity of commercial preparations of trypsin. They contain at least two active forms (a- and /?-trypsin) which are not structurally identical [41] and hence differ in stability [42]. The amination of tyrosine residues of trypsin slows down the inactivation of both fractions (Fig. 2). The observed stabilization is not a result of cross-linking [43] by tetranitromethane. If this were the case stabilization would appear immediately after nitration alone. Trypsin preparations, however, in which various numbers (up to four) of tyrosine residues have been transformed to nitrotyrosines are as labile as the native enzyme (Fig. 2). Moreover, using gel chromatography, we found no oligomers in the modified trypsin preparations. We also made sure that treatment of the native trypsin with sodium dithionite does not stabilize the enzyme and even slightly decreases its stability. These facts provide evidence that the stabilization is brought about only by incorporation of the amino group into tyrosine residues of trypsin. It should also be mentioned that the stabilizing effect depends exclusively on the number of aminotyrosines in trypsin and not on the state in which the other tyrosine residues of protein are (nitrated or unmodified) (Fig. 2). 100 r
' '"' ' ' ' ' ' '"01'
0
5
Dependence of the stabilizing effect on the degree ojmodfication The stabilizing effect increases with the number of aminotyrosines in trypsin (Fig. 2). As usually done [29,44], we define the stabilizing effect as a ratio of the first-order rate constants of thermoinactivation, observed for the more stable fractions (see Fig. 2) of the native and modified trypsin preparations (knat/kmOd). Then, for trypsin preparations, in which two or more surface tyrosine residues are converted into aminotyrosines, the stabilizing effect is equal to or exceeds 100, respectively (Table 1). Possible mechanism of trypsin stabilization The introduction of amino groups into the surface tyrosines does not change [30] the conformation of trypsin. This fact [30] leads us to postulate that the increase in stability (Table 1) is achieved by stabilization of the initial (native) conformation of the enzyme. As known [18, 451, the strength of hydrophobic interactions increases with temperature (at least up to 60 - 70 "C).Hence, the hydrophilization of the surface areas of the protein globule seems to hinder the formation (in the course of thermoinactivation) of new hydrophobic interactions which could upset the catalytically active conformation of the enzyme stable at low temperature only. Such a situation is schematically presented in Fig. 3. As seen from Table 1, the introduction of the NH, group into the surface layer of the trypsin globule (into a tyrosine residue) stabilizes the enzyme approximately tenfold. This stabilizing effect corresponds to an increase in about 6 kJ/mol (1.4 kcal/mol) of the activation barrier for the thermoinactivation process (the difference between the free energies of the transition and the ground states of the enzyme molecule
Table 1. Dependence of the stabilizing effect (k.,,/k,,d) on the number (n) ojaminotyrosine residues in trypsin As usual [29,44], we define the stabilizing effect as a ratio of the firstorder rate constants of thermoinactivation (kna,/kmod),observed for the more stable fractions (Fig. 2) of the native and modified trypsin preparations, respectively. The n values are determined with a precision of 10%
10
Time (h)
Fig. 2. Kinetic curves of thermoinactivation of native and modified trypsin preparations. ( A ) Unmodified trypsin; (A)trypsin modified with 2 or 4 nitrotyrosine residues; (0)trypsin modified with 1 aminotyrosine residue; ( 0 )trypsin modified with 1 aminotyrosine and 1 nitrotyrosine residue; (0) trypsin modified with 3 aminotyrosine residue. Experimental conditions: 56.5"C, pH 8.0 (3 mM KH2P04), 0.1 M KCI, initial enzyme concentration approximately 0.1 pM
n
Stabilizing effect
1 2 3
17 15 x 100 > 100
4
hydrophobic surface therrnoinactivation
-
native hydrophilic modifier
-97 u modified
Fig. 3. Schematic presentation ojpossible mechanism of protein stabilizaiion via hydrophilization of hydrophobic surface cluster
151
H2NwNH-C0
a
HOOC
CO \o
co’
0 I1
///
Fig. 4. Possible reaction of modijkation of amino groups in enzyme molecule by pyromellitic dianhydride
during its denaturation). This value corresponds well to the value of 5 kJ/mol (1.2 kcal/mol) [24] characterizing the hydrophilization increment as a result of amination of a tyrosine residue. Modification of a-chymotrypsin with cyclic anhydrides of aromatic carboxylic acids Cyclic anhydrides of aliphatic carboxylic acids (succinic, maleic, etc.) and of o-phthalic acid as well, are widely used as acylating reagents in protein chemistry [25] and, particularly, for the acylation of a-chymotrypsin [44, 461. As a result, one negatively charged carboxylic group is introduced, replacing one positively charged amino group. Hence, the total number of charged groups in the modified protein remains constant and the modification does not lead to a significant change in the hydrophilicity of the protein surface. Unlike anhydrides of aliphatic acids, modification by cyclic anhydrides of trimellitic, pyromellitic and mellitic acids (compounds C, D and E) should lead to the incorporation of 2, 3 and 5 additional carboxylic groups, respectively, into each Lys residue modified, thus significantly increasing the hydrophilicity of the protein surface. However, these reagents are seldom used in protein chemistry (for an example of modification by pyromellitic dianhydride, see [36]). Thus, the principal question is to establish the structure of the modification products. Fig. 4 shows the possible reaction pathways of protein modification by pyromellitic dianhydride. Acylation of hydroxyl groups of Ser, Thr and Tyr may proceed as well; however, the reaction products are much more labile than the modified NH2 groups [25], especially at pH 8.0 and at elevated temperatures (cf. our conditions of thermoinactivation of a-chymotrypsin preparations). Since pyromellitic anhydride is a bifunctional reagent intra- and inter-molecular cross-linking may in principle proceed simultaneously with monofunctional modification of protein (Fig. 4). However, both gel-filtration experiments on Sephadex G-75 and HPLC on TSKSW3000 have eliminated the possibility of formation of oligomers (product IV, Fig. 4).
ZOO 250 300 Wavelength (nm)
Fig. 5 . Absorption spectra of pyromellitic acid di-n-butyidiimide (I), pyromellitic acid dilysyldiamide (2), and qf cc-chymotrypsin modified with pyrornellitic dianhydride ( 3 )
Product I1 is excluded from the reaction pathways (Fig. 4) by a comparison of the ultraviolet spectra of modified chymotrypsin with some low-molecular-mass analogs. As seen from Fig. 5, there are characteristic absorption bands at 309 318 nm in the spectrum of pyromellitic acid di-n-butyldiimide (curve 1) which are absent in the spectra of both pyromellitic acid dilysyldiamide (curve 2) and modified a-chymotrypsin (curve 3). Basing on the ultraviolet spectra or gel-filtration experiments, it is impossible to discriminate product I and 111. However, compound 111 seems less probable if one considers the arrangement of lysine residues on the surface of the a-chymotrypsin molecule [28]. We supposed, therefore, that compound I is the main product of a-chymotrypsin acylation with pyromellitic dianhydride (see section below on possible mechanism of a-chymotrypsin stabilization). As a result of the modification, the molecule of a-chymotrypsin is enriched by a large number of negatively charged carboxylic groups (more than 50 in the case of pyromellitic dianhydride), which should lead to a substantial hydrophilization. The enzymatic activity after acylation by whichever derivative of aromatic carboxylic acids has never been lower than 30% in comparison with the activity of unmodified a-chymotrypsin. Evaluation of hydrophobicity of the modified a-chymotrypsin preparations To evalute the change in the hydrophilic character of a-chymotrypsin caused by its chemical modification we studied the partitioning of enzyme preparations in the aqueous biphasic polymeric system Ficoll/Dextran. In contrast to water/organic biphasic systems this system gives us a unique possibility to estimate the relative hydrophobicity of biological macromolecules under nondenaturing conditions [47]. The hydrophobicity of a biomacromolecule, determined by partitioning in the aqueous biphasic polymeric system, is usually expressed [47] by the equivalent number of CH2 groups (nCH2).For instance, if nCH2is equal to 1, the free energy of transfer of the biomacromolecule in question from
152 Table 2. Properties of modified preparations of a-chymotrypsin with nzaximal degree of modification The number of free amino groups was determined by spectrophotometric titration with picrylsulfonic acid. The conditions for determining kin were 6 0 T , pH 8.0 (1 mM KH2P04 + KOH), 0.1 M KC1, 1.6 pM a-chymotrypsin, the values for compounds C, D and E were obtained by extrapolation of the temperature dependence (Arrhenius plot) to 60°C Modifying reagent
Number of COOH groups replacing NHZgroups
None Benzoyl chloride (A) o-Phthalic anhydride (B) Trimellitic anhydride (C) Pyromellitic anhydride (D) Mellitic anhydride (E)
0 0
Relative hydrophobicity of modified enzyme, nfHl
r
Time (rnin)
kin
SC1
1
+
3.2+ 1.9 +25.4 3.7 -18.5+3.2
2
-
3 5
-72.3 f 4.0 -67.0 f 4.0
1 . 5 ~ 1 0 - ~ Fig. 6. Thermoinuctivation kinetics of a-chymotrypsin modified by 3.3 x pyrom~lliticdianhydride. A and A,, are the current and initial values of 5 . 6 ~ 1 0 - ~enzyme activity, respectively. Experimental conditions: 98 "C, pH 8.0 4.6 x ( 5 mM KHzPO,), 0.1 M KCl, initial enzyme concentration ap4.6 x proximately 1 pM 4.6 x
the hydrophilic to the hydrophobic phase is the same as for the transfer of the CH2 group. Table 2 shows the relative hydrophobicities of the modified preparations of a-chymotrypsin. As can be expected, the introduction of 12 nonpolar phenyl groups into the enzyme molecule brings about an appreciable hydrophobization (see the data for benzoyl chloride). On the other hand, the appearance of a great number of additional COOH groups on modification (approximately SO in the case of pyromellitic dianhydride) leads to significant hydrophilization. The extent to which the relative hydrophilic character of the protein surface is changed by chemical modification is really impressive. As a rule, structurally related proteins ( e g homologous proteins from different sources) differ in their nCH, values within a 12-fold range only [47]. For modified preparations of a-chymotrypsin the range of variations of nCH, units is much more pronounced, about 100-fold (see Table 2). It can also be seen from Table 2 that the relative hydrophilic character of the protein surface increases gradually with the number of carboxylic groups introduced. The only exception is a-chymotrypsin acylated with mellitic trianhydride which is less hydrophilic than the enzyme modified with pyromellitic dianhydride. The possible reason is that not all available amino groups of the protein can be acylated by mellitic trianhydride, since the shield of the negatively charged carboxylic anions already introduced into the protein makes some amino groups inaccessible both to the modifying agent and the titrant (trinitrobenzene sulfonic acid). Thermoinactivation ojmodified preparations oja-chymotrypsin
Fig. 6 shows thermoinactivation of a-chymotrypsin modified by pyromellitic dianhydride. It is clearly seen that irreversible thermoinactivation follows first-order kinetics. We can use therefore the first-order rate constant of irreversible thermoinactivation as a measure of the enzyme stability. Fig. 7 shows the temperature dependence (Arrhenius plot) of the rate constants of irreversible thermoinactivation for the native (curve 1) and modified (by pyromellitic dianhydride) chymotrypsins. The difference in thermostabilities is so great
.-c -
Y
m
-3.L
4.2 Temperature l°C)
Fig. I . Temperature dependence (Arrhenius plot) of thefirst-order rate constants of therrnoinactivution (kin) for native a-chymotrypsin and the enzyme modified with pyromellitic dianhydride. For experimental conditions, see Fig. 6
that it is practically impossible to compare the native and the modified enzymes under the same experimental conditions. We have therefore compared the extrapolated data (Fig. 7, dashed line) corresponding to some intermediate temperature, e.g. 60°C. The stabilizing effect (equal to the ratio of the rate constants of thermoinactivation of the native and the modified enzymes) is then about 300. Dependence of the stabilizing eifect on the structure o j the modiyying reagent
The lower column in Table 2 shows the rate constants of thermoinactivation for the native and modified enzymes at 60 "C. Modification of a-chymotrypsin with benzoyl chloride (the positive charge of each amino group is eliminated by modification) slightly decreases the stability. In a similar manner, acylation with o-phthalic anhydride (one negatively charged carboxylic group is introduced instead of one positively charged amino group) hardly increases the stability. The introduction of several carboxylic groups (2,3 or 5) into each amino group in the modification leads to dramatic stabilizing effects.
153 32
r
-0
4
2
6
a
CONCLUSION
In this study we succeeded in 100- 1000-fold stabilization of trypsin and a-chymotrypsin by chemical modification (hydrophilization). Such a large stabilizing effect has been obtained previously only be covalent multipoint binding of enzymes to polymeric supports [29, 481. The 1000-fold stabilizing effects are the highest ever achieved by treatment of protein with low-molecular-mass compounds; for review, see [5, 7, 8, 51, 521. The termostability of a-chymotrypsin preparations obtained by their modification with cyclic anhydrides of aromatic acids is practically equal to the stability of proteolytic enzymes from extremely thermophilic bacteria [53, 541 the most stable proteinases known to date.
PH
Fig. 8. p H dependence of the first-order rate constants of thermoinactivation (kin)for native a-chymotrypsin (at 50°C) and the enzyme modgied with pyromellitic dianhydride (at 80°C). For other experimental conditions, see Fig. 6
Possible mechanism of a-chymotrypsin stabilization The stabilization of a-chymotrypsin studied is not a result of intermolecular cross-linking, since oligomeric enzyme species are not formed in the course of the modification. Trimellitic anhydride is a monofunctional reagent which is not able to cross-link proteins at all. Nevertheless, the modification brings about the same stabilization of a-chymotrypsin as in the case of pyromellitic and mellitic anhydrides, which are potentially bifunctional reagents. Therefore, the stabilization is not a result of intramolecular cross-linking of a-chymotrypsin as well (compound 111 in Fig. 4). We assume that the most probable stabilization mechanism is the hydrophilization of the nonpolar surface areas of the protein as a result of covalent modification. First, the stabilizing effect depends greatly on the relative hydrophobicity of the modified enzyme preparations, see Table 2. It is highly probable that the enormous hydrophilization of the surface layer in an a-chymotrypsin molecule suppressed irreversible conformational changes in the enzyme (Fig. 3), in analogy to a-chymotrypsin covalently immobilized in polyacrylamide gel [48]. A slow inactivation of the modified enzyme at elevated temperatures may be due to one of the chemical processes stated by us before [48] and studied recently in detail by Klibanov [49, 501. To verify this assumption further studies are necessary. Second, the stability of a-chymotrypsin modified by pyromellitic dianhydride has a characteristic pH-dependence (Fig. 8): at pH < 3, which is close to the pK, values of carboxylic groups of pyromellitic acid, the apparent value of the first-order rate constant of thermoinactivation of the modified enzyme increases abruptly (the pH dependence of the rate constant of thermoinactivation of the native a-chymotrypsin is shown in the same figure for comparison). The decrease of therinostability of the modified enzyme at acidic pH is not a result of deacylation of modified lysine residues in the course of thermoinactivation, since we were not able to detect regeneration of free amino groups by spectrophotometric titration with trinitrobenzenesulfonic acid. Thus, we suppose that the decrease in the stabilizing effect at low pH values results from the decrease in hydrophilicity of the COOH group (and, therefore, of the protein surface) caused by its protonation.
REFERENCES 1. Mozhaev, V. V. & Martinek, K. (1982) Enzyme Microb. Technol. 4,299- 309. 2. Klibanov, A. M. (1983) Adv. Appl. Microbiol. 29, 1-23. 3. Martinek, K., Klibanov, A. M. & Berezin, I. V. (1977) J . SolidPhase Biochem. 2, 343 - 371. 4. Klibanov, A. M. (1979) Anal. Biochem. 92, 1-25. 5. Schmid, R. D. (1979) Adv. Biochem. Engin. 12,41-I 18. 6. Mozhaev, V. V. & Martinek, K. (1984) Enzyme Microb. Technol. 6, 50- 59. 7. Martinek, K. & Mozhaev, V. V. (1985) Adv. Enzymol. 57, 179245. 8. Mozhaev, V. V., Berezin, I. V. & Martinek, K. (1988) CRC Crit. Rev. Biochem., in the press. 9. Brock, T. D. (1985) Science (Wash. D C ) 230, 132-138. 10. Creighton, T. E. (1983) Proteins, Freeman, New York. 11. Lee, B. & Richards, F. M. (1971) J . Mol. Bid. 55, 379-400. 12. Chothia, C. (1984) Annu. Rev. Biochem. 53, 537-572. 13. Tanford, C. (1980) The hydrophobic ejyects: Formation of micelles and biological membranes, 2nd edn, Wiley, New York. 14. Burley, S. K. & Petsko, G. A. (1985) Science (Wash, DC) 229, 23-28. 15. Srere, P. A. (1984) Trends Biochem. Sci.9, 387-390. 16. Stellwagen, E. (1984) in Enzyme engineering (Laskin, A. I., Tsao, G. T. & Wingard, L. B. Jr, eds) vol. 7, pp. 1- 8, New York Academy of Sciences, New York. 17. Kauzmann, W. (1959) Adv. Protein. Chem. 14, 1-64. 18. Scheraga, H. A. (1963) in The proteins (Neurath, H., ed.) 2nd edn, vol. 1, ch. 6 , Academic Press, New York. 19. Argos, P., Rossmann, M. G., Grau, U. M., Zuber, H., Frank, G. & Tratschin, J. D. (1979) Biochemistry 18, 5698-5703. 20. Zuber, H. (1981) in Structural and functional aspects of enzyme catalysis (Eggerer, H. & Huber, R., eds) pp. 114- 127, Springer, Berlin. 21. Walker, J. E. (1978) in Proteins: Structure,function and industrial applications (Hoffmann, E., ed.) pp. 211 -225, Pergamon, Oxford. 22. Griitter, M. G. & Hawkes, R. B. (1983) Naturwissenschaften 70, 434-438. 23. Leo, A,, Hansch, C. & Elkins, D. (1971) Chem. Rev. 71, 525616. 24. Hansch, C. & Leo, A. (1979) Substituent constants for correlational analysis in chemistry and biology, Wiley, New York. 25. Lundblad, R. L. & Noyes, C. M. (1985) Chemical reagents for protein rnodijkation, vols 1-2, CRC Press, Boca Raton. 26. Sokolovsky, M., Riordan, J. F. & Vallee, B. L. (1967) Biochem. Biophys. Res. Comrnun. 27, 20 - 25. 27. Riordan, J . F. & Vallee, B. L. (1972) Methods Enzymol. 25, 515 521. 28. Birktoft, J. J. &Blow, D. M. (1972) J. Mol. Bid. 68, 187-240. 29. Martinek, K., Klibanov, A. M., Goldmacher, V. S. & Berezin, I. V. (1977) Biochim. Biophys. Acta 485, 1 - 12. 30. Kenner, R. A,, Walsh, K. A. & Neurath, H. (1968) Biochem. Biophys. Res. Comrnun. 33,353 -360.
31. Cole, R. D. (1967) Methods. Enzymol. 11,206-208. 32. Jellinek, K. (1911) Das Hydrosulfit, Verlag von Ferdinand Enke, Stuttgart. 33. Riordan, J. F., Sokolovsky, M. & Vallee, B. L. (1967) Biochemistry 6, 357 - 363. 34. Fields, R. (1971) Biochem. J. 124, 581 -589. 35. Riordan, J. F. &Vallee, B. L. (1967) Methods. Enzymol. 11,565570. 36. Bagree, A. & Sharma, I. K. (1980) FEES Lett. 120,275-280. 37. Zaslavsky, B. Yu., Mikheeva, L. M., Mestechkina, N. M. & Rogozhin, S. V. (1982) J. Chromatogr. 253,139-143. 38. Stroud, R. M., Kay, L. M. & Dickerson, R. E. (1974) J. Mol. B i d . 83, 185-208. 39. Villanueva, G. B. & Herskovitz, T. T. (1971) Biochemistry 10, 3358 - 3365. 40. Vincent, J . P., Lazdunski, M. & Delaage, M. (1970) Eur. J. Biochem. 12,250-258. 41. Schroeder, D. D. & Shaw, E. (1968) J. Biol. Chem. 243, 29432949. 42. Beardslee, R. A. & Zahnley, J. C. (1973) Arch. Biochem. Biophys. 158,806 - 81 1. 43. Doyle, R. J., Bello, J. & Roholt, 0. A. (1968) Biochim. Biophys. Acta 160, 274-276.
44. Torchilin, V. P., Maksimenko, A. V., Smirnov, V. N., Berezin, I. V., Klibanov, A. M. & Martinek, K. (1979) Biochim. Biophys. Acta.567, 1-11. 45. Brandts, J. F. (1964) J. Am. Chem. Soc. 86,4302-4313. 46. Shiao, D. D. F., Lumry, R. & Rajender, S. (1972) Eur. J. Biochem. 29,377-385. 47. Zaslavsky, B. Yu., Mestechkina, N. M. & Rogozhin, S. V. (1983) J. Chromatogr. 260, 329 - 332. 48. Mozhaev, V. V., &kinis, V. A,, Torchilin, V. P. & Martinek, K. (1983) Biotechnol. Bioeng. 25,1937-1945. 49. Ahern, T. J. & Klibanov, A. M. (1985) Science (Wash. DC) 228, 1280- 1284. 50. Zale, S. E. & Klibanov, A. M. (1986) Biochemistry 25, 54325444. 51. Torchilin, V. P. & Martinek, K. (1979) Enzyme Microh. Techno1 1, 74-81. 52. Kutuzova, G. D., Ugarova, N. N. & Berezin, I. V. (1984) Usp. Khim. 53, 1852- 1894. 53. Ohta, Y. (1967) J. Biol. Chem. 242, 509-515. 54. Taguchi, H., Hamaoki, M., Matsuzawa, H. & Ohta, T. (1983) J. Biochem. (Tokyo) 7,93- 101.
