BIOTECHNOLOGY AND BIOENGINEERING, VOL. XVIII, PAGES 105-118 (1976)
Studies on Immobilized Trypsin in High Concentrations of Organic Solvents HOWARD H. WEETALL and WILLIAM P. VANN, Corning Glass Works, Research and Development Labs, Corning, New York 14830
Summary Trypsin was covalently immobilized on porous glass in the presence and absence of a specific substrate and reacted in various organic solvents of different dielectric constants. Optimum solvent concentration, pH profile, K , (app)r V,,, (app), productivity versus temperature, activity, and reaction rates were determined. Reaction rates of six lysyl dipeptides were compared. Crystalline trypsin was dansylated for studies by nanosecond fluorescence techniques to determine the effects of introducing high concentrations of organic solvents on the molecule. The results indicated that greater reaction rates were observed with dipeptides having more acidic carboxyl terminal groups. The data also indicated that greater reaction rates were observed in higher concentrations of solvents of lower dielectric constants. Nanosecond fluorescence spectroscopy of trypsin in high concentrations of a low dielectric constant solvent indicated major dehydration even though maximal enzyme activity was achieved under these conditions.
INTRODUCTION It is well known that the rate of a chemical reaction involving ions or dipolar molecules can be modified by a change in the dielectric constant or the ionic strength of the reaction medium. Several methods have been applied to account for the solvent effects on reaction rates as it relates to the dielectric constant of the medium. These studies have been reviewed in great To our knowledge all the studies reported utilizing organic solvents with proteolytic enzymes have been, due to the nature of protein, limited to relatively low concentrations of solvents (usually below 50% by volume). By using enzymes immobilzed on inorganic support materials we have been able to extend the solvent concentration range studied to greater than 90% by volume. For our initial studies on the solvent effects on immobilized enzymes we have chosen trypsin as our model enzyme. Our choice 105
@ 1976 by John Wiley & Sons, Inc.
106
WEETALL AND VANN
of trypsin was prompted by the large amount of information on this enzyme and some results recently p~blished.~It was shown that if trypsin is immobilized on controlled pore glass in the presence of saturating quantities of benzoylarginine-ethylester (BAEE) , several changes in the kinetic parameters of the immobilized enzyme occur, including decreased K , and increased stability. The authors attribute these changes to an enzyme conformational change, which occurs in the presence of the substrate and becomes frozen during the immobilization process. This study describes the resulting effects on several kinetic parameters using trypsin immobilized on porous glass, in the presence and absence of BAEE, and when exposed to high concentrations of organic solvents of differing dielectric constants.
MATERIALS AND METHODS All reagents used in this study were reagent grade or better. The porous glass and porous glass derivatives are available from Pierce Chemical Company, Rockland, Illinois.
Preparation of Controlled Pore Glass Derivatives Controlled pore glass (CPG) was silanized with y-aminopropyltriethoxy~ilane.~
Immobilization of Trypsin
Without substrate (E) Ten grams of alkylamine 550 A, 40/80mesh CPG glass was added to 100 ml of 2.5% glutaraldehyde in 0.1M phosphate buffer, pH 7. The mixture was reacted under vacuum for 1 hr with occasional stirring and was then washed with 2 liters of distilled water. One gram of crystalline trypsin (3326 N.F. units/mg) was dissolved in 10 ml of 0.1M phosphate buffer, pH 7.0, and added to the CPG. The reaction was continued for 2 hr and the mixture was washed with 2 liters of distilled water and stored in an airtight container at 5°C. With substrate (ES) Conditions were as described above except that the CPG, after being reacted with glutaraldehyde and washed, was placed in a column filled with distilled water. One gram of benzoylarginineethylester (BAEE) was added to 5 ml of 0.1M phosphate buffer, pH 7.0. The BAEE and trypsin solutions were added in increments
IMMOBILIZED TRYPSIN I N ORGANIC SOLVENTS
107
and allowed to flow through the column. The immobilized enzyme (IME) was then removed from the column and washed as described above.
Determination of Immobilized Trypsin Activity Activity for the coupled E and ES preparations was determined by hydrolysis with BAEE.
Determination of Optimum Solvent Concentration One milliliter of an aqueous dipeptide solution containing 10-2M L-lysylglycine (lys-gly) (Sigma Chemical Company) was added to 9.0 ml of a solvent-distilled water mixture containing 0.01M CaClz so that final solvent concentrations were between 10 and 9570. Final dipeptide concentration was 10-3M. Trypsin IME, either the E or ES preparation (150 mg wet wt), was added to each flask and the pH was adjusted to pH 7.0 with dilute NaOH. Samples were sealed, placed on a mechanical shaker bath, and incubated for 72 hr at 23°C. After 72 hr, 0.5 ml was removed and dried under a heat lamp followed by three distilled water washing evaporation cycles. Final suspension was made in pH 2.2 acetate buffer. The samples were centrifuged and then 20 p1 of each sample was loaded into amino acid analyzer sample holder units. Controls for all experiments described in this study consisted of separate trypsin IME samples of identical weights and samples of equal dipeptide concentrations both reacted under appropriate conditions. Hydrolysis was monitored by the loss of L-lysylglycine. The amino acid analyses were carried out on a Durrum D-500 Analyzer (Palo Alto, California). Solvents used in these studies were : DMSO, methanol, ethanol, 1-propanol, and acetone. p H Profile in Ethanol
A pH profile of IME activity from pH 4 through 10 was prepared by adding 100 mg wet wt immobilized trypsin to each of seven flasks containing 10 ml of 76% ethanol buffer (v/v), 0.01M CaCL, and 10-3M lys-gly. The buffers were previously adjusted to the desired pH values. Controls contained only immobilized trypsin or dipeptides at the appropriate pH and ethanol concentrations. Samples were placed on a shaker bath and reacted for 120 hr at 23°C and then they were removed and processed as previously described. Since pH values as measured by a pH meter may not be valid in 76% ethanol solutions, one must simply view these values as relative differences in H+ and OH- concentrations.
108
WEETALL AND VANN
Seventy-six percent ethanol was used for this and most of the other studies because this concentration was found to give the greatest hydrolysis rate for L-lysylglycine. Similar studies were carried out in the other solvents.
Effect of Temperature o n Reaction Rate L-Lysylglycine (10-3M) was reacted a t 5, 10, 23, 30, and 40°C with 200 mg wet wt immobilized trypsin in 10 ml of 76% ethanol containing 0.01M CaCI2. The pH was adjusted to 7.0 and the samples were sealed and placed on a shaker bath for 72 hr. Samples were handled as previously described. Half-Life of Immobilized T r y p s i n in Ethanol Two immobilized trypsin columns were operated for 240 hr and 196 hr, respectively. Each column contained 1.1 g wet wt (0.5 g dry) of immobilized enzyme (IME). The substrate was prepared by adding 0.38 g p-toluenesulfonyl-L-arginine methyl ester (TAME) to 200 ml 0.046M Tris buffer containing 0.01M CaC12. The substrate solution was brought to 76% with 95% ethanol and the solution was adjusted t o p H 7.5. Flow rates were initially adjusted to indicate a n OD A,,, of about 0.8. Apparent K , and V,,, Determination for Immobilized T r y p s i n in Ethanol Apparent K , and V,,, were determined using L-lysylglycine as substrate. The substrate concentration ranged from 2.5 x 10-4M5 X 10-2M in 76y0 ethanol containing 0.01M CaCL Forty-five milligrams (dry wt) of immobilized trypsin derivative was added and the pH was adjusted to 7.0. The samples were then sealed and incubated with shaking for 24 hr. First Order Reaction Rates The first order rate constants were calculated from data obtained by continually sampling the hydrolyzing samples versus time. The substrate concentration was lOWM. Over the course of a n experiment no more than a total of 15y0of the substrate volume was used for sampling.
Additional Dipeptides Hydrolyzed in Ethanol Thirteen dipeptides (5 X 10-3M) were individually reacted with immobilized trypsin in 76y0 ETOH, pH 7.5, 23°C. Hydrolysis,
IMMOBILIZED TRYPSIN I N ORGANIC SOLVENTS
109
if any, and reaction rates of lysyl dipeptides was then determined. The reactions were terminated after 24 hr.
Polarization of Fluorescence Crystalline trypsin was dansylated to determine the effects on the molecule by introducing high concentrations of organic solvents. Since maximum activity was observed in 1-propanol a t a concentrapropanol, it was decided to use this solvent. The tion of dansylated trypsin was examined in pure water and in 90% 1-propanol (v/v) . The degree of polarization of fluorescence and the lifetimes were measured using a nanosecond fluorescence spectrometer. The sample was excited by polarized radiation a t 350 nm f 10 nm and the emitted single photons were counted a t 460 nm f 7 nm. The details of the equipment are presented e1sewhe1-e.~The experimental error for the polarization and lifetime values are about f5YO.
RESULTS Activity of Immobilized T r y p s i n The activity of the immobilized trypsin was determined by assay with BAEE. The soluble enzyme gave approximately 1.8 X lo5 BAEE units/g. Table I compares the activity with the total protein coupled AS determined by ninhydrin determination on the acid hydrolyzed derivatives. Based on activitylmg of protein coupled, the ES shows 260y0 greater activity than the E preparation.
Determination of Optimal p H in Organic Solvents The immobilized trypsin was studied in 76% ethanol. Maximum activity for the immobilzed enzyme was observed at pH 8.0 and TABLE I Preparation Eb ESc
Activity" Protein (IU/g IME) (mg/g IME)
489 1250
Activity units represent BAEE units. E, coupled without substrate present. ES, coupled with substrate present.
14.7 14.0
I10
WEETALL AND VANN 100 -
90 00
-
70
-
E
ul
60-
0
a
50I
* 40-
4
5
6 7 PH
8
9
Fig. 1. pH profile of immobilized trypsin prepared in the presence of BAEE ES and in the absence of BAEE E. The assays were carried out at room temperature in 76% ethanol.
fell off sharply above and below this value (Fig. 1). A screen of other solvents produced similar pH optima. Most of the further characterization was performed in ethanol.
Effect of Temperature on Reaction Rate Complete inactivation of the immobilized trypsin was observed at 40°C after incubation with l F 3 M lys-gly in 76% ethanol (Fig. 2). Because we observed no hydrolysis in the 40°C sample, we must assume that inactivation was very rapid. Maximum productivity was observed at 10°C. Immobilized trypsin preparations when reacted in an aqueous system gave maximum activity at 50-55"C.4 A t these temperatures, immobilized trypsin in 76y0 ethanol was totally inactive.
Determination of K , and V,,, in 76% Ethanol Table I1 gives the kinetic values obtained by Lineweaver-Burk plots for both the E and ES preparations. Results indicate that the K , values for both preparations are two orders of magnitude greater than that observed in aqueous solution. However, as is the
IMMOBILIZED TRYPSIN IN ORGANIC SOLVENTS
111
TEMPERATURE ("C)
Fig. 2. Substrate hydrolyzed by immobilized trypsin in 76% ethanol, pH 7.0, 72 hr a t increasing temperature. Substrate: blysylglycine 10-3M.
case for K , values in water, the ES preparation gives the smaller value. In the case of V,,, it is quite apparent that the ES preparation shows the greater Vma,.
Determination of Optimal Solvent Concentration A series of experiments were run at pH 7 using solvents of increasing dielectric constant to determine the optimal concentration for hydrolysis versus the dielectric constant of the solvent. These TABLE I1 Apparent K , and Vmaxfor E and ES Immobilized Trypsin
K,
E ES Eb ESb
Vrn,X8
(W
(pmol/min)
Solvent
0.025 0.011 0.0041 0.0024
1.2 4.5 -
76% Ethanol 76% Ethanol Water Water
These values were calculated using 45 mg of immobilized trypsin derivative. These values taken from ref. 9. ES preparation was coupled in the presence of saturating concentration of BAEE. 8
b
WEETALL AND VANN
112
30t 7
E
20
E 0
-b Y
15
10
05
0
0
20
30 40 50 60 70 00 90 SOLVENT CONCENTRATON (%V/V)
100
Fig. 3. Relationship between solvent concentration and first order rate constants compared t o water for the E preparations a t room temperature pH 7.0. Solvent: 0 DMSO; 0, methanol; A, acetone; 0,propanol; X, ethanol, +--+, water. Substrate: clysylglycjne.
TABLE I11 Comparison of Optimum Solvent Concentration for Hydrolysis with Dielectric Constant and First Order Rate Constants
Solvent at maximum rate Solvent
(%)
Water DMSO Methanol Ethanol Acetone 1-Propanol
100 50 60
76 80 92
Dielectric constant of solution
Dielectric constant of pure solvent
First order rate constants a t maximum rate K(lO-4 min-1)
76.0 57.7 41.9 43.5 29.3
78.0 45.0 32.6 24.3 20.7 20.1
0.95 0.91 1.91 2.53 2.41 3.04
IMMOBILIZED TRYPSIN IN ORGANIC SOLVENTS
113
solutions were compared with each other as shown in Figure 3. It was observed that for solvents of lower dielectric constant, the optimum for hydrolysis was a t greater solvent concentrations. These results are summarized in Table 111. The E preparation was used in all these studies. The actual dielectric constant of the solution at optimal concentration was calculated. These data are also given in Table 111.
First Order Rate Constants Figure 3 shows the relationship between solvent concentration and first order rate constants compared to water for the E preparations. All parameters including enzyme concentration and pH were held constant. The data indicated that at optimum solvent concentration only DMSO was lower than water. Solvents with lower dielectric constants had higher rate constants (Table 111). Half-Lives of Immobilized Trypsin Figure 4 shows regression analysis half-life data for trypsin columns operated at 10 and 23°C in 95% ethanol. As expected, the longest half-life observed (in hours) was at the lower temperature. 8-
\
7-
->
6-
2
5-
a
t m a a
-ln
4-
t
z 3 >
5 35a
2
U
I
l
l
l
l
l
l
l
l
l
l
20 40 60 80 100 120140 160 160 200220240 TIME (HOURS)
Fig. 4. Half-lives of immobilized trypsin using TAME (0.001M) in 76% ethanol as a substrate, pH 7.5; 23°C Half-life: T3,83.9; (0.95) UCL, 128.5 (UCL and LCL: upper and lower confidence limits; SE: standard error); (0.95) LCL, 62.3; SE, 0.1. 10°C Half-life: T3, 269.2; (0.95) UCL, 441.6; (0.95) LCL, 193.7; SE, 0.05.
114
WEETALL AND VANN TABLE IV Rates of Lysyl Dipeptide Hydrolysis ~
Dipeptide" Lys-Asp Lys-Gly Lys-Val Lys-Leu Lys-Phe
Charge of C-terminal amino acid
% hydrolysis
Uncharged Nonpolar Least hydrophobic Nonpolar More hydrophobic Nonpolar Most hydrophobic
~~
~~
Relative rate compared t o Lys-Gly
27.1 18.5
146 100
10.5
57
7.3
39
5.1
28
The following dipeptides were not hydrolyzed when reacted with immobilized trypsin: Met-Ah, Leu-Met, Tyr-Tyr, Phe-Leu, Phe-Tyr, Tyr-Leu, and Tyr-Lys.
Reaction Rates of Lysyl Dipeptides Table IV compares lysyl dipeptide hydrolysis with the charge of the carboxyl terminal amino acid residue. Substrate hydrolysis was greater for the negatively charged carboxyl terminal residue (Asp) and decreased progressively as the charge on the carboxyl terminal residue went from uncharged (Gly) to nonpolar and most hydrophobic (Phe).
Polarization of Fluorescence The results of measuring the degree of polarization P of the aqueous and 90% 1-propanol solutions containing DNS-trypsin indicated nearly identical values, 0.200 f 10 and 0.202 f 10, respectively. The value of P for a spherical molecule is given by p 1 - 31= ( $ - ; ) ( l + n v RT
t
)
where R is the gas constant, T is the absolute temperature, 9 is the viscosity of the solvent, V is the molecular volume of the fluorescent molecule, and T is the lifetime of the excited state of the fluorescence.6 Pois the value of P when ( T / q )+ 0, that is, when no depolarization by molecular rotation occurs. Thus, if the polarization of the same protein measured in different solvents is nearly identical, the ratio of the molecular volumes in the two solvents should follow the relation: v2 -
v1
- 7" 91 - x71
92
IMMOBILIZED TRYPSIN IN ORGANIC SOLVENTS
115
The average fluorescence lifetimes of DNS-trypsin were found to be 10.3 nsec and 8.7 nsec in 90% n-propanol and aqueous solution, respectively. The viscosity ratio of the propanol to the aqueous DNS-trypsin solution obtained using the Ostwald method was 2.6. Therefore, the ratio of molecular volumes from eq. (2) is found to be
( V ) 90% propanol
=
0.45 ( V ) HzO
(3)
This implies that the molecular volume of the DNS-trypsin in 90% propanol is about 3 the molecular volume in the aqueous solution. The molecular volume of a protein in aqueous solution is related to the molecular weight M , specific volume v (in cm3/g), and hydration h (cm3of water per gram weight of marcromolecular substance) by the equation:
V =
M(v
+ h) N
(4)
where N is Avogadro’s number. Generally, v and h are assumed to be 0.75 and 0.2, respectively.’ Thus, the hydration itself contributes about 20y0 to the molecular volume of proteins in aqueous solutions. This means that if the DNS-trypsin lost all its hydration going from water to 90% 1-propanol, eq. (4) requires an additional 25% reduction in the molecular volume of DNS-trypsin.
DISCUSSION Although the total protein coupled to the E and ES preparations was similar (Table I), the ES preparation gave twice the activity. Our work and that of Royerss9indicates that coupling trypsin in the presence of saturating concentrations of substrates such as BAEE, induces a conformational change which freezes or traps the enzyme in this configuration. This permanent change is reflected in K,, V,.,, and in the comparative rate constants for the E and ES preparations. The K , values for the E and ES preparations in both water and 76% ethanol were similar to Royer’s data9in that K , decreased upon immobilization in the presence of BAEE by approximately 5 0 z . The increased reactivity of the ES preparation was also observable in V,,,. Since enzyme molecules possess ionizable groups, some of which play an important role in the catalytic function, it is expected that the reaction rate of the enzymes might be affected by both the dielectric constant as well as the ionic strength of the reaction medium. Benderlo explained the effect of nucleophiles such as methanol,
116
WEETALL AND VANN
ethanol, and hydroxylamine on a-chymotrypsin catalyzed reactions in terms of competition of the solvent and water for the acyl-enzyme intermediate compound. For aprotic solvents Clement and Bender" proposed dielectric effects combined with competitive inhibition. Laidler12has predicted that an enzymatic reaction will be affected by a n increase in dielectric constant depending on the polarity of the activated complex. It has been noted that trypsin catalyzed hydrolysis of BAEE within a given range of dielectric constant is a linear function when one plots the log of the rate constant versus the reciprocal of the dielectric constant (0).13 The linearity is in accordance with the expressions derived by Amis'* from Coulombs law and the Arrhenius equation to account for three types of chemical reactions : ion-ion, ion-dipole, and dipole-dipole. This being the case, we would expect to observe even greater differences with larger changes in the dielectric constant of the medium. Calculation of the first order rate constants (Table 111) shows the stimulating effect of each solvent a t the optimal concentration. All solvents studied with the exception of DMSO showed hydrolytic rates greater than that of water a t the solvent optimum. The relative rates appear to be related to the dielectric constant of the reaction medium. The observation that there is a solvent optimum concentration for each solvent studied (Table 111, Fig. 3) and that increasing optimal concentrations and reaction rates in these solvents appear to be related to decreasing dielectric constant led us to consider the possibility that if charge interaction between the substrate and enzyme was the important factor, at dielectric constants below that of water, we might expect to see some effect if we varied the charge on the substrate. Since trypsin is rather specific in its ability to hydrolyze peptides, we chose to look a t lysyl peptides in which the second amino acid residue (carboxyl terminal) was varied from acid to base. The reaction rates of five lysyl peptides (Table IV) were compared and indicated that a direct relationship does indeed exist between the substrate and enzyme with respect to overall charge. The reaction rates were directly proportional to the relative acidity of the carboxyl terminal amino acid, the most acidic carboxyl terminal amino acid showing the greatest reactivity and the most basic one, showing the least. It is obvious that the substrate charge strongly affects the enzyme activity. It has been suggested that there are two sites involved in
IMMOBILIZED TRYPSIN IN ORGANIC SOLVENTS
117
the catalytic process in trypsin. One is the active site containing the reactive serine and imidazole groups. The other is a noncatalytic site which is involved in determining the polarization and the fit of the substrate. It has been observed that if the substrate acetylglycine ethyl ester (AGEE) is substituted for BAEE, a 1000-fold decrease in trypsin activity is observed.'"17 By adding a methylguanidinium ion to the reaction mixture a sevenfold increase in activity is observed over that of AGEE alone. Inagarnil' claims this indicates a noncatalytic site is involved in the catalytic process and he assigns a net negative charge to this site (Fig. 5). The data presented in Table IV would indicate that if there is a noncatalytic site it must be positively charged to account for the observations. There are, however, other alternative hypotheses for the observed results. The imidazole residue in the active site is a positively charged molecule. The organic solvents, as already stated, increase the ionic interaction between the substrate and enzyme by decreasing the insulating effects of water. The observed results could be due to the direct interaction between the positive charged group in the active site and the substrate, and thus have nothing to do a t all with the secondary site. The polarization of fluorescence data does indicate that a drastic change in the tertiary structure of the molecule occurs. I n fact, in 90% 1-propanol the trypsin is for the most part completely dehydrated and loses a large portion of its volume. This means that both polar and nonpolar portions of the molecule are 0
II
H2N>C= H2N
RCN o I II NHCH2 CH2 CH2 CHCOEt
2
Fig. 5. Schematic representation of Inagami's hypothesis concerning a noncatalytic site on trypsin.17 (1) Benzoylarginine ethyl ester in proposed catalytic and noncatalytic sites of trypsin. (2) Acetylglycine ethyl ester and methylguanidinium group in proposed catalytic and noncatalytic sites of trypsin.
118
WEETALL AND VANN
brought into closer contact with each other and with the substrate. The immobilization process obviously prevents complete collapse of the tertiary structure, thus allowing the molecule to retain activity. Although we cannot at this time distinguish between the many possible mechanisms at work in the systems studied, it does seem obvious that with dehydration by the solvents, the greater activities observed at lower dielectric constants, and the definite effectof charge on the carboxyl terminal amino acid of the lysyl peptides all point to direct interaction between the charge or polarity of the substrate and the charge or polarity in or near the active site of the enzyme. It is our belief that the use of an organic solvent system for the study of immobilized enzymes will permit researchers to probe the mechanisms of enzyme action even further than is possible at present. The authors acknowledge the kind assistance of Dr. L. S. Hersh for his work on the polarization of fluorescence and for the many discussions they had with him concerning this work.
References 1. L. M. del Castillo and M. Casteneda-Agullo, NCZ Monograph, No. 27, M. P. Stulberg, Ed., pp. 141-152. 2. E. S. Amis, Solvent EfJkcts on Reaction Rates and Mechanisms, Academic Press, New York, 1966. 3. G. Royer and R. Uy, J . Biol. Chem., 248, 3278 (1972). 4. H. H. Weetall and A. M. Filbert, Methods i n Enzymology, vol. 34, M. Wilcheck, Ed., Academic Press, New York, 1974, pp. 59-72. 5. L. Hersh and V. Ortabasi, Biopolymers, in press. 6. F. Perrin, J . Phys. Radium, 7 , 390 (1926). 7. c. Tanford, Physical Chemistry of Macromolecules, Wiley, New York, 1961, Ch. 6. 8. G. Royer and R. Uy, J . Biol. Chem., 248, 3278 (1973). 9. G. Royer in Immobilized Enzymes, Antigens, Antibodies and Peptides, H. H. Weetall, Ed., Marcel Dekker, New York, 1975. 10. M. L. Bender, G. E. Clement, C. R. Gunter, and F. C. Kezdy, J . Amer. Chem. SOC.,86, 3697 (1964). 11. G. E. Clement and M. L. Bender, Biochemistry, 2, 836 (1963). 12. K. J. Laidler, Faraday SOC.Discussions, 20, 83 (1965). 13. M. Castaneda-Agullo and L. M. del Castillo, J . Gen. Physiol., 42, 617 (1959). 14. E. S. Amis, J . Chem. Educ., 29, 337 (1952); and J . Chem. Educ., 30, 351 (1953). 15. T. Inagmi, J . Biol. Chem., 240, PC3463 (1965). 16. W. P. Jencks, Catalysis i n Chemistry and Enzymology, McGraw-Hill, New York, 1969, pp. 298-99. 17. T. Inagmi and H. Hatano, J . Biol. Chem., 244, 1176 (1969).
Accepted for Publication October 14, 1975