PROTEINS: Structure, Function, and Genetics 24:370-378 (1996) Experimental and Theoretical Study of Electrostatic Effects on the Isoelectric pH and the pK, of the Catalytic Residue His-102 of the Recombinant Ribonuclease From Bacillus amyloliquefaciens (Barnase) Katrin Bastyns,' Matheus Froeyen,' Jose Fernando Diaz,' Guido Volckaert? and Yves Engelborghs' 'Laboratory of Chemical and Biological Dynamics and 'Laboratory of Gene Technology, University of Leuuen, B-3001 Leuuen, Belgium ABSTRACT Barnase, the guanine specific ribonuclease of Bacillus amyloliquefaciens, was subjected to mutations in order to alter the electrostatic properties of the enzyme. Ser-85 was mutated into Glu with the goal to introduce an extra charge in the neighborhood of His-102. A double mutation (Serd5Glu and Asp-86-Asn)was introduced with the same purpose but without altering the global charge of the enzyme. A similar set of mutations was made using Asp at position 85. For all mutants the PI was determined using the technique of isoelectric focusing and calculated on the basis of the Tanford-Kirkwood theory. When Glu was used to replace Ser-85, the correlation between the experimental and the calculated values was perfect. However, in the Ser-85-Aspmutant, the experimental PI drop is bigger than the calculated one, and in the double mutant (Ser-85-Asp and Asp-86-Asn) the compensation is not achieved. The effect of the mutations on the pK, of His102 can be determined from the pH dependence of the k,,,/K, for the hydrolysis of dinucleotides, e.g., GpC. The effect can also be calculated using the the method of Honig. In this case the agreement is very good for the Glu-mutants and the single Asp-mutant, but less for the double Asp-mutant. The global stability of the Aspmutants is, however, the same as the wild type, as shown by stability studies using urea denaturation. Molecular dynamics calculations, however, show that in the double Asp-mutant His-102 (H+) swings out of its pocket to make a hydrogen bridge with Gln-104 which should cause an additional pK, rise. The effect of the Glu-mutations was also tested on all the kinetic parameters for GpC and the cyclic intermediate G > p at pH 6.5, for RNA at pH 8.0, and for poly(A) at pH 6.2. The effect of the mutations is rather limited for the dinucleotide and the cyclic intermediate, but a strong increase of the K, is observed in the 0 1996 WILEY-LISS, INC. case of the single mutant (extra negative charge) with polymeric substrates. These results indicate that the extra negative charge has a strong destabilizing effect on the binding of the polymeric substrates in the ground state and the transition state complex. A comparison with the structure of bound tetranucleotides (Buckle, A.M. and Fersht, A.R., Biochemistry 3316461653,1994) shows that the extra negative charge points towards the P2 site. 0 1996 Wiley-Liss, Inc. Key words: barnase, site-specific mutagenesis, RNAse, RNA, poly(A), pH-profile, enzyme kinetics, electrostatics, Delphi, Tanford-Kirkwood INTRODUCTION Barnase is a guanine-specific endoribonuclease from B . amyloliquefaciens. The crystal structure has been solved.' The mechanism of action consists of two steps: in a first step of transesterification, a cyclic intermediate is formed, which is hydrolysed in a second step. His-102 acts as a general acid, and Glu-73 as a general Like many enzymes that use polymeric substrates, barnase hydrolyses RNA much more efficiently than dinucleotides, indicating the presence of multiple subsites on the enzyme, and the influence of the interactions at these subsites on the conformation at the primary site. A detailed study with oligonucleotides of the type Zp,Gp,Xp,Y was made (Day et al., 1992) and this study revealed that: (1) the minimal substrate is GplX (kcat/KM lo4 M-ls-'); (2) a considerable improvement is obtainedhhen pz is present (kcat/KM lo6 M-ls-' 1; (3) the limit of diffusion control is ob- - - Received May 5, 1995; revision accepted October 3, 1995. Address reprint requests to Yves Engelborghs, Laboratory of Chemical and Biological Dynamics, Celestijnenlaan 200 D, B-3001 Leuven, Belgium. 371 ELECTROSTATIC EFFECTS IN BARNASE Fig. 1. Stereographic view of the ribbon representation of the main chain of barnase showing the position of the three important residues: Ser-85, Asp-86, and His-102. (The drawing was generated using the program MOLSCRIPr'). - tained when Y is also present (kcat/K, los M-ls-'); (4) Zp, does not influence the kinetic parameters. Furthermore it is remarkable that the pH value of maximal kCat/K, is a t a much higher pH for RNA hydrolysis than for the hydrolysis of dinucle~tides.~ In our previous study6 we mutated Glu-60 into Gln and proved that this mutation was deleterious for the activity of barnase on dinucleotides but not on polymeric substrates, indicating a different conformation of the substrate in the binding site when dinucleotide and polymeric substrates are compared. In this article, mutations are made with the goal to introduce a n extra charge in the neighborhood of the catalytically important His-102. This goal can be achieved by replacing Ser-85 by Glu or Asp. In the case of the Ser-85-Glu mutant, the COO- is at a distance of 5 A form the His-l02(Ne3) and 11A from the negative charge of the Pi of the dinucleotide GpG (own modelling). The position of the three residues considered is shown in Figure 1. This operation can also be achieved by the simultaneous mutation of Asp-86 into an Asn, with the goal to keep the total charge of the protein constant. The negative charge of Asp-86 is at 9 A from His-102 ( N E ~ ) . The results show that this is very important for polymeric substrates at pH 8.0 and 6.2, but much less for dinucleotides and for the cyclic intermediate at pH 6.5. A study of long range (between 12 and 24 A) surface charge-charge interactions on the His-18 of barnase was performed by Loewenthal et a1.l' showing that a good agreement was obtained between calculated and experimentally determined ApK, values, using the DELPHI program. The experimental ApK, values were obtained from fluorescence titrations, where the fluorescence of Trp-94 is strongly quenched by Hi~-18.~-' In this study we concentrate on the electrostatic effects at the catalytically important His-102 residue, and we want to discriminate between local and global charge effects, and between dinucleotide and polymeric substrates. MATERIALS AND METHODS Restriction enzymes were purchased from Boehringer Mannheim (Mannheim, Germany). Dideoxy sequencing was done with the T, sequencing kit from Pharmacia (Uppsala, Sweden). SP-Trisacryl was purchased from IBF Villeneuve (La Garenne, France) and the Mono S and Phenyl Superose column were purchased from Pharmacia. The dialysis tubing (cut off 3,500 and 13,000), was obtained from Spectrum Medical Industries (Los Angeles, CA). The purified samples were concentrated with Centricon microconcentrators (Amicon, Danvers, MA). The substrates GpA, GpC, ApA, and RNA were obtained from Sigma Chemical Co. (Dorset, UK). Poly(A) was obtained from Pharmacia. Reagent-grade buffer materials and distilled and Millipore filtered (0.45 pm) water were used in the preparation of buffers. The bacterial E . coli strains used in this study were WK6 (A(lac-proAB) galE strA [F'lucP ZAM15 proA+B+]) and WK6 mutS (A(1ac-proAB)gulE strA rnutSl25::TnlO [Flaclq ZAM15 proA B I). The WK6 strain was used for production of barnase, while the mutS derivative served for the mutagenesis experiment." For inducible expression, cells were grown in Hartley medium," otherwise LB (Luria Broth) was used. The construct barnasebarstar has been previously cloned in plasmid pMT416 by Hartley and Paddon" and this recombinant plasmid was obtained from Dr. Hartley. The plasmid pMa/c used for the mutagenesis and the expression of barnase was obtained from Dr. P. Stanssens.1° The expression cassette of barnase consists of the structural gene, the tuc promoter and phoA signal sequence. Hence the expression is induced with IPTG and secreted into the periplasmic space. An acid shock for separation of the periplasmatic material and the first two chromatografic purification steps with SP-Trisacryl and a Mono S column, were performed as described by Mossakowska et al.5 Further purification was achieved on the Pharmacia FPLC-system with a prepacked Phenyl Superose column. Barnase was dialyzed overnight against 50 + + K. BASTYNS ET AL. 372 TABLE I. Molar Extinction Coefficients (E) Used to Calculate the Concentrations of the Substrates, and Change in Molar Extinction Coefficients (Ac)Used to Calculate the Activities Substrate PH GPC Poly(A) 7 RNA 6.2 7.5 A 280 260 260 E (M-lcrn-l) PH* 12.6OOzs 5 10.00029~t 8.000309t 6.2 AE (M-lcrn-l) A* 280 260 298 8 2.130" 5.0OOz9 -39* *pH and at which Ae was determined. +Expressedas formal mononucleotidemolarity. *Owndetermination at pH 8.0. mM imidazole-HC1 pH 6.6 containing 1.7 M (NH,),SO,. The protein was loaded onto the Phenyl Superose column preequilibrated with the same buffer. Barnase was eluted with a linear gradient of 50 mM imidazole-HC1 pH 6.6 and then dialyzed against an appropriate buffer. After further concentration of the purified sample, the homogeneity of barnase was checked with SDS-polyacrylamideelectrophoresis (Midget system, Pharmacia). The concentration of the enzyme and substrates were determined spectrophotometrically. The concentration of the enzyme was determined at 280 nm with the extinction coefficient rZsonm = 27.411 M-l.cm-' calculated from the amino acid composition (3 Trp and 7 Tyr) and the molar extinction coefficients for Trp and The molar extinction coefficient corresponds to a value of 2.18 cm'mg-l. The concentration of GpC, poly(A) and RNA were determined using the molar extinction coefficients given in Table I. It should be noted that the concentration of poly(A) and RNA is thus expressed as the formal concentration of mononucleotides. The steady state kinetic studies of the cleavage of the substrates GpC, poly(A), and RNA were performed with a Kontron Uvicon 81OP spectrophotometer. All the kinetic measurements were performed at 25°C. The kinetic parameters were calculated using the changes in molar extinction coefficients Ar given in Table I. The poly(A) cleavage reaction was followed by an increase in absorption at a wavelength of 260 nm. The change in absorption upon complete digestion (A raBOnm) was determined. The kinetics of RNA cleavage by barnase were studied by following the decrease in absorbance a t 298.5 nm.14 An amount of Torula yeast RNA was dissolved in distilled water. After filtration (0.22 Km) the solution was extensively dialysed (cut-off = 13,000) to remove short fragments. The RNA stock solution was then diluted into the appropriate buffer, just before the kinetic measurements. The change in absorbance upon complete digestion was experimentally determined. Where the pH dependence was determined the following buffer systems were used: formic acid-NaOH (pH 3.0-4.51, acetic acid-NaOH (pH 4.5-5.8), imid- azole-HCl (pH 6.0-7.Q and Tris-HC1 (pH 7.8-9). In these pH-dependency studies, the ionic strength was always 0.1 M. No appreciable discontinuities were observed between buffer systems of comparable pH. The hydrolysis of the cyclic intermediate was followed by the pH-stat method. A radiometer titrator TTTla and a titrigraph SBR 2C were used to add NaOH (1 mN) during the hydrolysis reaction. The substrate was dissolved in a 5 mM imidazole-HC1 buffer a t pH 6.5 and the ionic strength was made 0.3 M with Na,SO,. The substrate concentration was 0.2 to 3.5 mM, and the reaction was started by adding barnase to a concentration of 10 pM. The total volume of the solution was 700 pL. The solution was stirred continuously and a micro-electrodewas used to measure the pH. The stoichiometry was calculated using a pK, of 6.0 for the second phosphate ionization of 3'GMP.15 Isoelectrofocussing was done with Immobiline Dry strips pH interval 3.5-10 on a precooled (15°C)Multiphor I1 electrophoresis unit (Pharmacia). A commercial calibration mixture was used. The conformational stability of the wild type and the Asp mutants was checked by measuring the ratio of the fluorescence intensities y = 13,0nm/1330nm after excitation a t 295 nm, as a function of urea concentration (data not shown). The transition curves obtained can be described by the following equation: (YN Y= + ~ N . [ D ]-t) ("D + ~D.[D]).K l + K with K = [Ul/[N] = exp (- AGO(H20) - m[D])/RT where yN (yD)is the relative fluorescence of the native (denatured) form, mN (m,) is the intrinsic change of this parameter with denaturant (D) in the native (denatured U) state, m is the cooperativity of the transition, K is the equilibrium constant for denaturation and AGD(H20) is the free energy of denaturation in water. Wild type Barnase was characterized by K = 1a t [urea],, = 4.64 M and m = 2374 +. 70 kcaUmo1.M. All Asp-mutants had a [urea],, = 373 ELECTROSTATIC EFFECTS IN BARNASE TABLE 11. Isoelectric Point (PI)for Wild Type and Mutant Barnase as Determined by Isoelectric Focusing (pIeF,,) or by Calculation (pIcd) Accordmg to the Tanford-Kirkwood Model* Protein Wild-type S85E S85E,D86N S85D S85D,D86N E60Q PI,,, 9.20 ? 0.22 7.97 f 0.11 9.30 f 0.22 7.76 2 0.11 8.70 f 0.22 9.45 f 0.16 PIC,, 9.3 7.9 9.25 8.10 9.25 9.6 *Ionic strength was 0.01 M. * 4.65 0.05 and a cooperativity m = 2314 2 380 showing that the stability of the mutants was not significantly different from the wild type protein. All curve fittings were done using SigmaplotR (Jandel Scientific, Erkrath, Germany). Electrostatic Calculations The DELPHI-program was used for the macroscopic electrostatic calculations.16In the interior of the protein a dielectric constant of 4 was used,17 while the external value was 78.5. To the X-ray coordinates of barnase H-atoms were added with the aid of the Brugel-programme. In the initial model of the mutants the same x1,x2, and x3 angles are used as in the wild type. Then the added atoms were energy minimized using the CHARMM force field. The allowed range of Apk, is calculated using the extremes of the charg-harge distances of the different allowed rotamers of xZ and x3 for Glu and of x2 for Asp. In the case of the Asp mutants additional molecular dynamics calculations were performed (see further). For the calculation of the isoelectric point, the solvent accessible Tanford-Kirkwood model was applied." The charged groups present in barnase are tabulated by Loewenthal et a1.l' Molecular dynamics calculations were performed using the GROMOS 87 package:' which was obtained from Biostructure S.A. (Strasbourg, France). The input files for the GROMOS 87 package were mainly generated using the program WHATIFFl which was also used for the visualization of the calculated results and the generation of the structure of the mutants. The energy of the structure was first minimized for 100 steps in vacuum (steepest descent"), and then placed in a truncated octahedral box of SPC wateP3 where a minimum distance of 8 A was kept between the protein and the border of the box. This results in a 61.78 A wide water box with 3,403 water molecules. We counted as charged residues Arg+ , Lys+,His+, Glu- and Asp-. Counterions (C1-) were added to compensate for these charges. The energy of the molecule was again minimized for another 1000 I I I I I I I I I 3 4 5 6 7 8 I I 800 ' h '2- 600 v z Y \ ; 400 1 u 200 01 DH Fig. 2. pH profile of the parameter k&K, for the S85E mutant with the substrate GpC. Ionic strength was 0.1 M. Temperature was 25°C. The dotted line represents the least squares fit of the data to equation (1). The parameters obtained are given in Table 111. 500 steps in the water box. The velocities of the atoms were assigned following a maxwellian velocity distribution a t 100 K. The system was warmed up in five consecutive steps of 1 ps to 300 K while the position of the protein atoms were harmonically restrained to their startup positions. Then a free molecular dynamics simulation was performed for 50 ps using a constant pressure of 1bar and a constant temperature of 300 K. The calculations were performed using an Indy Silicon Graphics workstation. The data were analyzed using the programs WHATIFZ1and SIMLYS.24 RESULTS AND DISCUSSION Isoelectric Point The results of the experimental determination of the isoelectric pH for wild-type barnase and the mutants using the IEF-technique are shown in Table 11. The values calculated using the Tanford-Kirkwood model are also given. Introducing an extra charge by changing a Ser into an Glu at position 85 causes the PI of the protein to decrease by about 1.2 pH units. Moving a charge on the surface by the simultaneous removal of Asp-86 and the introduction of Glu-85, causes no change of the PI. An almost perfect agreement can be found with the calculated PI values. When Asp is introduced as the additional charge a t position 85 instead of Glu, the experimental PI decrease is, however, more pronounced. When additionally Asp-86 is changed into Asn-86, a decrease of the PI with 0.5 pH units is observed. Clearly this cannot be accounted for by the calculations. The isoelectric point of our previously studied 374 K. BASTYNS ET AL. TABLE 111. Comparison of the ApK,, Determined From the p H Profile of k,,dK, for the Hydrolysis of GpC, and the ApK,, Calculated According to Gilson and Honig.ls* Protein Wild-type S85E S85E,D86N S85D S85D,D86N PKa, 4.21 k 0.26 3.90 k 0.25 4.14 0.36 3.92 0.32 3.97 k 0.34 * * PKa, 5.56 ? 0.07 6.36 0.11 5.97 k 0.15 6.53 k 0.10 6.87 f 0.04 APK,, (exp) * ApK,, (call - - 0.80 k 0.18 0.41 0.22 0.97 ? 0.17 1.31 0.11 1.0-0.6 0.6-0.11 1.04-0.86 0.54-0.36 * * *The ionic strength was 0.1 M. The error on ApKacexp,is determined by summing the error of the individual pK, values, while the range of ApK,,,,,, is obtained by taking into account all the sterically allowed rotamers of xz and x3 for Glu and xz for Asp. / E PHE 82 A B Fig. 3. Stereo view of the structure of the wild-type protein (A) and the double mutant ( 8 )after 50 ps of equilibration. The position of His-102, Phe-82, are largely different in the double mutant, and the hydrogen bridge between His-I02 and Gln-104 is clearly visible. E60Q mutant was also determined and calculated.6 Here the isoelectric point is increased to 9.60 by the removal of a charge. Also here the agreement between calculations and experiment is very good. pKa-Values of Wild Type Barnase and Mutants for the hydrolysis of GpC, and were also calculated using the model of Honig. The assignments of the observed pKa to Glu-73 and to His-102 was done before5 on the basis of site directed mutagenesis and NMR measurements. The pH profile of the parameter kc,&, for the hvdrolvsis is bell-shaDed. Since the substrate does not show any ionization in the pH region studied: the bell-shaped curve reflects the titration of active " Shifts in the pKa of His-102 were experimentally determined from the pH dependence of the kc,&, I 375 ELECTROSTATIC EFFECTS IN BARNASE > 0.16 0.08 1 700 0.00 0 100 200 300 400 500 600 [ G P C I (uM) Fig. 4. Michaels-Menten curve for wild-type barnase (0). and the mutants S85E,A86N (o), and S85E (0)for the substrate GpC. The pH was 6.5, the ionic strength was 0.1 M, and the temperature was 25°C. The parameters obtained by nonlinear least squares fitting to the Michaels-Menten equation are shown in Table IV. groups on barnase. The bell-shaped curve can be described with an equation that takes into account the ionization of two groups: vmax KM (2) ( 0 1 1+-+-) [H+l K1 Kz (1) [H+l These are presumed to be Glu-73 (with ionization constant K,) and His-102, with ionization constant K,.5 (VmJKM)o is the pH independent maximum. The parameters can be obtained by fitting the data to equation (1) using the general non-linear least squares curve fitting program Sigmaplot. A typical curve is shown in Figure 2. These experiments were repeated three times and the average values and the errors of the acid dissociation constants are shown in Table 111. The error of pk,, of Glu-73 is much larger than of the pK,, because a smaller number of ex- perimental points was determined in the acid pH range. The ApK,, values were calculated using the DELPHI algorithm at an ionic strength of 0.1 M and a uniform dielectric constant of the protein of 4 was used.17 The agreement between the calculated and experimentally determined values for ApK,, is very good for the Glu mutants. When an extra charge (Glu) is built in the protein, the pK,, rises with 0.8 (t 0.2) units, when the charge of residue 86 is moved closer to His-102, the pK,, rises with only 0.4 ( 2 0.2) units. When Asp is introduced at position 85, the experimentally observed increase of ApK,, is still in agreement with the calculated one, but in the case of the double mutant, the experimental increase is about 0.7 units larger. The deviating result for the S85D, D85N double mutant indicates some complications which are not seen by the structure calculation procedures used. The overall stability of the mutant, however, is the same as the wild-type as deduced from the similar value of the [urea],, concentration and the cooperativity parameter. Therefore, a molecular dyngmics calculation was applied to the Asp-mutants and-the wild type as reference. Although the molecular dynamics calculations were only performed for 50 ps, the system evolves very rapidly towards a new conformation in which a hydrogen bridge is formed between Asp-85 and Asn86. This reorganization seems to have a strong influence on the backbone bearing His-102. This residue rotates out of his normal position to make a hydrogen bridge with Gln-104, and Phe-82 rotates towards the hole previously occupied by His-102 (See Figure 3). The additional hydrogen bridge stabilizes the protonated form of His-102 and could be responsible for the unexpected rise of the pK,. An extra rise of the pK, of 0.7 units corresponds to a AGO = 1 kcal/mole, which is very reasonable for a hydrogen bridge.25 Although this conformation is not necessarily the only conformation accessible, a contribution from it explains the observed deviations in the electrostatic calculations. The swinging out of His-102 is most certainly a dynamic process TABLE IV.Kinetic Parameters for the Transesterification of GpC (I = 0.1 M) and the Hydrolysis of the Cyclic Intermediate G > p (I = 0.3 M)* Protein Wild-type Substrate GPC kCatW1) 0.38 2 0.03 Wild-type GPC GPC GP' S85E S85E,D86N G>P G>P 0.15 f 0.03 0.19 f 0.03 0.22 f 0.04 0.26 f 0.05 0.14 f 0.01 S85E S85E,D86N &(dW 652 2 91 858 -t 150 689 -t 130 1840 f 550 2390 f 800 2280 f 400 kcat/KM(M-ls-') 582 f 127 175 2 65 276 f 95 119 -t 57 109 f 57 61 k 15 I *Temperature = 25"C,pH = 6.5. The errors for K,, and K, were calculated using the nonlinear fitting program Sigmaplot, and the errors for kJK, were calculated using the previous errors by applying the rules for the propagation of errors. 376 K. BASTYNS ET AL. which is reversed in the unprotonated state since the enzyme is active with dinucleotides. I I I I I A Comparison of the Kinetic Parameters for Dinucleotides and Cyclic Intermediates The kinetic parameters can be obtained from the full Michaelis-Menten curve which is shown in Figure 4 for dinucleotide substrates. These values were obtained at 6.5 pH. The parameters obtained by fitting the data to the Michaelis-Menten equation are shown in Table IV. Only the Glu mutants are studied in detail. The parameters change but not dramatically. The change is most pronounced in kCat/KMof the S85E mutant which decreases by a factor of 6 for GpC In all other situations the changes are smaller. The kinetics of the hydrolysis of the cyclic intermediate were studied using the pH-stat method. The characteristic parameters are found in Table IV. Here again the changes of the catalytic parameters are relatively limited. Hydrolysis of Polymeric Substrates The activity of barnase WT and mutants is also determined with the natural substrate RNA and the artificial substrate poly(A). Barnase is a much better enzyme for RNA due to the presence of subsitesz6 The pH optimum for RNA is a t pH 8.0 instead of pH 5 for the dinucleotides. From Figure 5 and Table V it can be deduced that WT barnase and the double mutant have kCat/K, values that are comparable. However, the , k and KM values of the double mutant are two times larger than the corresponding parameters of the wild type. In contrast, the S85E mutant has a KMthat is dramatically increased. In fact, for this mutant only the parameter kcat/KMcan be calculated. This value is decreased by a factor of 13. For poly(A) (measured at pH = 6.2) a similar overall situation is found, except that for the single mutant all the parameters could be determined. Here it is clear that KM increases by a factor of 9.2. The question can be asked whether the strong increase of the KM for polymeric substrates due to the Fig. 5. Michaelis-Menten curve for the hydrolysis of RNA by the wild-type enzyme (a), and the mutants S85E,A86N (o),and the mutant S85E (0). pH = 8.0, I = 0.1 M, temperature = 25°C. The parameters obtained by nonlinear least squares fitting to the Michaelis-Menten equation, or to a straight line, are shown in Table V. introduction of the extra negative charge a t position 85 can be explained on the basis of the overall charge of the protein. The experiments with RNA seem to indicate this. These experiments were done at pH 8.0 where WT barnase is positively charged (PI = 9.3) while the S85E mutant, which has an experimental PI of 7.97 shows a charge that is almost zero. However, the same deleterious effect of the extra charge is also observed for poly(A) at pH 6.2, despite the fact that both proteins have a net positive charge. It is therefore clear that the observed deleterious effect of this extra charge on the KM of polymeric substrates is due to a local effect only visible with polymeric substrates. In fact, when the structure of the mutants, as obtained from molecular dynamics calculations, is compared to the structure of the barnase-tetranucleotide complexz7 it becomes clear TABLE V. Kinetic Parameters for the Wild-Type and Mutant Barnase for the Natural Substrate RNA (pH = 8 ) and the Artificial Substrate Poly(A) (pH = 6.2)* Substrate kcat (S-') KM (PM) kCat&q(PM-~s-') RNA 3998 f 870 2130 f 560 1.87 f 0.9 RNA 0.15 f 0.01 RNA 6765 f 790 3412 f 1100 1.98 f 0.9 poly(A) 26.7 f 1.2 263 f 30 0.12 f. 0.02 2415 f 340 0.007 f. 0.002 poly(A) 17.6 f 4.0 406 f 42 0.094 f 0.014 poly(A) 38.6 f 1.7 *Temperature = 25"C,I = 0.1 M.The errors fork,,, and K, were calculated by Sigmaplot, and for k J K , were calculated by the rules for the propagation of errors. In the case of the mutant S85E and RNA as substrate, only Protein Wild-type S85E S85E,D86N Wild-type S85E S85E,D86N k,,JK, could be determined because a linear relation between v and the substrate concentration was obtained. For this substrate the errors for k,,dKM were obtained from linear least squares fitting to the 16 data points. (Mean values and standard deviations are given.) 377 ELECTROSTATIC EFFECTS IN BARNASE that the extra negative charge points towards the P2 site. This explains immediately its specific effect on the polymers. CONCLUSION Mutations were made with the goal to introduce an extra charge or to move a charge on the surface closer to His-102. The effect of these mutations on the isoelectric point of the enzyme can be calculated with the Tanford-Kirkwood theory and when Glu is used to replace Ser-85 the correlation with the experimental values is perfect. However, when Asp is used, the correlation is less. The effect on the pK, of His-102 can be determined from the pH dependence of the k,,,/KM for the hydrolysis of dinucleotides, e.g., GpC. The effect can also be calculated using the the method of Honig. Again the agreement is very good for the Glu mutants and the single Asp-mutant, but less for the double Asp mutants. The aberrant behavior of the double Asp mutant becomes clear from molecular dynamics calculations: the protonated state of His-102 is stabilized by a new hydrogen bridge with the oxygen of Gln-104. This hydrogen bridge becomes possible thanks to a reorientation of the His-102, itself due to another new hydrogen bridge between Asn-85 and Asp-86. The reorientation of the His ring also explains the lower activity of the enzyme. The effect of the Glu-mutations was also tested on all the kinetic parameters for GpC, G > p, at pH 6.5, for RNA at pH 8.0, and for poly(A) a t pH 6.2. The effect of the mutations is rather limited for the dinucleotide and the cyclic intermediate, but a strong increase of the KM is observed in the case of the single mutant (extra negative charge) and the polymeric substrates. These results prove that the KM of barnase for polymeric substrates is strongly influenced by electrostatic charge effects and these effects are equally operative in the ground as in the transition state. We suggest that this deleterious repulsion is operating on P2, in view of the position of Asp-86 relative to the structure of the barnase-tetranucleotide complex proposed by Buckle and Fersht.” ACKNOWLEDGMENTS This research was supported by the Belgian national incentive program on fundamental research in life sciences initiated by the Belgian State, Prime Minister’s Office, Science Policy Program (Bioimpuls/O5) and by the Research Council of the University of Leuven. G.V. is supported by contract ETC-007 of the “Vlaams Actieprogramma Biotechnologie.” The authors thank Dr. R.W. Hartley (N.I.H., Bethesda, MD) and Dr. P. Stanssens (Plant Genetic Systems, Gent, Belgium) for the gift of the expression system, and B. Keyers, C.E. for starting up the expression system. The authors thank Dr. Y. Mauguen for the availability of the coordinates of barnase and Dr. S. Wodak (University of Brussels, Belgium) for the use of the DELPHI program. Prof. J. Vanderleyden (F.A. Janssens Laboratory for Genetics) is acknowledged for making available the isoelectrofocusing unit. Dr. B. Wroblowski is acknowledged for valuable discussions. REFERENCES 1. Mauguen, Y., Hartley, R.W., Dodson, E.I., Dodson, G.G., Bricogne, G., Chothia, C., Jack, A. Molecular structure of a new family of ribonucleases. Nature 297:162-164,1982. 2. Rushizky, G.W., Greco, A.E., Hartley, R.W., Sober, H.A. Studies on bacillus subtilis ribonuclease. I. Characterization of enzymatic specificity. Biochemistry 2787-793, 1963. 3. Nishikawa, S., Morioka, H., Kim, H.J., Fuchimura, K., Tanaka, T., Uesugi, S., Hakoshima, T., Tomika, K., Oh& suka, E., Ikehara, M. Two histidine residues are essential for ribonuclease T, activity as is the case for ribonuclease A. Biochemistry 268620-8624, 1987. 4. Breslow, R. How do imidazole groups catalyse the cleavage of RNA in enzyme models and in enzymes? Evidence from negative catalysis. Acc. Chem. Res. 24:317-324, 1991. 5. Mossakowska,D.E., Nyberg, K., Fersht A.R. Kinetic characterization of the recombinant ribonuclease from Bacillus amyloliquefwiens and investigation of the key residues in catalysis by site-directed mutagenesis. Biochemistry 28: 3843-3850, 1989. 6. Bastyns, K., Froeyen, M., Volckaert, G., Engelborghs, Y. The role of Glu-60 in the suecificitv of the recombinant ribonuclease from Bacillus amyloliqiefmiens (barnase) towards dinucleotides, poly(A) and RNA. Biochem. J. 300 737-742,1994. 7. Loewenthal, R., Sancho, J., Fersht, A.R. The fluorescence spectrum of barnase: Contributions of three tryptophan residues and a histidine-related pH-dependence.Biochemistry 30:6775-6779, 1991. 8. Willaert, K., Loewenthal, R., Sancho, J., Froeyen, M., Fersht, A,, Engelborghs, Y. Determination of the excitedstate lifetimes of the tryptophan residues in barnase, via multifrequency phase fluorometry of tryptophan mutants. Biochemistry 31:711-716, 1992. 9. Vos, R., Engelborghs, Y. A fluorescence study of tryptophan-histidine interactions in the peptide anantin and in solution. Photochem. Photobiol. 6024-32, 1994. 10. Stanssens, P., Opsomer, C., McKeown, Y.M., Kramer, W., Zabeau, M., Fritz, H.J. Efficient oligonucleotide-directed construction of mutations in expression vectors by the gapped duplex DNA method using alternating selectable markers. Nucleic Acids Res. 17:4441-4454, 1989. 11. Hartley, R.W., Rogerson, D.L., Jr. Production and purification of the extracellular ribonuclease of Bacillus amyloliquefuciens (barnase) and its intracellular inhibitor (barstar). I. Barnase. Prep Biochem 2:229-242, 1972. 12. Hartley, R.W., Paddon, C.J. Expression of Bacillus amyloliquefwiens extracellular RNAse (barnase) in E. coli following an inactivating mutation. Gene 53:ll-19, 1987. 13. Mach, H., Middaugh, C.R., Lewis, R.V. Statitical determination of the average values of the extinction coefficients of tryptophan and tyrosine in native proteins. Anal. Biochem. 200:74-80,1992. 14. Oshima, T., Ueniski, N., Imahori, K. Simple assay methods for ribonuclease T,, T,, and nuclease P,. Anal. Biochem. 71:623-634,1976. 15. Jordan, D.A. “The Chemistry of Nucleic Acids.” London: Butterworths, 1960~135-137. 16. Gilson, M.K., Honig, B.H. Calculation of electrostatic potentials in an enzyme active site. Nature 33084-86,1987. 17. Harvey, S.C., Hoekstra, P. Dielectric relaxation spectrum of water adsorbed on lysozyme. J. Phys. Chem. 76:29872994,1972. 18. Matthew, J.B., Richards, F.M. Anion binding and pHdependent electrostatic effects in ribonuclease. Biochemistry 21:4989-4999, 1982. 19. Loewenthal, R., Sancho, J., Reinikainen, T., Fersht, A.R. 378 20. 21. 22. 23. 24. 25. K. BASTYNS ET AL. Long-range surface charge-charge interactions in proteins. Comparison of experimental results with calculations from a theoretical method. J. Mol. Biol. 232574-583, 1993. van Gunsteren W.F., Berendsen H.J.C: Program system GROMOS 87. Distributed by: BIOMOS biomolecular software b.v., Laboratory of Physical Chemistry, University of Groningen, NL, 1987. Vriend, G. A molecular modeling and drug design program. J. Mol. Graph. 8:52-56, 1990. Levitt, M., Lifson, S. (1969) Refinement of protein conformations using a macromolecular energy minimization procedure. J. Mol. Biol. 46:269-279. Berendsen, H.J.C., Postma, J.P.M., Van Gunsteren, W.F., Hermans, J. Interaction models for water in relation to protein hydration. In Pullman, B. (ed.): “Intramolecular Forces.” Dordrecht: Reidel, 1981:331-342. Kriiger, P., Luke, M., Szameit, A. Structure of crystaline a-chymotrypsin. V. The atomic structure of tosil, a-chymotrypsin a t 2A. Comp. Phys. Comm. 62:371-380, 1991. Pace, C.N., Heinemann, U., Hahn, U., Saenger, W. Ribonuclease T1: Structure, function, and stability. Angewandte Chemie 30:343-454, 1991. 26. Day, A.G., Parsonage, D., Ebel, S., Brown, T., Fersht, A. Barnase has subsites that give rise to large rate enhancements. Biochemistry 31:6390-6395,1992. 27. Buckle, A.M., Fersht, A.R. Subsite binding in an RNase: Structure of a barnase-tetranucleotide complex at 1.76-w resolution. Biochemistry 33:1644-1653, 1994. 28. Zabinski, M., Walz, F.G. Subsites and catalytic mechanism of ribonuclease TI: Kinetic studies using GpC and GpU as substrates. Arch. Biochem. Biophys. 175:558-564, 1976. 29. Yakovlev, G.I., Chepurnova, N.K., Moiseyev, G.P., Bocharov, A.L., Lopatnev, S.V. Specificity of Bacillus interrnedius 7P RNase in polynucleotide cleavage reactions. [Translated from] Bioorganicheskaya Khimiya 13:338343, 1987. 30. Steyaert, J., Opsomer C., Wyns L., Stanssens, P. Quantitative analysis of the contribution of Glu46 and Am98 to the guanosine specificity of ribonuclease TI. Biochemistry 30:494-499, 1991. 31. Kraulis, P.J. MOLSCRIPT: A program to produce both detailed and schematic plots of protein structures. J. Appl. Crystallogr. 24:946-950, 1991.