Biochimica et Biophysiea Acta, 1079 (1991) 229-237 © 1991 Elsevier Science Publishers B.V. All rights reserved 0167-4838/91/$03.5(1 ADONIS 0167483891(XI276T
229
BBAPRO 33989
The importance of precise positioning of negatively charged carboxylate in the catalytic action of human lysozyme Michiro Muraki i, Kazuaki H a r a t a 2, Yasuhiro Hayashi ~, Masayuki Machida 1 and Yoshifumi Jigami ' I Biological Chemistry Dil'ision, National Chemical Laborato~" ]'or Industry, Tsukuba, lbaraki (Japan) and 2 Bioengineering Department, Research Institute for Polymers and Textiles, Tsukuba, lbaraki (Japan)
Key words: Lysozyme; Active site; Site-directed mutagenesis; (Human)
The role of aspartic acid 53 of human lysozyme (peptidoglycan N-acetylmuramoylhydrolase, EC 3.2.1.17) has been investigated by a site-directed mutagenesis. In order to clarify the importance of precise positioning of the negatively charged carboxylate group in the active site geometry, both the three-dimensional structure and the enzymatic function of glutamic acid 53 human lysozyme (Glu-53 human lysozyme) have been characterized in comparison with those of wild type enzyme. Glu-53 human lysozyme was crystallized and analysed by X-ray crystallography. No remarkable difference in the conformation of whole molecule except the side chain of 53th residue was observed. In spite of full retention of the binding activities against either ~-l,4-1inked trisaccharide of N-acetylglucosamine ((GIcNAc) 3) or the corresponding hexasaccharide ((GlcNAc)6), the conversion of Asp-S3 to Glu reduced the enzymatic activities against both bacterial cell substrate and p-nitrophenyl penta-N-acetyl.~(l --, 4)-chitopentaoside fp-NO2-(GicNAc) s) to a few percent of the activities of wild type enz~m~e. Calculation of electrostatic potential around the reaction center predicted that no significant change in pK, of Glu-35 was caused by the mutation. These results indicate that the precise positioning of the negatively charged carboxylate in the geometry of reaction center is essential for the rate enhancement in the catalytic action of lysozyme, and suggest that Asp-S3 of human lysozyme participates in the catalytic action not simply in an electrostatical manner but partly in a nucleophilical manner.
Introduction Since the model building study based on the X-ray crystallographic analysis of hen egg-white lysozyme(GlcNAc) 3 complex, a couple of carboxylate group bearing residues (Giu-35 and Asp-53 in human lysozyme) have been pointed out to embrace the glycosidic bond of substrate to be cleaved and to form a reaction center in the active site cleft [1,2]. This type of acid-base catalysis [3,4] is expected to work in a variety of other glycoside hydrolases, such as cellulases according to the similarities in both the active site structure [5,6] and the kinetic behavior [7]. Between these two residues, one residue with higher pK~ value (pK~ = 6.8,
Abbreviations: GIcNAc, N-acei~l-t,-~,lacosamine; (GIcNAe)3, /3-1,41inked trimer of N-acetyl-b-glucosamine; (GIcNAc)~,, fl-l,4-1inked hexamer of N-acetyI-D-glucosamine; p-NOz-(GIcNAc):,, pnitrophenyl g-l,4-1inked pentamer of N-acet)'l-t~-glucosamine. Correspondence: M. Muraki, Biological Chemistry Division, National Chemical Laboratory for Industry, Tsukuba, ibaraki 305, Japan.
Glu-35 in human lysozyme [8]) has been generally considered to act as a general acid to protonate O l oxygen atom [9-11] or O5 oxygen atom [12] of the sugar -esidue that occupies subsite D of active site. In contrast, there has been no general agreement on the role of the other carboxylate bearing residue with lower pK a value (pK~ = 3.4, Asp-53 in human lysozyme
I81). For the latter residue, two possibilities have been suggested in its role in the bond rearrangement in the catalytic action of iysozyme (Fig. 1) [4,101. The upper course in Fig. 1 illustrates the participation of Asp-53 as a residue stabilizing the positively charged oxocarbonium ion ¢iectrostaticaily. On the other hand, the lower course in Fig. 1 illustrates the participation of Asp-53 as a nucleophile to give a double displacement at the anomerie center. Although there has been no direct evidence for the existence of an intermediate compound in the catalysis by lysozyme, the presence of a negatively charged carboxylate group in the reaction center is considered to be essential to perform an efficient catalysis in either case [11,121.
230 Aspartic acid 52 in chicken lysozyme, which corresponds to Asp-53 in human lysozyme, has been modified in various ways either by chemical modifications [13-17] or by sitc-dircctcd mutagcncsis [18]. A m o n g them Asn-52 chicken lysozyme reported by Kuroki et al. [1"/] and Malcolm et al. [18] is the most interesting mutant, because only a very small change (1 dalton) in size has been caused by the mutation. Asparagine 52 chicken lysozyme exhibited a few percent of lytic activity with no remarkable change in binding activity against either (GIcNAc) 3 or (GIcNAc)r,. This rcsult clearly gave direct proof of the involvement of the carboxylate group of Asp-52 in the catalytic action of chicken lysozyme. However, as the character of carboxylic acid group, which can ionize into negatively charged carboxylate anion, has been lost in asparagine residue, no direct evidence of the importance of precise positioning of negative charge in the catalytic action of lysozymc can be obtained using Asn-52 mutant. In order to clarify this point, we introduced the mutation, Asp-53 to Glu, which changes the spatial position of carboxylate group of 53th residue without affecting both the ionizable character of itself and the chemical properties of other residues composing catalytic cleft, into h u m a n lysozyme. In the previous p a p e r [19], we described some enzymatic properties of Glu-53 human lysozyme, but the actual effect of the mutation on the three-dimensional structure of active site has not been elucidated. In the present paper, we report the results of structural characterization of Glu-53 human lysozyme by X-ray crystallography as well as the further examination of enzymatic properties of Glu-53 human iysozyme. We also describe the results of calculation that predicts the effect of mutation on the electrostatic field in the active site of human lysozyme.
Materials and Methods
MateriaL~ Authentic human lysozyme was purchased from G r e e n Cross, Japan. Its concentration was determined on the basis of the absorption coefficient t' , A ~/" 2 / 0 , 0 n m --25.65 cm - l ) [20]. Lyophilized Micrococcus haeus cells were from Sigma. p-Nitrophenyl penta-N-acetyl-/9chitopentaoside ( p - N O z - ( G I c N A c ) . ~ ) ( m o r e than 97% purity) and N-acetylhexosaminidase from jack been were obtained from Seikagaku, Japan. All other chemical reagents were of biochemical or analytical reagent grade. Preparation and crystallization o f G&-53 human lysozyme All procedures concerning the production and purification of Glu-53 human lysozyme were the same a s reported previously except the Saccharomyces cerel'isiae strain, KSC22-1C (MATa, ssll, leu2, his, ura3) was used as a host cell, The ssl I is a recessive single gene mutation causing a supcr~ecretion of lysozyme [21]. The concentration of purified mutant protein w a s determined by dye-binding method using a protein concemration assay kit (Bio-Rad). Authentic h u m a n lysozyme was used as the standard. T h e elution behavior of the equimolar mixture of purified Glu-53 mutant and authentic wild type enzyme on a cation-exchange column (Mono-S, Pharmacia) was examined. Elution condition was the samc as described previously [22]. Both iysozymes co-eluted as a sharp single peak at p H 7.6, indicating no difference in molecular charge at this pi-i. Chromatographically purified Glu-53 mutant human lysozyme, 11 mg in 50 ml of 50 mM sodium phosphate (pH 8.0) plus 0.22 M NaCI, was extensively dialyzed against 0.13 mM sodium acetate buffer (pH 4.5) at
HO • (GIcNAc)3 " O ' ~
HO
0
H
HO
-
(GicNA¢)3-o~oOx0 (GIcNAc~
(GIcNArd"0 ~ ~ , / 0
"-,
H
/ "
Fig. I. Alle: - ; "" calalylic mechanisms for hydrolysis o f (GIcNAc)~, by human ly~)zyme.
231 Asp-53. Detailed information on X-ray work will be described clscwhcre.
Meast~rement of ('D spectra CD spectra were measured with a JASCO J-6(~l spectropolarimeter at 25°C. The proteins were dissolved in 50 mM sodium phosphate buffer (pH 8.0) plus 0.2 M NaCI and the concentrations were adjusted to 0.33 m g / m l . Thc data are expressed in terms of mean residue ellipticity.
L~timation of dissociation constant with (GIcNAc)~ and (GIcNAc)~ Fig. 2. Crystals of Gtu-53 human lysozyme. C~.~tals were grown under lhe conditions as de~ribed in the text.
4 ° C and then lyophilized to an amorphous powder. Crystallization was performed basically according to the method of Osserman et al. [23]. The lyophilized sample was dissolved in 25 mM sodium acetate (oH 4.5) plus 3 M ammonium nitrate to give the concentration of 20 m g / m l . The undissolved material was removed by centrifugation and after filtration with 0.45 g,m pore size filter (Millipore) the supernatant was used as the sample for crystallization. When the protein solution (50 #1) alone was placed as a sitting drop in the concave well of a plastic plate and equilibrated with a reservoir solution (1 ml) of 25 mM sodium acetate (pH 4.5) plus 5 M ammonium nitrate, several aggregated crystals appeared as needles, which were not suitable for X-ray investigation. On the other hand, in the same condition except that one small (0.05 mm x 0.05 mm × 0.03 mm) prism o f wild type human lysozyme was introduced as a seed crystal, many badly shaped prisms of Glu-53 mutant appeared as well as needles. In a second cycle one of these prisms was used as a seed crystal, which produced many well-shaped long prisms (Fig. 2). This seeding procedure was rcpeated once again and one of the resulting well-shaped prisms (0.3 mm × 0.3 m m × 0.3 ram) was used as the sample for the measurement of X-ray ~iffraction.
Structural determination by X-ray analysis The crystal of Giu-53 human lysozyme obtained as above was isomorphous with the orthorhombic crystal of wild type human ly.s~3zyme [24] and diffracted to the resolution beyond 1.7 A. The crystal form was P2~2~2 ~, tl with the cell dimensions of a = 57.13 A, b = 61.05 A and c = 33.16 A. X-ray diffraction data were collected on an Enraf-Nonius FAST diffractometer up to 1.7 ,~ resolution. The orientation of side chain group of Glu-53 was determined from a difference electron density map calculated with phases of wild type human lysozyme structure without the side chain group of
Dissociation constants of wild type and Glu-53 mutant human iysozyme were estimated l]~,)rometricaily accoroing to the method of Chipman et al. [25] by plotting log( F , - F ) / ( F - F~) against log[S], where F., F and F~ are the relative fluorescence intensity of human iysozyme alone, that of human lysozyme in the presence of a concentration [S] of ligand and that of human lysozyme saturated with ligand, respectively. The concentration of protein was adjusted to 3.0 ~tM. Fluorescence measurements were performed with a Kontron SFM25 spcctrofluorimetcr at 25 ° C in 4-times diluted Macllvaine's buffer (pH 7.2) [26] composed of 25 mM citric acid and 50 mM di-sodium hydrogen phosphate as final concentrations. The excitation wave length was 285 nm and the fluorescence intensity of the emission spectrum was measured at 335 nm.
Assay of enzTmatic actit'ity (1) Against M. luteus cells. The activity against M. iuteas cells was determined spectrophotometrically basically according to the turbidometric method of Locquct et al. [27]. For the comparison of lysis curves, the lysis of M. luteus cells in 4-times diluted Maellvaine buffer (pH 6.2) was monitored. Measurements were carried out at 2 5 ° C in a thermosatted cell with a Shimadzu UV-160 recording spectrophotometer. For obtaining pH activity profile, the same conditions except the buffers at various pHs ranging from pH 4.4 to pH 8.0 were u ~ d . As to Glu-53 mutant, the lysis profile of the sample pretreated with M. luteus cells under the same condition used for the measurement of the lysis curve was also examined at the enzyme concentration of 10/~g/ml. (2) Agafllst p-NOe-(GIcNAc) ~. The activity against p-NO2-(GlcNac) 5 was determined basically according to the method of Nanjo et al. [9,8] by measuring the color intensity at 405 nm of liberated p-nitrophenol. The reaction was performed at 37 ° C for 0.5, 1 and 2 h. The amount of enzyme used for each assay was 30 p.g. T o examine the stability of wild type and Glu-53 mutant enzyme in the reaction conditions, they were incubated under the same conditions (37 ° C, 2 h) as used for the assay of activity (without substrate), and the
232 v
residual activity was evaluated b: comparing the lysis curves against M. hdteus cells. No lpp cciablc inactivation was found for both wild type and Glu-53 mutant under the assay conditions.
Calculation of electrostatic potentml Electrostatic potentials of wild type and Glu-53 mutant were calculated by the Klapper algorithm [29] at 298 K on a 65 × 65 × 65 three-dimensional grid. The software package, Del Phi [30], was used for the computations. Coordinates for wild type and Glu-53 mutant human lysozyme were obtained from the Protein Data Bank (entry set 1LZ1 [24]) and from our X-ray work in this study, respectively. Since the uncharged state has been considered as catalytically active form of Glu-35, Glu-35 was assumed to be electrically neutral for both molecules. All other appropriate side chains including histidine and the N and C termini were assumed to be fully charged. The dielectric constant of protein region was set to either 2 or 4. The dielectric constant of water region was set to 80 and the ionic strength of solvent was 0.145. The Coulombic approximation is applied to the boundary condition. Iterations were terminated when the maximum change in potential in the final iteration is < 10 -5 k T / e . pK~ shifts of Glu-35 residue were calculated from the change in mean electrostatic potential at the two carboxylate oxygen atoms of Glu-35 residue between wild type and mutant enzyme by the method of Tanford and Roxby [31]. For displaying the graphics of electrostatic potential maps, a software package, Insight I1 (Biosym Technologies) was used.
Results
Effect of mutation on protein stntcture The three-dimensional structure of Glu-53 human lysozyme was very similar to the structure of wild type human lysozyme [24]. An ( F o - Fc) electron density map indicating the strong electron density for Glu-53 residue fixed the orientation of the side chain group definitely (Fig. 3). Plate 1 illustrates the three-dimensional structure of Glu-53 human lysozyme superimposed on that of wild type enzyme. As shown in Plate 1, neither global change in the conformation of main chain nor significant change in the conformation of side chain, especially in the active site region, was observed between wild type and Glu-53 mutant except the difference of 53th residue. This indicates that the effect of mutation on the three-dimensional structure has been almost confined to the mutation point. The structure of Glu-53 human lysozyme has been crystallographically refined at 1.77 A by restrained least-squares procedures including similated annealing procedures to an R factor of 0.19. The r.m.s, difference between corresponding atomic positions in wild
Glu53
Fig. 3. Difference electron density map ( Fo ~ ~. - Ft._., ~t), in t l ~ immediate vicinityof Glu-53 super reposed on the model ofwdd type. human lysozyme.The side chain conformation of Glu-53 was estimated so as to fit the electron density. iulU-..,J
I--,I
.
type and Glu-53 mutant are 0.14 A, 0.26 ,g, and 0.12 for all a-carbons, the side chain atoms of the major: residues composing active site (Glu-35, Tyr-63, Trp-64, Asl,-102 and Trp-109), and the side chain atoms o f Glu-35, respectively. CD spectra in near-ultraviolet region (245-320 n m ) were compared between wild type and Glu-53 mutant (Fig. 4). The shapes of both spectra were essentially identical, suggesting the conformational similarity of both molecules in solution. However, a small but distinct difference in molecular ellipticity was observed, indicating some difference in optical activity between wild type and Glu-53 mutant. In contrast to little change in the strength of positive peak at 285-300 nm, fairly large differences were observed in the strength negative peak at 245-285 nm. The similar effect o n strength of CD hand was also observed in the bindinl[~ of either GlcNAc[32] or (GIcNAc)3 [33] to wild enzyme. The change in electrostatic potential can aft: feet the electric vector of polarized light which directly related to circular dichronism. Therefore, it speculated that the change in electrostatic environm©nt due to the movement of the spatial position of nega, tively charged carboxylate group of 53th residue contribute to the difference in optical activity described here, as well as the slight change in aromatic side chain conformation.
233
| ~ ,-i
~
E~
.u
234 O 65
I
I
I
I
O
o
_
~
z~
~¢
- 2 0 s "%*'~
#'
r~l
•
,' 04
L}
I
I
t
I
0
3O0
T~me -40
I 245
1 260
t
I 2~0
I
L
300
I 320
Fig. 4. CD spectra of wild type and Glu-53 mutant human lyso.,smc. (---), wild type: 1--), Glu-53 mutant. En~rne concentrations are (1.33 mg/ml in 5(I mM sodium phosphate bul-fcr (pit 8.0) plus O.2 M NaCI.
Effect of m,tation
on enzymatic fi#,ction
i n T a b l e I, the dissociation c o n s t a n t s for b i n d i n g of (GIcNAc) x and (GIcNAc)6 are s u m m a r i z e d . N~ rem a r k a b l e change in affinities against e i t h e r ( G I c N A c ) 3 or ( G I c N A c ) 6 were f o u n d b e t w e e n wild type a n d Glu-53 m u t a n t , suggesting the integrity of substrate recognizing ability in spite of the m u t a t i o n . T h e lytic activity of Glu-53 h u m a n lysozyme against M. l u t e u s cells, which is a high-molecular-weight polymeric substrate with highly negative charge, was examined in c o m p a r i s o n with wild type enzyn~e (Fig. 5). As described briefly in the previous p a p e r [19], the activity of Glu-53 m u t a n t was proved to be m u c h lower than that of wild type enzyme. 1"he lysis profile in Glu-53 m u t a n t was markedly different from that in wild type enzyme. In contrast to the linear decrease of :tbsorbance in wild type enzyme, the lysis curve of Glu-53 m u t a n t was apparently e o n s t i t u t c d from the first rapidly decreasing period and the fl~llowing fairly slowly decreasing period (Fig. 5). However. p H d e p c n d e n c c s , f
(s)
Fig. 5. Lysisof M, I, tcu,~ cells by wild type and Glu-53 mutant human lyso~me. Enzyme concentrations: a. ill ~g/ml (Glu-53 mutant); b, 0.1 ~tg/ml (wild type); c. IO ,~g/ml (Glu-53 mutant); d, 10/xg/ml (v,'ild type) and sub'~lratc concentration: 11.25 mg/ml, in 4-times diluted Macl!vaine', btfffcr (pH 6.21. Mcasuremenb, were performed at 25" C.
activity against this substratc were not significantly different from each o t h e r (Fig. 6). G l u - 5 3 m u t a n t pret r e a t e d with cell s u b s t r a t e showed thc same type of biphasic lysis profile as that of n o n - t r e a t e d G l u - 5 3 m u t a n t (data not shown). T h e hydrolytic activity against p - N O 2 - ( G I c N A c ) s was also e x a m i n e d , i n c o n t r a s t to M . h~teus cell s u b s t r a t e , this is a n o n - c h a r g e d oligomeric s u b s t r a t e with a welld e f i n e d u n i f o r m chemical structure. Recently, a reproducible colorimetric assay m e t h o d to detect fairly small activity of lysozyme has b e e n d e v e l o p e d by N a n j o ct al. using this substrate [28]. T h e relative activity of Glu-53 m u t a n t a:, c o m p a r e d with wild type e n z y m e is s u m m a !
I
!
1
1C~Q
u
TABLE I Di,~sociation constants of (GIcNAc ) ¢ and ((;h'NAc A,, for wild type and Ght-53 mutant human l v.~oz)'me K d values were determined by the fluort,nctric titralion method oI Chipman el al. [25]. All measLlrcmcnlswere pcrfl~rmcd al 25 ° C m 4-times diluted Macllevaine's buffer (pll 7.2t. Flu-re,tenet ~as measured at 335 nm (excitation wa~eicnglh at 285 rim). Enzyme concentration, 3 .uM.
En~me Wild type Glu-53 mutant
Ko(~M) (GIcNAch
(GIcNAc),,
27 28
II 15
o~
0 ....
I--
I I:>H
Fig. 6. Lffcct o, ptl on lyric acliVily against M. /utet~s cells by wild type and Glu-53 nlUt;llllhuman iysozymc. Activity was estimated as
the decrease of ahsorbance al 650 nm in the inillal 2 rain and expressed as a percentage of that fi)r each sample measured in pit 6.2. Enzyme concentration. I1.1 #g/ml fl~r wild type and |0 p.g/ml fl)r Glu-53 mutant: subMrate conccnlralkm, 0.25 mg/ml: measuremerit,, were poll,wined at 25 ° C. Wild lype, ( I - - - I l l ): Glu-53 mutant, ( t - - -
o).
23s ]'ABLE It Aciitit~" o]" (;ht-53 truman h'~o:vm~' agah~t p-mtropl~en~l F~'rtla-~~ atet) #chttopentao.~ide
r a i n e d suggesting that the catalytic ::cti~ity-pH pro'file might not be influcnccd substantially by the mutation.
Activity ~ a s determined by the colorimclric ~s~ay method of Nanjo et al. 128] and expressed as lhe percentage of that of wild type
Discussion
enzyme in the same react;on condition. Reactions were peril rmcd at 37°C and pit 5.(I. Initial conccntrati~ns of enzyme and substratc were 1.3 ~M and 1~.22mM. respcclivcl~
In ;hc catalytic action o f lysoz: c a negativel~ c h a r g e d carboxylate ~,roup (Asp-53 in h u m a n lysozyme) is constricted to c o n t r i b u t e t~ t h t r a t e e n h a n c e m e n t of catalysis most significantly by f a m r i n g the d e v e l o p m e n t o f the o x o c a r b o n i u m i(,n i n t e r m e d i a t e by substrate. A variety of biochemical studies [11,36-38] and of theoretical studies I36,37] have r e v e a l e d the i m p o r t a n c e of the p r e s e n c e of negative charge i t , l L although direct evidence on the necessity o f precise 0ositioning o f negative c h a r g e has not b c c n o b t a i n e d . In this respect, Glu-53 h u m a n lysozyme would be one o f the most valuable m u t a n t s to p e r t u r b only the location of negative c h a r g e of side chain of 53th residue, if no global c h a n g e in c o n f o r m a t i o n has o c c u r r e d during the mula-
Reaction time (h)
Relative acti~it~ ("; 1
0.5
3.4
1.0
1.0
2.(I
11.7
rized in T a b l e I!. In spite of no t h e r m a l inactivation d u r i n g the r e a c t i o n p e r i o d the relative activity o f Glu-53 m u t a n t d e c r e a s e d as the r e a c t i o n time was prolonged. Effect site
of m u t a t i o n
o n electrostatic p o t e / t t i a l i , t h e a c t i r e
Plate 2 shows the c o u n t e r e d images o f e l e c t r o s t a t i c p o t e n t i a l arou~td the active site o f wild type and G l u - 5 3 m u t a n t . Both images have the s a m e f e a t u r e in which a distinct large g r a d i e n t o f e l e c t r o s t a t i c p o t e n t i a l lies across t h e reaction c e n t e r c o m p o s e d of a c o u p l e o f carboxyl g r o u p b e a r i n g residues. However. t h e r e is a small d i f f e r e n c e in the s h a p e o f isopotential c o n t o u r s especially a r o u n d 53th r e s i d u e d u e to the m u t a t i o n . T h e influence o f m u t a t i o n on t h e e l e c t r o s t a t i c environm e n t r e a c h e d several active site r e s i d u e s including Glu-35, Tyr-63, T r p - 6 4 a n d Trp-109. which have b e e n c o n s i d e r e d to be i m p o r t a n t in the catalytic action o f h u m a n lysozyme [11]. A s p r o t e i n s a r c c o n s i d e r e d to have the di~'lectric constanls, 2 - 4 [3d 351, the p K , shifts at Glu-35 of h u m a n lysozyme c a u s e d by the m u t a t i o n of A s p - 5 3 to Glu was e s t i m a t e d u n d e r the c o n d i t i o n s that thc d i c l c c t r i c c o n s t a n t o f p r o t e i n region was a s s u m e d to bc e i t h e r 2 or 4 ( T a b l c I!!). Relatively small pK~, shifts (0.17 or 0.141 of Glu-35 were oh-
TABLE II! Culcutat~,d pK,, shifts at GI,-35 bvt,~ecn ~ihl type and (;h~ 53 muta,it /lllt/ltl/i Iv.~o~}'lll£"
Electrostatic I~>tenlials of wild type and Glu-53 mut~lnl human lysozyme were calculated on a 65 × 6,5 ~ 65 grid by a finite difference melhod as described in the text. Ct~rdinates were scaled ~ that the percentage of the grid filled by molecule in the longest Iinear dimension is set to 67c~-, resulting in a grid spacing of approx. 1.1 7k. pK a shifts were calculated from the change in l~)tential due lo the mutation by the method of Tanford and Roxby [311. Dielectric constant of protein region
Change in pK~
2 4
0.17
0.14
lion.
T h e relative enzymatic activity to wild type enzyme in Ght-53 h u m a n lysozyme d e c r c a s e d to the same Icvel as Ash-52 chicken lysozymc (5.5 + 2.5~'~ o f wtld type enzymc. [18]). which posscss no c o r r e s p o n d i n g hmizable earboxylate g r o u p in the reaction center. O n the o t h e r hand. X-ray crystallographic analysis r e v e a l e d that no r e m a r k a b l e change has o c c u r r e d in the active site c o n f o r m a t i o n includin~ s u s p e c t e d s u b s t r a t e recognizing r e s i d u e s except the m u t a t i o n point (Plate 1). T h e s e results t o g e t h e r with the fact that the b i n d i n g activity against e i t h e r (GlcNAc)~ or (GlcNAc~,, was not c h a n g e d significantly ( T a b l e !). strongly suggest that not only the p r e s e n c e of negatively c h a r g e d carboxylate in the r e a c t i o n center, but also the precise positianirlg of it is essential to e n h a n c e the catalytic reaction rate. T h e difference of lysis profilc against ;14. lute'us ceils b e t w e e n wild type and Glu-53 m u t a n t is poteworthy. In contrast to the iysis curve of wild type e a t ' m e , thc a b s o r b a n e e o f reaction mixture c o n t a i n i n g Glu-53 mutant did not d e c r e a s e lincarly even in the early s~ate o i lysis (Fig. 5). A similar behavior was a l ~ o b s e r v e d in the hydrolysis of p - N O 2 - ( G I c N A e ) s- The relative activity of Glu-53 m u t a n t to wild type e n z y m e d e c r e a s e d as the reaction time was p r o l o n g e d [ l ' a b l e 11). T h e s e results suggest that some dtffcrences may exist in the catalytic m e c h a n i s m s b e t w e e n wild type h u m a n lysozyme and Glu-53 mutant. In c o n n e c t i o n with this, M a l c o l m et al. r e p o r t e d the similar a n o m a l o u s behavior o f Asn-52 m u t a n t chicken I~,so~,me, w h e r e a p p a r ent b i p h a s i c kinetics with a fairly r a p i d p e r i o d o f !y~is followed by a ' s t e a d y s t a t e ' ir~ which virtually c o n s t a n t activity was o b ~ r v e d [18]. T h e p r e t r c a t m e n t with ce~.! s u b s t r a t e did not affect the lysis profile of Glu-53 m u t a n t This suggests that the a p p a r e n t inactivation a c c o m p a n i e d by cell lysis occurs in a ~eversible way. W h e n a couple of carboxylate g r o u p s are l o c a t e d
236 In contrast to the mutation of Asp-52 to Asn, the size of the side chain in the residue at issue becomes larger by a single methylene group in the mutation of Asp-53 to Glu, Asp-52 in chicken lysozyme is considered to be situated at about 3 .~, from C1 atom of the sugar residue that occupies subsite D and too far apart to form a full covalent bond [1,9,11]. The larger size in s;de chain of 53th residue in Glu-53 human lysozyme should permit a close location of the carboxylate group of 53th residue against C1 carbon atom of the sugar residue that occupies subsite D. Hence, in the case of Glu-53 mutant it is more likely that the earboxylate group of 53th residue forms a more complete covalent bond with substrate (Fig. 1), which may be too stable to perform the following bond rearrangements efficiently, as c o m p a r e d to that in wild type enzyme. Further works will be needed to clarify what mechanism is operated in the atomic level. The investigation on this line including the determination of the three dimensional structure of complex of Olu-53 mutant human lysozyme and substrate (analogs) is no v in progress. Kirby proposed that Asp-52 in chicken lysozyme is involved in the catalytic action by forming a 'less-thancomplete' covalent bond with the glycosidic reaction center (Fig. 7) as a possible explanation that satisfies both the nucleophilic assistaace of A s p - 5 2 - C O O - to the cleavage of glycosidic bond and the observed retention of configuration of anomeric center [9]. O u r data in the present study indicated the precise positioning of the negatively charged carboxylate group at the catalytic center is critical in the catalytic action of lysozyme. The nucleophilic attack (Fig. 1, lower course) to anomeric center by the carboxylate anion of 53th residue might require a more precise positioning than simple electrostatic interaction (Fig. 1, upper course), because more strict alignment of electron orbital will be necessary in the former case than in the latter case. The nucleophilic attack mechanism seems to explain our results well in the present study, in respect to the strictness in the requirement of the carbox'ylate group positioning. However, too strong bonding will result in
ctosc in each other, such as the carboxylatc group of Glu-35 (pK:, = 6.8) and that of Asp-53 (pK~, = 3.4) in human lysozyme, a movemcnt of the location of one carboxylate group may largely affect the pK~ of the other one. A!though pK~, values of ionizable groups in proteins have been generally estimated frem the p H dependence of parameters obtained in either spectroscopic experiments or kinetic experiments, it is difficult to estimate the p K a value of Glu-53 human lysozymc experimentally, because the productivity of human lysozyme by S. ceret,isiae is fairly small and the Glu-53 human lysozyme shows the anomalous catalytic behavior as described above. Gilson and Honig rcported that p K , shifts obtained by a calculation, which sloved Poisson's equation for continuum models concerning the interaction 9f charges in macromolecules, agreed well with the experimental values for mutant subtilisms [34]. Recently, the importatzce of large electrostatic potential gradient between two catalytic carboxylates in the catalytic action of lysozymes was pointed out by Dao-Pin et al. using the calculation of electrostatic fields in the active sites of lysozymes [35]. To investigate the effect of the mutation, Asp-53 to Glu, on the electrostatic potential around the catalytic residues, the calculation using the same algorithm derived from linealized Poisson-Boltzmann equation has been performed. The calculation indi.cated a similar feature of the electrostatic potential m a p around the activc site of Glu-53 mutant as compared with that of wild type enzyme (Plate 2) and the relatively small change in p K , value of Glu-35 during the mutation (Table liD. This relatively small effect of mvtation on the pK~ ~alue of Glu-35 seemed to be consistent with the experimental results that the pH dependence of lytic activity against M. luteus cells in Glu-53 mutant was similar to that in wild type enzyme (Fig. 6). Therefore, the large decrease of enzymatic activity in Glu. 53 mutant against either bacterial cell substrate or p - N O 2(GlcNAc)s may not be derived from the large change in pK~ of Glu-35 due to the approach of ionized carboxylatc of 53th residue to unionized Glu-35 residue.
I
H
~O-(GIcNAc)z EF
HO-.-.. (C~cblAc)3.~ 0 . . ~
,-c
H
.. II O.~
\oJ,
O,.(GIcNAc)z
N ~ HO~(GIc ~Ac,~.0 ~ . . . ~ .,,K. a.~
'
n
H,,N~-~
'
H.
E.F
.40 H/
Fig. 7. Plausible part cipation mechanism of Glu-35 and Asp-53 in the bond rearrangement steps of hydrolysisof (GIcNAch, by human lysozyme.
237 a relatively stable acyi-glycoside compound which is subject to be hydrolyzed at the ester bond rather than at the acetal center. This is obviously not the case in lysozyme catalysis, because the experimental data by others indicate the retention of configuration at the anomeric center [11]. Consequently, the above considerations lead us to the conclusion that Asp-53 in human lysozyme participates in stabilizing the development of oxocarbonium ion intermediate not simply in an electrostatical manner, but partly in a nucleophilicat manner by forming a highly tuned incomplete covalent bond as shown in Fig. 7. Acknowledgement We thank Dr. Y. lshizuka for the help in the measurement of CD spectra. References 1 Blake, C.C.F.. Johnson. LN., Malt. G.A.. North. A.C.T., Phillips, D.C. and Sarma, V R . (19671 Proc. R. Soc. Lond. B!67, 378-388. 2 Joll~s, P. and Joll~:s. J. (19841 Mol. Cell. Biochem. 63, 165~189. 3 Vernon. C.A. (19671 Proc. R. Soc. Lond. B167. 389-401. 4 Chipman, D.M. and Sharon. N. (19691 Science 165, 454-465. 5 Paice. H.G., Desrochers, M.. Rho, D.. Jusarek. L. Roy. C., Rollin, C.F., DeMiguel. E. and Yaguchi. M. (19841 Biotechnology 2, 535-539. 6 Rouvinen, J., Berglors, T.. Teeri, T., Knowles. J.K.C. and J,~nes. T.A. (19901 Science 264, 380-386. 7 Hoj, P.B., Rodriguez. E.B., Stick. R.V. and Stone, B.A. (19891J. Biol. Chem. 264, 4939-4947. 8 Kuramitsu, S., Ikeda, K., Hamaguchi, K., Fujio, H.. Amano, T., Miwa, S. and Nishina, T. (19741J. Biochem. (Tokyo) 76. 671 -e,83. 9 Kirby, A.J. (19871 CRC Crit. Rev. Biochem. 22, 283-315. | 0 Fersht. A.R, (19851 in Enzyme Structure and Mechanism, 2nd Edn. (Freeman, W.H., ed.), New York, Ch. 2 and 15. I 1 lmoto, T., Johnson, LN,, North. A.C.T., Phillips. D.C. and Rupley. J.A. (19721 in The E ~ ' m e s (Boyer. P.D.. ed.). Vol. 7, Ch. 21, Academic Press. New York. 12 Post, C.B. and Karplus. M. (19861 J. Am. Chem. Soc, 108. 1317-1319. 13 Parsons. S.M. and Raftery. J.A. (I9691 giochemistn 21. 21~72192.
14 Parsom,. S.M.. Jao, L., Dahlquist_ F.W., Borders. C.I... Jr.. Gr~ffl. T., Racs. J. and Rafters. M.A. (19691 Biochemist~ ~, 7(){J- 7i2. 15 Eshdat, Y.. Dunn, A. and Sharon. N. (197..~) Proc. N a i l A c a d Sci. USA 73. 1658-1662. tt~ Yamada, t]., lmoto. T. and Noshita. S. (1~,~,2; Bi~lchc,nistr 3 21. 2187-2192. 17 Kumki. R., Yamada, H., Moriyama. T. and lmoto, r. (19881 J. Biol. Chem. 261. 13571-13574. 18 Malcolm. B.A.. Rosenberg, S.. Core~,. MJ., Allen. J.S.. DeBaeterlier. A. and Kirsch. LF. (19891 Proc. Natl. Acud. Sci. USA 86. 133-137. 19 Muraki, M.. Jigami. Y., Morikawa, M. and Tanaka, It. (19~7~ Biochim. Biophys. Aeta :~11,. 376-3~0. 2ll Parry. R.M,. Jr.. Chandan, R.C. and Shahani. K.M. (19691 Arch. Biochem. Biophys. 103, 50-65. 21 Suzuki, K.. Iehikawa. K, and Jigami. Y. (19891 Mol. Gen. Genet. 219. 58-64. 22 Jigami. Y.. Muraki. M.. Harada. N. and Tanaka. H. (19861 Gene 43. 273-279. 23 Osserman. E.F.. Cole, S.J.. Swan. I.D.A. and Blake. C.C.F. ( 19691 J. Mol, Biol. 46, 211-212. 24 Artymiuk. P.J. and Blake. C.C,F. ( 1981 ) J. Mol. Biol. 152. 737-7,52. 25 Chipman, D.M. and Sharon, N. (19691 Science 165. 454-465. 26 Macllvaine. T.C. (19211J. Biol. Chem. 49. 183-186. 27 Locquet. J.P.. Saint-Blanccard. J. and J(~lh?s, P. (19681 Biochim. Biophys. Acta 167, 150-153. 28 Nanjo. F.. Sakai. K. and Usui. T. (198S1J. Biochem (Tok3o) 1()4, 255-258. 29 Klapper. I., Hagstrom. R.. Fine, F.. Sharp, K. and llonig, g (19861 Proteins I. 47-59. 30 Gilson. M.K., Sharp, K.A. and Honig, B H (19871 J. Comp. C h e m 9. 327-335. 31 Tanfnrd. C. and Roxby. R. (19721 Biochemistr':, l h 21'-92-2198. 32 Halper, J.P.. Lato~,'itzki, N.. Bernstein. Fi. and Be',chok. S. 119711 Proc. Natl. Acad. Sci. USA 68. 517-522. 33 lkeda, K., Hamaguchi. K.. Miwa. S. and Ni~,hina. T_ (1972~ J. Biochem ( T o ~ ' o ) 71. 371-378. 34 Gilson, M.K. and Iqonig, B.H. (19871 Nature 3311, S4-8fl. 35 Dao-Pin. S,. kilo, D-I. and Remington. S.J, (19N91 Ploc. Natl. Acad. Sei. USA 86. 53fll-53(~5. 3b Rupley. J.A.. Gates. V. and Bilbrey. R. (196S) J. Am. Chem, S(;c. t~l. 5633-5635. 37 Lowe. G. (19,671 Prt~¢. R. Soc. Lond. Bit57. 431-434. 38 Tsai. C.S., Tang. J.~. and Subbara ~. S.C ([t~hg~ Biochcm J. 114_ 52t)-534. 39 Warshel, A. and ke~,itt, M, (1976)J. Mol. Biol. 103, 227-249 41t Bakthavahalam. V. and Czarnik. A,W. 11987/Tctrahedr~m kctt. 2~4. 2925-2928.
229
BBAPRO 33989
The importance of precise positioning of negatively charged carboxylate in the catalytic action of human lysozyme Michiro Muraki i, Kazuaki H a r a t a 2, Yasuhiro Hayashi ~, Masayuki Machida 1 and Yoshifumi Jigami ' I Biological Chemistry Dil'ision, National Chemical Laborato~" ]'or Industry, Tsukuba, lbaraki (Japan) and 2 Bioengineering Department, Research Institute for Polymers and Textiles, Tsukuba, lbaraki (Japan)
Key words: Lysozyme; Active site; Site-directed mutagenesis; (Human)
The role of aspartic acid 53 of human lysozyme (peptidoglycan N-acetylmuramoylhydrolase, EC 3.2.1.17) has been investigated by a site-directed mutagenesis. In order to clarify the importance of precise positioning of the negatively charged carboxylate group in the active site geometry, both the three-dimensional structure and the enzymatic function of glutamic acid 53 human lysozyme (Glu-53 human lysozyme) have been characterized in comparison with those of wild type enzyme. Glu-53 human lysozyme was crystallized and analysed by X-ray crystallography. No remarkable difference in the conformation of whole molecule except the side chain of 53th residue was observed. In spite of full retention of the binding activities against either ~-l,4-1inked trisaccharide of N-acetylglucosamine ((GIcNAc) 3) or the corresponding hexasaccharide ((GlcNAc)6), the conversion of Asp-S3 to Glu reduced the enzymatic activities against both bacterial cell substrate and p-nitrophenyl penta-N-acetyl.~(l --, 4)-chitopentaoside fp-NO2-(GicNAc) s) to a few percent of the activities of wild type enz~m~e. Calculation of electrostatic potential around the reaction center predicted that no significant change in pK, of Glu-35 was caused by the mutation. These results indicate that the precise positioning of the negatively charged carboxylate in the geometry of reaction center is essential for the rate enhancement in the catalytic action of lysozyme, and suggest that Asp-S3 of human lysozyme participates in the catalytic action not simply in an electrostatical manner but partly in a nucleophilical manner.
Introduction Since the model building study based on the X-ray crystallographic analysis of hen egg-white lysozyme(GlcNAc) 3 complex, a couple of carboxylate group bearing residues (Giu-35 and Asp-53 in human lysozyme) have been pointed out to embrace the glycosidic bond of substrate to be cleaved and to form a reaction center in the active site cleft [1,2]. This type of acid-base catalysis [3,4] is expected to work in a variety of other glycoside hydrolases, such as cellulases according to the similarities in both the active site structure [5,6] and the kinetic behavior [7]. Between these two residues, one residue with higher pK~ value (pK~ = 6.8,
Abbreviations: GIcNAc, N-acei~l-t,-~,lacosamine; (GIcNAe)3, /3-1,41inked trimer of N-acetyl-b-glucosamine; (GIcNAc)~,, fl-l,4-1inked hexamer of N-acetyI-D-glucosamine; p-NOz-(GIcNAc):,, pnitrophenyl g-l,4-1inked pentamer of N-acet)'l-t~-glucosamine. Correspondence: M. Muraki, Biological Chemistry Division, National Chemical Laboratory for Industry, Tsukuba, ibaraki 305, Japan.
Glu-35 in human lysozyme [8]) has been generally considered to act as a general acid to protonate O l oxygen atom [9-11] or O5 oxygen atom [12] of the sugar -esidue that occupies subsite D of active site. In contrast, there has been no general agreement on the role of the other carboxylate bearing residue with lower pK a value (pK~ = 3.4, Asp-53 in human lysozyme
I81). For the latter residue, two possibilities have been suggested in its role in the bond rearrangement in the catalytic action of iysozyme (Fig. 1) [4,101. The upper course in Fig. 1 illustrates the participation of Asp-53 as a residue stabilizing the positively charged oxocarbonium ion ¢iectrostaticaily. On the other hand, the lower course in Fig. 1 illustrates the participation of Asp-53 as a nucleophile to give a double displacement at the anomerie center. Although there has been no direct evidence for the existence of an intermediate compound in the catalysis by lysozyme, the presence of a negatively charged carboxylate group in the reaction center is considered to be essential to perform an efficient catalysis in either case [11,121.
230 Aspartic acid 52 in chicken lysozyme, which corresponds to Asp-53 in human lysozyme, has been modified in various ways either by chemical modifications [13-17] or by sitc-dircctcd mutagcncsis [18]. A m o n g them Asn-52 chicken lysozyme reported by Kuroki et al. [1"/] and Malcolm et al. [18] is the most interesting mutant, because only a very small change (1 dalton) in size has been caused by the mutation. Asparagine 52 chicken lysozyme exhibited a few percent of lytic activity with no remarkable change in binding activity against either (GIcNAc) 3 or (GIcNAc)r,. This rcsult clearly gave direct proof of the involvement of the carboxylate group of Asp-52 in the catalytic action of chicken lysozyme. However, as the character of carboxylic acid group, which can ionize into negatively charged carboxylate anion, has been lost in asparagine residue, no direct evidence of the importance of precise positioning of negative charge in the catalytic action of lysozymc can be obtained using Asn-52 mutant. In order to clarify this point, we introduced the mutation, Asp-53 to Glu, which changes the spatial position of carboxylate group of 53th residue without affecting both the ionizable character of itself and the chemical properties of other residues composing catalytic cleft, into h u m a n lysozyme. In the previous p a p e r [19], we described some enzymatic properties of Glu-53 human lysozyme, but the actual effect of the mutation on the three-dimensional structure of active site has not been elucidated. In the present paper, we report the results of structural characterization of Glu-53 human lysozyme by X-ray crystallography as well as the further examination of enzymatic properties of Glu-53 human iysozyme. We also describe the results of calculation that predicts the effect of mutation on the electrostatic field in the active site of human lysozyme.
Materials and Methods
MateriaL~ Authentic human lysozyme was purchased from G r e e n Cross, Japan. Its concentration was determined on the basis of the absorption coefficient t' , A ~/" 2 / 0 , 0 n m --25.65 cm - l ) [20]. Lyophilized Micrococcus haeus cells were from Sigma. p-Nitrophenyl penta-N-acetyl-/9chitopentaoside ( p - N O z - ( G I c N A c ) . ~ ) ( m o r e than 97% purity) and N-acetylhexosaminidase from jack been were obtained from Seikagaku, Japan. All other chemical reagents were of biochemical or analytical reagent grade. Preparation and crystallization o f G&-53 human lysozyme All procedures concerning the production and purification of Glu-53 human lysozyme were the same a s reported previously except the Saccharomyces cerel'isiae strain, KSC22-1C (MATa, ssll, leu2, his, ura3) was used as a host cell, The ssl I is a recessive single gene mutation causing a supcr~ecretion of lysozyme [21]. The concentration of purified mutant protein w a s determined by dye-binding method using a protein concemration assay kit (Bio-Rad). Authentic h u m a n lysozyme was used as the standard. T h e elution behavior of the equimolar mixture of purified Glu-53 mutant and authentic wild type enzyme on a cation-exchange column (Mono-S, Pharmacia) was examined. Elution condition was the samc as described previously [22]. Both iysozymes co-eluted as a sharp single peak at p H 7.6, indicating no difference in molecular charge at this pi-i. Chromatographically purified Glu-53 mutant human lysozyme, 11 mg in 50 ml of 50 mM sodium phosphate (pH 8.0) plus 0.22 M NaCI, was extensively dialyzed against 0.13 mM sodium acetate buffer (pH 4.5) at
HO • (GIcNAc)3 " O ' ~
HO
0
H
HO
-
(GicNA¢)3-o~oOx0 (GIcNAc~
(GIcNArd"0 ~ ~ , / 0
"-,
H
/ "
Fig. I. Alle: - ; "" calalylic mechanisms for hydrolysis o f (GIcNAc)~, by human ly~)zyme.
231 Asp-53. Detailed information on X-ray work will be described clscwhcre.
Meast~rement of ('D spectra CD spectra were measured with a JASCO J-6(~l spectropolarimeter at 25°C. The proteins were dissolved in 50 mM sodium phosphate buffer (pH 8.0) plus 0.2 M NaCI and the concentrations were adjusted to 0.33 m g / m l . Thc data are expressed in terms of mean residue ellipticity.
L~timation of dissociation constant with (GIcNAc)~ and (GIcNAc)~ Fig. 2. Crystals of Gtu-53 human lysozyme. C~.~tals were grown under lhe conditions as de~ribed in the text.
4 ° C and then lyophilized to an amorphous powder. Crystallization was performed basically according to the method of Osserman et al. [23]. The lyophilized sample was dissolved in 25 mM sodium acetate (oH 4.5) plus 3 M ammonium nitrate to give the concentration of 20 m g / m l . The undissolved material was removed by centrifugation and after filtration with 0.45 g,m pore size filter (Millipore) the supernatant was used as the sample for crystallization. When the protein solution (50 #1) alone was placed as a sitting drop in the concave well of a plastic plate and equilibrated with a reservoir solution (1 ml) of 25 mM sodium acetate (pH 4.5) plus 5 M ammonium nitrate, several aggregated crystals appeared as needles, which were not suitable for X-ray investigation. On the other hand, in the same condition except that one small (0.05 mm x 0.05 mm × 0.03 mm) prism o f wild type human lysozyme was introduced as a seed crystal, many badly shaped prisms of Glu-53 mutant appeared as well as needles. In a second cycle one of these prisms was used as a seed crystal, which produced many well-shaped long prisms (Fig. 2). This seeding procedure was rcpeated once again and one of the resulting well-shaped prisms (0.3 mm × 0.3 m m × 0.3 ram) was used as the sample for the measurement of X-ray ~iffraction.
Structural determination by X-ray analysis The crystal of Giu-53 human lysozyme obtained as above was isomorphous with the orthorhombic crystal of wild type human ly.s~3zyme [24] and diffracted to the resolution beyond 1.7 A. The crystal form was P2~2~2 ~, tl with the cell dimensions of a = 57.13 A, b = 61.05 A and c = 33.16 A. X-ray diffraction data were collected on an Enraf-Nonius FAST diffractometer up to 1.7 ,~ resolution. The orientation of side chain group of Glu-53 was determined from a difference electron density map calculated with phases of wild type human lysozyme structure without the side chain group of
Dissociation constants of wild type and Glu-53 mutant human iysozyme were estimated l]~,)rometricaily accoroing to the method of Chipman et al. [25] by plotting log( F , - F ) / ( F - F~) against log[S], where F., F and F~ are the relative fluorescence intensity of human iysozyme alone, that of human lysozyme in the presence of a concentration [S] of ligand and that of human lysozyme saturated with ligand, respectively. The concentration of protein was adjusted to 3.0 ~tM. Fluorescence measurements were performed with a Kontron SFM25 spcctrofluorimetcr at 25 ° C in 4-times diluted Macllvaine's buffer (pH 7.2) [26] composed of 25 mM citric acid and 50 mM di-sodium hydrogen phosphate as final concentrations. The excitation wave length was 285 nm and the fluorescence intensity of the emission spectrum was measured at 335 nm.
Assay of enzTmatic actit'ity (1) Against M. luteus cells. The activity against M. iuteas cells was determined spectrophotometrically basically according to the turbidometric method of Locquct et al. [27]. For the comparison of lysis curves, the lysis of M. luteus cells in 4-times diluted Maellvaine buffer (pH 6.2) was monitored. Measurements were carried out at 2 5 ° C in a thermosatted cell with a Shimadzu UV-160 recording spectrophotometer. For obtaining pH activity profile, the same conditions except the buffers at various pHs ranging from pH 4.4 to pH 8.0 were u ~ d . As to Glu-53 mutant, the lysis profile of the sample pretreated with M. luteus cells under the same condition used for the measurement of the lysis curve was also examined at the enzyme concentration of 10/~g/ml. (2) Agafllst p-NOe-(GIcNAc) ~. The activity against p-NO2-(GlcNac) 5 was determined basically according to the method of Nanjo et al. [9,8] by measuring the color intensity at 405 nm of liberated p-nitrophenol. The reaction was performed at 37 ° C for 0.5, 1 and 2 h. The amount of enzyme used for each assay was 30 p.g. T o examine the stability of wild type and Glu-53 mutant enzyme in the reaction conditions, they were incubated under the same conditions (37 ° C, 2 h) as used for the assay of activity (without substrate), and the
232 v
residual activity was evaluated b: comparing the lysis curves against M. hdteus cells. No lpp cciablc inactivation was found for both wild type and Glu-53 mutant under the assay conditions.
Calculation of electrostatic potentml Electrostatic potentials of wild type and Glu-53 mutant were calculated by the Klapper algorithm [29] at 298 K on a 65 × 65 × 65 three-dimensional grid. The software package, Del Phi [30], was used for the computations. Coordinates for wild type and Glu-53 mutant human lysozyme were obtained from the Protein Data Bank (entry set 1LZ1 [24]) and from our X-ray work in this study, respectively. Since the uncharged state has been considered as catalytically active form of Glu-35, Glu-35 was assumed to be electrically neutral for both molecules. All other appropriate side chains including histidine and the N and C termini were assumed to be fully charged. The dielectric constant of protein region was set to either 2 or 4. The dielectric constant of water region was set to 80 and the ionic strength of solvent was 0.145. The Coulombic approximation is applied to the boundary condition. Iterations were terminated when the maximum change in potential in the final iteration is < 10 -5 k T / e . pK~ shifts of Glu-35 residue were calculated from the change in mean electrostatic potential at the two carboxylate oxygen atoms of Glu-35 residue between wild type and mutant enzyme by the method of Tanford and Roxby [31]. For displaying the graphics of electrostatic potential maps, a software package, Insight I1 (Biosym Technologies) was used.
Results
Effect of mutation on protein stntcture The three-dimensional structure of Glu-53 human lysozyme was very similar to the structure of wild type human lysozyme [24]. An ( F o - Fc) electron density map indicating the strong electron density for Glu-53 residue fixed the orientation of the side chain group definitely (Fig. 3). Plate 1 illustrates the three-dimensional structure of Glu-53 human lysozyme superimposed on that of wild type enzyme. As shown in Plate 1, neither global change in the conformation of main chain nor significant change in the conformation of side chain, especially in the active site region, was observed between wild type and Glu-53 mutant except the difference of 53th residue. This indicates that the effect of mutation on the three-dimensional structure has been almost confined to the mutation point. The structure of Glu-53 human lysozyme has been crystallographically refined at 1.77 A by restrained least-squares procedures including similated annealing procedures to an R factor of 0.19. The r.m.s, difference between corresponding atomic positions in wild
Glu53
Fig. 3. Difference electron density map ( Fo ~ ~. - Ft._., ~t), in t l ~ immediate vicinityof Glu-53 super reposed on the model ofwdd type. human lysozyme.The side chain conformation of Glu-53 was estimated so as to fit the electron density. iulU-..,J
I--,I
.
type and Glu-53 mutant are 0.14 A, 0.26 ,g, and 0.12 for all a-carbons, the side chain atoms of the major: residues composing active site (Glu-35, Tyr-63, Trp-64, Asl,-102 and Trp-109), and the side chain atoms o f Glu-35, respectively. CD spectra in near-ultraviolet region (245-320 n m ) were compared between wild type and Glu-53 mutant (Fig. 4). The shapes of both spectra were essentially identical, suggesting the conformational similarity of both molecules in solution. However, a small but distinct difference in molecular ellipticity was observed, indicating some difference in optical activity between wild type and Glu-53 mutant. In contrast to little change in the strength of positive peak at 285-300 nm, fairly large differences were observed in the strength negative peak at 245-285 nm. The similar effect o n strength of CD hand was also observed in the bindinl[~ of either GlcNAc[32] or (GIcNAc)3 [33] to wild enzyme. The change in electrostatic potential can aft: feet the electric vector of polarized light which directly related to circular dichronism. Therefore, it speculated that the change in electrostatic environm©nt due to the movement of the spatial position of nega, tively charged carboxylate group of 53th residue contribute to the difference in optical activity described here, as well as the slight change in aromatic side chain conformation.
233
| ~ ,-i
~
E~
.u
234 O 65
I
I
I
I
O
o
_
~
z~
~¢
- 2 0 s "%*'~
#'
r~l
•
,' 04
L}
I
I
t
I
0
3O0
T~me -40
I 245
1 260
t
I 2~0
I
L
300
I 320
Fig. 4. CD spectra of wild type and Glu-53 mutant human lyso.,smc. (---), wild type: 1--), Glu-53 mutant. En~rne concentrations are (1.33 mg/ml in 5(I mM sodium phosphate bul-fcr (pit 8.0) plus O.2 M NaCI.
Effect of m,tation
on enzymatic fi#,ction
i n T a b l e I, the dissociation c o n s t a n t s for b i n d i n g of (GIcNAc) x and (GIcNAc)6 are s u m m a r i z e d . N~ rem a r k a b l e change in affinities against e i t h e r ( G I c N A c ) 3 or ( G I c N A c ) 6 were f o u n d b e t w e e n wild type a n d Glu-53 m u t a n t , suggesting the integrity of substrate recognizing ability in spite of the m u t a t i o n . T h e lytic activity of Glu-53 h u m a n lysozyme against M. l u t e u s cells, which is a high-molecular-weight polymeric substrate with highly negative charge, was examined in c o m p a r i s o n with wild type enzyn~e (Fig. 5). As described briefly in the previous p a p e r [19], the activity of Glu-53 m u t a n t was proved to be m u c h lower than that of wild type enzyme. 1"he lysis profile in Glu-53 m u t a n t was markedly different from that in wild type enzyme. In contrast to the linear decrease of :tbsorbance in wild type enzyme, the lysis curve of Glu-53 m u t a n t was apparently e o n s t i t u t c d from the first rapidly decreasing period and the fl~llowing fairly slowly decreasing period (Fig. 5). However. p H d e p c n d e n c c s , f
(s)
Fig. 5. Lysisof M, I, tcu,~ cells by wild type and Glu-53 mutant human lyso~me. Enzyme concentrations: a. ill ~g/ml (Glu-53 mutant); b, 0.1 ~tg/ml (wild type); c. IO ,~g/ml (Glu-53 mutant); d, 10/xg/ml (v,'ild type) and sub'~lratc concentration: 11.25 mg/ml, in 4-times diluted Macl!vaine', btfffcr (pH 6.21. Mcasuremenb, were performed at 25" C.
activity against this substratc were not significantly different from each o t h e r (Fig. 6). G l u - 5 3 m u t a n t pret r e a t e d with cell s u b s t r a t e showed thc same type of biphasic lysis profile as that of n o n - t r e a t e d G l u - 5 3 m u t a n t (data not shown). T h e hydrolytic activity against p - N O 2 - ( G I c N A c ) s was also e x a m i n e d , i n c o n t r a s t to M . h~teus cell s u b s t r a t e , this is a n o n - c h a r g e d oligomeric s u b s t r a t e with a welld e f i n e d u n i f o r m chemical structure. Recently, a reproducible colorimetric assay m e t h o d to detect fairly small activity of lysozyme has b e e n d e v e l o p e d by N a n j o ct al. using this substrate [28]. T h e relative activity of Glu-53 m u t a n t a:, c o m p a r e d with wild type e n z y m e is s u m m a !
I
!
1
1C~Q
u
TABLE I Di,~sociation constants of (GIcNAc ) ¢ and ((;h'NAc A,, for wild type and Ght-53 mutant human l v.~oz)'me K d values were determined by the fluort,nctric titralion method oI Chipman el al. [25]. All measLlrcmcnlswere pcrfl~rmcd al 25 ° C m 4-times diluted Macllevaine's buffer (pll 7.2t. Flu-re,tenet ~as measured at 335 nm (excitation wa~eicnglh at 285 rim). Enzyme concentration, 3 .uM.
En~me Wild type Glu-53 mutant
Ko(~M) (GIcNAch
(GIcNAc),,
27 28
II 15
o~
0 ....
I--
I I:>H
Fig. 6. Lffcct o, ptl on lyric acliVily against M. /utet~s cells by wild type and Glu-53 nlUt;llllhuman iysozymc. Activity was estimated as
the decrease of ahsorbance al 650 nm in the inillal 2 rain and expressed as a percentage of that fi)r each sample measured in pit 6.2. Enzyme concentration. I1.1 #g/ml fl~r wild type and |0 p.g/ml fl)r Glu-53 mutant: subMrate conccnlralkm, 0.25 mg/ml: measuremerit,, were poll,wined at 25 ° C. Wild lype, ( I - - - I l l ): Glu-53 mutant, ( t - - -
o).
23s ]'ABLE It Aciitit~" o]" (;ht-53 truman h'~o:vm~' agah~t p-mtropl~en~l F~'rtla-~~ atet) #chttopentao.~ide
r a i n e d suggesting that the catalytic ::cti~ity-pH pro'file might not be influcnccd substantially by the mutation.
Activity ~ a s determined by the colorimclric ~s~ay method of Nanjo et al. 128] and expressed as lhe percentage of that of wild type
Discussion
enzyme in the same react;on condition. Reactions were peril rmcd at 37°C and pit 5.(I. Initial conccntrati~ns of enzyme and substratc were 1.3 ~M and 1~.22mM. respcclivcl~
In ;hc catalytic action o f lysoz: c a negativel~ c h a r g e d carboxylate ~,roup (Asp-53 in h u m a n lysozyme) is constricted to c o n t r i b u t e t~ t h t r a t e e n h a n c e m e n t of catalysis most significantly by f a m r i n g the d e v e l o p m e n t o f the o x o c a r b o n i u m i(,n i n t e r m e d i a t e by substrate. A variety of biochemical studies [11,36-38] and of theoretical studies I36,37] have r e v e a l e d the i m p o r t a n c e of the p r e s e n c e of negative charge i t , l L although direct evidence on the necessity o f precise 0ositioning o f negative c h a r g e has not b c c n o b t a i n e d . In this respect, Glu-53 h u m a n lysozyme would be one o f the most valuable m u t a n t s to p e r t u r b only the location of negative c h a r g e of side chain of 53th residue, if no global c h a n g e in c o n f o r m a t i o n has o c c u r r e d during the mula-
Reaction time (h)
Relative acti~it~ ("; 1
0.5
3.4
1.0
1.0
2.(I
11.7
rized in T a b l e I!. In spite of no t h e r m a l inactivation d u r i n g the r e a c t i o n p e r i o d the relative activity o f Glu-53 m u t a n t d e c r e a s e d as the r e a c t i o n time was prolonged. Effect site
of m u t a t i o n
o n electrostatic p o t e / t t i a l i , t h e a c t i r e
Plate 2 shows the c o u n t e r e d images o f e l e c t r o s t a t i c p o t e n t i a l arou~td the active site o f wild type and G l u - 5 3 m u t a n t . Both images have the s a m e f e a t u r e in which a distinct large g r a d i e n t o f e l e c t r o s t a t i c p o t e n t i a l lies across t h e reaction c e n t e r c o m p o s e d of a c o u p l e o f carboxyl g r o u p b e a r i n g residues. However. t h e r e is a small d i f f e r e n c e in the s h a p e o f isopotential c o n t o u r s especially a r o u n d 53th r e s i d u e d u e to the m u t a t i o n . T h e influence o f m u t a t i o n on t h e e l e c t r o s t a t i c environm e n t r e a c h e d several active site r e s i d u e s including Glu-35, Tyr-63, T r p - 6 4 a n d Trp-109. which have b e e n c o n s i d e r e d to be i m p o r t a n t in the catalytic action o f h u m a n lysozyme [11]. A s p r o t e i n s a r c c o n s i d e r e d to have the di~'lectric constanls, 2 - 4 [3d 351, the p K , shifts at Glu-35 of h u m a n lysozyme c a u s e d by the m u t a t i o n of A s p - 5 3 to Glu was e s t i m a t e d u n d e r the c o n d i t i o n s that thc d i c l c c t r i c c o n s t a n t o f p r o t e i n region was a s s u m e d to bc e i t h e r 2 or 4 ( T a b l c I!!). Relatively small pK~, shifts (0.17 or 0.141 of Glu-35 were oh-
TABLE II! Culcutat~,d pK,, shifts at GI,-35 bvt,~ecn ~ihl type and (;h~ 53 muta,it /lllt/ltl/i Iv.~o~}'lll£"
Electrostatic I~>tenlials of wild type and Glu-53 mut~lnl human lysozyme were calculated on a 65 × 6,5 ~ 65 grid by a finite difference melhod as described in the text. Ct~rdinates were scaled ~ that the percentage of the grid filled by molecule in the longest Iinear dimension is set to 67c~-, resulting in a grid spacing of approx. 1.1 7k. pK a shifts were calculated from the change in l~)tential due lo the mutation by the method of Tanford and Roxby [311. Dielectric constant of protein region
Change in pK~
2 4
0.17
0.14
lion.
T h e relative enzymatic activity to wild type enzyme in Ght-53 h u m a n lysozyme d e c r c a s e d to the same Icvel as Ash-52 chicken lysozymc (5.5 + 2.5~'~ o f wtld type enzymc. [18]). which posscss no c o r r e s p o n d i n g hmizable earboxylate g r o u p in the reaction center. O n the o t h e r hand. X-ray crystallographic analysis r e v e a l e d that no r e m a r k a b l e change has o c c u r r e d in the active site c o n f o r m a t i o n includin~ s u s p e c t e d s u b s t r a t e recognizing r e s i d u e s except the m u t a t i o n point (Plate 1). T h e s e results t o g e t h e r with the fact that the b i n d i n g activity against e i t h e r (GlcNAc)~ or (GlcNAc~,, was not c h a n g e d significantly ( T a b l e !). strongly suggest that not only the p r e s e n c e of negatively c h a r g e d carboxylate in the r e a c t i o n center, but also the precise positianirlg of it is essential to e n h a n c e the catalytic reaction rate. T h e difference of lysis profilc against ;14. lute'us ceils b e t w e e n wild type and Glu-53 m u t a n t is poteworthy. In contrast to the iysis curve of wild type e a t ' m e , thc a b s o r b a n e e o f reaction mixture c o n t a i n i n g Glu-53 mutant did not d e c r e a s e lincarly even in the early s~ate o i lysis (Fig. 5). A similar behavior was a l ~ o b s e r v e d in the hydrolysis of p - N O 2 - ( G I c N A e ) s- The relative activity of Glu-53 m u t a n t to wild type e n z y m e d e c r e a s e d as the reaction time was p r o l o n g e d [ l ' a b l e 11). T h e s e results suggest that some dtffcrences may exist in the catalytic m e c h a n i s m s b e t w e e n wild type h u m a n lysozyme and Glu-53 mutant. In c o n n e c t i o n with this, M a l c o l m et al. r e p o r t e d the similar a n o m a l o u s behavior o f Asn-52 m u t a n t chicken I~,so~,me, w h e r e a p p a r ent b i p h a s i c kinetics with a fairly r a p i d p e r i o d o f !y~is followed by a ' s t e a d y s t a t e ' ir~ which virtually c o n s t a n t activity was o b ~ r v e d [18]. T h e p r e t r c a t m e n t with ce~.! s u b s t r a t e did not affect the lysis profile of Glu-53 m u t a n t This suggests that the a p p a r e n t inactivation a c c o m p a n i e d by cell lysis occurs in a ~eversible way. W h e n a couple of carboxylate g r o u p s are l o c a t e d
236 In contrast to the mutation of Asp-52 to Asn, the size of the side chain in the residue at issue becomes larger by a single methylene group in the mutation of Asp-53 to Glu, Asp-52 in chicken lysozyme is considered to be situated at about 3 .~, from C1 atom of the sugar residue that occupies subsite D and too far apart to form a full covalent bond [1,9,11]. The larger size in s;de chain of 53th residue in Glu-53 human lysozyme should permit a close location of the carboxylate group of 53th residue against C1 carbon atom of the sugar residue that occupies subsite D. Hence, in the case of Glu-53 mutant it is more likely that the earboxylate group of 53th residue forms a more complete covalent bond with substrate (Fig. 1), which may be too stable to perform the following bond rearrangements efficiently, as c o m p a r e d to that in wild type enzyme. Further works will be needed to clarify what mechanism is operated in the atomic level. The investigation on this line including the determination of the three dimensional structure of complex of Olu-53 mutant human lysozyme and substrate (analogs) is no v in progress. Kirby proposed that Asp-52 in chicken lysozyme is involved in the catalytic action by forming a 'less-thancomplete' covalent bond with the glycosidic reaction center (Fig. 7) as a possible explanation that satisfies both the nucleophilic assistaace of A s p - 5 2 - C O O - to the cleavage of glycosidic bond and the observed retention of configuration of anomeric center [9]. O u r data in the present study indicated the precise positioning of the negatively charged carboxylate group at the catalytic center is critical in the catalytic action of lysozyme. The nucleophilic attack (Fig. 1, lower course) to anomeric center by the carboxylate anion of 53th residue might require a more precise positioning than simple electrostatic interaction (Fig. 1, upper course), because more strict alignment of electron orbital will be necessary in the former case than in the latter case. The nucleophilic attack mechanism seems to explain our results well in the present study, in respect to the strictness in the requirement of the carbox'ylate group positioning. However, too strong bonding will result in
ctosc in each other, such as the carboxylatc group of Glu-35 (pK:, = 6.8) and that of Asp-53 (pK~, = 3.4) in human lysozyme, a movemcnt of the location of one carboxylate group may largely affect the pK~ of the other one. A!though pK~, values of ionizable groups in proteins have been generally estimated frem the p H dependence of parameters obtained in either spectroscopic experiments or kinetic experiments, it is difficult to estimate the p K a value of Glu-53 human lysozymc experimentally, because the productivity of human lysozyme by S. ceret,isiae is fairly small and the Glu-53 human lysozyme shows the anomalous catalytic behavior as described above. Gilson and Honig rcported that p K , shifts obtained by a calculation, which sloved Poisson's equation for continuum models concerning the interaction 9f charges in macromolecules, agreed well with the experimental values for mutant subtilisms [34]. Recently, the importatzce of large electrostatic potential gradient between two catalytic carboxylates in the catalytic action of lysozymes was pointed out by Dao-Pin et al. using the calculation of electrostatic fields in the active sites of lysozymes [35]. To investigate the effect of the mutation, Asp-53 to Glu, on the electrostatic potential around the catalytic residues, the calculation using the same algorithm derived from linealized Poisson-Boltzmann equation has been performed. The calculation indi.cated a similar feature of the electrostatic potential m a p around the activc site of Glu-53 mutant as compared with that of wild type enzyme (Plate 2) and the relatively small change in p K , value of Glu-35 during the mutation (Table liD. This relatively small effect of mvtation on the pK~ ~alue of Glu-35 seemed to be consistent with the experimental results that the pH dependence of lytic activity against M. luteus cells in Glu-53 mutant was similar to that in wild type enzyme (Fig. 6). Therefore, the large decrease of enzymatic activity in Glu. 53 mutant against either bacterial cell substrate or p - N O 2(GlcNAc)s may not be derived from the large change in pK~ of Glu-35 due to the approach of ionized carboxylatc of 53th residue to unionized Glu-35 residue.
I
H
~O-(GIcNAc)z EF
HO-.-.. (C~cblAc)3.~ 0 . . ~
,-c
H
.. II O.~
\oJ,
O,.(GIcNAc)z
N ~ HO~(GIc ~Ac,~.0 ~ . . . ~ .,,K. a.~
'
n
H,,N~-~
'
H.
E.F
.40 H/
Fig. 7. Plausible part cipation mechanism of Glu-35 and Asp-53 in the bond rearrangement steps of hydrolysisof (GIcNAch, by human lysozyme.
237 a relatively stable acyi-glycoside compound which is subject to be hydrolyzed at the ester bond rather than at the acetal center. This is obviously not the case in lysozyme catalysis, because the experimental data by others indicate the retention of configuration at the anomeric center [11]. Consequently, the above considerations lead us to the conclusion that Asp-53 in human lysozyme participates in stabilizing the development of oxocarbonium ion intermediate not simply in an electrostatical manner, but partly in a nucleophilicat manner by forming a highly tuned incomplete covalent bond as shown in Fig. 7. Acknowledgement We thank Dr. Y. lshizuka for the help in the measurement of CD spectra. References 1 Blake, C.C.F.. Johnson. LN., Malt. G.A.. North. A.C.T., Phillips, D.C. and Sarma, V R . (19671 Proc. R. Soc. Lond. B!67, 378-388. 2 Joll~s, P. and Joll~:s. J. (19841 Mol. Cell. Biochem. 63, 165~189. 3 Vernon. C.A. (19671 Proc. R. Soc. Lond. B167. 389-401. 4 Chipman, D.M. and Sharon. N. (19691 Science 165, 454-465. 5 Paice. H.G., Desrochers, M.. Rho, D.. Jusarek. L. Roy. C., Rollin, C.F., DeMiguel. E. and Yaguchi. M. (19841 Biotechnology 2, 535-539. 6 Rouvinen, J., Berglors, T.. Teeri, T., Knowles. J.K.C. and J,~nes. T.A. (19901 Science 264, 380-386. 7 Hoj, P.B., Rodriguez. E.B., Stick. R.V. and Stone, B.A. (19891J. Biol. Chem. 264, 4939-4947. 8 Kuramitsu, S., Ikeda, K., Hamaguchi, K., Fujio, H.. Amano, T., Miwa, S. and Nishina, T. (19741J. Biochem. (Tokyo) 76. 671 -e,83. 9 Kirby, A.J. (19871 CRC Crit. Rev. Biochem. 22, 283-315. | 0 Fersht. A.R, (19851 in Enzyme Structure and Mechanism, 2nd Edn. (Freeman, W.H., ed.), New York, Ch. 2 and 15. I 1 lmoto, T., Johnson, LN,, North. A.C.T., Phillips. D.C. and Rupley. J.A. (19721 in The E ~ ' m e s (Boyer. P.D.. ed.). Vol. 7, Ch. 21, Academic Press. New York. 12 Post, C.B. and Karplus. M. (19861 J. Am. Chem. Soc, 108. 1317-1319. 13 Parsons. S.M. and Raftery. J.A. (I9691 giochemistn 21. 21~72192.
14 Parsom,. S.M.. Jao, L., Dahlquist_ F.W., Borders. C.I... Jr.. Gr~ffl. T., Racs. J. and Rafters. M.A. (19691 Biochemist~ ~, 7(){J- 7i2. 15 Eshdat, Y.. Dunn, A. and Sharon. N. (197..~) Proc. N a i l A c a d Sci. USA 73. 1658-1662. tt~ Yamada, t]., lmoto. T. and Noshita. S. (1~,~,2; Bi~lchc,nistr 3 21. 2187-2192. 17 Kumki. R., Yamada, H., Moriyama. T. and lmoto, r. (19881 J. Biol. Chem. 261. 13571-13574. 18 Malcolm. B.A.. Rosenberg, S.. Core~,. MJ., Allen. J.S.. DeBaeterlier. A. and Kirsch. LF. (19891 Proc. Natl. Acud. Sci. USA 86. 133-137. 19 Muraki, M.. Jigami. Y., Morikawa, M. and Tanaka, It. (19~7~ Biochim. Biophys. Aeta :~11,. 376-3~0. 2ll Parry. R.M,. Jr.. Chandan, R.C. and Shahani. K.M. (19691 Arch. Biochem. Biophys. 103, 50-65. 21 Suzuki, K.. Iehikawa. K, and Jigami. Y. (19891 Mol. Gen. Genet. 219. 58-64. 22 Jigami. Y.. Muraki. M.. Harada. N. and Tanaka. H. (19861 Gene 43. 273-279. 23 Osserman. E.F.. Cole, S.J.. Swan. I.D.A. and Blake. C.C.F. ( 19691 J. Mol, Biol. 46, 211-212. 24 Artymiuk. P.J. and Blake. C.C,F. ( 1981 ) J. Mol. Biol. 152. 737-7,52. 25 Chipman, D.M. and Sharon, N. (19691 Science 165. 454-465. 26 Macllvaine. T.C. (19211J. Biol. Chem. 49. 183-186. 27 Locquet. J.P.. Saint-Blanccard. J. and J(~lh?s, P. (19681 Biochim. Biophys. Acta 167, 150-153. 28 Nanjo. F.. Sakai. K. and Usui. T. (198S1J. Biochem (Tok3o) 1()4, 255-258. 29 Klapper. I., Hagstrom. R.. Fine, F.. Sharp, K. and llonig, g (19861 Proteins I. 47-59. 30 Gilson. M.K., Sharp, K.A. and Honig, B H (19871 J. Comp. C h e m 9. 327-335. 31 Tanfnrd. C. and Roxby. R. (19721 Biochemistr':, l h 21'-92-2198. 32 Halper, J.P.. Lato~,'itzki, N.. Bernstein. Fi. and Be',chok. S. 119711 Proc. Natl. Acad. Sci. USA 68. 517-522. 33 lkeda, K., Hamaguchi. K.. Miwa. S. and Ni~,hina. T_ (1972~ J. Biochem ( T o ~ ' o ) 71. 371-378. 34 Gilson, M.K. and Iqonig, B.H. (19871 Nature 3311, S4-8fl. 35 Dao-Pin. S,. kilo, D-I. and Remington. S.J, (19N91 Ploc. Natl. Acad. Sei. USA 86. 53fll-53(~5. 3b Rupley. J.A.. Gates. V. and Bilbrey. R. (196S) J. Am. Chem, S(;c. t~l. 5633-5635. 37 Lowe. G. (19,671 Prt~¢. R. Soc. Lond. Bit57. 431-434. 38 Tsai. C.S., Tang. J.~. and Subbara ~. S.C ([t~hg~ Biochcm J. 114_ 52t)-534. 39 Warshel, A. and ke~,itt, M, (1976)J. Mol. Biol. 103, 227-249 41t Bakthavahalam. V. and Czarnik. A,W. 11987/Tctrahedr~m kctt. 2~4. 2925-2928.
