Determination of the Roles of Glu-461 in ,&Galactosidase (Escherichia coli) Using Site-specific Mutagenesis*

Site-directed substitutions (Asp, Gly, Gln, His, and Lys) were made for Glu-461 of beta-galactosidase (Escherichia coli). All substitutions resulted in loss of most activity. Substrates and a substrate analog inhibitor were bound better by the Asp-substituted enzyme than by the normal enzyme, about the same for enzyme substituted with Gly, but only poorly when Gln, His, or Lys was substituted. This shows that Glu-461 is involved in substrate binding. Binding of the positively charged transition state analog 2-aminogalactose was very much reduced with Gly, Gln, His, and Lys, whereas the Asp-substituted enzyme bound this inhibitor even better than did the wild-type enzyme. Since Asp, like Glu, is negatively charged, this strongly supports the proposal that one role of Glu-461 is to electrostatically interact with a positively charged galactosyl transition state intermediate. The substitutions also affected the ability of the enzyme to bind L-ribose, a planar analog of D-galactose that strongly inhibits beta-galactosidase activity. This indicates that the binding of a planar "galactose-like" compound is somehow mediated through Glu-461. The data indicated that the presence of Glu-461 is highly important for the acid catalytic component of kappa 2 (glycosylic bond cleavage or "galactosylation"), and therefore Glu-461 must be involved in a concerted acid catalytic reaction, presumably by stabilizing a developing carbonium ion. The kappa 2 values with o- and p-nitrophenyl-beta-D-galactopyranoside as substrates varied more or less as did the K8 values, indicating that most of the glycolytic bond breaking activity found for the enzymes from the mutants with these substrates was probably a result of strain or other such effects. The kappa 3 values (hydrolysis or "degalactosylation") of the substituted enzymes were also low, indicating that Glu-461 is important for that part of the catalysis. The enzyme with His substituted for Glu-461 had the highest kappa 3 value. This is probably a result of the formation of a covalent bond between His and the galactosyl part of the substrate.

Site-directed substitutions (Asp, Gly, Gln, His, and Lys) were made for Glu-461 of &galactosidase (Escherichia coli). All substitutions resulted in loss of most activity.
Substrates and a substrate analog inhibitor were bound better by the Asp-substituted enzyme than by the normal enzyme, about the same for enzyme substituted with Gly, but only poorly when Gln, His, or Lys was substituted.
This shows that Glu-461 is involved in substrate binding. Binding of the positively charged transition state analog 2-aminogalactose was very much reduced with Gly, Gln, His, and Lys, whereas the Asp-substituted enzyme bound this inhibitor even better than did the wild-type enzyme. Since Asp, like Glu, is negatively charged, this strongly supports the proposal that one role of Glu-461 is to electrostatically interact with a positively charged galactosyl transition state intermediate. The substitutions also affected the ability of the enzyme to bind L-ribose, a planar analog of D-galactose that strongly inhibits /!Igalactosidase activity. This indicates that the binding of a planar "galactose-like" compound is somehow mediated through  fi-Galactosidase from Escherichiu coli is a disaccharidase which catalyzes the hydrolysis and transgalactosylis (Huber et al., 1976) of P-D-galactopyranosides.
Its mechanism of action has not been firmly established, but it has been suggested that it functions in a somewhat analogous way to lysozyme (Sinnott, 1978). Thus, it has been proposed that pgalactosidase has a group which acts as a general acid (donating a proton to the glycosidic oxygen) and a group which stabilizes a galactosyl intermediate, allowing HZ0 to react. Studies have shown that the group which acts as the general acid is probably Tyr-503 (Ring et al., 1985. Wallenfels and Malhotra (1961) suggested that there is an imidazole group at the active site of ,&galactosidase that covalently stabilizes a galactosyl intermediate for reaction with water, but other researchers (Tenu et al., 1971;Sinnott and Withers, 1974;Sinnott and Souchard, 1973;Sinnott, 1978) proposed that a carboxyl group acts as a counter ion to a transition state carbonium ion form of galactose. The presence of a negative counter ion was further supported by studies (Huber and Gaunt, 1982;Legler and Herrchen, 1983) which showed that amino sugars and amino alcohols are very good competitive inhibitors of &galactosidase (especially if the structures of the sugars or alcohols resemble D-galactose).
In 1984, Herrchen and Legler used an irreversible active site-directed inhibitor (conduritol-C-cis-epoxide) to identify Glu-461 as a residue with a carboxyl group that might be involved at the active site of @-galactosidase.
In addition to stabilizing a transition state carbonium ion, studies of Sinnott and Souchard (1973), Rosenberg and Kirsch (1981), and Withers et al. (1988) have provided some very good evidence that the putative active site residue with the negative charge also forms a transient covalent bond with the transition state carbonium ion form of galactose. Sinnott and Souchard (1973) proposed that this is necessary to prevent back reaction with aglycon products which are not rapidly released from the enzyme. Bader et al. (1988) showed that site-specific substitution of Glu-461 with Gln results in the loss of >99% of the pgalactosidase activity in a purified preparation. Cupples and Miller (1988) introduced 12 additional substitutions for Glu-461 and showed that all of them decreased the activity of /3galactosidase (by at least 90%) in crude preparations. We describe here the purification and kinetic analysis of five of these P-galactosidases. The purpose of this study was to clarify the role of Glu-461 in P-galactosidase.

Chemicals-
The purest reagents possible were always obtained. Enzyme Production and Purification-+Galactosidases with amino acid substitutions for Glu-461 were produced as described previously Cupples and Miller, 1988)  It is thought that Tyr-503 acts as a general acid catalyst in this step (Ring et al., 1985. S' mce  is probably at the active site (Herrchen and Legler, 1984;Bader et al., 1988;Cupples and Miller, 1988), it could be important for stabilization of a positively charged carbonium ion form of galactose. In addition, Sinnott and Souchard (1973) and Rosenberg and Kirsch (1981)  or PNPG as substrate, the following equations hold: k,,, = k2k3/(k2 + k:s) and K, = K, (k,/(k, + ka)). It follows that when k, is rate-determining, the values become k,,, = k, and K, = KS, and when ks is rate-determining, the values become k,., = ka and K, approaches 0.
The effect of nucleophils (which replace water) on the rate of reaction can be used to determine whether k3 is rate-determining (Deschavanne et al., 1978;Huber et al., 1984). This can be seen from inspection of the following equation which predicts how the apparent k,., value changes as a function of the added nucleophil [N]: apparent k,,, = (&(& + k~([N]/K,"))/(k, + ka + (k, + k,)([N]/K,")). If kz and k, are large relative to kS, the rate of the reaction will increase as a function of the nucleophil concentration.
If, on the other hand, k, is small relative to kJ, the rate of the reaction will not increase even if k, is large. We found that azide ion was a very diagnostic nucleophil for the purpose of finding whether k, or kB was rate-determining. The kinetics showed that k4 for azide ion was large; and gas chromato- results showed that when 100 mM sodium azide was present in the assay, the D-galactose peak could no longer be detected, but a new peak appeared which we assumed to be /3-D-galactosylazide. That showed that the azide was completely replacing water in the reaction. When kS is rate-determining, the addition of azide could be used to determine a lower limit for the value of kf since the above equation predicts that if the reaction with azide is much faster than kS, the apparent kc,, will approach k,k,/(k, f kd) as a function of nucleophil concentration. This value is a lower limit for k, since it is impossible for kl to be smaller than k,k,/(k, + k,). The above definition of the K,,, value one would obtain when k, is rate-determining (i.e. K, approaches 0 when kB is rate-limiting) indicates that if the K,,, of an enzyme from a mutant is very low (approaches 0) relative to the K, of the wild-type enzyme, the ka is probably rate-determining.
(This was found to be the case for some of the enzymes from the mutants when ONPG was the substrate.) Some caution must be used, however, since the K, can also be small because K, has a low value.
Enzyme Assays with ONPG and PNPG-fl-Galactosidase was assayed using two synthetic substrates (ONPG and PNPG) in addition to the natural substrate lactose (see below). Both ONPG and PNPG were used because they react very differently when used as substrates of the wild-type enzyme.
The reaction with ONPG using normal /3galactosidase is such that the values of k, and k, are quite similar, whereas with PNPG, k, is much smaller than ks. The assay solution was a TES buffer (30 mM, pH 7.0) with NaCl (140 mM) and MgS04 (1 mM). Because the enzymes from the mutants were usually quite unreactive, large amounts of enzyme solution had to be added for assay. Since the enzymes were stored in buffer solutions with azide (to prevent bacterial growth) and mercaptoethanol (to keep the enzymes reduced), the transfer of the large amounts of enzyme solution to the assay solutions resulted in the addition of significant amounts of azide and mercaptoethanol into the assay. These were potent nucleophils (replacing water), and even additions of small amounts of these reagents affected the rate dramatically. Therefore, before assaying, the enzyme solutions were always desalted into assay buffer (without substrate) using a fast protein liquid chromatography desalting column (Pharmacia LKB Biotechnology Inc.). For the assays, tubes with assay solution were preincubated in a 25 "C water bath for 10 min before addition of the enzyme. After the enzyme was added, the absorbance at 420 nm was followed at 25 "C as a function of time. The value obtained were converted to units/milligram (1 unit is defined as 1 pmol of nitrophenol produced per min). The kinetic data were analyzed by the simple weighting method described by Cornish-Bowden (1976). The K, values were calculated by the methods described by Deschavanne et al. (1978) and by Huber and Gaunt (1982), which take into account the fact that some inhibitors can also act as nucleophils (replacing water as discussed above). The V values were converted to k,,, (second-').
For routine assay during purification, only 1 mM ONPG was used. To determine whether k, or kj was rate-determining and to obtain estimates of kp, sodium azide (10 or 100 mM) was added.
Enzyme Assays with Lactose-Enzyme (0.1 mg/ml) was added to 20 mM lactose in the TES assay buffer.
Samples (200 ~1) were taken at various times and added to plastic centrifuge tubes (1.5 ml) immersed in liquid NP (to instantly stop the reaction). The tubes also contained 100 ~1 of internal standard (1.0 mM inositol and 3.5 mM pphenyl-D-glucopyranoside).
Holes were drilled into the caps of the tubes to allow lyophilization.
After lyophilization, 450 ~1 of dimethylformamide, 300 ~1 of hexamethyldisilazane, and 150 ,.d of trimethylchlorosilane were added to silylate the sugars. The silylation reaction was allowed to proceed for 2 days, after which two layers had formed. The top layer contained the silylated sugars (Ellis, 1969).  (Tenu et al., 1972)), and the reaction was then started by adding substrate. Stability Studies-The enzyme (0.5 mg/ml) was incubated in the TES assay buffer in a 55 "C water bath. Aliquots (50 ~1) were removed at intervals and placed into tubes which were in ice. After assaying, the percent activity remaining was determined as a function of time.

RESULTS
Purification- Fig.  1 is an SDS-PAGE analysis of aliquots from each of the enzymes. The enzymes were highly pure and migrated to the same position as wild-type P-galactosidase. The /3-galactosidases were found to precipitate in the same ammonium sulfate cuts in which the wild-type enzyme precipitates, and they eluted from columns at similar positions as did the wild-type enzyme.
Stability- Fig.  2 shows enzyme stability at 55 "C. The enzymes from the mutants were less stable than the wild-type jY-galactosidase.
All of the enzymes from the mutants were entirely stable when kept in assay medium for 1 day at 25 "C.
Kinetic Constant Values- Table  I  Aliquots were removed at various times and added to tubes in ice. The tubes were later incubated at 25 "C for 5 min, and the reaction was started by the addition of the assay buffer which was kept at 25 "C. 0, wild-type enzyme; 0, E461Q-@-galactosidase; +, E461H-B-galactosidase; 0, E461D-@-galactosidase; A, E461G-fi-galactosidase; n , E461K-&galactosidase.
PNPG. The enzymes from mutant sources had much lower activities than did wild-type p-galactosidase.
Except for the Km of E461H-P-galactosidase (which was a little larger than that of the wild-type enzyme), the Km values with ONPG were all much lower than the K,,, of the normal enzyme. The K,,, value for E461Q-@-galactosidase was particularly low. When PNPG was substrate, the K,, values were also lower (except again for E461H-P-galactosidase), but with this substrate, they were not decreased by such a large factor. Since K,,, = K,(k:,/( k2 + k:,)), the low K,, values were probably due to the fact that the k3 values were much lower than the k2 values rather than because K, was small. Because only lower limits of k, were found in some cases, only lower limits of K, were also obtained.
With ONPG as substrate, lower limits of k2 were obtained in every case from the maximum rate that was observed upon addition of a high concentration of azide. The lower limit for E461Q-P-galactosidase shown is probably a substantial underestimate of the actual value obtained since the Km was very low in that case. With PNPG as substrate, kp was ratelimiting for E461D-, E461G-, and E461H-p-galactosidases.
Therefore, for these enzymes, actual k2 values with PNPG could be obtained.
Only a lower limit of k2 for E461Q+galactosidase could be established.
In the case of E461K-Pgalactosidase, the PNPG k,,, in the presence of 10 mM azide was essentially one-tenth of the k,,, value with ONPG as substrate when 100 mM azide was present. This indicates that kd is at least 10 times as large as the PNPG k,,, found with 10 mM azide, and the k,,, (being 10 times lower) would therefore essentially be equal to k2. Values of k,,, were assumed to be equal to kn if adding azide increased the k,,, more than IO-fold because this meant that the value of k, was at least 10 times as large as that of k.3. This was the case for E461G-, E461&-, and E461K-P-galactosidase with ONPG as substrate.
When PNPG was the substrate, this was the case for E461Q-and E461K-P-galactosidases. As expected, the ONPG k,,, values for these latter two enzymes were essentially equal to the PNPG k,,, values (because k,,, = kl in each case). For E461D-and E461H-/3galactosidases, we were able to calculate ranges of values the Iz3 values would fall. The lower limits of the ranges (0.10 s-' for E461D-fi-galactosidase and 42 s-l for E461H-fi-galactosidase) were determined using the fact that the k,,, values of the enzymes for ONPG were 0.10 SC' for E461D-/3-galactosidase and 42 s-l for E461H-@-galactosidase. Since k,,, is equal to k2kR/(k2 + k3), it is impossible for k:i to be less than kc,,. The upper limits for kS were determined using the fact that the k2 for E461D-P-galactosidase is greater than 0.16 s-' and greater than 71 s-' for E461H+-galactosidase (k,,, values with 100 mM azide). From calculations using ky 2 0.16 s-' for E461D-P-galactosidase and kz 2 71 s-l for E461H-/3-galactosidase, k,,, values of 0.10 and 42 s-i for the two enzymes, and the equation for k,,, (kcat = k2k3/(k2 + kJ), the value of k3 has to be less than 0.27 s-l for E461D-Pgalactosidase and less than 103 s-' for E461H-@-galactosidase. Thus, the value of kB for E461D-p-galactosidase is between 0.10 and 0.27 s-', whereas that for E461H-P-galactosidase is between 42 and 103 s-l. A good overall indication of the values of k3 for each of the enzymes was therefore obtained; and except for E461H+galactosidase, the kS values obtained were small. Even in that case, the kS value was at least an order of magnitude smaller than the kn value of the wild-type enzyme.
The K, values for IPTG are shown in Table II. Also given are the I& values either obtained directly (in the cases where k2 was rate-determining) or calculated from the Km value using the equation for K,,, (K,,, = KJk,/(k, + ks)) and using the kZ and kS values (or range of values) obtained above. For E461Q-P-galactosidase, the calculated KS value (a lower limit) is probably a substantial underestimate of the actual value since the lower limit for k2 is probably much lower than the true k, value. In general, the Ki values for IPTG and the calculated or true values of KS were small compared to normal for E461D-galactosidase, about the same as normal for E461G-@-galactosidase, and large for the other enzymes. Table II also  determine accurately (large concentrations of inhibitor have to be added with only small rate effects) and since they only indicate that binding is very poor, we have not given precise values in those cases. E461D-@galactosidase was the only enzyme that was inhibited well by 2-aminogalactose (it was actually inhibited better than was the normal enzyme). E461H-P-galactosidase was inhibited quite well by L-ribose, but the other enzymes were not.
Effect of EDNA-The effect of EDTA is presented in Table  III. E461D-/I-galactosidase was highly activated by removal of M$+, E461H$-galactosidase was slightly activated, and E461Q-P-galactosidase was strongly inhibited. E461K-P-galactosidase was inactivated only -3O%, whereas E461G-flgalactosidase was inhibited to about the same extent as normal /3-galactosidase.
pH Profiles- Fig.  3 shows the k,,, and Km values for the enzymes as functions of pH. P-galactosidase is unstable and hard to assay at low pH even with normal p-galactosidase; and thus, in most cases, the kinetic constants were not evaluated at pH values below 6.0. Note that the kcat values for E461D-/3-galactosidase with ONPG increased dramatically as the pH was increased, with a midpoint of -8.4, but that there was not a similar increase when PNPG was the substrate. The k,,, values for E461G-, E461Q-, and E46lK-P-galactosidases with ONPG were more or less constant as functions of pH. E461H-/3-galactosidase had a high k,,, value at pH 6, but the values dropped in a complex manner, with a midpoint of -7.0. At pH values higher than 8.0, the data with E461H-,&galactosidase represent the activities found after "substrate inactivation" had occurred (as will be described below). For E461G-, E461Q-, and E461K+galactosidase, the Km values increased at pH values less than 7.0, but the values were essentially constant at pH values between 7.0 and 10.0. The K,,, values for E461D-P-galactosidase increased somewhat as the pH increased. The K,,, values for E461H-P-galactosidase increased and then decreased as the pH was increased. Fig. 4 shows that the rate of catalysis with E461H-@galactosidase decreased as a function of time in the presence of substrate at pH 8.0. Fig. 4 also shows that the activity at pH 8.5 was very low even after only 1 min, but this was because the activity in that case dropped very rapidly to low values. This enzyme was, however, stable at high pH because it did not irreversibly lose activity at high pH (8 to 10) in the absence of substrate (enzyme was incubated without substrate at high pH and then assayed at pH 7.0).
Action on Lactose-The enzymes all had very low activities when lactose was the substrate. E461K-P-galactosidase had no activity even though 0.1 mg of enzyme was incubated with 20 mM lactose for 65 h. For the other enzymes, the activities found were 0.0026 unit/mg for E461D-P-galactosidase, 0.0059 unit/mg for E461G-P-galactosidase, 0.0015 unit/mg for E461Q+galactosidase, and 0.0007 unit/mg for E461H-P-galactosidase (compared to 28.4 unit/mg for the wild-type enzyme). In the presence of 50 mM lactose, very similar values were obtained. It was not practical to determine V and K,,, values with lactose because very large amounts of the enzyme were needed, and the assays required very long incubation times (>60 h). However, the values obtained with 20 mM lactose are essentially equal to kz for the following two reasons. First, the rates with 50 mM lactose were not very different from those with 20 mM lactose, suggesting that 20 mM was well over the Km values for the enzymes for lactose; and thus, the rates are equal to V. Second, the k3 values for P-galactosidase should be the same for every substrate; and since the KS values obtained with ONPG and/or PNPG (Table I) were all much higher than the rates found here with lactose, k2 must be rate-determining, and these rates must therefore be equal to kz.

DISCUSSION
It is important, when doing site-directed studies, that several substitutions be made at a position to definitively establish that a particular residue is important for catalysis. A loss of activity when only one substitution is made may be due to a simple conformation change rather than to the fact that the group is essential for catalysis. Also, if the changes in activity that result from a series of substitutions follow some rational pattern, it is easier to determine the role of the residue. Ideally, every possible change should be made, and the effects studied in detail. Obviously, that is not usually practical. The specific substitutions made in this study were chosen both on biochemical considerations and on the basis of preliminary screening Cupples and Miller, 1988). The substitution of Asp for Glu-461 introduced another side chain with an identical negative charge. However, Asp is shorter than Glu, and there could be a "space" or "cavity" present. The substitution with Gly removed the entire side chain at position 461. This would also open up a space or cavity. (In both of these cases, the space or cavity could, of course, collapse and fill in.) Substitution with Gln was made to have a neutral substitution without a space or cavity. The substitution by His was carried out for two purposes: 1) to determine how a group without a negative charge but which could donate electrons would interact, and 2) to see how a group which is a better nucleophil than Glu (because the electron densities at the nitrogens of His are greater than those at the oxygens of Glu) interacts. Finally, the Lys substitution introduced a positive charge in place of a negative charge. Lysine, of course, also has a longer side chain than has Glu.
The study reported here shows that Glu-461 is highly important for the overall catalytic action of /I-galactosidase. All of the enzymes from mutant sources (except E461H-Pgalactosidase) had very low activity with each of the substrates studied (~0.3% of normal in even the best case). However, sufficient residual activity remained with each of the enzymes purified from the mutants for us to define the roles of Glu-461 in @-galactosidase.
Physical Characteristics of Enzymes-The normal and substituted enzymes were >95% pure in every case (see Fig. 1). The specific activities of the enzymes isolated from the mutant cells were low therefore because the activity was adversely affected, not because the enzymes were impure. The enzymes from the mutants appeared in the same fractions as normal P-galactosidase throughout purification. This suggests that the overall physical properties which relate to purification (e.g. size, charge, aggregation, etc.) of each of the enzymes from the mutants were similar to those of the wild-type enzyme. Especially important is the fact that the enzymes were all in tetrameric form (all eluted from the Superose-12B size fractionation column at similar positions to the elution position of the wild-type enzyme).
Each of the substituted enzymes was less stable to heat (55 "C) than was the wild-type enzyme. This was not entirely unexpected. The state of the active site of most enzymes is very important in establishing their stability (Yao et al., 1984;Chothia and Lesk, 1985). Despite their instability at 55 "C, the enzymes were all stable for at least 5 months at 4 "C and for at least 1 day at 25 "C! (when incubated in the assay buffer).
Substrate and Inhibitor Biding-The enzymes had differing abilities to bind both IPTG (a substrate analog inhibitor) and the substrates. E461D-fi-galactosidase bound IPTG and the substrates better than did normal @galactosidase. E461G-P-galactosidase bound IPTG and the substrates in a more or less comparable manner to the normal enzyme. E461Q-, E461H-, and E461K-P-galactosidases bound these compounds much less well. This is the first time that Glu-461 has been implicated in substrate binding. Gln, His, and Lys are either the same size or larger than Glu, whereas Asp and Gly are smaller. Substrates are often held in a constrained (destabilized) state at enzyme active sites (Jencks, 1975). If the substitution of Asp for Glu creates an extra space or cavity, this may relieve the constraint to a certain extent and allow better binding. Size, however, does not seem to be the only critical factor for binding since Gln is about the same size as Glu; and yet binding is disrupted.
Also, Lys is no more disruptive to binding than is Gln despite the considerable difference in size and, of course, charge.
The binding capacity seems to be inversely correlated to the kz values with both ONPG and PNPG. This implies that some sort of steric effect (e.g. strain, desolvation) is important for the kz values with these substrates and is mediated in some way by the position occupied by Glu-461.
E461D-galactosidase was the only enzyme from among the mutants which bound 2-aminogalactose (a positively charged transition state analog) well. This strongly implies that the negative charge of Glu-461 is important for binding amino sugars and amino alcohols tightly (Huber and Gaunt, 1982;Legler and Herrchen, 1983) and supports the proposal that one of the roles of Glu-461 is as a counterion for stabilization of a carbonium ion galactosyl transition state intermediate. It is interesting that E461D-P-galactosidase is so unreactive despite binding 2-aminogalactose so well. The other inhibitor studied was L-ribose (Huber and Brockbank, 1987). On the whole, the enzymes from the mutant sources bound L-ribose much more poorly than did the wildtype enzyme. The only exception was E461H-P-galactosidase, and it had the highest kB value (see below). These results mean that Glu-461 must play a role of some kind in destabilizing galactose into a planar conformation and that this is important for the hydrolysis (k3) step. Effect on Galactosylution-The k2 values (representing galactosylation) were all low compared to those of the wild-type enzyme. However, relatively speaking, they were quite large with ONPG, smaller with PNPG, and very small with lactose for every substitution tested. These results are best explained by assuming that Glu-461 is necessary for the acid catalytic component of kp (in a concerted fashion along with the protondonating residue). Sinnott and Souchard (1973) and Sinnott et al. (1978) reported that glycosidic cleavage of P-galactosidase substrates will occur to a certain extent without acid assistance if the leaving aglycon is acidic (i.e. a good leaving group). Presumably catalytic forces, such as strain or desolvation, are responsible for the nonacidic part of the cleavage in those cases. The aglycon moiety of lactose (glucose) is clearly much more basic than o-or p-nitrophenol; and thus, one would expect that acid catalysis would be (comparatively) a much more important part of the catalysis with lactose (since other effects are unimportant). The results with the substituted enzymes relate to this. The k2 values with ONPG were quite high despite the substitutions for Glu-461 because o-nitrophenol is a good leaving group, and catalytic forces other than acid assistance have significant effects. The kz values were very low with lactose because acid assistance is highly essential and is poor without Glu-461. With PNPG, factors such as strain or desolvation are probably not as significant as with ONPG. E461D-P-galactosidase had low kS values with each substrate tested. Distance differences could be important.
The k2 for E461K-P-galactosidase with lactose as substrate was essentially 0. Lysine, having a positive charge, would probably destabilize the transition state. Therefore, the values of kS for this enzyme with ONPG and PNPG must be totally due to factors other than acid assistance.
Effect on Degalactosykztion-The results showed that Glu-461 is also highly important for the step in which water reacts with the enzyme-galactosyl intermediate to form galactose (hydrolysis or degalactosylation).
The kS value was dramatically lowered as a result of every substitution made except for the substitution with His (the effects on k3 for E461H$galactosidase will be discussed separately). The most logical explanation for the effect is that Glu-461 optimally stabilizes a carbonium ion intermediate for reaction with water. The action of Glu-461 in degalactosylation is one of the roles that a carboxyl group at the active site of P-galactosidase was assumed to have (Sinnott, 1978). This, however, is the first hard evidence for this.
Effect of M$+--The data indicate that Mg2+ strongly modulates the effect of the substitutions.
It is thought that Mg" is important in maintaining proper structure at the active site of /3-galactosidase and does not play a direct role in catalysis (Case et al., 1973), and it has been proposed that it is important in aligning the proton-donating group for acid catalysis . The data presented here suggest that Mg2+ may also be important in aligning the carboxyl group of Glu-461, a fact that has not been realized before.
Effect of pH-The k,,, values for E461G-, E461Q-, and E461K-P-galactosidases were unaffected by pH with ONPG (Fig. 3). This means that the rate-determining step was independent of pH in the range studied. Since the rate-determining step with ONPG is k3 (at least at pH 7), neither general acid nor general base catalysis seems to be involved in kF for these three enzymes. In the normal enzyme, there is good evidence that general base catalysis is involved in the value of 123 (Ring et al., 1985). Since Glu-461 seems to be important for general acid catalysis, the absence of a group that could effectively function in place of Glu-461 should have the effect seen.
The activity of E461D-P-galactosidase with ONPG was strongly dependent upon pH (midpoint was -8.4). One possible reason could be that the pH may cause a conformation change that shifts the Asp carboxyl group nearer to the carbonium ion; and thus, the rate could increase. Another plausible reason could be that the k,., increase with pH may reflect the ionization of Tyr-503 (Ring et al., 1985, acting as a general base catalyst in removing a proton from water. The pH profile with PNPG supports this idea. Since k2 is rate-determining for PNPG, there should not be an increase because base catalysis is not involved in glycosidic bond cleavage. E461H-P-galactosidase (Possible Covalent Znteraction)-The properties of E461H-@galactosidase were so different from those of the other enzymes that they deserve separate comment. E461H-P-galactosidase was quite reactive with both