Uridine Diphosphate Galactose-4-Epimerase - jbc.org · Uridine Diphosphate Galactose-4-Epimerase...

8
THE Jonma~ OF Bmworca~ CHEMISTRY Vol. 250, No. 18, Issue of September 25, pp. 7099-7105, 1975 Printed in U.S.A. Uridine Diphosphate Galactose-4-Epimerase URIDINE MONOPHOSPHATE-DEPENDENT REDUCTION BY (Y- AND P-D-GLUCOSE* (Received for publication, January 14, 1975) UAN G. KANG, LYNN D. NOLAN, AND PERRY A. FREY$ From the Department of Chemistry, The Ohio State University, Columbus, Ohio .@‘2lO SUMMARY Rates of UMP-dependent reduction of the DPN+ associ- ated with Escherichia co2i UDP-galactose-4-epimerase at 27” and 0.2 M ionic strength in 0.1 M Tris .HCl buffer, pH 8.5, are reported. The reaction exhibits excellent pseudo-first order behavior when D-glucose is at anomeric equilibrium. The effects of [UMP] and [glucose] on the observed first order rate constants are consistent with the following equa- tion. The symbols I$ are empirical parameters. 41 42 “” = ” ’ LUMP] ’ (glucose] ~ -+ h [UMP][glucosel The data indicate that the pathway involves random equi- librium binding of UMP and glucose followed by rate-limiting decomposition of the ternary complex to epimerase .DPNH. The binding parameters indicate that the principal activating effect of UMP is not simply to increase the affinity of the enzyme for glucose. UMP appearsto increase the reactivity or availability of enzyme-bound DPN+. The kinetic isotope effect for the reaction of D-[l-2H]glucose (kH/kD) is 4.2, which confirms that C-l is oxidized and that hydride transfer is rate limiting. Both of the purified anomers, CY- and /3-D-glucose, reduce the enzyme-bound DPN+. As indicated by the deviations from pseudo-first order kinetics because of concurrent mutarotation, the /3 anomer is the more reactive, reacting about 4 to 5 times faster than the o(anomer at concentrations well below saturation. It is suggested that the lack of stereo- specificity in this reaction may be attributed to the two anomers being productively bound with their opposite faces projecting toward C-4 of bound DPN+. Nonstereospecific oxidation of CY- and P-D-glucose may be a model for the mechanism of UDP-hexose epimerization, which also in- volves nonstereospecifichydride transfer. The tightly bound DPNf associated with the UDP-galac- tose-4-epimerases (EC5.1.3.2) purified from Escherichia coli * This research was supported by (;rant AM 13502 from the National Institute ol Arthritis, Metabolism, and Digestive IXs- eases. $ To whom correspondence concerning this paper should be ad- dressed. or Saccharomyces jragilis can be reduced to DPNH by a variety of aldohexoses and aldopentoses in a reaction which is absolutely dependent upon the presence of UMP (l-4). The reduced holo- enzyme is catalytically inactive on normal substrates, because in the epimerization pathway, the pyridine nucleotide must initially be in its oxidized form. The epimerase.DPNH complex produced in this reaction can be reoxidized and reactivated by nucleoside diphosphate 4-keto-6-deoxyhexoses, by myo-inosose-2, or by 2-ketoglucose (5-7). It has been suggested that this reaction may be a model for the first step in the epimerization of substrates (5). Several re- cent findings concerning hydrogen transfer specificity exhibited by sugars effective as reducing agents in this reaction have cast doubt upon its relevance as a model. In particular, the reaction has been shown not to involve highly specific oxidation of the reactive sugars at carbon 4. Rather, in radiochemical experi- ments, glucose has been found to be oxidized at carbon 1, the reducing carbon, but not at carbon atoms 2, 3, or 4, while galac- tose has been shown to be oxidized at either carbon 1 or carbon 4 (8, 9). Arabinose has been found to be oxidized to arabonic acid (8). The published radiochemical data clearly show that the re- action is not valid as a model for establishing which carbon atom of substrates undergoes reversible oxidation during epimeriza- tion. Nonetheless, it is likely that the reaction is in some sense relevant to the epimerization mechanism. In this paper, we pre- sent the results of some rate measurements on UlMP-dependent reduction by glucose, which somewhat clarify the role of UMP as an activator. Our data also suggest that the reaction may be more fundamentally related to the epimerization mechanism than heretofore suspected. EXPERIMENTAL PROCEDURE Enzyme-We purified and assayed UDP-galactose-4-epimerase by the published procedure (1Oj from a regulatory mut,ant of Escherichia coli. ATCC 27797. We made one modification in the procedure. We carried out the hydroxylapatite chromatography step at pH 7.2 rather than at pH 6.5. The column equilibration, charging, and elution were exactly as described (lo), except that all buffers were at pH 7.2. We found that at pH 6.5. the enzyme would not be eluted from the column with the gradient described, but at pH 7 to 7.2, the elution results were equivalent to the pub- lished data (10). Chemicals-We obtained UMP from commercial soIIrces. It was chromatographically pure, and we used it without further purifi- cation. ol-D-Crlucose was ACS certified commercial dextrose. Its specific rotation was +108.2” in 10% aqueous solution at 20”. For certain 7099 by guest on June 17, 2018 http://www.jbc.org/ Downloaded from

Transcript of Uridine Diphosphate Galactose-4-Epimerase - jbc.org · Uridine Diphosphate Galactose-4-Epimerase...

THE Jonma~ OF Bmworca~ CHEMISTRY Vol. 250, No. 18, Issue of September 25, pp. 7099-7105, 1975

Printed in U.S.A.

Uridine Diphosphate Galactose-4-Epimerase

URIDINE MONOPHOSPHATE-DEPENDENT REDUCTION BY (Y- AND P-D-GLUCOSE*

(Received for publication, January 14, 1975)

UAN G. KANG, LYNN D. NOLAN, AND PERRY A. FREY$

From the Department of Chemistry, The Ohio State University, Columbus, Ohio .@‘2lO

SUMMARY

Rates of UMP-dependent reduction of the DPN+ associ- ated with Escherichia co2i UDP-galactose-4-epimerase at 27” and 0.2 M ionic strength in 0.1 M Tris .HCl buffer, pH 8.5, are reported. The reaction exhibits excellent pseudo-first order behavior when D-glucose is at anomeric equilibrium. The effects of [UMP] and [glucose] on the observed first order rate constants are consistent with the following equa- tion. The symbols I$ are empirical parameters.

41 42 “” = ” ’ LUMP] ’ (glucose] ~ -+

h [UMP][glucosel

The data indicate that the pathway involves random equi- librium binding of UMP and glucose followed by rate-limiting decomposition of the ternary complex to epimerase .DPNH. The binding parameters indicate that the principal activating effect of UMP is not simply to increase the affinity of the enzyme for glucose. UMP appears to increase the reactivity or availability of enzyme-bound DPN+. The kinetic isotope effect for the reaction of D-[l-2H]glucose (kH/kD) is 4.2, which confirms that C-l is oxidized and that hydride transfer is rate limiting.

Both of the purified anomers, CY- and /3-D-glucose, reduce the enzyme-bound DPN+. As indicated by the deviations from pseudo-first order kinetics because of concurrent mutarotation, the /3 anomer is the more reactive, reacting about 4 to 5 times faster than the o( anomer at concentrations well below saturation. It is suggested that the lack of stereo- specificity in this reaction may be attributed to the two anomers being productively bound with their opposite faces projecting toward C-4 of bound DPN+. Nonstereospecific oxidation of CY- and P-D-glucose may be a model for the mechanism of UDP-hexose epimerization, which also in- volves nonstereospecific hydride transfer.

The tightly bound DPNf associated with the UDP-galac- tose-4-epimerases (EC5.1.3.2) purified from Escherichia coli

* This research was supported by (;rant AM 13502 from the National Institute ol Arthritis, Metabolism, and Digestive IXs- eases.

$ To whom correspondence concerning this paper should be ad- dressed.

or Saccharomyces jragilis can be reduced to DPNH by a variety of aldohexoses and aldopentoses in a reaction which is absolutely dependent upon the presence of UMP (l-4). The reduced holo- enzyme is catalytically inactive on normal substrates, because in the epimerization pathway, the pyridine nucleotide must initially be in its oxidized form. The epimerase.DPNH complex produced in this reaction can be reoxidized and reactivated by nucleoside diphosphate 4-keto-6-deoxyhexoses, by myo-inosose-2, or by 2-ketoglucose (5-7).

It has been suggested that this reaction may be a model for the first step in the epimerization of substrates (5). Several re- cent findings concerning hydrogen transfer specificity exhibited by sugars effective as reducing agents in this reaction have cast doubt upon its relevance as a model. In particular, the reaction has been shown not to involve highly specific oxidation of the reactive sugars at carbon 4. Rather, in radiochemical experi- ments, glucose has been found to be oxidized at carbon 1, the reducing carbon, but not at carbon atoms 2, 3, or 4, while galac- tose has been shown to be oxidized at either carbon 1 or carbon 4 (8, 9). Arabinose has been found to be oxidized to arabonic acid (8).

The published radiochemical data clearly show that the re- action is not valid as a model for establishing which carbon atom of substrates undergoes reversible oxidation during epimeriza- tion. Nonetheless, it is likely that the reaction is in some sense relevant to the epimerization mechanism. In this paper, we pre- sent the results of some rate measurements on UlMP-dependent reduction by glucose, which somewhat clarify the role of UMP as an activator. Our data also suggest that the reaction may be more fundamentally related to the epimerization mechanism than heretofore suspected.

EXPERIMENTAL PROCEDURE

Enzyme-We purified and assayed UDP-galactose-4-epimerase by the published procedure (1Oj from a regulatory mut,ant of Escherichia coli. ATCC 27797. We made one modification in the procedure. We carried out the hydroxylapatite chromatography step at pH 7.2 rather than at pH 6.5. The column equilibration, charging, and elution were exactly as described (lo), except that all buffers were at pH 7.2. We found that at pH 6.5. the enzyme would not be eluted from the column with the gradient described, but at pH 7 to 7.2, the elution results were equivalent to the pub- lished data (10).

Chemicals-We obtained UMP from commercial soIIrces. It was chromatographically pure, and we used it without further purifi- cation.

ol-D-Crlucose was ACS certified commercial dextrose. Its specific rotation was +108.2” in 10% aqueous solution at 20”. For certain

7099

by guest on June 17, 2018http://w

ww

.jbc.org/D

ownloaded from

7100

experiments, we used recrystallized a-n-glucose (11). The specific rotation of our recrvstallized a-n-glucose was +106.4” in 10% aqueous solution at 20”. By comparing our rotation data with the published values of 110” and 112” under these conditions (11, 12), and assuming that any impurity was fl-D-glUCOSC, we estimated that our samples of ol-n-glucose could have contained 2.6 to 6% p-o-glucose. p-n-Glucose was purchased from Calbiochem and its specific rotation was +18.0” in 10% aqueous solution at 20”, in reasonable agreement with the published value of +18.8” (12).

n-[1-*HIGlucose was purchased from Merck Sharp & Dohme of Canada Ltd. For isotope effect measurements at anomeric equilib- rium, it was recrystallized from 90% ethanol/lOyc water, v/v. For isotope effect measurements on or-o-[l-2H]glucose, it was recrystal- lized as described (11) for obtaining a-D-glUCOSC of high anomeric purity.

Rate Measurements-We measured rates of UMP-dependent re- ductive inactivation of UDP-galactose-4.epimerase by D-glUCOSC at anomeric equilibrium at 27” in 0.10 M Tris.HCl buffer at pH 8.5 and 0.2 M ionic strength adjusted with KCl. The rates were inde- pendent of whether the reactions were started by addition of en- zyme, UMP, or glucose, however, most react,ions were started by addition of enzyme. Typically, the otherwise complete reaction mixture was prepared in a cuvette minus enzyme and thermally equilibrated at 27”. The cuvette was then positioned in the ther- mostated cell block of a Norelco-Unicam SP800 spectrophotometer equipped with a scale expansion accessory and external recorder. After recording a steady base-line for several minutes, the cell compartment was opened and the reaction was initiated by quickly adding 4000 to 8000 units of enzyme/ml of solution and mixing quickly with a Pasteur pipette. The increase in 4345 was monitored continuously with time. Data points were taken from the chart record for analysis. All rates were pseudo-first order in enzyme, and the observed rate constants, k ohs, were evaluated by computer fitting the data to the first order rate law. We measured the isotope effect for the reaction of n-glucose and n-(l-QHjglucose by the same procedure, but at 10” instead of 27”.

We measured rates of UMP-dependent reductive inactivation bye and by p-o-glucose in 20 mM UMP buffer at pH 7.0, 27”, and 0.2 M ionic strength adjusted with KCl. The otherwise complete reaction mixtures, minus a- or @-n-glucose, were prepared and equilibrated at 27”. Weighed samples of (Y- or o-o-glucose in small volumetric flasks were simultaneously thermally equilibrated. At zero time, 1.0 ml of reaction mixture was used to dissolve a sample of solid glucose, and the solution was quickly transferred to a cuvette positioned in the thermostated cell block of the spectro- photometer. This required no longer than 1 min. The reaction was then monitored with time as above. Many of these rates did not follow the first order rate law, although most of the faster rates with O-D-glUCOSe appeared to, because mutarotation was proceed- ing concurrently, and the two anomers did not react at the same rate. Data from the progress curves were computer fitted to an arbitrary equation in such a way that we could estimate the pseudo-first order rate constants at zero time. This is described under “Results.”

We measured mutarotation rates for 01. and ~-n-glucose in 20 mM UMP or 20 mM potassium phosphate buffer at pH 7.0, 27”, and 0.2 M ionic strength maintained with KCl. The reactions were followed polarimetrically, and the data were computer fitted to the first order rate law. The observed mutarotation rate constants were the same within error for the two anomers.

RESULTS

Rate Law for UMP-dependent Reduction by Glucose-In pre- liminary experiments,’ we measured pseudo-first order rate

constants at UMP concentrations between 0.4 mM and 10 mM

at 50 mM glucose. The double reciprocal plot l/&n, versus l/[UMP] was a straight line with a positive intercept, character-

istic of ordinary hyperbolic saturation kinetics. The apparent K, for UMP was 1.24 mM, which was in good agreement with

the inhibition constant, 1.2 mM, for UMP acting as a competitive reversible inhibitor of this eneyme.2 In another experiment, we

r F. Steven Davis and P. A. Frey, unpublished data.

I I I I I 0062mM UMP

0 IllmM UMP

0 645mM UMP

1::; 505mM UMP

20 SO IOC

FIG. 1. Kinetics for UMP-dependent reductive inactivation of UDP-galactose-4-epimerase by n-glucose at anomeric equilibrium. The pseudo-first order rate constants were measured as described under “Experimental Procedure.”

24-

2 o-

0 I I I I 0 4 a 12 16

[UMP]-’ (mM-‘I

FIG. 2. Slope and intercept replots of Fig. 1 data. The least squares slopes and intercepts from Fig. 1 are plotted Versus [IIMP]-1. O--O, slopes; O--O, intercepts.

varied [glucose] between 50 and 500 mM at 0.1 mM UMP, and again obtained a straight line with a positive intercept.

A more extensive collection of data over a similar range of [UMP] but a larger range of [glucose] is shown in Fig. 1. The points are mean k&s values from three to six determinations, while the lines are the best least squares straight lines through the data points. Replots of the slopes and intercepts in Fig. 1 versus [UMP]-l are given in Fig. 2, which shows that the data fit Equation 1.

1 Gbs

The data are consistent with the rate law d[EH]/dt = kobs ([Eo] - [EH]), where [Eo] is the total enzyme concentration, [EH] is the concentration of reduced holoenzyme at time t, and kobs is given by Equation 1. The experimental parameters $J, evalu-

ated by fitting the data to Equation 1, are given in Table I. * P. A. Frey, unpublished data. Saturation with glucose was not closely approached in the

by guest on June 17, 2018http://w

ww

.jbc.org/D

ownloaded from

7101

TABLE I Parameters for UMP-dependent reduction of epimerase.DPN+

to epimerase.nPNH by u-glucose. The experimental parameters were obtained by fitting the data in Fig. 1 to Equation 1. The analytical parameters refer to Scheme 2 and were calculated from the experimental parameters 4 and Equations 1 and 3.

Experimental Analytical

do (0.159 f 0.020) min ks (6.3 •I= 0.8) min-1 ~$1 (0.461 f 0.145) X lo-’ K, (1.86 f 0.72) X 10-M

M.min 42 (0.0772 f 0.0207) M.min K’, (0.290 f 0.098) X 10e3 M 412 (0.144 f 0.032) X lo+ K, (3.12 f 1.85) M

M2. min K', (0.49 f 0.14) M

kl EtA= EA

k2

k3 EA+Be EAB kg_ EH

k4

SCHEME 1

experiments shown in Fig. 1, which is probably the explanation for the relatively large experimental errors associated with the Table I parameters. Close approach to glucose saturation proved to be impractical because of the large concentrations required. We found definite downward curvature in some double reciprocal plots as the glucose concentrations reached 2 and 3 M. This might have been due to higher order glucose binding or to a medium effect on enzyme structure. Inasmuch as distinguishing between these possible effects would be complex, we chose to confine our studies to glucose concentrations 1.6 1\1 and below, where straight line plots were obtained.

Molecular Pathway for UMP-dependent Reduction by Glucose- Of the many pathways consistent with saturation kinetics which might be followed in this reaction, most are excluded by Figs. 1 and 2. The rapid equilibrium-ordered binding pathways con- sidered by Segal et al. (13) for enzyme activation are inconsistent with Equation 1. Analogous ordered pathways, in which it is assumed that the first ligand-binding step is locally equilibrated while the second is in a steady state, are also inconsistent. The rate laws require that 41 or & or both in Equation 1 be zero, whereas Figs. 1 and 2 show that they are not. It is also pertinent to consider whether the pathway may be analogous to the first two steps of a Theorell-Chance pathway not involving a kinet- ically significant ternary complex. Again, the rate laws for such pathways would require that 40 be zero in Equation 1, which it is not.

Conventional pathways consistent with Equation 1 are Schemes 1 and 2. Scheme 1 is a compulsory binding order path- way in which E symbolizes the native epimerase .DPN+ com- plex, A and B are either UMP or glucose, and EH is the enzyme. DPNH complex. The latter contains tightly bound UMP which stabilizes it against auto-oxidation (14). When the steady state approximation is made, it can be shown that Equation 2 gives

(k4+k5) k2(k4tk$

k3k5 b + klk3kgsb (2)

kobs for Scheme 1, where a and b refer to concentrations of either UMP or glucose, depending upon their binding sequence. Al- though Equation 2 is identical in form with Equation 1, it is

very unlikely that Scheme 1 can be correct, because the experi- mental data, when evaluated against the background of pub- lished data on rates of ligand binding by enzymes, do not sustain it. The rate constants kl and k2 in Equation 2 can be calculated from the experimental parameters 4. For example, if ,4 is UMP and B is glucose, kl is 41-1 and kz is c$&L-~&-~. M7hen kl and kz are calculated in this way, their values are as follows. If A is UMP, kl = 360 M-I s-l and kp = 0.67 s-l; if A is glucose, kl = 0.22 M-I s-l and k2 = 0.67 s-r. These would be the association and dissociation rate constants for the binding interactions of the enzyme with UMP or glucose. Such constants have been measured for interactions of substrates or inhibitors with a wide variety of enzymes (15) ; they are found to be lo5 to lo8 M-l s-1 for association rate constants, and lo2 to lo5 s-i for dissoci- ation rate constants in cases where there is no reason to believe that binding may be complicated by such factors as possible covalency. Our calculated values are all smaller than the smallest measured values by factors of 2 X IO2 to 5 X 105. Inasmuch as the calculated values are unrealistically small, the ordered steady state pathway is unlikely to be correct.

It is possible that the binding of UMP is coupled with a con- formational change as suggested in a later section. If such a con- formation change were slow, it might limit the association rate and give an anomalously small apparent binding rate constant. Although this possibility cannot be definitively excluded at this time, we believe it to be unlikely to be correct for the following reason. The interactions of UMP and UDP-sugar substrates with this enzyme appear to be similar. UMP is a competitive inhibitor, both UMP and UDP-hexoses activate the reduction of enzyme- bound DPN+ by NaBH*, and both UMP and UDP-hexose sub- strates are very tightly bound by the epimerase .DPNH complex, although not by the epimerase.DPN+ complex. The dissociation rate constant for substrates diffusing away from the active site is not less than 500 s-1 at pH 8.5 and 27” (lo), while the calculated dissociation rate constant for UMP under these conditions, as- suming the ordered steady state pathway with UlMP binding first, is only 0.67 0, about 1000 times smaller. We think it unlikely that UMP would dissociate so much more slowly than UDP-glucose.

Scheme 2 is the rapid equilibrium random binding pathway, in which E and EH are the native and reduced enzyme forms, re- spectively, and the KS are dissociation constants.

E

E*UMP

K5 E*Glc l UMP - EH

SCHEME 2

For this pathway, kobs is given by Equation 3. It is the simplest scheme consistent with the data. All of the analytical parameters can be calculated from the experimental parameters 4 and are

I Kb $+&q+-+

Ku K; - =

k obs k&cl k&lc][lJMP] (3)

by guest on June 17, 2018http://w

ww

.jbc.org/D

ownloaded from

7102

given in Table I. The binding parameters for glucose refer to the anomeric mixture of (Y- and /?-o-glucose at equilibrium. In Scheme 2, K, is the dissociation constant for the enzyme .UMP complex, which should be the same as the inhibition constant for UMP acting as a competitive reversible inhibitor. The value given in Table I is within error of 1.2 mM, the independently measured inhibition constant2 Scheme 2 is the simplest reaction pathway that is consistent with Figs. 1 and 2, however, there are other more complex pathways involving nonproductive binding interactions that are equally consistent.

Isotope Effect for UMP-dependent Reduction by [l -2H]Glucose- Radiochemical evidence indicates that glucose is oxidized by direct hydrogen transfer from C-l to enzyme-bound DPNf (9), which is in conflict with the report of a primary kinetic isotope effect in the reaction of [3JH]glucose (5). The rate constants in Table II establish that [l-2H]glucose reacts with a primary kinetic isotope effect of about 4.2, which confirms that C-l is oxidized.

The rates in Table II were measured at 10” rather than at 27” for the convenience of measuring slower rates, [l-2H]glucose being available in limited quantities and required at high con- centrations. We hoped in this way to work at near saturating [UMP] and up to half-saturating [glucose]. This probably was not successful because the analysis of the Table II data showed the apparent K, values for glucose and [1-*HIglucose to be 6.2 and 6.7 M, respectively, well above the highest [glucose] in Table II. Nevertheless, the apparent isotope effect of 4.2 f 0.3 was fairly constant over the range of [glucose] in Table II, which would be consistent with the proposed mechanism and rate-limit- ing hydrogen transfer from C-l of glucose to the pyridine nucleo- tide.

Anomeric Specijicify for Glucose-The rate data show that at

large [UMP] and [glucose] and pH 8.5, the half-time is substan- tially shorter than the known half-time for mutarotation of glu- cose at pH 7, which is 20 min (12). Of course, mutarotation is much faster under Fig. 1 conditions, because hydroxide ions and buffers are highly effective mutarotation catalysts, so that anomeric specificity would be difficult to detect. However, pre- liminary rate measurements at pH 7 showed that the half-time for UMP-dependent reduction by glucose can be less than 1 min; therefore, it should be possible to establish anomeric specificity at this pH.

TABLE II

Kinetic isotope effect for UMP-dependent reduction of epimer- ase.DPN+ by [1-2Hlglucose. The rates were measured as described under “Experimental Procedure” at 10.0 =t 0.5” and pH 8.5 in 0.1 M Tris.HCl buffer, pH 8.5, and 30 mM UMP with the ionic strength maintained at 0.2 M with KCl. The glucose solutions were per- mitted to reach mutarotation equilibrium before beginning the rate measurements. Three to six rates were measured with un- labeled glucose to establish the precision of the data. Then one rate was measured with [l-2H]glucose at each of the concentra- tions.

[Glucose] Y ka min-’ ko min-1 RR/AD

0.186 0.330 It 0.009 0.081 4.07 0.243 0.446 f 0.011 0.097 4.59 0.378 0.656 f 0.024 0.148 4.58 0.487 0.834 f 0.038 0.191 4.37 0.752 1.22 f 0.08 0.277 4.40 1.12 1.73 f 0.24 0.481 3.59 1.52 2.32 zk 0.12 0.596 3.89

In establishing anomeric specificity, it is necessary to know what would be the mutarotation rate at pH 7 in the presence of UMP and the enzyme. Table III shows that, at pH 7 and in the presence of 0.02 M UMP or Pi as buffer and epimerase concen- trations near those employed in rate measurements, the mutaro- tation half-time for n-glucose is not shorter than 10 min. This assures that specificity will be detectable provided the rates are fast enough, that is that they are measured at sufficiently large UlMP and glucose concentrations. Table III also indicates that mutarotation may be, to a small degree, catalyzed by epimerase in a IMP-dependent process, because the mutarotation rate with UMP buffer and epimerase is slightly faster than with UNlP buffer and serum albumin, although there is little or no difference when the buffer is Pi. The data are not considered to be reliable enough to establish this point with certainty.

First order plots of rate data for cY-n-glucose and P-n-glucose show by visual inspection that P-n-glucose reacts preferentially. This is exemplified in Fig. 3, where it is seen that these plots curve downward with time when oc-n-glucose is the reducing agent, and slightly upward when &n-glucose reacts. Such plots are excellent straight lines when glucose is at anomeric equilib- rium. Our interpretation is that the curvature results from muta-

TABLE III

Mutarotation rates for n-glucose. All rates were measured polarimetrically at 27” starting with each anomer. The tabulated observed rate constants are mean values for the two anomers. The published value for mutarotation in Hz0 is 0.0145 mini at 20” (12) which can be corrected to 0.028 min-’ at 27” assuming an activation energy of 17 kcal/mol (12).

Mutarotation conditions kobs

Hz0 20 m&l K Pi, pH 7.0, = 0.2 M /J

+ 0.5 mg/ml of serum albumin + 0.5 mg/ml of epimerase

20 mM UhtP, pH 7.0, = 0.2 M p

+ 0.5 mg/ml of serum albumin + 0.5 mg/ml of epimerase

mid 0.0304 f 0.008 0.0768 f 0.0005 0.1020 f 0.0003 0.1077 f 0.0001 0.0561 f 0.0008 0.0560 f 0.0001 0.0644 3~ 0.0001

FIG. 3. First order plots of kinetic data for reduction of UDP- galactose 4.epimerase by a-n-glucose and fl-n-glucose. The rates were measured as described under “Experimental Procedure.” These are semilogarithmic plots of the changes in A,45 during re- duction by the two anomers of n-glucose. The points are experi- mental data taken from the chart records of the reaction progress. The solid lines are fitted curves calculated as described in the text, and the dashed lines are the initial slopes resulting from the fitting procedure. A, reduction by 0.05 M ol-n-glucose. B, reduction by 0.01 M B-D-ghCOSE!.

by guest on June 17, 2018http://w

ww

.jbc.org/D

ownloaded from

7103

TABLE IV

Rates of UMP-dependent reduction of epimerase.DPN+ by (Y- and P-n-glucose. The rates were measured at 27.0 z!= 0.5” in 0.02 M

UMP buffer at pH 7.00 f 0.02 and at 0.2 M ionic strength main- tained with KCI. The procedure was that described under “Ex- perimental Procedure,” and the initial kobr values were evaluated as described in the text. The rates were measured in duplicate or triplicate. a-n-Glucose and ol-o-[l-2H]glucose were recrystallized under identical conditions before being used for rate measure- ments.

Reducing sugar Concentration Initial k,bs

ol-D-glucose a-D-glucose a-D-[I-‘%]glucose ol-D-glucose ol-D-glucose ol-D-glucose a-D-[l-2H]glucose P-D-glucose P-D-glucose p-D-glucose p-D-glucose

M min-’

0.05 0.103 f 0.018 0.10 0.216 f 0.044 0.10 0.117 f 0.004 0.20 0.377 f 0.010 0.30 0.426 f 0.017 0.50 0.585 f 0.192 0.50 0.191 f 0.032 0.01 0.076 f 0.003 0.02 0.158 + 0.001 0.05 0.386 zk 0.008 0.10 ,0.937 f 0.015

rotation and that the direction of curvature is determined by the fact that &n-glucose is the preferred anomer. In Fig. 3A, the P-n-glucose concentration gradually increases because mutaro- tation is occurring. This causes a gradual increase in the ap- parent rate constant which is reflected in the downward curva- ture. In Fig. SB, mutarotation gradually decreases the /3-n-glu- case concentration, which results in a progressive decrease in the apparent rate constant and slope.

To determine whether ol-n-glucose also reacts, but at a slower rate than the /3 anomer, we performed a number of additional rate measurements with both cr- and fl-n-glucose at various con- centrations and fixed [UMP]. The data obtained were computer fitted to the equation y = a + bx + cx2, where y = ln(A, - A,), x = time (mm), and the constants a, b, and c were evaluated to give a best fit. The second coefficient b was taken to be k&s at zero time, that is, before mutarotation had proceeded to a sig- nificant extent.

The data are given in Table IV, and they show that both o(- and P-n-glucose reduce epimerase .DPN+, the /3 anomer being the more reactive. The conclusion that oc-n-glucose reacts depends implicitly upon the assumption that our samples are not con- taminated by sufficient fl anomer to give a small zero time rate. We believe this to be the case, because the optical rot.ations of our glucose samples show that our ol-u-glucose cannot contain more than 2.6 to 6% of the fi anomer. In order to account for the zero time rate seen with a-n-glucose on the basis of con- tamination by fi-n-glucose, our samples of a-o-glucose would have to contain at least 20 to 30% &n-glucose. Even then, we would have to assume that the presence of 70 to 80% a-n-glucose would not inhibit the rate at which the fi anomer reacts. The conclusion that both anomers react appears to be inescapable.

Inasmuch as the /? anomer is greatly preferred, the small fraction of the reaction involving a-n-glucose in an anomer mix- ture at equilibrium, 62% @- and 38% cu-n-glucose (la), might proceed by oxidation at some carbon atom other than C-l. This would have been difficult to detect in the radiochemical experiments of Ketley and Schellenberg (9), and would not have been detected in Table II. To resolve this question, we measured

the rate constants for reactions of cr-n-[l-2H]glucose and a-~-

glucose under identical conditions. As shown in Table IV, cY-n-[l-2H]glucose reacts significantly

more slowly than cu-n-glucose. We conclude that both (Y- and /&n-glucose are oxidized at C-l .3

DISCUSSION

Role of UMP in Reduction by Glucose-The most striking as- pect of the action of UMP is the large magnitude of its rate-ac- celerating effect. Any explanation of UMP activation must account for the fact that reduction by glucose has not yet been detected in the absence of UMP even at very large glucose con- centrations. There are several ways in which UMP might act to promote this reaction. Its association with the enzyme may pro- mote the over-all binding of glucose. Its association may promote the productive binding of glucose, which is preferentially non- productively bound by the enzyme itself. UMP in the ternary complex may interact directly with bound glucose to increase its reactivity toward epimerase .DPN+. Finally, UMP may, upon binding, increase the chemical reactivity of epimerase .DPN toward bound glucose. This would imply that UMP induces a structural change in epimerase . DPN+.

The present results are not consistent with the interpretation that the principal effect of UhilP is to promote the binding of glucose. The analytical parameters for Scheme 2 in Table I indi- cate that, while UMP does increase the affinity of the enzyme for glucose, the effect is modest, certainly not large enough to ac- count for absolute UMP dependence. If this were the principal activating function of UMP, the kinetic pathway would be or- dered sequential, at least experimentally, whereas the data are not consistent with this.

We cannot exclude the possibility that IMP may promote the productive binding of glucose at the expense of nonproduc- tive binding orientations. In order for this to account fully for the activating effect of UMP, it would be necessary to assume that glucose, and each of the other reactive sugars as well, is bound essentially exclusively nonproductively by epimerase . DPN. Given the fact that the reaction is not highly specific for sugars, or even for which carbon atom is the hydride donor in galactose, it does not appear to be very likely that the principal activating effect of UMP is to promote productive binding. More- over, it might be expected that, if this were the case, UMP would guide the binding of glucose and arabinose in such a way as to lead to oxidation at C-4, the reactive carbon atom in substrates, rather than at C-1. Nevertheless, this possibility cannot be ex- cluded at this time, and as discussed below, it is probable that several binding orientations are accessible to sugars.

Our data do not permit us to decide whether UMP may in- crease the reactivity of bound glucose toward epimerase .DPN+ or of epimerase.DPN+ toward bound glucose. However, inde- pendent findings suggest that a major effect of UMP is to in- crease the reactivity of epimerase .DPN+ toward reducing agents. We have found that the reduction of epimerase.DPN+

3 A reviewer points out that the smaller isotope effect in Table IV compared with that in Table 11 may indicate that the oxidation of o-D-glUCOSe occurs only in part at C-l and in part at some other position. This may be correct, although it is still large enough that it would not alter our conclusion that oc-n-glucose can be oxidized at C-l. We note further. however. that the isotope effect in Table IV refers tool-o-glucose: whereas that in Table II probably largely reflects the react,ion of fl-D-glUCOSe, given that the p anomer both reacts preferentially and dominates the equilibrium. The isotope effects for the reactions of the two anomers may well differ if the transition state geometries differ.

by guest on June 17, 2018http://w

ww

.jbc.org/D

ownloaded from

7104

to epimerase.DPNH by NaBH&N, a nonspecific hydride-re- ducing agent, is very similar to reduction by glucose, in that it exhibits an absolute requirement for the presence of UMP. The reaction cannot be detected in the absence of UMP, whereas, upon addition of UMP, it proceeds more than 100 times faster than the reaction of NaBH&N with free DPN+ itself (16). Again, the rate acceleration by UMP is large, from no detectable rate in its absence to rates comparable with those reported here upon adding UMP. It appears that UMP binding may induce a struc- tural change in enzyme.DPN+, which has the effect of greatly increasing its reactivity toward hydride donors in general, in- cluding glucose.

Specijkity for Oxidation of Glucose at Carbon 1-A stereochemi- cal binding model has been proposed to account for the fact that glucose is oxidized at C-l, while galactose can be oxidized either at C-l or C-4 (9). This model, as it relates to the oxidation of glucose, is not compatible with our results. The model is based implicitly upon the premise that only a-n-glucose reacts, whereas we find that P-n-glucose is actually the preferred anomer. We believe that the explanation for glucose and other sugars being oxidized at C-l is not based upon a high degree of binding spec- ificity. It appears to be more likely that several binding orienta- tions are possible, and that oxidation at C-l is preferred largely because C- 1 is the reducing carbon atom. Carbon 1, being bonded to two oxygen atoms, is the best hydride donor and reacts pref- erentially. Binding specificity may be invoked to account for galactose reacting in part at C-4.

Possible Model for UMP-dependent Reduction by Glucose and Epimerization of Substrates-The fact that both (Y- and B-D-

glucose are oxidized at C-l suggests that this reaction may be fundamentally related to the mechanism by which the enzyme catalyzes epimerization of substrates, because the epimerization pathway involves reversible nonstereospecific oxidation of nucleo- tide sugars at glycosyl-C-4 (17-20), producing UDP-4-keto- sugars, which have been shown to occur as free intermediates under certain special conditions (19). How can this enzyme act without stereospecificity at C-4 of nucleotide sugars and C-l of glucose? A characteristic property of pyridine nucleotide-depend- ent alcohol dehydrogenases is that they catalyze stereospecific hydrogen transfers, yet in these reactions the analogous processes are nonstereospecific.

On the presumption that the DPNf associated with UDP- galactose-4-epimerase is fairly deeply buried and essentially im- mobile within its binding site, the nonstereospecific oxidation of LY- and /3-n-glucose can be understood if, in their enzyme-bound states, the two anomers are placed in such orientations that their C-l hydrogen atoms can occupy nearly the same space ad- jacent to nicotinamide C-4 of DPN+. The presumption of im- mobility for enzyme-bound DPN+ is supported by two well established properties of the holoenzyme. First, the DPN+ is so tightly bound that it survives repeated chromatography during purification (10). Second, the reduction of enzyme-bound DPN+ with NaBaH is essentially stereospecific (17, 19). Up to 90% [4-B-aH]DPNH and about 1% [4-A-3H]DPNH has been re- ported in such experiments, which is consistent with about 99% stereoselectivity, because the tritiated DPNH was isolated with carrier after heating at 100” for 5 mm to denature the protein, conditions which would cause DPNH to undergo mutarotation to anomeric equilibrium, 90% /3- and 10% (r-DPNH (21). We would not expect such a high degree of stereoselectivity from a highly mobile nicotinamide ring.

The simplest mechanism by which the anomeric hydrogen atoms of (Y- and &n-glucose could occupy nearly the same space

would be for the sugar-binding interactions to be sufficiently flexible to permit glucose to bind with either face of the molecule projecting into the binding site toward DPN+. Space-filling models show that a 180” rotation about the C-l to C-4 axis of one anomer relative to the other would place the anomeric C-l hy- drogen atoms and hydroxyl groups in similar spatial orientations. Because of the fact that the remaining hydroxyl and hydroxy- methyl groups are equatorial, the over-all molecular topographies of the anomers in the two rotational states are not grossly dif- ferent; still, they differ by enough that the binding interactions would not be very specific, in accord with the fact that glucose is very weakly bound.

We have also examined molecular models of these anomers in conformations other than Cl. The models showed the over-all molecular topographies of the unstable boat and 1C conforma- tions to be radically different from that of the Cl conformation, far more different than the topographical differences resulting from simple rotation of the Cl conformation to opposite faces. The models also showed that if one anomer reacts from a boat conformation and the other from a stable Cl chair, the stereo- chemistry is such that the a! anomer should react from the stable chair conformation. It would, therefore, be the more reactive anomer, whereas the fi anomer is actually the more reactive.

The oxidations of LY- and &n-glucose may be a model for epi- merization of substrates. Nonstereospecificity in epimerization might then result from the epimeric substrates binding with opposite faces of their glycosyl moieties projecting into the active site toward DPN+, as shown in Fig. 4 with space-filling models of the epimeric substrates UDP-n-xylose and UDP-L-arabinose (22). The models in Fig. 4 are placed in extended conformations with the UDP portions in approximately the conformation sug- gested for UDP-glucose (23) but with opposite faces of the D-

xylosyl and L-arabinosyl moieties projecting downward. The operation of projecting the opposite faces downward involves a single rotation of approximately 180” about the bond connecting the glycosyl oxygen atom and the P-phosphorous atom in the pyrophosphoryl linkage. This rotation is fairly unhindered, and it places the C-4 hydrogen atoms as well as the C-3 and C-4 hydroxyl groups in similar spatial dispositions. The C-3 hydroxyl group is known to be required from the fact that UDP-3 deoxy- glucose is not a substrate (24). This model suggests that the binding interactions between the C-3 and C-4 hydroxyl groups

FIG. 4. Space-filling models of UDP-n-xylose (right) and UDP- L-arabinose (left). Atoms marked with the numeral 4 are the hy- drogen atoms bonded to xylosyl- and arabinosyl-C-4.

by guest on June 17, 2018http://w

ww

.jbc.org/D

ownloaded from

7105

and the enzyme may be reversed in the two conformations, and 8. SEYAMA, Y., AND KALCKAR. H. M. (1972) Biochemistru 11.

it would permit the-C-4 hydrogen atom from either isomer to be transferred to DPN+. The resultant UDP-4-ketosugar would then undergo the required conformation change culminating in the projection of the opposite face of the ketosugar toward DPNH. During this process, the nucleotide moiety remains firmly anchored to the epimerase*DPNH complex, which is known to bind uridine nucleotides very tightly (14, 17, 19). An alternative hypothesis postulates that one epimer reacts from a chair and the other from a boat conformation (25).

While this hypothesis is attractive in certain ways, it is by no means proved. Nor do we intend it to imply anything about the exact conformations assumed by the substrates in their bound states or about the exact details of the conformation changes which occur either in the enzyme or the UDP-4-ketosugar inter- mediate during epimerization. Such a detailed picture will de- pend upon the availability of additional information on the con- formations of substrates, which have been postulated to exist in extended (23) or folded (26) solution conformations, and on the structure of the active site.

9.

10.

11.

12.

36-40. ” ,

KETLEY, J. N., AND SCHELLENBERG, K. A. (1973) Biochemistry 12, 315-320

REFERENCES

1. BHADURI, A., CHRISTENSEN, A., AND KALCK~R, H. M. (1965) Biochem. Biophys. Res. Commun. 21, 631-637

2. BGRTLAND, A. II., BUGGE, B., AND KALCKAR, H. M. (1966) Arch. Biochem. Biophys. 116, 280-283

3. BERTLAND, A. II., AND KALCKAR, H. M. (1968) Proc. N&Z. Acad. Sci. U. S. A. 61, 629-635

4. KALCKAR, H. M., BERTLAND, A. U., AND BUGGE, B. (1970) Proc. Natl. Acad. Sci. U. 8. A. 66, 1113-1119

5. DAVIS, L., AND GLASER, L. (1971) Biochem. Biophys. Res. Commun. 43, 1429-1435

6. KETLEY, J. N., AND SCHELLENBERG, K. A. (1972) Biochim. Bzophys. Actu 284, 549-551

7. KALC~AR, H. M., BERTLAND, A. II., JOHANSEN, J. T., AND OTTF,SEN. M. (1969) in The Role of Nucleotides for the Func- tion and ‘Confkmakon of Enzymes (KALCKAR, H. M., KLE- NOW, H., MUNCH-PETERSEN, A., OTTESEN, M., AND THAY- SEN, J. H., eds) pp. 247-275, Munksgaard, Copenhagen

13.

14.

15.

16.

17.

18.

19. 20.

21.

22.

23.

24.

25.

26.

WILSON, D. B., ,\ND HOGNESS, D. S. (1964) J. Biol. Chem. 239, 2469-2481

ISBELL, H. S., AND WADE, C. W. R. (1967) J. Res. Nat. Bureau Stand. 71A, 137-148

PIGMAN, W., AND ANET, E. F. L. J. (1972) in The Carbohydrates, Chemistry/Biochemistry (HORTON, D., AND PIGM.IN, W., eds) 2nd Ed., Vol. IA, pp 166-194, Academic Press, New York

&GAL, H. L., KACHM~R, J. F., AND BOYER, P. D. (1951) En- zymologia 16, 187-198

BERTLAND, A. U., II, SEYAMA, Y., AND KALCKAR, H. M. (1971) Biochemistry 10, 1545-1551

HAMMF,S, G. G., AND SCHIMMEL, P. R. (1970) in The Enzymes (BOYER, P. D., ed) 3rd Ed, Vol. 2, pp. 109-110, Academic Press, New York

DAVIS, J. E., NOLAN, L. D., AND FREY, P. A. (1974) Biochim. Biophys. Acta 334, 442-447

NELSESTUEN, G. L., AND KIRKWOOD, S. (1971) J. Biol. Chem. 246, 7533-7543

MAITRA, U. S., AND ANKEL, H. (1971) Proc. Natl. Acad. Sci. 77. S. A. 68, 2660-2663

WEE, T. G., AND FREY, P. A. (1973) J. Biol. Chem. 248, 33-40 ADAIR, W. L., JR., GABRIEL, O., ULLREY, D., AND KALCKAR,

H. M. (1973) J. Biol. Chem. 248, 4635-4639 OPPENHEIMER, N. J., ARNOLD, L. J., AND KAPLAN, N. 0.

(1971) Proc. Natl. Acad. Sci. U. S. A. 68, 3200-3205 ANKEL, H., AND MAITRA, U. S. (1968) Biochem. Biophys. Res.

Commun. 32, 526-532 SARMA, R. H., LEE, C-H., HRUSKA, F. E., AND WOOD, D. J.

(1973) FEBS Lett. 36, 157-162 KOCHETOV, N. K., BUDOWSKI, E. I.. DRUZHININA, T. N..

GABRIELYAN, N. D., KOMLE~, I. V.; Kusov, Yu. Yu., AND SHIBAEV. V. N. (1969) Carboh.udr. Res. 10. 152-156

GLASER, L:, AND G,RD( L. (1976) Biochim. kiophys. Acta 196, 613-615

BUDOWSKY, E. I., DRUSHININA, T. N., ELISEEVA, G. I., GABRIEL~AN, N.. D., KOCHETK~V, N. K., NOVIKOV~, M. A.; SHIBAEV. V. N.. AND ZHDANOV. G. L. (1966) Biochim. Bio- phys. Acta 122, 213-224 ’ .

by guest on June 17, 2018http://w

ww

.jbc.org/D

ownloaded from

U G Kang, L D Nolan and P A Freyreduction by alpha- and beta-D-glucose.

Uridine diphosphate galactose-4-epimerase. Uridine monophosphate-dependent

1975, 250:7099-7105.J. Biol. Chem. 

  http://www.jbc.org/content/250/18/7099Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/250/18/7099.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on June 17, 2018http://w

ww

.jbc.org/D

ownloaded from