The use of fluoro- and deoxy-substrate analogs to examine binding specificity and catalysis in the enzymes of the sorbitol pathway

Article abstract of DOI:10.1016/S0008-6215(98)00266-3

The carbohydrate specificity of the two enzymes that catalyze the metabolic interconversions in the sorbitol pathway, aldose reductase and sorbitol dehydrogenase, has been examined through the use of fluoro- and deoxy-substrate analogs. Hydrogen bonding has been shown to be the primary mode of interaction by which these enzymes specifically recognize and bind their respective polyol substrates. Aldose reductase has broad substrate specificity, and all of the fluoro- and deoxysugars that were examined are substrates for this enzyme. Unexpectedly, both 3-fluoro- and 4-fluoro-d-glucose were found to be better substrates, with significantly lower K(m) and higher k(cat)/K(m) values than those of D-glucose. A more discriminating pattern of substrate specificity is observed for sorbitol dehydrogenase. Neither the 2-fluoro nor the 2-deoxy analogs of D-glucitol were found to be substrates or inhibitors, suggesting that the 2-hydroxyl group of sorbitol is a hydrogen bond donor. The 4-fluoro and 4-deoxy analogs are poorer substrates than sorbitol, also implying a binding role for this hydroxyl group. In contrast, both 6-fluoro- and 6-deoxy-D-glucitol are very good substrates for sorbitol dehydrogenase, indicating that the primary hydroxyl group at this position is not involved in substrate recognition by this enzyme. Copyright (C) 1998 Elsevier Science Ltd.

Full text of DOI:10.1016/S0008-6215(98)00266-3

Carbohydrate Research 313 (1998) 247–253
The use of fluoro- and deoxy-substrate analogs to
examine binding specificity and catalysis in the enzymes
of the sorbitol pathway
Mary Ellen Scott 1, Ronald E. Viola *
Department of Chemistry, The Uni6ersity of Akron, Akron, OH 44325-3601, USA
Received 15 April 1998; revised 15 August 1998; accepted 15 September 1998
Abstract
The carbohydrate specificity of the two enzymes that catalyze the metabolic interconversions in the sorbitol
pathway, aldose reductase and sorbitol dehydrogenase, has been examined through the use of fluoro- and
deoxy-substrate analogs. Hydrogen bonding has been shown to be the primary mode of interaction by which these
enzymes specifically recognize and bind their respective polyol substrates. Aldose reductase has broad substrate
specificity, and all of the fluoro- and deoxysugars that were examined are substrates for this enzyme. Unexpectedly,
both 3-fluoro- and 4-fluoro-
cat/Km values than those of
dehydrogenase. Neither the 2-fluoro nor the 2-deoxy analogs of
D
-glucose were found to be better substrates, with significantly lower Km and higher
-glucose. A more discriminating pattern of substrate specificity is observed for sorbitol
-glucitol were found to be substrates or inhibitors,
k
D
D
suggesting that the 2-hydroxyl group of sorbitol is a hydrogen bond donor. The 4-fluoro and 4-deoxy analogs are
poorer substrates than sorbitol, also implying a binding role for this hydroxyl group. In contrast, both 6-fluoro- and
6-deoxy-D-glucitol are very good substrates for sorbitol dehydrogenase, indicating that the primary hydroxyl group
at this position is not involved in substrate recognition by this enzyme. © 1998 Elsevier Science Ltd. All rights
reserved.
Keywords: Aldose reductase; Sorbitol dehydrogenase; Sorbitol pathway; Substrate specificity; Fluoro sugars; Deoxy sugars
The physiological role of the sorbitol path-
way in most tissues is not completely under-
stood. There is evidence, however, that this
pathway functions as a bypass between glycol-
ysis and the pentose phosphate shunt in the
metabolism of glucose [1]. Several hypotheses
have been proposed which link the increased
activity of the sorbitol pathway to tissue in-
jury under conditions of hyperglycemia [2,3],
where this pathway can account for more than
30% of glucose utilization [4]. Complications
or tissue injury in diabetes can be caused by
osmotic stress, by imbalance in myo-inositol
metabolism, by modification of protein func-
tion through glycation, and by alterations in
1. Introduction
The sorbitol pathway consists of two en-
zymes, aldose reductase and sorbitol
dehydrogenase:
* Corresponding author. Tel.: +1-330-972-6065; fax: +1-
330-972-6687; e-mail: rviola@uakron.edu
1 Present address: Department of Nutrition, Case Western
Reserve University, Cleveland, OH, USA.
0008-6215/98/$ - see front matter © 1998 Elsevier Science Ltd. All rights reserved.
PII: S0008-6215(98)00266-3

248
M.E. Scott, R.E. Viola / Carbohydrate Research 313 (1998) 247–253
drogenase and an alignment of the amino acid
sequences [20]. From this model, the ligands
to zinc have been identified as a histidine,
cysteine, and glutamate, and these assign-
ments have been supported by X-ray absorp-
tion fine structure (EXAFS) studies [21],
affinity labeling [22], metal chelation studies
[23], and by site-directed mutagenesis [24].
There has been progress on a determination
of the mechanism of catalysis of the enzymes
in the sorbitol pathway, however, several as-
pects of these mechanisms remain unresolved.
This paper describes the use of alternative
substrates, specifically fluorosugars and de-
oxy-sugars, to probe the active site binding
and catalytic requirements of these sorbitol
pathway enzymes.
redox state [5]. Under these conditions, tissues
which do not require insulin-mediated glucose
uptake, such as lens, kidney and nerve, are the
most susceptible to cellular damage. The accu-
mulation of sorbitol interferes with myo-
inositol transport, resulting in reduced Na+/
K+-ATPase activity which is critical for cellu-
lar homeostasis [6]. Increases in sorbitol also
cause swelling of the epithelial cells of the
lens, which has been linked to cataract forma-
tion [7]. Increased levels of NAD+ lead to
permeability changes in vessel formation and
alterations in glucose and lipid metabolism [1].
Aldose reductase (ALR) is the rate-limiting
enzyme in the sorbitol pathway. This enzyme,
which converts aldoses to their corresponding
alditols, has a broad specificity for a variety of
aromatic and aliphatic aldehydes [8], including
three-carbon aldehydes and ketones [9]. ALR
has been shown to exist in an unactivated (low
activity) form [8] that can be converted to an
activated (high activity) form by the oxidation
of a specific cysteine [10]. Structural studies on
human ALR have revealed an a/b-barrel
structure, with the coenzyme binding domain
located near the carboxyl terminal [11,12].
Tyrosine-48 has been proposed to be the pro-
ton donor to the aldehyde substrate, with the
participation of a lysine (lys-77) to lower the
pKa of the phenolic hydroxide [13,14]. An
active site histidine (his-110) has been assigned
a substrate binding and orientation role [13],
and chemical modification, peptide mapping,
and site-directed mutagenesis studies have
identified a lysine and several arginines that
are involved in coenzyme binding [15,16].
The second enzyme in the sorbitol pathway
is sorbitol dehydrogenase (SDH), which oxi-
2. Experimental
Materials.—Purified aldose reductase from
bovine kidney (high activity form, 3.5 U
mg−1) [8] was provided by Dr. Charles
Grimshaw (Igen International) and sheep liver
sorbitol dehydrogenase (26–28 U mg−1) was
purchased from Sigma Chemical Co. (St.
Louis, MO). The 2-, 3-, 4-, and 6-fluorodeoxy-
D
-glucoses were obtained from Dr. Stephen
Withers (University of British Columbia). Ad-
ditional fluoro- and deoxy- -glucose com-
pounds used for synthesis of the
corresponding -glucitol derivatives were ob-
D
D
tained from Lancaster-PCR, Inc. (Gainsville,
FL) or from Sigma. The corresponding sor-
bitol analogs were synthesized by chemical
reduction with sodium borohydride by a stan-
dard procedure [25], that briefly involves dis-
solving the various fluoro- and deoxyglucoses
in water at 4 °C followed by the dropwise
addition of a several fold molar excess of
sodium borohydride over a period of 30 min.
The reaction mixtures were then allowed to
warm to room temperature and were incu-
bated overnight. Aliquots of the reaction mix-
ture were tested for the presence of the aldose
starting material by using Fehling’s solution
or by o-toluidine determination, and the iso-
lated yield of products ranged from 60 to
90%. The identities of the products were confi-
rmed by proton and carbon NMR spec-
dizes
ductase, sorbitol dehydrogenase has broad
specificity; -xylitol is a substrate with a lower
Km than that of -glucitol, but -galactitol is
D
-glucitol to
D-fructose. Like aldose re-
D
D
D
not a substrate of this enzyme [4]. Aromatic
polyols [17] and 3-carbon alcohols [18] have
also been found to be substrates for sorbitol
dehydrogenase. SDH contains a single zinc
per subunit that is essential for catalytic activ-
ity [19]. There is no high-resolution structure
of SDH, however, a model of the three-dimen-
sional structure of SDH has been constructed
from the structure of horse liver alcohol dehy-

M.E. Scott, R.E. Viola / Carbohydrate Research 313 (1998) 247–253
249
troscopy. In all cases the expected carbon-13
upfield shift was observed for the deoxy car-
bon atom, and there was no evidence of any
remaining starting material. The purity of the
products was confirmed by comparison of the
melting points, where crystalline, to reported
values, and by thin-layer chromatography
(EM silica plate, EtOH/water/ammonium hy-
droxide, 21/4/1). Detection was achieved by
anisidine–HCl spray [26] with 10% sulfuric
acid in methanol. All of the compounds
yielded a single spot by TLC analysis, except
Package, SciTech International, Chicago, IL)
of the kinetics programs originally written by
Cleland [28].
Spectroscopic studies.—Circular dichroism
studies were conducted on a JASCO J600 CD
spectrometer. Samples were monitored near
the maximum carbonyl ellipticity peak (285–
290 nm) following a method for the quantita-
tion of the proportion of acyclic form [29].
Temperature studies were performed with a
thermally heated cuvette block. Spectra were
recorded when samples were equilibrated at
each temperature, as indicated by no further
changes in the measured ellipticities. The ellip-
ticities (in mdeg) were reproducible, leading to
less than a 10% error in the calculated percent
acyclic form.
for 6-deoxy- -glucitol. The impurity that was
D
detected in this compound was nonreactive to
p-anisidine, and therefore was not a reducing
sugar. This product mixture was purified on a
AG-50WX8 equilibrated with 1 M CaCl2.
TLC analysis after chromatography revealed a
single, non-reducing product whose identity
was confirmed by NMR spectroscopy.
3. Results
Enzyme assays.—Aldose reductase kinetics
were examined in 100 mM phosphate buffer,
pH 7.2, 27 °C, with 160 mM NADPH (Km=1
mM) by following the decrease in absorbance
at 340 nm resulting from the oxidation of
NADPH. Sorbitol dehydrogenase kinetics
were determined in 100 mM HEPES buffer,
pH 8, 30 °C, with 2 mM NAD (Km =11 mM)
by following the increase in absorbance at 340
nm resulting from the formation of NADH.
One unit of activity is defined as 1 mmol of
product produced min−1 mg−1 of protein un-
der the standard reaction conditions. The ki-
netic parameters for each substrate were
determined by varying the substrate concen-
tration at saturating concentration (\20
times Km) of the cofactor and fitting the mea-
sured rates to the equation for Michaelis–
Menten kinetics. Kinetic studies with a variety
of substrates for each enzyme have shown that
the affinity for the coenzyme is not signifi-
cantly affected by the nature of the substrate
[4,27]. When substrate inhibition was ob-
served, the data were modeled by Eq. (1):
Substrate specificity of aldose reductase.—
ALR catalyzes the reduction of the physiolog-
ical substrate,
D
-glucose, to
D
-glucitol.
Structural analogs of
D
-glucose, in which one
of the sugar hydroxyl groups was replaced
with either a fluorine or a hydrogen, were
examined as possible alternative substrates for
the reaction catalyzed by ALR. ALR binds
only the acyclic form of glucose and does not
catalyze ring-opening [30]. Therefore it is nec-
essary to correct the substrate concentrations
for the amount of acyclic form present in
order to compare the relative kinetic parame-
ters of these alternative substrates. Table 1
gives the percent acyclic form of these sub-
strates, as determined by circular dichroism
measurements. The 4-fluoro analog contains
2.5 times the amount of acyclic form, and the
3-fluoro analog about the same percentage as
glucose. The 6-deoxy-
D
-glucose has been de-
termined to have 0.002% acyclic form [29],
while 2-deoxy-D-glucose has about 2.5 times
this amount.
Given the broad specificity of ALR, it was
not surprising that all of the fluoro- and de-
oxysugar analogs that were tested were found
to be substrates for the enzyme. However,
Vmax[A]
w=
(1)
Km +[A]+[A]2/Ki
where w is the measured rate, and Km and Ki
are the Michaelis and the substrate inhibition
constants, respectively. The kinetic data were
fit by using BASIC versions (Enzyme Kinetics
unexpectedly, 3-fluoro- and 4-fluoro- -glucose
D
were found to be better substrates, with sig-
nificantly lower Km values and higher kcat/Km

250
M.E. Scott, R.E. Viola / Carbohydrate Research 313 (1998) 247–253
Table 1
Kinetic parameters for aldose reductasea
Substrate
kcat (min−1
)
Km (mM)
% acyclic Kamdj (mM)b
k
cat/Kmadj (mM−1 min−1 b
)
DDG‡ (kcal mol−1
)
D
-Glucose
15.195.4
-glucose 3.0090.22
11.190.4
-glucose 16.591.8
-glucose 15.890.4
12.991.3
12.592.6
10.490.7
0.3190.05 0.0024
0.5490.07 0.0073
6.391.8
3.190.4
9.891.6
0.0025
n.d.
0.0073
32.393.3
n.d.
75.996.5
467936
n.d.
146911
2-Fluoro-
D
2-Deoxy-
3-Fluoro-
4-Fluoro-
5-Thio-
6-Fluoro-
6-Deoxy-
D
-glucose
0.7
−2.3
−1.3
0.1
D
D
0.7490.12 2230092900
3.990.5
31.599.0
n.d.
40509360
400970
n.d.
D
-glucose
12.691.4
-glucose 15.190.7
0.005c
n.d.
D
D
-glucose
11.590.4
0.002c
19.693.2
587972
−0.1
a Conditions: 100 mM phosphate buffer, pH 7.2, 160 mM NADPH, 10 mg aldose reductase, 27 °C.
b The Km and kcat/Km values were adjusted by multiplying the apparent Km by the percent of acyclic form present.
c Ref. [29].
values than that of
D
-glucose (Table 1). 2-Flu-
mM, indicating that interactions between the
enzyme and the substrate at this position are
important for binding. The 3-fluoro and 4-
oro- -glucose has a comparable Km to that of
D
D
-glucose, but the kcat for this substrate
analog has decreased by 5-fold, in contrast to
the similar kcat for the 2-deoxy analog. The
5-thio-
substrate to
fluoro analogs are poorer substrates than
glucitol, with an observed increase in the free
energy of activation of 1.5–1.6 kcal mol−1
D
-
D
-glucose analog is also a comparable
,
D-glucose and, when adjusted for
suggesting that hydrogen bonding at these
positions also plays a role in substrate recog-
nition and binding. This is confirmed by the
2.6 kcal mol−1 increase in activation energy
the amount of acyclic form present at equi-
librium, has identical kinetic parameters to
D-glucose (Table 1). The 1-thio- -glucose
D
analog is not a substrate for ALR, but was
found to be an inhibitor of the enzyme.
Changes in the free energy of activation
(DDG‡) are observed when an individual sub-
strate hydroxyl group is substituted by either
a hydrogen or fluorine [31]. These changes can
be calculated from a comparison of the kcat/
Km values for the substrate and the analog
according to Eq. (2):
for 4-deoxy- -glucitol. However, there is a
D
recovery in Km and kcat/Km when the 3-hy-
droxyl group is removed. The 6-fluoro- and
6-deoxy- -glucitols are very good substrates
D
for SDH (Table 2), with their free energies of
activation lowered by about 0.5 kcal mol−1
for each of these analogs. These analogs also
exhibit substrate inhibition at higher concen-
trations, with a Ki value of 65 mM for sub-
strate inhibition by 6-fluoro-
D
-glucitol, while
DDG‡=RT ln[(kcat/Km)substrate/(kcat/Km)analog
]
the Ki for 6-deoxy- -glucitol is 20 mM. The
D
(2)
Km for 6-deoxysorbitol is about 2-fold lower
than that of the 6-fluoro compound, and SDH
has a higher affinity for each of these com-
These values show a 1.3–2.3 kcal mol−1 de-
crease in the free energy of activation for the
4-fluoro and 3-fluoro analogs of
pounds than for
D
-glucitol.
D
-glucose
compared to the physiological substrate
(Table 1).
Substrate specificity of sorbitol dehydrogen-
ase.—A more discriminating pattern of sub-
strate specificity has been observed for
sorbitol dehydrogenase. Neither the 2-fluoro
4. Discussion
Substrate recognition and the role of hydro-
gen bonding.—Binding and orientation of the
substrate at the active site of an enzyme plays
an important role in catalysis. Recognition
and binding of a substrate requires that the
enzyme active site provide a complementary
nor the 2-deoxy analogs of -glucitol are sub-
D
strates for the enzyme (Table 2). However,
surprisingly, these analogs do not inhibit SDH
when examined at concentrations up to 50

M.E. Scott, R.E. Viola / Carbohydrate Research 313 (1998) 247–253
251
Table 2
Kinetic parameters for
D
-glucitol dehydrogenasea
kcat (min−1
Substrate
)
Km (mM)
1.590.7
k
cat/Km (mM−1 min−1
)
DDG‡ (kcal mol−1
)
D
-Glucitol
506953
3379103
b
2-Fluoro-
2-Deoxy-
3-Fluoro-
3-Deoxy-
4-Fluoro-
4-Deoxy-
D
-glucitol
b
D-glucitol
D
-glucitol
35098
441915
354930
129911
55994
27498
15.590.8
1.990.3
11.793.0
26.495.7
0.7490.02
0.4890.06
22.690.8
231924
30.395.3
4.990.5
75594
1.6
0.2
1.5
2.6
−0.5
−0.4
D-glucitol
D
-glucitol
-glucitol
-glucitol
-glucitol
D
6-Fluoro-
6-Deoxy-
D
D
685920
a Conditions: 100 mM Hepes buffer, pH 8.0, 0.1 mM EDTA, 1 mM DTT, 2 mM NAD, 0.2 mg SDH, 30 °C.
b No activity observed.
environment, which can include hydrogen
bonding, electrostatic, and/or hydrophobic in-
teractions, and requires the correct steric
placement of the active site groups to interact
with the available substrate functional groups.
Hydrogen bonding is expected to be the pri-
mary mode by which carbohydrate-utilizing
enzymes specifically recognize and bind their
polyol substrates. Each sugar hydroxyl group
can potentially participate in three hydrogen
bonds, as acceptor of two hydrogen bonds
through the lone pairs on oxygen, and as a
donor of one hydrogen bond. Hydrogen
bonds are highly directional, and are stable
enough to contribute significantly to substrate
affinity. At the same time hydrogen bonds are
transient enough to allow rapid dissociation of
product. The hydrogen bonds that are formed
with enzyme functional groups are essential
for active site binding since these interactions
must compete with and replace the hydrogen
bonds that are lost upon removal of the sugar
substrates from aqueous solution.
Alternati6e deoxy- and fluorosugar sub-
strates.—Kinetic studies of alternative sub-
strates can examine the role of individual
hydroxyl groups in binding and in catalysis.
The use of deoxygenated and fluorinated car-
bohydrate substrate analogs to probe sub-
strate binding is based on the different
capacities of these analogs to form hydrogen
bonds, while maintaining similar steric inter-
actions. A deoxysugar cannot be involved in a
hydrogen bond at the modified position. How-
ever, a fluorine substituent in this position can
act as a potential proton acceptor due to the
electronegativity of fluorine and the inherent
polarity of the C–F bond. Reductions in the
catalytic rate and in the substrate affinity
provide an estimate of the cost of removing
specific interactions at each hydroxyl group,
and therefore an estimate of the extent to
which those specific site interactions stabilize
the transition state for the normal substrate.
Substrate recognition by aldose reductase.—
ALR is a broad-spectrum reductase with sub-
strates that include a variety of carbohydrates
as well as hydrophobic aromatic compounds.
In contrast to the binding sites of most carbo-
hydrate-binding proteins, in which an intricate
network of hydrogen bonds form the basis for
carbohydrate recognition [32], the active site
of ALR is quite hydrophobic [11]. Only the
polar nature of histidine-110 is involved in the
binding and orientation of the polar carbohy-
drate substrates [13]. Despite the paucity of
well-defined hydrogen bonding interactions,
ALR still shows a preference for aldoses with
both 2-
The 2- and 6-deoxy-
slightly poorer substrates with ALR than
D
and 3-
L
stereochemistry [27].
D
-glucoses are only
D
-
glucose, suggesting that any interactions that
may occur at these positions are not critical
binding determinants. The 6-fluoro analog is a
better substrate than the 6-deoxy sugar, sug-
gesting that the 6-hydroxyl group is involved
in substrate binding as a hydrogen bond ac-
ceptor. The 3- and 4-fluoro analogs are very
good substrates, possessing some 40- and 20-
fold better kcat/Km values than
D-glucose, re-
spectively. Fluorine substitution at these
positions can participate as a potential hydro-

252
M.E. Scott, R.E. Viola / Carbohydrate Research 313 (1998) 247–253
for 6-fluoro-
lower, and that for the 6-deoxy analog de-
creases by 3-fold compared to -glucitol.
These results suggest that the presence of a
primary hydroxyl group at carbon-6 is not
important for substrate binding and
orientation.
D
-glucitol is a factor of two
gen bond acceptor, but fluorine substitution
also increases the acidity of the adjacent hy-
droxyl groups. The 2-fluoro compound is a
poor substrate for aldose reductase, but the
D
2-deoxy compound is as good as
D
-glucose.
This result argues that hydrogen bonding at
this position is not critical, and points to an
important inductive effect in acyclic sub-
The free energies of activation for cataly-
sis of the fluoro- and deoxy-substrate
analogs of ALR and SDH show significant
differences when compared to the physiolog-
ical substrates. These changes have helped
in the assignment of the roles of the indi-
vidual substrate hydroxyl groups in binding
and catalysis. This approach has been used
to examine substrate hydrogen bonding in-
teractions in other enzyme systems. In those
cases, the activation energies of the fluoro-
and deoxy-analogs were all found to be
substantially higher than that of the native
substrate. These values ranged from 2.7 to
4.5 kcal mol−1 for analogs of glucose-1-
phosphate interacting with phosphoglucomu-
tase [35], to as high as 8 kcal mol−1 for
the same substrate analogs interacting with
glycogen phosphorylase [31]. The smaller
changes in activation energies observed with
the enzymes of the sorbitol pathway reflect
the broader substrate specificities of these
enzymes, as a consequence of the smaller
number of strong and specific hydrogen
bonding interactions with the substrates.
strate formation. The 1-thio- -glucose has
D
little tendency to mutarotate, since the sul-
fur anion does not favor electron release to
the ring oxygen to form the acyclic com-
pound. Thus this analog is not expected to
be a substrate, but the binding of this cyclic
inhibitor is consistent with the lack of bind-
ing discrimination in substrate recognition
by ALR.
Substrate recognition by sorbitol dehydro-
genase.—The overall conformation of sor-
bitol in water is not a common zigzag
conformation, but is a non-linear bent chain
conformation [33] resulting from the steric
interaction between substituents on alternate
carbon atoms and interactions with solvent
water molecules. This conformation is also
observed in the crystal structure, where C1
˚
is 1.06 A out of plane from the other five
coplanar carbons [34]. Substitution by either
fluorine or hydrogen at the 2- or 4-positions
would alleviate some of the steric hindrance
in this structure, and could therefore lead to
enhanced binding to SDH.
Neither the 2-deoxy nor the 2-fluoro
analogs of
D-glucitol are substrates or in-
hibitors, implying that this position is an
important substrate binding determinant and
that the 2-hydroxyl group of sorbitol is a
hydrogen bond donor to an active site
group of SDH. The 10-fold increase in Km
Acknowledgements
This work was supported by a Faculty
Research Grant from the University of
Akron. These studies were aided by the
generous gifts of fluoroglucose samples by
Dr. Stephen Withers (University of British
Columbia) and aldose reductase by Dr.
Charles Grimshaw (Igen International). Dr.
Withers also provided helpful advice on the
synthesis of several sugar analogs. Dr. Bryce
Plapp (University of Iowa) provided the co-
ordinates of the modeled sorbitol dehydro-
genase structure, and the CD spectra were
collected in the laboratory of Dr. Joyce Jen-
toft (Case Western Reserve University).
Their assistance is gratefully acknowledged.
for the 3-fluoro- and 4-fluoro- -glucitols
D
also implies a binding role for the hydroxyl
groups at these positions. The results with
the corresponding deoxysorbitols show a
dramatic effect on the kinetic parameters
upon removal of the 4-hydroxyl group,
confirming a binding role for this position.
However, the recovery of activity with the
3-deoxy analog eliminates this position from
involvement in substrate binding, and sug-
gests some additional effects from fluorine
substitution at this position. The Km value

M.E. Scott, R.E. Viola / Carbohydrate Research 313 (1998) 247–253
253
[17] W. Maret, Ad6. Exp. Med. Biol., 284 (1991) 327–336.
[18] R.I. Lindstad, J.S. McKinley-McKee, FEBS Lett., 330
(1993) 31–35.
[19] J. Jeffery, J. Chesters, C. Mills, P.J. Sadler, H. Jornvall,
EMBO J., 3 (1984) 357–360.
[20] H. Eklund, E. Horjales, H. Jornvall, C.I. Branden, J.
Jeffery, Biochemistry, 24 (1985) 8005–8012.
[21] M.C. Feiters, J. Jeffery, Biochemistry, 28 (1989) 7257–
7262.
[22] H. Reiersen, K. Sletten, J.S. McKinley-McKee, Eur. J.
Biochem., 211 (1993) 861–869.
[23] H. Reiersen, R.I. Lindstad, J.S. McKinley-McKee, Arch.
Biochem. Biophys., 311 (1994) 450–456.
[24] C. Karlsson, J.O. Ho¨o¨g, Eur. J. Biochem., 216 (1993)
103–107.
[25] M.L. Wolfrom, A. Thompson, in R.L. Whistler (Ed.),
Methods in Carbohydrate Chemistry, Vol. II: Reactions of
Carbohydrates, Academic Press, New York, 1963, pp.
65–71.
[26] R.M. Dawson, D.C. Elliot, W.H. Elliot, K.M. Jones,
Data for Biochemical Research, Oxford Science Publica-
tions, Oxford, UK, 1986.
References
[1] J. Jeffery, H. Jornvall, Proc. Natl. Acad. Sci. USA, 80
(1983) 901–905.
[2] K.H. Gabbay, New Engl. J. Med., 288 (1973) 831–836.
[3] S.D. Varma, S.S. Schocket, R.D. Richards, In6est. Oph-
thal., 18 (1979) 237–241.
[4] J. Jeffery, H. Jornvall, Ad6. Enzymol., 61 (1988) 47–106.
[5] T.A. Beyer, N.J. Hutson, Metabolism, 35 (1986) 1–3.
[6] G. Pugliese, R.G. Tilton, J.R. Williamson, Diab. Metab.
Re6., 7 (1991) 35–59.
[7] G.J. Zenon, C.V. Abobo, B.L. Carter, D.W. Ball, Clin.
Pharm., 9 (1990) 446–457.
[8] C.E. Grimshaw, M. Shahbaz, G. Jahngiri, C.G. Putney,
S.R. McKercher, E.J. Mathur, Biochemistry, 28 (1989)
5343–5353.
[9] D.L. Vander Jagt, B. Robinson, K.K. Taylor, L.A.
Hunsaker, J. Biol. Chem., 267 (1992) 4364–4369.
[10] J.M. Petrash, T.M. Harter, C.S. Devine, P.O. Olins, A.
Bhatnagar, S. Liu, S.K. Srivastava, J. Biol. Chem., 267
(1992) 24833–24840.
[11] D.K. Wilson, K.M. Bohren, K.H. Gabbay, F.A. Quio-
cho, Science, 257 (1992) 81–84.
[12] D.W. Borhani, T.M. Harter, J.M. Petrash, J. Biol.
Chem., 267 (1992) 24841–24847.
[27] C.E. Grimshaw, Biochemistry, 31 (1992) 10139–10145.
[28] W.W. Cleland, Ad6. Enzymol., 29 (1967) 1–32.
[29] L.D. Hayward, S.J. Angyal, Carbohydr. Res., 53 (1977)
13–20.
[13] K.M. Bohren, C.E. Grimshaw, C.J. Lai, D.H. Harrison,
D. Ringe, G.A. Petsko, K.H. Gabbay, Biochemistry, 33
(1994) 2021–2032.
[30] C.E. Grimshaw, Carbohydr. Res., 148 (1986) 345–348.
[31] I.P. Street, K. Rupitz, S.G. Withers, Biochemistry, 28
(1989) 1581–1587.
[14] I. Tarle, D.W. Borhani, D.K. Wilson, F.A. Quiocho,
J.M. Petrash, J. Biol. Chem., 268 (1993) 25687–25693.
[15] T. Yamaoka, Y. Matsuura, K. Yamashita, T. Tanimoto,
C. Mshimura, Biochem. Biophys. Res. Commun., 183
(1992) 327–333.
[16] T.J. Kubiseski, N.C. Green, D.W. Borhani, T.G. Flynn,
J. Biol. Chem., 269 (1994) 2183–2188.
[32] F.A. Quiocho, Annu. Re6. Biochem., 55 (1986) 287–315.
[33] G E. Hawkes, D. Lewis, J. Chem. Soc. Perkin Trans., 2
(1984) 2073–2078.
[34] Y.J. Park, G.A. Jeffrey, W.C. Hamilton, Acta Crystal-
logr., B27 (1971) 2393–2401.
[35] M.D. Percival, S.G. Withers, Biochemistry, 31 (1992)
498–505.
.