Design and synthesis of highly potent and selective (2-arylcarbamoyl- phenoxy)-acetic acid inhibitors of aldose reductase for treatment of chronic diabetic complications

Article abstract of DOI:10.1016/j.bmc.2004.07.062

Recent efforts to identify treatments for chronic diabetic complications have resulted in the discovery of a novel series of highly potent and selective (2-arylcarbamoyl-phenoxy)-acetic acid aldose reductase inhibitors. The compound class features a core template that utilizes an intramolecular hydrogen bond to position the key structural elements of the pharmacophore in a conformation, which promotes a high binding affinity. The lead candidate, example 40, 5-fluoro-2-(4-bromo-2-fluoro-benzylthiocarbamoyl)-phenoxyacetic acid, inhibits aldose reductase with an IC₅₀ of 30 nM, while being 1100 times less active against aldehyde reductase, a related enzyme involved in the detoxification of reactive aldehydes. In addition, example 40 lowers nerve sorbitol levels with an ED₅₀ of 31 mg/kg/d po in the 4-day STZ-induced diabetic rat model.

Full text of DOI:10.1016/j.bmc.2004.07.062

Bioorganic & Medicinal Chemistry 12 (2004) 5661–5675
Design and synthesis of highly potent and selective
(2-arylcarbamoyl-phenoxy)-acetic acid inhibitors of
aldose reductase for treatment of chronic diabetic complications
a
a
a
Michael C. Van Zandt,a, Evelyn O. Sibley, Erin E. McCann, Kerry J. Combs,
*
Brenda Flam,a Diane R. Sawicki,a Al Sabetta,a Anne Carrington,a Janet Sredy,a
Eduardo Howard,b Andre Mitschlerb and Alberto D. Podjarnyb
aThe Institute for Diabetes Discovery, LLC, 23 Business Park Drive, Branford, CT 06405, USA
bLaboratoire de Biologie et Genomique Structurales, UPR 9004 du CNRS, IGBMC,
1 rue Laurent Fries, B.P. 163, 67404 Illkirch Cedex, France
Received 14 May 2004; revised 28 July 2004; accepted 28 July 2004
Available online 1 September 2004
Abstract—Recent efforts to identify treatments for chronic diabetic complications have resulted in the discovery of a novel series of
highly potent and selective (2-arylcarbamoyl-phenoxy)-acetic acid aldose reductase inhibitors. The compound class features a core
template that utilizes an intramolecular hydrogen bond to position the key structural elements of the pharmacophore in a confor-
mation, which promotes a high binding affinity. The lead candidate, example 40, 5-fluoro-2-(4-bromo-2-fluoro-benzylthiocarba-
moyl)-phenoxyacetic acid, inhibits aldose reductase with an IC50 of 30nM, while being 1100 times less active against aldehyde
reductase, a related enzyme involved in the detoxification of reactive aldehydes. In addition, example 40 lowers nerve sorbitol levels
with an ED50 of 31mg/kg/d po in the 4-day STZ-induced diabetic rat model.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1.Introduction
mal pathways of glucose metabolism and a dramatic in-
crease in flux through the polyol pathway results
(Scheme 1). Glucose entering the polyol pathway is re-
duced to sorbitol by aldose reductase (ALR2) and
NADPH. Sorbitol is subsequently oxidized to fructose
by sorbitol dehydrogenase and NAD+. This increased
flux through the polyol pathway results in a change in
redox potential for these cofactors. The increased ratio
of cytosolic NADH to NAD+ is called reductive stress,
which has been linked to depleted intracellular levels
of glutathione, increased nonenzymatic glycation, and
activation of protein kinase C.4,5 In addition to reduc-
tive stress, accumulation of sorbitol in certain cells re-
sults in osmotic stress, which can lead to cell swelling
and eventually cell death.6 Inhibitors of aldose reductase
can prevent these metabolic and biochemical changes,
which have been linked to the pathogenesis of diabetic
complications.7
Diabetes mellitus and its disabling complications, which
include blindness, renal failure, neuropathy, limb ampu-
tation, myocardial infarction, and stroke affect some 17
million people in the United States with an estimated
cost of over 100 billion dollars annually.1 Through var-
ious clinical studies, including the Diabetes Control and
Complications Trial (DCCT) and the United Kingdom
Prospective Diabetes Study (UKPDS), the development
and progression of these complications in type I and
type II patients has been clearly linked to elevated blood
glucose levels.2,3 During euglycemic conditions glucose
is preferentially metabolized through the glycolytic
pathway where it is phosphorylated with ATP by hexo-
kinase to form glucose-6-phosphate. However, during
conditions of hyperglycemia, as observed in diabetes
mellitus, elevated blood glucose levels saturate the nor-
In addition to these biochemical endpoints, recent
reports provide a strong genetic correlation between
over expression of aldose reductase and the development
of diabetic complications.8–10 As a result, a high ALR2
Keywords: Aldose reductase; Aldehyde reductase; Diabetes;
Complications.
*
Corresponding author. Tel.: +1 203 315 5951; fax: +1 203 315
4002; e-mail: michael.vanzandt@ipd-discovery.com
0968-0896/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.07.062

5662
M. C. Van Zandt et al. / Bioorg. Med. Chem. 12 (2004) 5661–5675
CH2OH
O
H
OH
OH
CH2OH
OH
H
OH
OH
CH2OH
CHO
OH
H
OH
OH
CH2OH
NADP+
NAD+
NADPH
NADH
H
HO
H
H
HO
H
HO
H
H
sorbitol
dehydrogenase
aldose
reductase
H
H
CH2OH
Glucose
Sorbitol
Fructose
Scheme 1. Polyol pathway.
level is now considered to be an important risk factor for
chronic diabetic complications including neuropathy,
retinopathy, and nephropathy.
terminal loop, a section of the protein lining the active
site cleft, are responsible for various substrate and inhib-
itor specificities.20
Although many potent aldose reductase inhibitors
(ARIs) have been identified and developed, none are
currently marketed for worldwide use.4,7 Many of these
candidates failed to gain acceptance due to an inade-
quate therapeutic index. In some cases this toxicity
may have resulted from a lack of selectivity relative to
aldehyde reductase (ALR1). The physiological impor-
tance of ALR1 has been clearly demonstrated.11,12 It
converts highly reactive 2-oxoaldehydes like 3-deoxy-
glucosone and methyl glyoxal to their corresponding
nonreactive alcohols.13,14 These aldehydes, which are in-
volved in the formation of various protein cross-links
and other advanced glycation end products (AGEs),
are formed as degradation products from glucose and
fructose.15 During conditions of prolonged hyperglyce-
mia, high concentrations of these aldehydes are formed
resulting in the development of chronic diabetic compli-
cations and accelerated aging.16–18
Efforts to design novel, potent, and selective ARIs have
resulted in the discovery of a compound class that fea-
tures a 2-arylcarbamoyl-phenoxyacetic acid as a core
template. Within this template, an intramolecular
hydrogen bond is used to position the acetic acid moiety
and aromatic side chain, the key structural elements of
the pharmacophore, in the necessary orientation for
high binding affinity.
2.Inhibitor design and synthesis
The target compounds were synthesized using methods
illustrated in Schemes 2–4. (2-Arylcarbamoyl-phen-
oxy)-acetic acids 6, 8, and 13 in Scheme 2 illustrate
methods used to prepare most examples. Treatment of
commercially available 2,4-difluorobenzoic acid (1) with
sodium hydroxide at 135ꢁC affords 4-fluorosalicyclic
acid.21 Activation of the carboxylic acid with thionyl
chloride provides the corresponding acid chloride, which
is coupled with 3-nitrobenzylamine to give amide 4.
Alkylation with ethyl bromoacetate followed by hydro-
As a member of the aldo-ketoreductase super family,
ALR1 shares greater than 85% homology with ALR2
(rat and man).19 Specific amino acid changes in the C-
F
F
F
F
OH
R1
F
OH
HCl•H2N
HS
F
g
a,b
H
N
+
OH
N
O
O
O
F
9
h
10
1
2, R = OH
3, R = Cl
F
F
F
OH
c
N
S
H
N
F
F
OH
H
11
O
N
NO2
d,e
O
d
4
O
O
O
OR2
F
F
F
OR2
OR2
f,e
F
O
O
F
O
F
O
O
N
S
H
N
H
N
H
N
NO2
NO2
S
5, R2 = Et
6, R2 = H
7, R2 = Et
8, R2 = H
12, R2 = Et
13, R2 = H
e
Scheme 2. Reagents and conditions: (a) NaOH, DMI, 135ꢁC (68%); (b) SOCl2, cat. DMF, heptane, 60ꢁC, (85%); (c) 3-nitrobenzylamine, TEA,
CH2Cl2 (67%); (d) ethyl bromoacetate, aq K2CO3, acetone; (e) aq NaOH, ethanol; (f) P4S10, pyridine (98%); (g) aminoacetonitrile hydrochloride,
TEA, CH2Cl2 (65%); (h) ethanol, heat (21%).

M. C. Van Zandt et al. / Bioorg. Med. Chem. 12 (2004) 5661–5675
5663
H2N
H3C
OCH3
NO2
F
OCH3
NO2
F
OCH3
NH2
F
OCH3
Br
a
c
b
H3C
H3C
H3C
14
15
16
17
18
d
F
OCH3
Br
F
OCH3
OH
F
OCH3
CN
f,g
H
e
N
H3C
H3C
H3C
O
F
O
20
19
O
h,i
OH
F
OR
Br
F
OR
Br
F
O
Br
j,k
k
H
H
H
N
N
N
H3C
H3C
H3C
S
F
O
F
O
F
23, R = CH2CO2Et
24, R = CH2CO2H
21, R = H
22, R = CH2CO2Et
25
Scheme 3. Reagents and conditions: (a) NaNO2, HF-pyridine (76%); (b) H2, 10% Pd–C, ethanol (99%); (c) 48% aq HBr, NaNO2, CuBr (47%); (d)
CuCN, DMF, 160ꢁC (86%); (e) 2N NaOH, 90ꢁC, (96%); (f) (COCl)2, cat. DMF, CH2Cl2; (g) 4-bromo-2-fluorobenzylamine hydrochloride, DIEA,
CH2Cl2 (97%, two steps); (h) 25% aq HBr, AcOH, 120ꢁC (73%); (i) ethyl bromoacetate, aq K2CO3, acetone (93%); (j) P4S10, pyridine (89%); (k) aq
NaOH, ethanol.
Diazotization of commercially available 5-methoxy-2-
methyl-4-nitroaniline (14) with sodium nitrite in the
presence of HF-pyridine provided the corresponding
fluoride22 (15) in 76% yield. Subsequent hydrogenation
followed by diazotization and treatment with a CuBr
provided bromide 17.23 Treatment with CuCN in
DMF at 160ꢁC gave nitrile 18 (86%), which was
hydrolyzed with aq NaOH to give carboxylic acid 19.
Formation of the acid chloride with oxalyl chloride
and catalytic DMF followed by coupling with 4-bro-
mo-2-nitrobenzylamine provided amide 20 in nearly
quantitative yield. Deprotection of the methyl ether
with 25% HBr in acetic acid at 120ꢁC followed by
alkylation with ethyl bromoacetate and hydrolysis pro-
vides target amide 25. Alternatively, treatment of ester
22 with phosphorous pentasulfide and subsequent
hydrolysis provides the corresponding thioamide target
(24).
F
OR1
OR2
a,b
2
Br
O
26, R1,R2 = H
c,d
f,d
F
27, R1,R2 = CH3
e,d
F
OCH3
OR2
F
OCH3
OR2
OCH3
OR2
N
NC
Ph
O
O
O
O
28, R2 = CH3
29, R2 = H
30, R2 = CH3
32, R2 = CH3
31, R2 = H
33, R2 = H
Scheme 4. Reagents and conditions: (a) Br2, AcOH, H2SO4 (95%); (b)
CH3I, K2CO3, acetone (80%); (c) morpholine, Cs2CO3, Pd(OAc)2,
BINAP, toluene, 100ꢁC (47%); (d) aq NaOH, ethanol; (e) CuCN,
DMF, 150ꢁC (75%); (f) PhB(OH)2, K2CO3, Pd(PPh3)4, toluene, 110ꢁC
(91%).
The syntheses of additional 4-fluoro analogs with other
substituents in the 5-position were conveniently
prepared via bromide 26 as illustrated in Scheme 4.
Bromination of 4-fluorosalicyclic acid with Br2 in a
solution of acetic acid and sulfuric acid followed by
alkylation with iodomethane and aq K2CO3 provided
the bis-alkylated intermediate 27 in 76% yield (two
steps). This highly functionalized salicylic acid inter-
mediate was used to prepare examples with a variety
of substituents in the 4-position. For example, the
morpholine intermediate 28 was prepared from bromide
lysis of the ethyl ester gives phenoxyacetic acid 6. Alter-
natively, treatment of amide 5 with phosphorous penta-
sulfide followed by hydrolysis provides thioamide
derivative 8. If the benzothiazole derivatives are desired,
aminoacetonitrile can be used in place of the substituted
benzylamine to provide nitrile 9, which can subsequently
be condensed with the desired 2-aminothiophenol (10)
to provide amide 11. Alkylation of the phenol with ethyl
bromoacetate followed by hydrolysis of the ethyl ester
provides the target benzothiazole derivative 13 in high
yield.
27 using standard Buckwald conditions.24 The nitrile
¨
intermediate 30 was prepared by treating bromide 27
with CuCN in DMF at 150ꢁC (75%). Additionally, a
Suzuki coupling of phenylboronic acid with bromide
27 provided phenyl derivative 32. These intermediates
were used to prepare the corresponding target com-
pounds using methods already described in Schemes 2
and 3.
Additional examples, where the desired substituted sali-
cylic acid or corresponding 2-fluorobenzoic acid starting
materials were not commercially available, were pre-
pared separately and are described in Schemes 3 and
4. The 2-fluoro-3-methyl examples were prepared using
chemistry outlined in Scheme 3.

5664
M. C. Van Zandt et al. / Bioorg. Med. Chem. 12 (2004) 5661–5675
3.Results and discussion
new, more refined, structure with exceptional resolution
˚
(0.66A). These results will undoubtedly advance our
The major objective of this program has been to identify
highly potent, selective and efficacious aldose reductase
inhibitors for treatment of chronic diabetic complica-
tions. All compounds in the program were initially
tested for potency against human aldose reductase
(hALR2) and selectivity relative to human aldehyde
reductase (hALR1). Selected compounds with signifi-
cant in vitro activity against hALR2 were also tested
in the 4-day streptozotocin (STZ) induced diabetic rat
model. The results from these experiments are listed in
Tables 1–3.
understanding of the catalytic mechanism. The details
of this work has been published separately.29
With a clear picture of the inhibitor-enzyme complex
early in the lead optimization program, little was done
with the (2-arylcarbamoyl-phenoxy)-acetic acid core
structure. New analog synthesis was primarily directed
to those areas perceived to have the greatest potential
for significant improvement. In keeping with our belief
that selectivity relative to aldehyde reductase would play
an important role in the success of any clinically useful
ARI, a significant portion of the optimization effort
was directed toward the aromatic side chain. The side
chain is the portion of the inhibitor that interacts with
the C-terminal loop, the region of the enzyme most dif-
ferentiated between hALR2 and hALR1. In general, the
structure–activity relationships associated with the
structural class are consistent with binding information
derived from the enzyme-inhibitor complexes. Large
variations in hALR2 activity are observed with different
substitution patterns on the aromatic side chain. In gen-
eral, the compounds with the strongest binding affinities
contained electron deficient aromatic side chains. All
examples with strong electron donating substituents
had poor binding affinities with IC50ꢁs > 2lM (35, 7-
methoxy; 43, 6,7-methylenedioxo; 50, 6,8-dimethoxy).
Examples with substituents only in the para-position
(R7) were, at best, only weakly active with Br being
preferred with an IC50 of 550nM (37, Br > 41, Cl > 34,
H > 35, OCH3 > 36, CF3). Addition of an ortho- or
meta-substituent often provides a dramatic increase in
potency. This is illustrated in example 42 where intro-
duction of a second chlorine (6,7-dichloro), results in
an IC50 of 34nM. This is nearly a 40-fold increase in
potency relative to example 41 (7-choro), which has an
IC50 of 1300nM. A similar effect is observed when
comparing example 37 (7-bromo) with example 38 (5-
fluoro-7-bromo). Introduction of the 5-fluoro group
brings the IC50 from 550 to 30nM, nearly a 20-fold
increase in potency. In general, electron withdrawing
groups capable of being hydrogen bond acceptors in
the ortho or meta-positions appear to be critical for high
in vitro potency. Preferred substituents include fluorine,
trifluoromethyl, and nitro. The IC50ꢁs for examples with
a 6-fluoro (46), 6-trifluoromethyl (47), and 6-nitro (44)
are 820, 39, and 6nM, respectively. Introduction of sec-
ond meta-substituted group can provide some improve-
ment as seen in example 48, with a 6-triflouromethy and
8-fluoro-substituent, which has an IC50 of 29nM. All
activity is lost however when two large trifluormethyl
substituents are present in the meta-positions as seen
in example 49 with an IC50 of only 22lM. In this exam-
ple, the side chain is simply too large for the specificity
pocket.
The program benefited significantly from the early use
of X-ray crystallography. The structures provided a
clear understanding of the important interactions that
make up the enzyme-inhibitor complexes. This informa-
tion was used to identify the specific fragments of the
inhibitor template that provided the greatest opportuni-
ties for optimization of potency and selectivity relative
to aldehyde reductase.
The complexes were obtained using hALR2 expressed in
E. coli and crystallized with the oxidized form of the
coenzyme b-NADP+ at pH5 and 277K. Initial results
with example 40 provided a clear picture of the inhibi-
˚
tor-enzyme complex (1.8A). As illustrated in Figure 1,
the inhibitor is oriented in the active site of hALR2 in
a manner such that the hydrophilic carboxylate head
forms tight hydrogen bonds with the OH of Tyr 48
˚
˚
(2.74A), the NE2 of His 110 (2.67A), and the NE1 of
˚
Trp 111 (3.07A). The carboxylate also forms an optimal
˚
intramolecular hydrogen bond (3.01A) with the amide
N–H forming a unique nine-membered ring. In addition
to the hydrogen bonding interactions, the benzyl side
chain forms a nearly perfect p-stacking interaction with
the indole moiety of Trp 111. These interactions anchor
the inhibitor deep within the enzyme active site.
As this class of inhibitors bind to the hALR2 active site,
a conformational change occurs opening a pocket local-
ized between Phe 122, Trp 111, Leu 300, and Ala 299.25
The specific opening of the pocket varies to accommo-
date each inhibitor, producing an ꢀinduced fitꢁ. Since
the residues lining this pocket are not conserved in alde-
hyde reductase, the interactions are specific for
ALR2.26–28 As a result, inhibitors with binding interac-
tions in this ꢀspecificity pocketꢁ are generally highly selec-
tive for aldose reductase. The specific interactions within
the specificity pocket, for example, 40 include a signifi-
cant p-stacking interaction of the benzyl side chain with
the indole moiety of Trp 111, H-bonding interactions
between the side chain fluorine and amide N-Hꢁs of
˚
˚
Leu 300 (3.30A) and Ala 299 (3.27A), and an unusually
short contact between the benzyl Br and Thr 113-OG
˚
(2.98A). This Br-Thr-OG interaction is likely important
for selectivity since in aldehyde reductase Thr 113 is re-
placed by a much bulkier tyrosine, which would cause
steric hindrance.
As consistent with previously identified ARIs, the pre-
ferred substitution patterns appear to be the 3-nitro ben-
zyl (8) and the 2-fluoro-4-bromo benzyl (40).30,31 These
examples are highly potent and selective with IC50ꢁs of 6
and 30nM for hALR2 and 35,000 and 33,000nM for
hALR1, respectively.
After significant optimization, subsequent co-crystalliza-
tion experiments with example 40 have resulted in a

M. C. Van Zandt et al. / Bioorg. Med. Chem. 12 (2004) 5661–5675
5665
Table 1. Physical and in vitro inhibitory properties of (2-benzylcarbamoyl-phenoxy)-acetic acid ARIs
O
R1
OH R6
R5
R2
R3
O
X
R7
R8
H
N
R4
R9
Ex.
Substituents
X
Mp (ꢁC)
Empirical formula
Analysis
Inh. IC50 (nM)
hALR2 hALR1
34
35
36
37
38
39
40
41
42
43
44
6
2-Cl
2-Cl, 7-OCH3
O
O
O
O
O
O
S
145–146
178–179
184–185
172–173
184–185
143–145
154–157
184–186
177–178
208–209
200
C
16H14ClNO4
C20H22ClNO5
C, H, N, Cl
C,b H, N, Cl
C,b H, N, Clb
C, H,b N, Cl
C, H, N, Cl
C, H, N
3300
8800
10,000
550
30
NTa
NT
NT
NT
2-Cl, 7-CF3
2-Cl, 7-Br
C17H13ClF3NO5
C16H13BrClNO4
C16H12BrClFNO4
C16H12BrF2NO4
2-Cl, 5-F, 7-Br
2-F, 5-F, 7-Br
14,000
6600
33,000
NT
55
30
2-F, 5-F, 7-Br
2-Cl, 7-Cl
C
16H12BrF2NO3S
C, H, N, S, Br
C, H, N, Cl
C, H, N, Cl
C, H, N, Cl
C, H, N, Cl
C, H, N
O
O
O
O
O
S
C16H13Cl2NO4
C16H12Cl3NO4
C17H14ClNO6
C16H13ClN2O6
C16H13FN2O6
C16H13FN2O5S
C17H15FN2O6
1300
34
2-Cl, 6-Cl, 7-Cl
2-Cl, 3-4-methylenedioxo
2-Cl, 6-NO2
18,000
NT
NT
2800
6
2-F, 6-NO2
2-F, 6-NO2
148–151
147–150
159–160
155
8
34,000
35,000
5000
NT
8
C, H, N
6
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
25
24
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
Tolrestat
2-F, 6-NO2, 7-CH3
2-Cl, 6-F
2-Cl, 6-CF3
O
O
O
O
O
O
O
O
O
O
O
O
O
O
S
C, H, N
8
C
16H13ClFNO4ÆH2O
C, H,b N, Cl
C, H, N, Cl
C, H, N, Cl
C, H, N, Cl
C, H, N, Cl
C, H, N
820
39
179–181
160–162
191–193
163
C17H15ClF3NO4
C17H12ClF4NO2
C18H12ClF6NO4
C18H18ClNO6
6700
7300
NT
2-Cl, 6-CF3, 8-F
2-Cl, 6-CF3, 8-CF3
2-Cl, 6-OCH3, 8-OCH3
2-Cl, 5-F, 9-F
29
22,000
4800
730
176
630
2,400
550
83
NT
NT
189–191
144–145
188–189
169–170
203–204
196–199
193–194
156–158
206–209
169
C16H12ClF2NO2
C
5-F, 7-Br
16H13BrFNO4
C, H, N
C, H, N
2400
NT
2-CH3, 5-F, 7-Br
2-CF3, 5-F, 7-Br
2-OCH3, 5-F, 7-Br
2-SCH3, 5-F, 7-Br
2-SO2CH3, 5-F, 7-Br
2-F, 3-F, 5-F, 7-Br
2-F, 3-F, 5-F, 7-Br
2-F, 3-CH3, 5-F, 7-Br
2-F, 3-CH3, 5-F, 7-Br
2-F, 4-F, 5-F, 7-Br
3-F, 5-F, 7-Br
C17H15BrFNO4
C17H12BrF4NO4
C17H15BrFNO5
C17H15BrFNO4S
C17H15BrFNO6S
C16H11BrF3NO4
C16H11BrF3NO3S
C17H14BrF2NO4
C17H14BrF2NO3S
C16H11BrF3NO4
C, H, N
C, H, N
NT
NT
C, H, N, S
C,b H, N, S
C, H, N
2900
NT
13,000
37
33
1100
2700
5900
31,000
11,000
5500
1800
1900
11,000
NT
C, H, N, S
C, H, N
O
S
24
24
162–164
158–159
145–146
153–155
153–155
145–146
204 dec
174–175
235 dec
244 dec
149–150
186–187
182–183
163–165
229–230
189–190
169–172
180
C, H, N, S
C, H, N
C, H, N
O
O
O
S
820
150
64
C
16H10BrFNO4
3-Br, 5-F, 7-Br
3-Br, 5-F, 7-Br
C16H12Br2FNO4
C16H12Br2FNO3S
C17H15BrFNO4
C, H, N, Br
C, H, N, Br, S
C, H, N
37
69
3-CH3, 5-F, 7-Br
3-NH2, 5-F, 7-Br
3-NO2, 5-F, 7-Br
3-NHAc, 5-F
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
S
C16H14BrFN2OÆ0.5H2O
C16H12BrFN2O6
C18H17FN2O5
C, H, N
C, H, N
8300
200
5000
470
100
3100
2700
>50lM
130
1100
6
7900
NT
C, H, N
C, H, N
3-OH, 5-F, 7-Br
3-OCH3, 5-F, 7-Br
3-OCH2CH2CH3, 5-F, 7-Br
3-OCH2CHCH2, 5-F, 7-Br
1,2-Benzo, 5-F, 7-Br
2,3-Benzo, 5-F, 7-Br
3,4-Benzo, 5-F, 7-Br
2-F, 3-Br, 6-NO2
2-F, 3-morpholino, 6-NO2
2-F, 3-CN, 6-NO2
2-F, 3-Ph, 6-NO2
2-F, 3-Ph, 6-NO2
2-F, 3-CH3, 6-NO2
2-F, 3-CH3, 6-NO2
3-CH3, 6-NO2
C
16H13BrFNO5
NT
NT
C17H15BrFNO5
C19H19BrFNO5
C19H17BrFNO5
C, H, N
C, H, N
NT
NT
C, H, N
C, H, N
C
20H15BrFNO4
NT
NT
C20H15BrFNO4
C20H15BrFNO4
C16H12BrFN2O6
C20H20FN3O7
C17H12FN3O6
C22H17FN2O6
C22H17FN2O5S
C17H15FN2O6
C17H15FN2O2S
C17H16N2O6
C, H, N
C, H, N
NT
C, H, N, Br
C, H, N
C, H, N
44,000
54,000
6500
14,000
12,000
2600
3100
NT
7
179–180
210–211
177–179
177–179
137–139
193–194
154–156
158–161
180–182
7
C, H, N
C, H, N
C, H, N
8
11
O
S
8
C, H, N, S
C, H, N
C, H
7
11
O
O
S
3-OCF3, 6-NO2
3-OCF3, 6-NO2
2-F, 6-NO2, N-CH3
C17H13F3N2O7
C17H13F3N2O6S
C17H15FN2O6
9
12
NT
C, H, N, S
C, H, N
3600
NT
O
5300
14
1900
a Not tested.
b Elemental analysis not within 4% of theoretical value.

5666
M. C. Van Zandt et al. / Bioorg. Med. Chem. 12 (2004) 5661–5675
Table 2. Physical and in vitro properties of {2-[(benzothiazol-2-ylmethyl)-carbamoyl]-phenoxy}-acetic acid ARIs
O
R5
R6
R8
OH
R2
O
X
N
H
N
S
Ex.
Substituents
X
Mp (ꢁC)
Empirical formula
Analysis
Inhibition IC50 (nM)
hALR2
hALR1
86
87
13
88
2-Cl, 6-CF3
2-F, 6-CF3
O
O
O
S
225–227
206–208
172–174
150
C18H12ClF3N2O4S
C19H13F4NO4S
C17H10F4N2O4S
C, H, N, Cl
C, H, N
C, H, N
NTa
35
29
11
15
8600
35,000
22,000
17,000
2-F, 5-F, 6-F, 8-F
2-F, 5-F, 6-F, 8-F
C17H10F4N2O3S2
a Not tested.
Table 3. In vivo activity of phenoxyacetic acid ARIs
Ex.
Inhibition of sorbitol accumulation in STZ-induced
diabetic rat nerve
Dose (mg/kg/d) po
% Inh.
ED50 (mg/kg/d) po
38
44
50
50
50
50
50
50
50
10
50
50
50
50
50
50
10
33
27
15
39
54
37
41
40
89
49
36
16
55
62
31
52
64
39
82
6
40
31
8
25
24
87
13
88
Tolrestat
Substitution on the core salicylic acid structure is gener-
ally well tolerated and can significantly improve in vitro
and in vivo activity. Exceptions include those examples
with large groups in the 1 and 2-positions where unfa-
vorable steric interactions occur with VAL 47 (examples
54, 57, and 72). When comparing examples 39 and 52,
introduction of fluorine in the 2-position provides about
a threefold increase in potency and selectivity. In vivo
efficacy is also significantly increased with the 2-fluoro
substituent. The 3-methyl derivative (64) also provided
a significant improvement in potency, selectivity, and
in vivo efficacy relative to the unsubstituted example
(52). When combined together, the 2-fluoro-3-methyl
derivative (25) provided an additive improvement rela-
tive to the monosubstituted examples in the primary in
vitro assay. Unfortunately, the same effect was not ob-
served in the in vivo model.
Figure 1. Environment of inhibitor example 40 in the subatomic
˚
resolution structure of aldose reductase. Contacts shorter than 3.5A
are indicated. Note the short contacts of the inhibitor carboxylic acid
head with residues Tyr 48, His 110, and Trp 111 and the perfect p-
stacking of the benzyl side chain with the indole moiety of Trp 111.
(R3 = morpholino) are also two of the most selective
with IC50ꢁs for hALR1 of 44,000 and 54,000nM, respec-
tively. With selectivity ratios (hALR1/hALR2) > 7000,
these are among the most selective inhibitors identified
to date. Other examples in this series including 77
(R3 = cyano), 78 (R3 = phenyl), 80 (R3 = methyl), and
the corresponding thioamide derivatives (79 and 81)
are also very potent, but somewhat less selective.
Examples with a nitro substituent in the meta-position
on the benzyl ring (R6) are consistently most potent in
the primary in vitro assay. Some of the most interesting
examples combine the 6-nitro substituent on the aro-
matic side chain with the 2-fluoro substituent on the sali-
cylic amide moiety (examples 6, 8, 44, 75–81). Within
this series a variety of substituents have been introduced
in the 3-position. Examples 75 (R3 = bromo) and 76
In addition to the (2-benzylcarbamoyl-phenoxy)-acetic
acids a few {2-[(benzothiazol-2-ylmethyl)-carbamoyl]-
phenoxy}-acetic acids were also prepared (Table 2). In
general the in vitro and in vivo activities were compara-
ble to the benzyl series. Example 87 with a 2-fluoro sub-
stituent has similar potency for hALR2 as the
corresponding example with the 2-chloro substituent
(86), but is approximately fourfold more selective. When

M. C. Van Zandt et al. / Bioorg. Med. Chem. 12 (2004) 5661–5675
5667
comparing substituents on the benzthiazole, the 5,6,8-
trifluoro example (13) is about twofold more potent
and selective relative to the 6-trifluoromethyl example
(87). The thioamide analog of example 13, (88) is some-
what less potent and selective, but both have good activ-
ity in the STZ-diabetic rat model.
using ultraviolet illumination, exposure to iodine va-
pors, or dipped in an aqueous potassium permanganate
solution. All starting materials were used without fur-
ther purification. All reactions were carried out under
an atmosphere of dried nitrogen.
4.1.2. General procedures for the synthesis of (2-arylcar-
bamoyl-phenoxy)-acetic acid derived ARIs. (2-Arylcarba-
moyl-phenoxy)-acetic acids 6, 8, 13, 34–56, 61–4, 66, 69,
72–74, and 82–88 (Tables 1 and 2) were prepared from
2,4-difluorobenzoic acid 1 or the corresponding com-
mercially available salicylic acids as outlined in Scheme
2 via the general experimental methods described, for
examples, 6, 8, and 13 below. The substituted 2-amino-
thiophenol hydrochloride salts used to prepare examples
13 and 86–88 as illustrated were prepared by literature
methods32 or purchased commercially.
Because of the obvious importance in positioning the
key elements of the pharmacophore, few modifications
of the amide linker were explored. The N-methyl deriv-
ative, example 85, is approximately 700-fold less active
than the corresponding N–H derivative, example 6. As
expected, this indicates the intramolecular hydrogen
bond between the phenolic ether and carboxylate is crit-
ical for activity. In contrast to the N-methyl derivative, a
significant improvement in oral efficacy was also ob-
tained when the amide was converted to the thioamide.
As evident from examples 6 and 8, oral efficacy im-
proved from 41% for the amide to 89% for the thio
amide at 50mg/kg/d. A similar but less dramatic effect
was also observed with the 2-fluoro-4-bromo derivatives
(examples 39 and 40).
4.1.3. [5-Fluoro-2-(3-nitro-benzylcarbamoyl)-phenoxy]-
acetic acid (6).
4.1.3.1. Step 1: synthesis of 4-fluoro-2-hydroxy-benzoic
acid (2). A solution of 2,4-difluoro-benzoic acid (100g,
0.63mol) in 1,3-dimethyl-2-imidazolidinone (1400mL,
0.45M) was treated with sodium hydroxide (88g,
2.2mol) and heated to 135ꢁC. After stirring for 4h,
the solution was cooled to 0ꢁC and dissolved in water
(100mL). After transferring the solution to a 5L Erlen-
meyer flask, aq HCl (2800mL, 2N) was carefully added.
After filtering off the crude product, the precipitate was
dissolved in ethyl acetate, dried over sodium sulfate and
decolorizing charcoal, and filtered. The solution was
concentrated under reduced pressure and recrystallized
from ethyl acetate and heptane to give 4-fluorosalicylic
acid (two crops, 67g, 68%) as off-white needles. Mp:
188–189ꢁC; Rf 0.26 (20% methanol in dichloro-
methane).
In conclusion, a novel series of (2-arylcarbamoyl-phen-
oxy)-acetic acid derivatives have been identified and
optimized as highly potent and selective inhibitors of al-
dose reductase. In vivo evaluation of example 40, 5-flu-
oro-2-(4-bromo-2-fluoro-benzylthiocarbamoyl)-phen-
oxyacetic acid, in the 4-day STZ-induced diabetic rat
model resulted in lowered nerve sorbitol levels with an
ED50 of 31mg/kg/d (po). This compares favorably with
tolrestat, an important aldose reductase inhibitor
already developed for treatment of chronic diabetic
complications. Continued development of these com-
pounds may result in an important new treatment for
chronic diabetic complications.
4.Experimental section
4.1. Chemistry
4.1.3.2. Step 2: synthesis of 4-fluoro-2-hydroxy-ben-
zoyl chloride (3). A suspension of 4-fluoro-2-hydroxy-
benzoic acid (15g, 96.1mmol) in heptane (190mL) was
treated with thionyl chloride (21mL, 288mmol) in a
dropwise manner over 30min. A drop of N,N-dimethyl-
formamide was added and the solution was heated for
4h at 60ꢁC. The excess thionyl chloride was distilled
off under reduced pressure. The remaining solution
was cooled to room temperature, filtered, and concen-
trated to give 4-fluoro-2-hydroxy-benzoyl chloride as a
pale yellow crystalline solid (14.2g, 85%).
4.1.1. General methods. Melting points were determined
in open capillary tubes on a Thomas–Hoover apparatus
and are uncorrected. Proton nuclear magnetic resonance
(1H NMR) spectra were determined on a Varian Gemini
2000 (300MHz) instrument. Chemical shifts are pro-
vided in parts per million (ppm) downfield from tetra-
methylsilane (internal standard) with coupling
constants in hertz (Hz). Multiplicity is indicated by the
following abbreviations: singlet (s), doublet (d), triplet
(t), quartet (q), multiplet (m), broad (br). Mass spectra
were recorded on a Perkin–Elmer API 100. Elemental
analyses (C, H, N, S, Cl, Br) were performed by Atlantic
Microlab, Inc. (Norcross, Georgia) and are within 4%
of theory unless otherwise noted. Products from all reac-
tions were purified by either flash column chromatogra-
phy or medium pressure liquid chromatography using
silica gel 60 (230–400mesh Kieselgel 60, EM Reagents)
unless otherwise indicated. Thin-layer chromatography
using glass-backed silica plates containing a fluorescent
indicator (0.25mm, Whatman, Merck) was used to
monitor reactions. The chromatograms were visualized
4.1.3.3. Step 3: synthesis of 4-fluoro-2-hydroxy-N-(3-
nitro-benzyl)-benzamide (4). A solution of 4-fluoro-2-hy-
droxy-benzoyl chloride (12.3g, 70.3mmol) in dichloro-
methane (140mL) was cooled to 0ꢁC, and treated with
N,N-diisopropylethylamine (31.0mL, 175mmol) and 3-
nitrobenzylamine hydrochloride (16g, 84.6mmol). After
stirring at room temperature for 24h, the solution was
concentrated in vacuo and diluted with ethyl acetate.
The organic layer was washed successively with aq 2N
HCl and saturated aq NaCl, dried over MgSO4, filtered,
and concentrated. Purification by MPLC (10–100%
ethyl acetate in heptane, 23mL/min, 75min) provided

5668
M. C. Van Zandt et al. / Bioorg. Med. Chem. 12 (2004) 5661–5675
4-fluoro-2-hydroxy-N-(3-nitro-benzyl)-benzamide as a
yellow solid (13.7g, 67%): 1H NMR(DMSO- d6
300MHz) d 12.75 (s, 1H), 9.43 (t, J = 6.0Hz, 1H), 8.11
(ddd, J1 = 8.3Hz, J2 = 2.1Hz, J3 = 1.2Hz, 1H), 8.18 (t,
J = 1.5Hz, 1H), 7.94 (dd, J1 = 8.9Hz, J2 = 6.5Hz, 1H),
7.78 (td, J1 = 7.8Hz, J2 = 1.4Hz, 1H), 7.63 (t,
J = 8.0Hz, 1H), 6.81–6.72 (m, 2H), 4.61 (d, J = 5.7Hz,
2H).
sion of phosphorous pentasulfide (0.77g, 1.73mmol) in
pyridine (6.9mL) was treated with [5-fluoro-2-(3-nitro-
benzylcarbamoyl)-phenoxy]-acetic acid ethyl ester
(example 6, 1.3g, 3.45mmol) and heated to 115ꢁC for
4h. After cooling to room temperature, the mixture
was diluted with water and ethyl acetate. The organic
layer was washed successively with 2N HCl and satu-
rated NaCl, dried over MgSO4, and concentrated. The
resulting dark orange solid was filtered through a short
pad of silica with ethyl acetate and concentrated to give
[5-fluoro-2-(3-nitro-benzylthiocarbamoyl)-phenoxy]-
acetic acid ethyl ester as an orange solid (1.2g, 89%): mp
118ꢁC; Rf 0.43 (40% ethyl acetate in heptane); 1H NMR
(DMSO-d6 300MHz) d 10.71 (br s, 1H), 8.23 (s, 1H),
8.12 (d, J = 8.1Hz, 1H), 7.83 (d, J = 7.8Hz, 1H), 7.71–
7.60 (m, 2H), 7.04 (dd, J1 = 11.1Hz, J2 = 2.4Hz, 1H),
6.87 (dt, J1 = 8.4Hz, J2 = 2.4Hz, 1H), 5.07 (d,
J = 3.3Hz, 2H), 4.89 (s, 2H), 4.13 (q, J = 6.7Hz, 2H),
1.17 (t, J = 5.7Hz, 3H). ESI-LC/MS m/z calcd for
C18H17FN2O5S: 394.1; found 393.0 (M + 1)+. Anal.
(C18H17FN2O5S) C, H, N.
4.1.3.4. Step 4: synthesis of [5-fluoro-2-(3-nitro-ben-
zylcarbamoyl)-phenoxy]-acetic acid ethyl ester (5). A
solution of 4-fluoro-2-hydroxy-N-(3-nitro-benzyl)-benz-
amide (5.00g, 17.2mmol) in acetone (86mL) was treated
with aq 2N K2CO3 (13.0mL, 25.8mmol) and ethyl bro-
moacetate (1.50mL, 9.66mmol), and heated to 50ꢁC for
2h. After cooling to 0ꢁC, the solution was acidified to
pH1 with 2N HCl, diluted with ethyl acetate and
washed with saturated aq NaCl. The organic layer was
dried over MgSO4, filtered, and concentrated. Purifica-
tion by MPLC (10–100% ethyl acetate in heptane,
23mL/min, 75min) gave [5-fluoro-2-(3-nitro-benzylcar-
bamoyl)-phenoxy]-acetic acid ethyl ester as a pale yellow
solid (6.1g, 93%): mp 78–80ꢁC; Rf 0.26 (40% ethyl ace-
4.1.4.2. Step 2: [5-Fluoro-2-(3-nitro-benzylthiocarba-
moyl)-phenoxy]-acetic acid (8). A suspension of [5-fluo-
ro-2-(3-nitro-benzylthiocarbamoyl)-phenoxy]-acetic acid
ethyl ester (4.39g, 11.2mmol) in ethanol (40mL) was
treated with aq 2N NaOH (11mL, 22.4mmol). After
stirring for 4h, the solution was concentrated under
reduced pressure until most of the ethanol was removed,
and the mixture was acidified to pH1 with 2N HCl.
After extracting with ethyl acetate, the organic layer
was washed with saturated aq NaCl, dried over MgSO4,
and concentrated to give [5-fluoro-2-(3-nitro-benz-
ylthiocarbamoyl)-phenoxy]-acetic acid (4.0g, 98%) as
an off-white solid: mp 147–150ꢁC; Rf 0.27 (20% metha-
nol in dichloromethane); 1H NMR(DMSO- d6
300MHz) d 13.23 (s, 1H), 10.79 (t, J = 5.7Hz, 1H),
8.23 (t, J = 1.8Hz, 1H), 8.11 (ddd, J1 = 8.3Hz,
J2 = 2.7Hz, J3 = 1.2Hz, 1H), 7.85 (br d, J = 7.5Hz,
1H), 7.73 (dd, J1 = 8.9Hz, J2 = 7.1Hz, 1H), 7.63 (t,
J = 8.1Hz, 1H), 7.03 (dd, J1 = 11.4Hz, J2 = 2.4Hz,
1H), 6.86 (dt, J1 = 8.4Hz, J2 = 2.4Hz, 1H), 5.07 (d,
J = 5.7Hz, 2H), 4.83 (s, 2H). ESI-LC/MS m/z calcd
for C16H13FN2O5S: 364.4; found 363.0 (Mꢀ1)ꢀ. Anal.
(C16H13FN2O5S) C, H, N.
1
tate in heptane); H NMR(DMSO- d6 300MHz) d 9.01
(t, J = 6.3Hz, 1H), 8.19 (t, J = 1.8Hz, 1H), 8.10 (dd,
J1 = 8.1Hz, J2 = 1.5Hz, 1H), 7.89 (dd, J1 = 8.7Hz,
J2 = 6.9Hz, 1H), 7.79 (d, J = 7.5Hz, 1H), 7.61 (t,
J = 8.1Hz, 1H), 7.11 (dd, J1 = 11.1Hz, J2 = 2.4Hz,
1H), 6.92 (dt, J1 = 8.4Hz, J2 = 2.4Hz, 1H), 5.0 (s, 2H),
4.63 (d, J = 6.3Hz, 2H), 4.15 (q, J = 7.2Hz, 2H) 1.17
(t, J = 6.5Hz, 3H). ESI-LC/MS m/z calcd for
C18H17FN2O6: 376.4; found 377.0 (M + 1)+. Anal.
(C18H17FN2O6) C, H, N.
4.1.3.5. Step 5: synthesis of [5-fluoro-2-(3-nitro-benz-
ylcarbamoyl)-phenoxy]-acetic acid (6). A suspension
of [5-fluoro-2-(3-nitro-benzylcarbamoyl)-phenoxy]-acetic
acid ethyl ester (3.3g, 8.77mmol) in ethanol (40mL) was
treated with aq 2N NaOH (24mL, 47.8mmol). After
stirring for 4h, the solution was concentrated in vacuo
until most of the ethanol was removed, and the mixture
was acidified to pH1 with 2N HCl. After extracting with
ethyl acetate, the organic layer was washed with satu-
rated aq NaCl, dried over MgSO4 and concentrated to
give
[5-fluoro-2-(3-nitro-benzylcarbamoyl)-phenoxy]-
acetic acid as an off-white solid (3.00g, 98%): mp 148–
151ꢁC; Rf 0.39 (20% methanol in dichloromethane);
1H NMR(DMSO- d6 300MHz) d 9.20 (t, J = 6.3Hz,
1H) 8.18 (s, 1H), 8.09 (dd, J1 = 7.2Hz, J2 = 2.7Hz,
1H), 7.90 (dd, J1 = 8.7Hz, J2 = 7.0Hz, 1H), 7.79 (d,
J = 7.5Hz, 1H), 7.61 (t, J = 7.8Hz, 1H), 7.08 (dd,
J1 = 10.8Hz, J2 = 2.1Hz, 1H), 6.91 (dt, J1 = 8.7Hz,
J2 = 2.4Hz, 1H), 4.89 (s, 2H), 4.62 (d, J = 6.3Hz, 2H).
ESI-LC/MS m/z calcd for C16H13FN2O6: 348.3; found
347.0 (Mꢀ1)ꢀ. Anal. (C16H13FN2O6) C, H, N.
4.1.5. {5-Fluoro-2[(4,5,7-trifluoro-benzothiazol-2-ylmeth-
yl)carbamoyl]-phenoxy}-acetic acid (13).
4.1.5.1. Step 1: synthesis of N-cyanomethyl-4-fluoro-2-
hydroxy-benzamide (9). A solution of 4-fluorosalicylic
acid chloride (example 6, 10g, 57.3mmol) in dichloro-
methane (114mL) was treated with N,N-diisopropyl-
ethyl amine (25mL, 143mmol) and acetonitrile
hydrochloride (7.95g, 85.9mmol). After stirring at
35ꢁC for 24h, the solution was concentrated under
reduced pressure, diluted with ethyl acetate, and washed
successively with 2N HCl and saturated aq NaCl.
The resulting solution was dried over MgSO4, filtered,
and concentrated. The resulting solid was suspended in
dichloromethane, filtered, and rinsed with heptane to
give N-cyanomethyl-4-fluoro-2-hydroxy-benzamide as
4.1.4. [5-Fluoro-2-(3-nitro-benzylthiocarbamoyl)-phen-
oxy]-acetic acid (8).
4.1.4.1. Step 1: synthesis of [5-fluoro-2-(3-nitro-benz-
ylthiocarbamoyl)-phenoxy]-acetic acid ethyl ester (7). In a
flame dried flask under a nitrogen atmosphere, a suspen-

M. C. Van Zandt et al. / Bioorg. Med. Chem. 12 (2004) 5661–5675
5669
1
a pale yellow solid (7.20g, 65%): H NMR(DMSO- d6
300MHz) d 12.16 (br s, 1H), 9.17 (t, J = 5.3Hz, 1H),
7.88 (dd, J1 = 8.7Hz, J2 = 6.3Hz, 1H), 6.81–6.73 (m,
2H), 4.32 (d, J = 5.7Hz, 2H).
(dd, J1 = 11.0Hz, J2 = 2.3Hz, 1H), 6.94 (dd,
J1 = 8.4Hz, J2 = 2.4Hz, 1H), 4.94–4.92 (m, 4H). ESI-
LC/MS m/z calcd for C17H10F4N2O4S: 414.3; found
413.0 (Mꢀ1)ꢀ. Anal. (C17H10F4N2O4S) C, H, N.
4.1.5.2. Step 2: synthesis of 4-fluoro-2-hydroxy-N-
(4,5,7,-trifluoro-benzothiazol-2-ylmethyl)-benzamide (11).
A solution of N-cyanomethyl-4-fluoro-2-hydroxy-benz-
amide (2.93g, 13.6mmol) and 2-amino-4,5,7-trifluoro-
thiophenol hydrochloride33,34 (6.64g, 13.6mmol) in
ethanol (27.2mL) was heated to reflux for 24h. After
cooling to room temperature, the mixture was concen-
trated in vacuo, diluted with water and extracted with
ethyl acetate. The organic layer was washed with satu-
rated aq NaCl, dried over MgSO4, filtered, and concen-
trated. Purification by MPLC (10–100% ethyl acetate in
heptane, 23mL/min, 75min) provided 4-fluoro-2-hydr-
oxy-N-(4,5,7,-trifluoro-benzothiazol-2-ylmethyl)-benz-
amide (1.00g, 21%) as a white crystalline solid: 1H
NMR(DMSO- d6 300MHz) d 12.35 (br s, 1H), 9.70 (t,
J = 5.4Hz, 1H), 7.95 (dd, J1 = 8.9Hz, J2 = 6.8Hz, 1H),
7.80–7.70 (m, 1H), 6.83–6.74 (m, 2H), 4.94 (t,
J = 3Hz, 2H).
The 2-fluoro-3-methyl examples 24, 25, 80, and 81 were
prepared from 5-methoxy-2-methyl-4-nitro-phenyl-
amine as illustrated in Scheme 3 and described in detail
below.
4.1.6. [2-(4-Bromo-2-fluoro-benzylcarbamoyl)-5-fluoro-4-
methyl-phenoxy]-acetic acid (25).
4.1.6.1. Step 1: synthesis of 1-fluoro-5-methoxy-2-
methyl-4-nitro-benzene (15). Under an atmosphere of
nitrogen in a Nalgene bottle, a stirring solution of pyr-
idine (17.1mL, 3.2M, ꢀ70ꢁC) was treated dropwise
with HF-pyridine (51.91mL). 5-Methoxy-2-methyl-4-
nitro-phenylamine (10.0g, 54.9mmol) was added and
the mixture was treated with sodium nitrite (6.4g,
92.76mmol). The reaction was allowed to warm to room
temperature over 45min and subsequently heated to
60ꢁC for 2h (until nitrogen evolution stops). After cool-
ing to room temperature, the Nalgene bottle was placed
in an ice bath and 375mL of water was slowly added to
the solution. The resulting orange precipitate was col-
lected by suction filtration and washed with 250mL of
water. Purification of the solid via silica gel flash chro-
matography (30% ethyl acetate in heptane) provided
4.1.5.3. Step 3: synthesis of {5-fluoro-2[(4,5,7-trifluoro-
benzothiazol-2-ylmethyl)carbamoyl]-phenoxy}-acetic acid
ethyl ester (12). A solution of 4-fluoro-2-hydroxy-
N-(4,5,7,-trifluoro-benzothiazol-2-ylmethyl)-benzamide
(1.0g, 2.8mmol) in acetone (14mL) was treated with aq
2N K2CO3 (2.1mL, 4.2mmol) and ethyl bromoacetate
(2mL, 19mmol) and heated to 45ꢁC for 5h. After cool-
ing to room temperature, the solution was acidified to a
pH1 with aq 2N HCl. The resulting solution was diluted
with ethyl acetate and washed with saturated NaCl. The
organic layer was dried over MgSO4, filtered, and con-
centrated. Purification by MPLC (10–100% ethyl acetate
in heptane 23mL/min, 75min) provided {5-fluoro-
2[(4,5,7-trifluoro-benzothiazol-2-ylmethyl)carbamoyl]-
phenoxy}-acetic acid ethyl ester (1.0g, 81%) as a white
1-fluoro-5-methoxy-2-methyl-4-nitro-benzene
76%) as a light orange solid: mp 71–74ꢁC; Rf 0.56
(30% ethyl acetate in heptane); 1H NMR(CDCl
(7.69g,
,
3
300MHz)
d 7.82 (d, J = 7.5Hz, 1H), 6.75 (d,
J = 10.5Hz, 1H), 3.93 (s, 3H), 2.25 (d, J = 2.1Hz, 3H);
ESI-LC/MS m/z calcd for C8H8FNO3: 185.1; found
186.0 (M + 1)+. Anal. (C8H8FNO3) C, H, N.
4.1.6.2. Step 2: synthesis of 4-fluoro-2-methoxy-5-
methyl-phenylamine (16). A solution of 1-fluoro-5-meth-
oxy-2-methyl-4-nitro-benzene (5.5g, 29.3mmol) and
10% Pd–C (1.56g) in ethanol (300mL, 0.1M) was
hydrogenated at 1atm. After stirring overnight, the
solution was flushed through a pad of silica gel using
400mL of ethanol as the eluant. The filtrate was concen-
trated under reduced pressure to afford 4-fluoro-2-meth-
oxy-5-methyl-phenylamine (4.5g, 99%) as a pale purple
solid: mp 108–110ꢁC; Rf 0.35 (30% ethyl acetate in hep-
tane); 1H NMR(DMSO- d6, 300MHz) d 6.62 (d,
J = 11.4Hz, 1H), 6.43 (d, J = 7.8Hz, 1H), 3.70 (s, 3H),
2.02 (d, J = 1.8Hz, 3H); ESI-LC/MS m/z calcd for
C8H10FNO: 155.1; found 156.0 (M + 1)+. Anal.
(C8H10FNO) C, H, N.
1
crystalline solid: H NMR(DMSO- d6 300MHz) d 9.32
(t, J = 5.9Hz, 1H), 7.94 (dd, J1 = 9.0Hz, J2 = 7.2Hz,
1H), 7.80–7.70 (m, 1H), 7.13 (dd, J1 = 11.1Hz,
J2 = 2.4Hz, 1H), 6.95 (dt, J1 = 8.4Hz, J2 = 2.5Hz,
1H), 5.02 (s, 2H), 4.94 (d, J = 6Hz, 2H), 4.17 (q,
J = 7.2Hz, 2H), 1.18 (t, J = 7.1Hz, 3H).
4.1.5.4. Step 4: synthesis of {5-fluoro-2[(4,5,7-trifluoro-
benzothiazol-2-ylmethyl)carbamoyl]-phenoxy}-acetic acid
(13). A suspension of {5-fluoro-2[(4,5,7-trifluoro-ben-
zothiazol-2-ylmethyl)carbamoyl]-phenoxy}-acetic acid
ethyl ester (1g, 2.3mmol) in ethanol (11mL) was treated
with aq 2N NaOH (6.8mL, 14mmol) and stirred at
room temperature. After 1h, the solution was concen-
trated in vacuo and acidified to pH1 with aq 2N HCl.
The resulting solution was diluted with ethyl acetate
and washed with saturated NaCl. The organic layer
was dried over MgSO4, filtered, and concentrated to give
{5-fluoro-2[(4,5,7-trifluoro-benzothiazol-2-ylmethyl)car-
bamoyl]-phenoxy}-acetic acid. (0.68g, 73%) as a white
crystalline solid. Mp 172–174ꢁC; Rf 0.38 (20% methanol
4.1.6.3. Step 3: synthesis of 1-bromo-4-fluoro-2-meth-
oxy-5-methyl-benzene (17). A suspension of 4-fluoro-2-
methoxy-5-methyl-phenylamine (25.75g, 0.17mol) in
180mL of HBr (48%, 0.9M) in an ice bath was carefully
treated with an aq solution of sodium nitrite (12.6g,
0.18mol, 3.6M). Throughout the addition, a brown
gas was evolved and the temperature of the reaction
was monitored so that the internal temperature did
not raise above 10ꢁC. In parallel, a suspension of CuBr
1
in dichloromethane); H NMR(DMSO- d6 300MHz) d
13.25 (br s, 1H), 9.49 (t, J = 6Hz, 1H), 7.95 (dd,
J1 = 9Hz, J2 = 7.2Hz, 1H), 7.78–7.69 (m, 1H), 7.11

5670
M. C. Van Zandt et al. / Bioorg. Med. Chem. 12 (2004) 5661–5675
(13.1g, 0.09mol) in 6.5mL of HBr (48%, 13.9M) was
heated to 110ꢁC and carefully treated with the first solu-
tion (over 20min). The combined reaction mixture was
heated for 2.5h at 110ꢁC, cooled to room temperature,
diluted with ethyl acetate, and treated with aq sulfuric
acid (50% v/v). The organic layer was separated and
washed successively with water, sulfuric acid (50% v/v),
water, 1.25M NaOH, water, and saturated aq NaCl.
The organic layer was dried over Na2SO4, filtered, and
concentrated under reduced pressure. Purification of
the resulting brown oil via silica gel flash chromatogra-
phy (5% ethyl acetate in heptane) provided 1-bromo-4-
fluoro-2-methoxy-5-methyl-benzene (17.1g, 47%) as a
colorless liquid: Rf 0.70 (30% ethyl acetate in heptane);
1H NMR(CDCl 3, 300MHz) d 7.34 (d, J = 8.4Hz,
1H), 6.61 (d, J = 10.8Hz, 1H), 3.85 (s, 3H), 2.18 (d,
J = 1.8Hz, 3H); Anal. (C8H8BrFO) C, H, Br.
4.1.6.6. Step 6: synthesis of N-(4-bromo-2-fluoro-
benzyl)-4-fluoro-2-methoxy-5-methyl-benzamide (20).
Under a nitrogen atmosphere, a solution of 4-fluoro-2-
methoxy-5-methyl-benzoic acid (1.19g, 6.47mmol) in
dichloromethane (16mL, 0.5M) was treated with oxalyl
chloride (1.7mL, 19.4mmol) and a drop of DMF at
0ꢁC. The mixture was allowed to gradually warm to
room temperature over 30min and subsequently concen-
trated. The resulting yellow powder was dissolved in
dichloromethane (16mL, 0.5M), cooled to 0ꢁC and
treated with diisopropylethylamine (2.8mL, 16.2mmol)
and 4-bromo-2-fluorobenzylamine hydrochloride salt
(2.34g, 9.71mmol). The mixture was warmed to room
temperature and stirred for 21h. The resulting solution
was washed successively with 1N HCl (3 · 50mL) and
saturated aq NaCl (50mL). The organic layer was dried
over Na2SO4, filtered, and concentrated to afford N-(4-
bromo-2-fluoro-benzyl)-4-fluoro-2-methoxy-5-methyl-
benzamide (2.33g, 97%) as a brown oil, which was used
4.1.6.4. Step 4: synthesis of 4-fluoro-2-methoxy-5-
methyl-benzonitrile (18). A solution of 1-bromo-4-flu-
oro-2-methoxy-5-methyl-benzene (5.0g, 22.8mmol) in
DMF (100mL, 0.2M) was treated with CuCN (4.7g,
52.5mmol) and heated to 160ꢁC for 20h. After cooling
to room temperature, the solution was poured into a 2L
Erlenmeyer flask. Ethyl acetate (400mL), saturated aq
LiCl (100mL), 1N HCl (100mL), iron(III) chloride
hexahydrate (11g), and 15mL of concd HCl were added
to the solution. The resulting green mixture was heated
at 70ꢁC for 2h (until emulsion dissipated). After cooling
to room temperature, the mixture was poured into a
separatory funnel and extracted with ethyl acetate
(3 · 200mL). The combined organic extracts were
washed successively with 1N HCl (200mL), saturated
aq LiCl (2 · 200mL), and saturated aq NaCl (100mL).
The organic layer was dried over Na2SO4, filtered, and
concentrated under reduced pressure. The resulting
powder was dissolved in ethyl acetate (50mL) and
washed with saturated aq LiCl (3 · 30mL), dried over
Na2SO4, filtered, and concentrated under reduced pres-
sure to afford 4-fluoro-2-methoxy-5-methyl-benzonitrile
without further purification: Rf 0.40 (30% ethyl acetate
1
in heptane); H NMR(CDCl , 300MHz) d 8.17 (br s,
3
1H), 8.06 (d, J = 9.3Hz, 1H), 7.33–7.22 (m, 3H), 6.65
(d, J = 11.1Hz, 1H), 4.63 (d, J = 6.3Hz, 2H), 3.92 (s,
3H), 2.23 (br d, J = 1.8Hz, 3H).
4.1.6.7. Step 7: synthesis of N-(4-bromo-2-fluoro-
benzyl)-4-fluoro-2-hydroxy-5-methyl-benzamide (21). A
stirring suspension of N-(4-bromo-2-fluoro-benzyl)4-flu-
oro-2-methoxy-5-methyl-benzamide (2.32g, 6.51mmol)
in a 25% solution of HBr in acetic acid (60mL,
0.11M) was heated to 120ꢁC for 3.5h. After cooling
to room temperature, saturated aq NaCl (50mL) was
added and the solution was extracted with ethyl acetate
(3 · 50mL). The combined organic extracts were dried
over Na2SO4, filtered, and concentrated. The resulting
orange powder was purified via silica gel flash chroma-
tography (30% ethyl acetate in heptane) to give N-(4-
bromo-2-fluoro-benzyl)-4-fluoro-2-hydroxy-5-methyl-
benzamide (1.69, 73%) as a white powder: mp 149–
150ꢁC; Rf 0.51 (30% ethyl acetate in heptane); 1H
(3.25g, 86%) as an off-white powder: mp 99–101ꢁC; Rf
NMR(CDCl , 300MHz) d 12.20 (d, J = 1.5Hz, 1H),
3
1
0.53 (30% ethyl acetate in heptane); H NMR(CDCl ,
7.30–7.29 (m, 2H), 7.26 (s, 1H), 7.14 (d, J = 8.1Hz,
1H), 6.64 (d, J = 10.8Hz, 1H), 6.48 (br s, 1H), 4.62 (d,
J = 5.7Hz, 2H), 2.19 (s, 3H); ESI-LC/MS m/z calcd
for C15H12BrF2NO2: 355.0; found 354.0 (Mꢀ1)ꢀ. Anal.
(C15H12BrF2NO2) C, H, N.
3
300MHz) d 7.39 (d, J = 8.1Hz, 1H), 6.65 (d,
J = 11.1Hz, 1H), 3.89 (d, J = 0.9Hz, 3H), 2.21 (s, 3H).
Anal. (C9H8FNO) C, H, N.
4.1.6.5. Step 5: synthesis of 4-fluoro-2-methoxy-5-
methyl-benzoic acid (19). A suspension of 4-fluoro-2-
methoxy-5-methyl-benzonitrile (3.0g, 18.2mmol) in 2N
NaOH (300mL, 0.06M) was heated to 90ꢁC for 19h.
The remaining starting material was recovered by filtra-
tion (0.86g). The aqueous solution was acidified to pH1
(concentrated HCl) and extracted with ethyl acetate
(2 · 250mL). The combined organic extracts were dried
over Na2SO4, filtered, and concentrated to afford 4-flu-
oro-2-methoxy-5-methyl-benzoic acid as a white powder
(2.28g, 96% based on recovered starting material): mp
4.1.6.8. Step 8: synthesis of [2-(4-bromo-2-fluoro-
benzylcarbamoyl)-5-fluoro-4-methyl-phenoxy]-acetic acid
ethyl ester (22). A stirring solution of N-(4-bromo-2-flu-
oro-benzyl)-4-fluoro-2-hydroxy-5-methyl-benzamide
(1.69g, 4.76mmol) in acetone (24mL, 0.2M) was treated
with a second solution of aq K2CO3 (3.6mL, 2M,
7.12mmol) and ethyl bromoacetate (0.63mL, 5.69
mmol) in acetone (24mL, 0.2M). After heating to
50ꢁC for 2.5h, the solution was cooled to room temper-
ature, concentrated, and acidified to pH1 with 2N HCl.
The resulting solution was diluted with ethyl acetate
(100mL) and washed with saturated aq NaCl (50mL).
The organic layer was dried over Na2SO4, filtered, and
concentrated to provide [2-(4-bromo-2-fluoro-benzyl-
carbamoyl)-5-fluoro-4-methyl-phenoxy]-acetic acid ethyl
ester (1.95, 93%) as a white solid: mp 128–129ꢁC; Rf
1
125–127ꢁC; Rf 0.15 (40% ethyl acetate in heptane); H
NMR(CDCl 3, 300MHz) d 10.20 (s, 1H), 8.05 (d,
J = 9.0Hz, 1H), 6.74 (d, J = 10.8Hz, 1H), 4.04 (s, 3H),
2.25 (d, J = 1.8Hz, 3H); ESI-LC/MS m/z calcd for
C9H9FO3: 184.1; found 185.0 (M + 1)+. Anal.
(C9H9F03) C, H, N.

M. C. Van Zandt et al. / Bioorg. Med. Chem. 12 (2004) 5661–5675
5671
1
0.42 (30% ethyl acetate in heptane); H NMR(CDCl ,
thiocarbamoyl)-5-fluoro-4-methyl-phenoxy]-acetic acid
ethyl ester (0.79g, 1.53mmol) in ethanol (7.6mL,
0.2M) was treated with aqueous NaOH (2N, 7.4mL,
9.15mmol). After stirring for 4h, the solution was
concentrated under reduced pressure until most of the
ethanol was removed. The remaining solution was acid-
ified to pH1 with 2N HCl and extracted with ethyl
acetate. The organic layer was washed with saturated
aq NaCl, dried over MgSO4 and concentrated to give
[2-(4-bromo-2-fluoro-benzylthiocarbamoyl)-5-fluoro-4-
methyl-phenoxy]-acetic acid (0.65g, 98%) as a yellow
solid: mp 162–164ꢁC; Rf 0.41 (20% methanol in dichlo-
3
300MHz) d 8.79 (br t, J = 4.5Hz, 1H), 8.10 (d,
J = 9.0Hz, 1H), 7.34 (t, J = 8.3Hz, 1H), 7.26–7.23 (m,
1H), 7.21 (t, J = 2.3Hz, 1H), 6.53 (d, J = 10.5Hz, 1H),
4.66 (d, J = 3.9Hz, 1H), 4.65 (s, 2H), 4.28 (q,
J = 7.2Hz, 2H), 2.23 (br d, J = 1.5Hz, 3H), 1.30 (t,
J = 7.2Hz, 3H); Anal. (C19H18BrF2NO4) C, H, N.
4.1.6.9. Step 9: synthesis of [2-(4-bromo-2-fluoro-
benzylcarbamoyl)-5-fluoro-4-methyl-phenoxy]-acetic acid
(25). A solution of [2-(4-bromo-2-fluoro-benzylcarba-
moyl)-5-fluoro-4-methyl-phenoxy]-acetic acid ethyl ester
(0.72g, 1.62mmol) in ethanol (8.1mL, 0.2M) was cooled
to 0ꢁC and treated with aq NaOH (1.25M, 7.8mL,
9.73mmol). After the addition was complete, the mix-
ture was warmed to room temperature and stirred an
additional 2h. The resulting solution was concentrated
under reduced pressure, acidified with 2N HCl (10mL)
and diluted with ethyl acetate. The separated organic
layer was washed with saturated aq NaCl, dried over
Na2SO4, filtered, and concentrated to afford [2-(4-bro-
mo-2-fluoro-benzylcarbamoyl)-5-fluoro-4-methyl-phen-
oxy]-acetic acid (0.66g, 98%) as a white solid: mp
169ꢁC; Rf 0.22 (20% methanol in dichloromethane);
1
romethane); H NMR(DMSO- d6, 300MHz) d 10.64 (t,
J = 5.0Hz, 1H), 7.61 (d, J = 9.0Hz, 1H), 7.52 (dd,
J1 = 9.6Hz, J2 = 1.7Hz, 1H), 7.45–7.36 (m, 2H), 7.00
(d, J = 11.4Hz, 1H), 4.89 (br d, J = 5.1Hz, 2H),
4.78 (s, 2H), 2.15 (d, J = 1.2Hz, 3H); ESI-LC/MS m/z
calcd for C17H14BrF2NO3S: 428.98; found 428.0
(Mꢀ1)ꢀ; Anal. (C17H14BrF2NO3S) C, H, N, S.
Examples 75–79 were prepared from salicylic acid deriv-
atives 26, 29, 31, and 33 using the general method de-
scribed, for examples, 24 and 25. Salicylic acid
derivatives 26, 29, 31, and 33 were prepared using the
method illustrated in Scheme 4 and described in detail
below.
1H NMR(DMSO- d6, 300MHz)
d 9.00 (br t,
J = 5.3Hz, 1H), 7.76 (d, J = 9.6Hz, 1H), 7.49 (d,
J = 9.6Hz, 1H), 7.39–7.33 (m, 2H), 7.04 (dd,
J1 = 11.4Hz, J2 = 1.4Hz, 1H), 4.84 (d, J = 1.8Hz, 2H),
4.48 (br s, 2H), 2.17 (s, 3H); ESI-LC/MS m/z calcd for
C17H14BrF2NO4: 413.0; found 412.0 (Mꢀ1)ꢀ. Anal.
(C17H14BrF2NO4) C, H, N.
4.1.8. 5-Bromo-4-fluoro-2-hydroxy-benzoic acid (26). A
solution of 4-fluoro-2-hydroxy-benzoic acid (5.58g,
35.7mmol) in dimethylformamide (72mL, 0.5M) was
treated with N-bromosuccinimide (7.08g, 39.3mmol)
and stirred for 24h at room temperature. The resulting
mixture was diluted with ethyl acetate (300mL) and
washed successively with water (3 · 330mL) and
saturated aq LiCl (4 · 200mL). The combined organic
layers were dried over Na2SO4, filtered, and concen-
trated to give 5-bromo-4-fluoro-2-hydroxy-benzoic acid
(8.1g, 96%) as an off-white powder. Rf 0.32 (20% meth-
anol in dichloromethane); 1H NMR(DMSO- d6,
4.1.7. [2-(4-Bromo-2-fluoro-benzylthiocarbamoyl)-5-fluo-
ro-4-methyl-phenoxy]-acetic acid (24).
4.1.7.1. Step 1: synthesis of [2-(4-bromo-2-fluoro-
benzylthiocarbamoyl)-5-fluoro-4-methyl-phenoxy]-acetic
acid ethyl ester (23). A solution of [2-(4-bromo-2-fluoro-
benzylcarbamoyl)-5-fluoro-4-methyl-phenoxy]-acetic
acid ethyl ester (example 25, 0.816g, 1.83mmol) in pyr-
idine (3.7mL, 0.5M) was treated with phosphorous pen-
tasulfide (0.41g, 0.92mmol) and heated to 115ꢁC for 3h.
The reaction was cooled to room temperature, diluted
with ethyl acetate and successively washed with 1M
HCl and saturated aq NaCl. The organic layer was dried
over Na2SO4, filtered, and concentrated under reduced
pressure. The resulting brown oil was dissolved in a min-
imal amount of methylene chloride and flushed through
a plug of silica using 40% ethyl acetate in heptane as the
eluant. The filtrate was concentrated to afford [2-(4-
bromo-2-fluoro-benzylthiocarbamoyl)-5-fluoro-4-meth-
yl-phenoxy]-acetic acid ethyl ester (0.75g, 89%) as a yel-
300MHz) d 8.00 (d, J = 8.1Hz, 1H), 7.05 (d,
J = 10.5Hz, 1H).
4.1.9. 5-Bromo-4-fluoro-2-methoxy-benzoic acid methyl
ester (27). A solution of 5-bromo-4-fluoro-2-hydroxy-
benzoic acid (15.0g, 63.8mmol) and anhydrous K2CO3
(19g, 140mmol) in acetone (128mL, 0.5M) was treated
with iodomethane (24mL, 380mmol) and heated to
60ꢁC for 4h. After cooling to room temperature, the
solution was concentrated and purified by MPLC (10–
100% ethyl acetate in heptane, 23mL/min, 70min) to
give 5-bromo-4-fluoro-2-methoxy-benzoic acid methyl
ester (13.52g, 80%) as a white solid: mp 73–77ꢁC; Rf
1
0.56 (40% ethyl acetate in heptane); H NMR(CDCl ,
3
low solid: mp 98–100ꢁC; Rf 0.45 (30% ethyl acetate in
300MHz) d 8.05 (d, J = 7.8Hz, 1H), 6.76 (d,
J = 10.5Hz, 1H), 3.89 (s, 3H), 3.88 (s, 3H). Anal.
(C9H6BrFO3) C, H, Br.
1
heptane); H NMR(CDCl , 300MHz) d 10.15 (br s,
3
1H), 8.26 (d, J = 8.4Hz, 1H), 7.38 (t, J = 8.4Hz, 1H),
7.26–7.23 (m, 2H), 6.53 (d, J = 10.8Hz, 1H), 5.08 (d,
J = 5.4Hz, 2H), 4.67 (s, 2H), 4.20 (q, J = 7.2Hz, 2H),
2.23 (s, 3H), 1.27 (t, J = 6.8Hz, 3H).
4.1.10.
4-Fluoro-2-methoxy-5-morpholin-4-yl-benzoic
acid methyl ester (28). A solution of oven-dried cesium
carbonate (4.33g, 13.3mmol), palladium acetate
(85.3mg, 0.380mmol), R-(+)-2,20-bis(diphenylphosph-
ino)-1,10-binaphthyl (0.355g, 0.570mmol) and 5-bro-
mo-4-fluoro-2-methoxy-benzoic acid methyl ester
4.1.7.2. Step 2: synthesis of [2-(4-bromo-2-fluoro-
benzylthiocarbamoyl)-5-fluoro-4-methyl-phenoxy]-acetic
acid (24). A solution of [2-(4-bromo-2-fluoro-benzyl-

5672
M. C. Van Zandt et al. / Bioorg. Med. Chem. 12 (2004) 5661–5675
(2.50g, 9.50mmol) in toluene (0.76mL) was treated with
morpholine (0.99mL, 11.4mmol) and heated to
100ꢁC for 24h. After cooling to room temperature the
solution was diluted with ether, filtered, and concen-
trated. Purification by MPLC (10–50% ethyl acetate in
heptane, 23mL/min, 35min) provided 4-fluoro-2-meth-
oxy-5-morpholin-4-yl-benzoic acid methyl ester (1.20g,
47%): 1H NMR(DMSO- d6 300MHz) d 7.33 (d,
J = 9.9Hz, 1H), 7.08 (d, J = 14.4Hz, 1H), 3.76 (s, 3H),
3.75 (s, 3H), 3.70 (t, J = 4.7Hz, 4H), 2.90 (t,
J = 4.5Hz, 4H).
afford 5-cyano-4-fluoro-2-methoxy-benzoic acid (1.9g,
70%) as a white solid. Rf = 0.34 (20% methanol in
1
dichloromethane); H NMR(CDCl , 300MHz) d 8.13
3
(d, J = 8.1Hz, 1H), 7.36 (d, J = 12.0Hz, 1H), 3.90 (s,
3H).
4.1.14. 6-Fluoro-4-methoxy-biphenyl-3-carboxylic acid
methyl ester (32). A solution of 5-bromo-4-fluoro-2-
methoxy-benzoic acid methyl ester (2.0g, 7.6mmol),
anhydrous K2CO3 (2.1g, 15mmol), and phenylboronic
acid (3.7g, 30.4mmol) in degassed toluene (15.2mL,
0.5M) was treated with Pd(PPh3)4 (0.88g, 0.76mmol)
and heated to 110ꢁC for 3.5h. The resulting solution
was cooled to 0ꢁC and carefully treated with H2O2
(30%, 10mL). After warming to room temperature
and stirring for an additional hour, the mixture was di-
luted with ether and washed successively with 2N HCl
and saturated aq NaCl. The organic layer was dried over
Na2SO4, filtered, and concentrated. Purification by
MPLC (10–100% ethyl acetate in heptane, 23mL/min,
75min) provided 6-fluoro-4-methoxy-biphenyl-3-car-
boxylic acid methyl ester (1.8g, 91%) as a pale yellow
4.1.11.
4-Fluoro-2-methoxy-5-morpholin-4-yl-benzoic
acid (29). A suspension of 4-fluoro-2-methoxy-5-mor-
pholin-4-yl-benzoic acid methyl ester (1.20g, 4.46mmol)
in ethanol (22.0mL) was treated with aq 2N NaOH
(13mL, 26.7mmol) and stirred at room temperature
for 2h. The resulting solution was concentrated until
most of the ethanol was removed and subsequently acid-
ified to pH1 with aq 2N HCl. After extracting with ethyl
acetate, the organic layer was washed with saturated aq
NaCl, dried over MgSO4 and concentrated to give 4-flu-
oro-2-methoxy-5-morpholin-4-yl-benzoic acid as a pale
yellow foam (0.90g, 79%): 1H NMR(DMSO- d6
300MHz) d 12.62 (s, 1H), 7.34 (d, J = 10.5Hz, 1H),
7.04 (d, J = 14.7Hz, 1H), 3.76 (s, 3H), 3.70 (t,
J = 4.7Hz, 4H), 2.90 (t, J = 4.7Hz, 4H).
1
solid: Rf 0.5 (40% ethyl acetate in heptane); H NMR
(CDCl3, 300MHz) d 8.00 (d, J = 9.0Hz, 1H), 7.53–
7.33 (m, 5H), 6.79 (d, J = 12.6Hz, 1H), 3.94 (s, 3H),
3.89 (s, 3H).
4.1.15. 6-Fluoro-4-methoxy-biphenyl-3-carboxylic acid
(33). A solution of 6-fluoro-4-methoxy-biphenyl-3-carb-
oxylic acid methyl ester (0.5g, 2.3mmol) in dioxane
(8mL, 0.3M) was treated with 2N NaOH (6.0mL,
12mmol) and stirred for 1h at room temperature. The
resulting reaction mixture was partially concentrated,
acidified with 2N HCl and extracted with ether. The or-
ganic layer dried over Na2SO4, filtered and concentrated
to afford a 6-fluoro-4-methoxy-biphenyl-3-carboxylic
4.1.12. 5-Cyano-4-fluoro-2-methoxy-benzoic acid methyl
ester (30). A solution of 5-bromo-4-fluoro-2-methoxy-
benzoic acid methyl ester (5.0g, 10.0mmol) in DMF
(38mL, 0.5M) was treated with CuCN (3.92g,
43.7mmol) and heated at 150ꢁC for 24h. After cooling
to room temperature, the solution was transferred to
an Erlenmeyer flask and diluted with ethyl acetate
(400mL), saturated aq LiCl (100mL), 1N HCl
(100mL), 11g of iron(III) chloride hexahydrate, and
15mL of concd HCl. The resulting green mixture
was heated to 70ꢁC until the emulsion dissipated (2h).
After cooling to room temperature, the mixture was
extracted with ethyl acetate (600mL) and successively
washed with 1N HCl (200mL), saturated aq LiCl
(2 · 200mL) and saturated aq NaCl (100mL). The
organic layer was dried over Na2SO4, filtered, and
concentrated under reduced pressure. Purification by
MPLC (10–100% ethyl acetate in heptane, 23mL/min,
75min) provided 5-cyano-4-fluoro-2-methoxy-benzoic
acid methyl ester (2.98g, 75%) as a white crystalline
acid (0.55g, 96%) as a pale yellow powder: Rf 0.18
(5% methanol in dichloromethane); H NMR(CDCl ,
1
3
300MHz) d 8.35 (d, J = 9.0Hz, 1H), 7.55–7.52 (m,
2H), 7.48–7.38 (m, 3H), 6.89 (d, J = 11.7Hz, 1H), 4.12
(s, 3H).
Examples 58, 59, and 60 were prepared in a manner
analogous to examples 6 and 8 except 2,4,5-trifluoro-
benzoic acid and 2,4,6-trifluoro-benzoic acid were used
in place of 2,4-difluoro-benzoic acid.
Examples 83 and 84 were prepared from 2-hydroxy-5-
trifluoromethoxy-benzoic acid using the method de-
scribed for examples 6 and 8.
solid: 1H NMR(CDCl
, 300MHz) d 8.15 (d,
J = 7.5Hz, 1H), 6.80 (d, J = 11.1Hz, 1H), 3.97 (s, 3H),
3
3.90 (s, 3H).
4.1.16. 2-Hydroxy-5-trifluoromethoxy-benzoic acid. Aq
NaOH (35mL, 5.8M, 204mmol) was combined with
an aq silver nitrate solution (35mL water, 2.9M,
102mmol) and stirred. The resulting suspension was
cooled to 0ꢁC and treated with 2-hydroxy-5-trifluoro-
methoxy-benzaldehyde (10.0g, 48.5mmol) in several
portions. After the addition was complete, the reaction
was stirred an additional 10min and filtered. The precip-
itate was washed with hot water and the combined
washings were acidified to pH1 with concd HCl. The
resulting precipitate was collected by vacuum filtration,
dissolved in ethyl acetate, and washed with saturated aq
4.1.13. 5-Cyano-4-fluoro-2-methoxy-benzoic acid (31). A
stirring suspension of 5-cyano-4-fluoro-2-methoxy-benz-
oic acid methyl ester (2.98g, 14.25mmol) in ethanol
(30mL, 0.5 M) was treated with 1.25M NaOH
(68mL, 85.5mmol) and stirred for 10min (solution
becomes clear). The reaction mixture was concentrated
to about 10% of its original volume and acidified to
pH1 with 2N HCl. The resulting white precipitate was
collected by suction filtration, dissolved in dioxane,
and washed with aq saturated NaCl. The organic layer
was dried over Na2SO4, filtered, and concentrated to

M. C. Van Zandt et al. / Bioorg. Med. Chem. 12 (2004) 5661–5675
5673
NaCl. The solution was dried over Na2SO4, filtered, and
concentrated to give 2-hydroxy-5-trifluoromethoxy-ben-
zoic acid (9.8g, 91%) as a white solid: Rf 0.38 (20%
LM-485 mouse/rat, Madison, WI). Water is supplied
ad libitum. After a 24h fast, 40 mg/kg STZ (S01230
Sigma, St. Louis, MO) prepared in 0.03M citrate buffer,
pH4.5 is administered ip. Control animals receive citrate
buffer.
methanol in dichloromethane); 1H NMR(CDCl
,
3
300MHz) d 10.30 (br s, 1H), 7.79 (d, J = 3.0Hz, 1H),
7.41 (dd, J1 = 9.3Hz, J2 = 3.0Hz, 1H), 7.05 (d,
J = 9.0Hz, 1H).
Four days after receiving STZ, tail blood glucose is
measured using a One Touch II blood glucose meter
(Lifescan, Inc., Johnson & Johnson, Milpitas, CA).
Diabetic rats whose blood glucose levels are P 300mg/
dl are randomized according to blood glucose levels into
diabetic control and diabetic-treatment groups. Each
study includes 5 nondiabetic controls, 10 diabetic
controls, and 7 diabetic rats for each drug-treated
group.
4.2. Pharmacology
4.2.1. Enzyme preparation. Recombinant human aldose
reductase (hALR2) was overexpressed in E. coli using
a T7 based expression system and purified using a metal
affinity column.35 The enzyme was precipitated in 3.5M
ammonium sulfate, 10mM Tris–HCl buffer and 1mM
DTT. The enzyme was centrifuged at 10,000rpm for
10min at 4ꢁC and the pellet was resuspended to an
approximate concentration of 10–15mg protein/mL in
(NH4)2SO4 in NaH2PO4, pH6.6. Protein content was
determined using Coomassie Protein Assay reagent
(Pierce, Rockford, IL). Aliquots were made and the en-
zyme was stored at ꢀ40ꢁC.
Beginning on day 4 after induction of diabetes with
STZ, the compound is administered for five consecutive
days po suspended in a vehicle of 2% Tween 80 in saline.
Nondiabetic and diabetic controls receive vehicle during
this period. After the initial dose of compound, animals
continue to receive the standard diet for the remainder
of the study. After receiving the final dose of compound,
the food is removed. Four hours after the final dose is
administered, the rats are euthanized by asphyxiation
with CO2 and sciatic nerves are removed, weighed, fro-
zen on porcelain plates on dry ice, and stored at ꢀ80ꢁC
for sorbitol analysis.
Recombinant human aldehyde reductase (hALR1) sup-
plied by Dr. K. Gabbay (Baylor College of Medicine,
Houston, TX) was prepared as previously described.36
The enzyme was stored in 5mM sodium phosphate,
pH7.4, and 0.1mM EDTA at 4ꢁC. Protein content
was determined using Coomassie Protein Assay reagent
(Pierce, Rockford, IL).
4.2.4. Tissue sorbitol analysis. Sciatic nerve sorbitol lev-
els are measured by a standard enzymatic method that
uses sorbitol dehydrogenase.40,41 A D-sorbitol kit pur-
chased from Boehringer Mannheim (670 057 Indianap-
olis, IN) was adapted to a microplate format.
4.2.2. Enzymatic assays. The method to examine the
inhibition of aldose reductase (ALR2) was similar
to those methods used by numerous other laborato-
ries.37–39 Enzymatic activity was measured by monitor-
ing the disappearance rate of NADPH. Briefly,
compound or solvent is placed in the microplate well
to which buffer is added and mixed for 1min. Enzyme
is added to the well and incubated at 37ꢁC and shaken
for 10min. DL-Glyceraldehyde and NADPH are added
to initiate the reaction. The microplate is placed in the
platereader for 30min at 340nm and read at 2min inter-
vals. The reaction contents in a final volume of 300lL
are 6.6% (NH4)2SO4, 33mM NaH2PO4 (pH6.6),
0.11mM NADPH, 4.7mM DL-glyceraldehyde, 0.59lg
enzyme, 1% DMSO, and compound. Percent inhibition
is calculated based on the enzyme activity in the pres-
ence or absence of compound. The IC50 is calculated
using SAS release 6.10 and a four-parameter logistic
regression of data.
Tissues are homogenized in ice cold 0.2% perchloric acid
in methanol using a DUALL tissue grinder and centri-
fuged at 10,000rpm for 10min at 4ꢁC. A sampling of
the supernatant is transferred to a microplate. A reac-
tion mixture combining buffer, diaphorase, NAD+,
and iodonitrotetrazolium chloride (INT) is added to
each sample followed by an initial absorbance reading
at 490nm. Sorbitol dehydrogenase (SDH) is added to
initiate the reaction and the plate is mixed for 90s on
an orbital shaker. Plates are incubated at room temper-
ature in the dark for 30min and the final absorbance
reading at 490nm is made. The reaction contents in a
final volume of 255 lL at pH8.7 are 14.1mM potassium
phosphate, 94mM triethanolamine, 0.94% Triton
X-100, 0.16U diaphorase, 1.66mM NAD, 0.034mM
INT and 0.82U SDH.
The inhibition of aldehyde reductase (ALR1) was as-
sessed using a standard spectrophotometric assay.36
The procedure is similar to the method described above
for determining ALR2 inhibition. The reaction contents
in a final volume of 300lL are 33mM NaH2PO4
(pH6.6), 0.11mM NADPH, 4.7mM DL-glyceraldehyde,
0.28lg enzyme, 1% DMSO, and compound.
Sciatic nerve sorbitol content is calculated as nmol sor-
bitol/mg wet tissue weight by taking the difference in the
final and initial absorbance readings and comparing to a
sorbitol standard curve. % Inhibition is calculated by
comparing treatment groups to both the nondiabetic
and diabetic control groups within the study as de-
scribed by the equation: (t ꢀ n)/(d ꢀ n) where n = nondi-
abetic control, d = diabetic control, and t = drug-treated
diabetic control. The ED50 was calculated using Sigma
plot, version 4.0, using a Hill slope four-parameter logis-
tic equation.
4.2.3. STZ-induced diabetic rat model. Male Sprague–
Dawley rats (Harlan, Indianapolis, IN), weighing
150g, are allowed to acclimate for one week and placed
on a standard diet (7012, Harlan Teklad certified

5674
M. C. Van Zandt et al. / Bioorg. Med. Chem. 12 (2004) 5661–5675
4.3. X-ray crystallography
The atomic coordinates and structure factors of the
hALR2-NADP+-example 40 (IDD 594) complex have
been deposited in the Protein Data Bank (1US0).
4.3.1. Overproduction and purification of recombinant
human ALR2. The ORF of the human ALR2 gene
(Accession GenBankTM/embl Data Bank Number
J05017) was amplified by PCRfrom its cDNA and
cloned into the T7 RNA polymerase-based vector
pET15b (Novagen). Expression of the (HIS)6-human
ALR2 in the E. coli strain BL21(DE3) (Novagen) is in-
duced by IPTG (Euromedex) during a 3h culture at
37ꢁC. The pellet from a 4L culture was disrupted by
sonication and centrifuged at 4ꢁC. The supernatant
was applied on a Talon metal affinity column (Clon-
tech). After thrombin cleavage of the hexahistidine
extension, the detagged protein was then loaded on a
DEAE Sephadex A-50 column (Pharmacia) and eluted
with a NaCl gradient.
Acknowledgements
This work was supported, in part, by the Centre Na-
tional de la Recherche Scientifique (CNRS), the Institut
´
´
National de la Sante et de la Recherche Medicale and
the Hopital Universitaire de Strasbourg (H.U.S.). We
ˆ
thank the personnel of the SBC, and in particular And-
rzej Joachimiak and Ruslan Sanishvili (Nukro), for their
help in data collection.
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.bmc.2004.
07.062.
4.3.2. Crystallization trials and preliminary crystallo-
graphic analysis. All trials were carried out in Linbro 24-
well tissue culture plates (Flow Laboratories) using the
hanging-drop vapor diffusion method. The enzyme was
dialyzed against 50mM ammonium citrate pH5.0. The
drops of a final volume of 12mL contained 15mg/mL
of human ALR2 with 2equiv of NADP+ (Sigma) and
5% of PEG 6000; they were equilibrated against a well
containing 120mM ammonium citrate, pH5.0, 20%
PEG 6000 and the same concentration of NADP+ as
above. Crystals were grown at 4ꢁC and reached the size
of 0.5 · 0.4 · 0.4mm.
References and notes
1. American Diabetes Associationꢁs homepage. http://
www.diabetes.org.
2. The Diabetes Control and Complications Trial Research
Group. The effect of intensive treatment of diabetes on the
development and progression of long-term complications
in insulin-dependent diabetes mellitus. New Engl. J. Med.
1993, 329, 977–986.
3. The United Kingdom Prospective Diabetes Study Group.
Intensive blood–glucose control with sulfonyl-ureas or
insulin compared with conventional treatment and risk of
complications in patients with type
(UKPDS33). Lancet 1998, 352, 837–853.
4. Textbook of Diabetes; Pickup, J. C., Williams, G., Eds.;
Blackwell Science: United Kingdom, 1997; Vol. 1, Chapter
42.
5. Williamson, J. R. et al. Hyperglycemic pseudohypoxia
and diabetic complications. Diabetes 1993, 42, 801–813.
6. Lee, A.; Chung, S. K.; Chung, S. S. Demonstration that
polyol accumulation is responsible for diabetic cataract by
use of transgenic mice expressing the aldose reductase gene
in the lens. Proc. Natl. Acad. Sci. U.S.A. 1995, 92,
2780–2784.
7. Oates, P. J.; Mylari, B. L. Aldose reductase inhibitors:
therapeutic implications for diabetic complications.
Expert Opin. Inv. Drug. 1999, 8, 1–15.
2
diabetes
4.3.3. Soaking with the inhibitor. Crystals of the ALR2-
inhibitor complex were obtained by soaking of the
native crystals with a solution containing 2mg/mL of
inhibitor dissolved in the mother liquor containing
120mM ammonium citrate, pH5.0, and 25% of PEG
6000. The soaking time was three weeks.
4.3.4. Data collection and refinement. The diffraction
analysis was performed using a rotating anode labora-
tory source and an image plate, and the data treated
with the HKL package.42 The inhibitor-soaked crystals
˚
diffracted to 1.8A and were isomorphous to the native
8. Kao, Y.-L.; Donahue, K.; Chan, A.; Knight, J.; Silink, M.
A novel polymorphism in the aldose reductase gene
promoter region is strongly associated with diabetic
retinopathy in adolescents with Type I diabetes. Diabetes
1999, 48, 1338–1340.
9. Heesom, A.; Hibberd, M.; Millward, A.; Demaine, A.
Polymorphism in the 50-end of the aldose reductase gene is
strongly associated with the development of diabetic
nephropathy in Type I diabetes. Diabetes 1997, 46,
287–291.
10. Moczulski, D. et al. The role of aldose reductase gene in
the susceptibility to diabetic nephropathy in Type II (non-
insulin-dependent) diabetes mellitus. Diabetologia 1999,
42, 94–97.
11. Shinoda, M.; Mori, S.; Shintani, S.; Ishikura, S.; Hara, A.
Inhibition of human aldehyde reductase by drugs for
testing the function of liver and kidney. Biol. Pharm. Bull.
1999, 22, 741–744.
crystals. The native structure was used as a model for
the molecular replacement; after a rigid-body minimiza-
tion, the inhibitor was placed in the electron density
shown by a Fourier difference map. The model of the
complex was used as starting point of the refinement,
which included a rigid-body refinement step, a slow-
cooling step, a Powell minimization, and a tempera-
ture-factor refinement. Water molecules were then
placed in a difference map and minimization was per-
formed as above without putting restraints on the water
molecules; bulk solvent correction was also applied. The
43
44
programs O and X-PLOR were used for model build-
ing and refinement.
Diffraction experiments were conducted on two crystals
at the X-ray beam of the undulator line 19ID of SBC-
CAT at the APS. All conditions were optimized for
the high resolution experiments.29
12. Suzuki, K.; Koh, Y. H.; Mizuno, H.; Hamaoka, R.;
Taniguchi, N. Overexpression of aldehyde reductase

M. C. Van Zandt et al. / Bioorg. Med. Chem. 12 (2004) 5661–5675
5675
protects PC12 cells from the cytotoxicity of methylglyoxal
or 3-deoxyglucosone J. Biochem. 1998, 123, 353–357.
13. Takahashi, M.; Fujii, J.; Teshima, T.; Suzuki, K.; Shiba,
T.; Taniguchi, N. Identity of a major 3-deoxyglucosone-
reducing enzyme with aldehyde reductase in rat liver
established by amino acid sequencing and cDNA expres-
sion. Gene 1993, 127, 249–253.
14. Kanazu, T.; Shinoda, M.; Nakayama, T.; Deyashiki, Y.;
Hara, A.; Sawada, H. Aldehyde reductase is a major
protein associated with 3-deoxyglucosone reductase activ-
ity in rat, pig and human livers. Biochem. J. 1991, 279,
903–906.
28. El-Kabbani, O.; Carper, D. A.; McGowan, M. H.;
Devedjiev, Y.; Rees-Milton, K. J.; Flynn, T. G. Studies
on the inhibitor-binding site of porcine aldehyde reduc-
tase: crystal structure of the holoenzyme-inhibitor ternary
complex. Proteins 1997, 29, 186–192.
29. Howard, E.; Sanishvili, R.; Cachau, R.; Mitschler, A.;
Chevrier, B.; Barth, P.; Lamour, V.; Van Zandt, M. C.;
Sibley, E.; Bon, C.; Moras, D.; Schneider, T. R.; Joachi-
miak, A.; Podjarny, A. Ultrahigh resolution drug design I:
details of interactions in human aldose reductase-inhibitor
˚
complex at 0.66A. Proteins: Struct., Funct., Bioinformatics
2004, 55, 792–804.
15. Hamada, Y.; Araki, N.; Koh, N.; Nakamura, J.; Horiuchi,
S.; Hotta, N. Rapid formation of advanced glycation end-
products by intermediate metabolites of glycolytic path-
way and polyol pathway. Biochem. Biophys. Res. Com-
mun. 1996, 228, 539–543.
30. Kotani, T.; Nagaki, Y.; Ishii, A.; Konishi, Y.; Yago, H.;
Suehiro, S.; Okukado, N.; Okamoto, K. Highly selective
aldose reductase inhibitors. 3. Structural diversity of 3-
(arylmethyl)-2,4,5-trioxoimidazolidine-1-acetic acids. J.
Med. Chem. 1997, 40.
16. Oya, T.; Hattori, N.; Mizuno, Y.; Miyata, S.; Maeda, S.;
Osawa, T.; Uchida, K. Methylglyoxal modification of
protein. J. Biol. Chem. 1999, 274, 18492–18502.
31. Malamas, M.; Holman, T.; Millen, J. Novel spirosuccin-
imide aldose reductase inhibitors derived from isoquinol-
ine-1,3-diones: 2-[(4-bromo-2-fluorophenyl)methyl]-6-fluo
rospiro[isoquinoline-4(1H),30-pyrrolidine]-1,20,3,50(2H)-
tetrone and congeners. J. Med. Chem. 1994, 37,
2043–2058.
32. Mylar, B.; Larson, E.; Beyer, T.; Zembrowski, W.;
Aldinger, C.; Dee, M.; Siegel, T.; Singleton, D. Novel,
potent aldose reductase inhibitors: 3,4-dihydro-4-oxo-
3-[[5-(trifluoromethyl)-2-benzothiazolyl]methyl]-1-phthal-
azine-acetic acid (zopolrestat and congeners). J. Med.
Chem. 1991, 34, 108–122.
33. Jones, M.; Gunn, D.; Jones, J.; Van Zandt, M. U.S. Patent
6,214,991, 2001.
34. Tomoji, A.; Nishio, T.; Hosono, H.; Nakamura, Y.;
Matsui, T.; Ishikawa, H. U.S. Patent 5,543,420, 1994.
35. Hayman, S.; Kinoshita, J. H. J. Biol. Chem. 1965, 240,
877–880.
36. Barski, O. A.; Gabbay, K. H.; Grimshaw, C. E.; Bohren,
K. M. Mechanism of human aldehyde reductase: charac-
terization of the active site pocket. Biochemistry 1995, 34,
11264–11275.
37. Stribling, D.; Mirrlees, D. J.; Harrison, H. E.; Earl, D. C.
N. Properties of ICI 128,436 a novel aldose reductase
inhibitor, and its effects on diabetic complications in the
rat. Metabolism 1985, 4, 336–344.
38. Ashizawa, N.; Yoshida, M.; Sugiyama, Y.; Akaike, N.;
Ohbayashi, S.; Aotsuka, T.; Abe, N.; Fukushima, K.;
Matsuura, A. Effects of a novel potent aldose reductase
inhibitor, GP-1447, on aldose reductase activity in vitro
and on diabetic neuropathy and cataract formation in
rats. Jpn. J. Pharmacol. 1997, 73, 133–144.
39. Ehrig, T.; Bohren, K. M.; Prendergast, G.; Gabbay, K.
Mechanism of aldose reductase inhibition: binding of
NADP+/NADPH and alrestatin-like inhibitors. Biochem-
istry 1994, 33, 7157–7165.
17. Cavalot, F.; Anfossi, G.; Russo, I.; Mularoni, E.; Mas-
succo, P.; Mattiello, L.; Burzacca, S.; Hahn, A.; Trovati,
M. Nonenzymatic glycation of fibronectin impairs adhe-
sive and proliferative properties of human vascular
smooth muscle cells. Metabolism 1996, 45, 285–292.
18. Fu, M.; Wells-Knecht, K.; Blackledge, J.; Lyons, T.;
Thorpe, S.; Baynes, J. Glycation, glycoidation, and cross-
linking of collagen by glucose. Diabetes 1994, 43, 676–683.
19. Flynn, T. G.; Green, N. C.; Bhatia, M. B.; El-Kabbani
Structure and mechanism of aldehyde reductase. In
Enzymology and Molecular Biology of Carbonyl Metabo-
lism 5; Weiner, et al., Eds.; Plenum: New York, 1995;
Chapter 25, pp 193–201.
20. Harrison, D. H. T.; Bohren, K. M.; Petsko, G. A.; Ringe,
D.; Gabbay, K. H. The alrestatin double-decker: binding
of two inhibitor molecules to human aldose reductase
reveals a new specificity determinant. Biochemistry 1997,
36, 16134–16140.
21. Umezu, K. Preparation of 4-fluorosalicylic acids from
difluorobenzoic acids in polar aprotic solvents. Japanese
patent JP 96-193878.
22. Wrobel, J.; Millen, J.; Sredy, J.; Dietrich, A.; Gorham, B.;
Malamas, M.; Kelly, J.; Bauman, J.; Harrison, M.; Jones,
L.; Guinosso, C.; Sestanj, K. Synthesis of tolrestat
analogues containing additional substituents in the ring
and their evaluation as aldose reductase inhibitors. Iden-
tification of poten, orally active 2-fluoro derivatives. J.
Med. Chem. 1991, 34, 2504–2520.
23. Hartwell, J. L. o-Chlorobromobenzene. Org. Syn. (Col-
lective Volume III), 1955, 185.
24. Wolfe, J.; Buchwald, S. Scope and limitations of the Pd/
¨
BINAP-catalyzed amination of aryl bromides. J. Org.
Chem. 2000, 65, 1144–1157.
ˆ
25. Urzhumtsev, A.; Tete-Favier, F.; Mitschler, A.; Barban-
40. Bergmeyer, H. U.; Gruber, W.; Gutmann, I. In Methods of
Enzymatic Analysis, 2nd ed.; Bergmeyer, H. U., Ed.;
Weinheim and Academic: New York, London, 1974; Vol.
3, pp 1323–1330.
41. Malone, J. I.; Knox, G.; Benford, S.; Tedesco, T. A. Red
cell sorbitol. An indicator of diabetic control. Diabetes
1980, 29, 861–864.
42. Otwinowski, Z.; Minor, W. In Methods in Enzymology,
Macromolecular Crystallography, part A; Carter, C. W.,
Jr., Sweet, R. M., Eds.; Academic, 1997; Vol. 276, pp
307–326.
43. Jones, T. A.; Zou, J. Y.; Cowan, S. W.; Kjeldgaard, M. O.
Acta Cryst. 1991, A47, 110–119.
ton, J.; Barth, P.; Urzhumtseva, L.; Biellmann, J.-F.;
Podjarny, A. D.; Moras, D. A. ꢀSpecificityꢁ pocket inferred
from the crystal structures of the complexes of aldose
reductase with the pharmaceutically important inhibitors
tolrestat and sorbinil. Structure 1997, 5, 601–612.
26. El-Kabbani, O.; Wilson, D. K.; Petrash, J. M.; Quiocho,
F. A. Structural features of the aldose reductase and
aldehyde reductase inhibitor-binding sites. Mol. Vis. 1998,
4, 19–25.
27. Rees-Milton, K. J.; Jia, Z.; Green, N. C.; Bhatia, M.; El-
Kabbani, O.; Flynn, T. G. Aldehyde reductase: the role of
C-terminal residues in defining substrate and cofactor
specificities. Arch. Biochem. Biophys. 1998, 355, 137–
144.
44. Brunger, A. T. Yale University Press: New Haven, CT,
USA, 1992.