Design of amino acid aldehydes as transition-state analogue inhibitors of arginase

Article abstract of DOI:10.1021/ja047788w

Arginase is a binuclear manganese metalloenzyme that catalyzes the hydrolysis of L-arginine to form L-ornithine and urea. Chiral L-amino acids bearing aldehyde side chains have been synthesized in which the electrophilic aldehyde C=O bond is isosteric with the C=N bond of L-arginine. This substitution is intended to facilitate nucleophilic attack by the metal-bridging hydroxide ion upon binding to the arginase active site. Syntheses of the amino acid aldehydes have been accomplished by reduction, oxidation, and Wittig-type reaction with a commercially available derivative of L-glutamic acid. Amino acid aldehydes exhibit inhibition in the micromolar range, and the X-ray crystal structure of arginase I complexed with one of these inhibitors, (S)-2-amino-7-oxoheptanoic acid, has been determined at 2.2 A resolution. In the enzyme-inhibitor complex, the inhibitor aldehyde moiety is hydrated to form the gem-diol: one hydroxyl group bridges the Mn²⁺₂ cluster and donates a hydrogen bond to D128, and the second hydroxyl group donates a hydrogen bond to E277. The binding mode of the neutral gem-diol may mimic the binding of the neutral tetrahedral intermediate and its flanking transition states in arginase catalysis.

Full text of DOI:10.1021/ja047788w

Published on Web 07/30/2004
Design of Amino Acid Aldehydes as Transition-State
Analogue Inhibitors of Arginase
Hyunshun Shin, Evis Cama, and David W. Christianson*
Contribution from the Roy and Diana Vagelos Laboratories, Department of Chemistry,
UniVersity of PennsylVania, Philadelphia, PennsylVania 19104-6323
Received April 16, 2004; E-mail: chris@xtal.chem.upenn.edu
Abstract: Arginase is a binuclear manganese metalloenzyme that catalyzes the hydrolysis of L-arginine to
form L-ornithine and urea. Chiral L-amino acids bearing aldehyde side chains have been synthesized in
which the electrophilic aldehyde CdO bond is isosteric with the CdN bond of L-arginine. This substitution
is intended to facilitate nucleophilic attack by the metal-bridging hydroxide ion upon binding to the arginase
active site. Syntheses of the amino acid aldehydes have been accomplished by reduction, oxidation, and
Wittig-type reaction with a commercially available derivative of L-glutamic acid. Amino acid aldehydes exhibit
inhibition in the micromolar range, and the X-ray crystal structure of arginase I complexed with one of
these inhibitors, (S)-2-amino-7-oxoheptanoic acid, has been determined at 2.2 Å resolution. In the enzyme-
inhibitor complex, the inhibitor aldehyde moiety is hydrated to form the gem-diol: one hydroxyl group bridges
the Mn2+2 cluster and donates a hydrogen bond to D128, and the second hydroxyl group donates a hydrogen
bond to E277. The binding mode of the neutral gem-diol may mimic the binding of the neutral tetrahedral
intermediate and its flanking transition states in arginase catalysis.
On the basis of Hammond’s postulate,9 the transition state
for the nucleophilic attack of hydroxide ion at the guanidinium
Introduction
Arginase is a 105-kD trimeric metalloenzyme that catalyzes
the hydrolysis of L-arginine to form L-ornithine and urea, and
two mammalian isozymes have been identified.1-3 Arginase I
catalyzes the final cytosolic step of the urea cycle in the liver,
whereas arginase II is nonhepatic and functions predominantly
in L-arginine homeostasis. The X-ray crystal structures of rat
arginase I and human arginase II, as well as the hexameric
arginase from Bacillus caldoVelox, reveal a binuclear manganese
cluster located at the base of a ∼15 Å-deep active site cleft.4-6
The pH-rate profile7 of arginase I exhibits an activity-linked
pKa of 7.9, consistent with the ionization of the likely catalytic
nucleophile, a metal-bridging solvent molecule.4 Structure-
activity relationships outlined for site-specific variants of
arginase I are consistent with a metal-activated hydroxide
mechanism in which the metal-bridging hydroxide ion attacks
the substrate guanidinium group oriented by hydrogen bonds
with E277 and the backbone carbonyl of H141 (Figure 1).4,8
carbon is expected to be close in structure and energy to the
metastable tetrahedral intermediate illustrated in Figure 1, and
this has motivated the investigation of tetrahedral transition-
state analogues. For example, in a recent study we reported the
X-ray crystal structure of arginase I complexed with an amino
acid sulfonamide: the NH2 group of the tetrahedral sulfonamide
ionizes to the NH- form, which bridges the binuclear manga-
nese cluster and donates a hydrogen bond to D128.10 In earlier
work, we reported boronic acid analogues of L-arginine in which
the trigonal planar boronic acid moiety is substituted for the
trigonal planar guanidinium group.11,12 Strictly speaking, these
boronic acid analogues are substrate analogues and not transi-
tion-state analogues. However, the electron-deficient boron atom
in each of these analogues facilitates nucleophilic attack by
hydroxide ion to form a tetrahedral boronate anion.5,12,13
Accordingly, boronic acid inhibitors of arginase are reactive
substrate analogues that, upon binding to the enzyme, undergo
a chemical transformation to better mimic the tetrahedral
intermediate and its flanking transition states in catalysis. Since
this chemical transformation mimics the first bond-making step
with an actual substrate, such reactive substrate analogues have
been designated “reaction coordinate analogues”.14 The chemical
* Corresponding author: phone 215-898-5714; fax 215-573-2201.
(1) Christianson, D. W.; Cox, J. D. Annu. ReV. Biochem. 1999, 68, 33-57.
(2) Ash, D. E.; Cox, J. D.; Christianson, D. W. Metal Ions in Biological Systems;
Sigel, A., Sigel, H., Eds.; M. Dekker: New York, 2000; Vol. 37 (Manganese
and its Role in Biological Processes), pp 408-428.
(3) Morris, S. M., Jr. Annu. ReV. Nutr. 2002, 22, 87-105.
(4) Kanyo, Z. F.; Scolnick, L. R.; Ash, D. E.; Christianson, D. W. Nature
1996, 383, 554-557.
(9) Hammond, G. S. J. Am. Chem. Soc. 1955, 77, 334-338.
(10) Cama, E.; Shin, H. S.; Christianson, D. W. J. Am. Chem. Soc. 2003, 125,
13052-13057.
(11) Baggio, R.; Elbaum, D.; Kanyo, Z. F.; Carroll, P. J.; Cavalli, R. C.; Ash,
D. E.; Christianson, D. W. J. Am. Chem. Soc. 1997, 119, 8107-8108.
(12) Kim, N. N.; Cox, J. D.; Baggio, R. F.; Emig, F. A.; Mistry, S.; Harper, S.
L.; Speicher, D. W.; Morris, S. M., Jr.; Ash, D. E.; Traish, A. M.;
Christianson, D. W. Biochemistry 2001, 40, 2678-2688.
(13) Cox, J. D.; Kim, N. N.; Traish, A. M.; Christianson, D. W. Nat. Struct.
Biol. 1999, 6, 1043-1047.
(5) Cama, E.; Colleluori, D. M.; Emig, F. A.; Shin, H.; Kim, S. W.; Kim, N.
N.; Traish, A. M.; Ash, D. E.; Christianson, D. W. Biochemistry 2003, 42,
8445-8451.
(6) Bewley, M. C.; Jeffrey, P. D.; Patchett, M. L.; Kanyo, Z. F.; Baker, E. N.
Structure 1999, 7, 435-448.
(7) Kuhn, N. J.; Talbot, J.; Ward, S. Arch. Biochem. Biophys. 1991, 286, 217-
221.
(8) Cama, E.; Emig, F. A.; Ash, D. E.; Christianson, D. W. Biochemistry 2003,
42, 7748-7758.
9
10278
J. AM. CHEM. SOC. 2004, 126, 10278-10284
10.1021/ja047788w CCC: $27.50 © 2004 American Chemical Society

Amino Acid Aldehyde Inhibitors of Arginase
A R T I C L E S
Figure 1. Proposed mechanism of arginase.
Scheme 1. Synthesis of (S)-2-Amino-7-oxoheptanoic Acid
Scheme 2. Synthesis of (S)-(E)-2-Amino-7-oxohept-5-enoic Acid
reactivity of a reaction coordinate analogue exploits the tendency
of an enzyme to bind the transition state (or transition-state
analogue) more tightly than the substrate (or substrate ana-
logue).15
Given the successful incorporation of the aldehyde moiety
into inhibitors of the zinc proteases,16-20 we hypothesized that
the substitution of the trigonal planar aldehyde moiety for the
trigonal planar guanidinium group of L-arginine would yield a
reactive substrate analogue capable of binding as a gem-diol
analogue of the tetrahedral intermediate in the arginase mech-
anism. Here, we report the design and synthesis of two amino
acid aldehyde inhibitors of arginase: (S)-2-amino-7-oxohep-
tanoic acid (AOH; 5, Scheme 1) and (S)-(E)-2-amino-7-oxohept-
5-enoic acid (unsaturated AOH; 8, Scheme 2). The X-ray crystal
structure of the arginase I-AOH complex reveals that, as
designed, the inhibitor binds to the enzyme active site as the
tetrahedral gem-diol.
Compounds containing the aldehyde moiety can similarly
serve as reaction coordinate analogues for hydrolytic enzymes
due to the enhanced susceptibility of the aldehyde carbonyl to
nucleophilic attack. Like the electron-deficient boronic acid, a
properly oriented aldehyde group in a metalloenzyme inhibitor
can undergo nucleophilic attack by a metal-bound solvent
molecule to form a tetrahedral gem-diol. Aldehyde moieties have
been incorporated into inhibitors of the zinc proteases leucine
aminopeptidase and carboxypeptidase A,16-18 and the crystal
structures of these enzyme-inhibitor complexes confirm the
binding of each inhibitor as a gem-diol in its target enzyme
active site.19,20
Experimental Section
Syntheses of Amino Acid Aldehydes. Reactions requiring anhy-
drous conditions were carried out under a nitrogen atmosphere in flame-
or oven-dried glassware, and solvents were freshly distilled. Diethyl
(14) Christianson, D. W.; Lipscomb, W. N. Acc. Chem. Res. 1989, 22, 62-69.
(15) Pauling, L. Chem. Eng. News 1946, 24, 1375-1377.
(16) Andersson, L.; Isley, T. C.; Wolfenden, R. Biochemistry 1982, 21, 4177-
4180.
(17) Andersson, L.; MacNeela, J.; Wolfenden, R. Biochemistry 1985, 24, 330-
(19) Christianson, D. W.; Lipscomb, W. N. Proc. Natl. Acad. Sci. U.S.A. 1985,
82, 6840-6844.
(20) Stra¨ter, N.; Lipscomb, W. N. Biochemistry 1995, 34, 14792-14800.
333.
(18) Galardy, R. E.; Kortylewicz, Z. P. Biochemistry 1984, 23, 2083-2087.
9
J. AM. CHEM. SOC. VOL. 126, NO. 33, 2004 10279

A R T I C L E S
Shin et al.
ether and tetrahydrofuran (THF) were distilled from sodium/benzophe-
none. Dichloromethane and triethylamine were distilled from calcium
hydride. Benzene was dried with molecular sieves. Ethyl acetate was
dried by passage through a column of alumina. All other reagents and
solvents were used without further purification from commercial
sources. Organic extracts were dried over MgSO4 or Na2SO4. Reactions
were monitored by thin-layer chromatography (TLC) with 0.25-mm
E. Merck precoated silica gel plates and visualized with ninhydrin
solution (0.1% ninhydrin in 95% n-butanol, 4.5% water, 0.5% glacial
acetic acid) and KMnO4 solution (3 g of KMnO4, 20 g of K2CO3, 5
mL of 5% NaOH, and 300 mL of water). Flash column chromatography
was carried out with E. Merck silica gel 60 (230-240 mesh ASTM).
The 1H and 13C NMR spectra were recorded on a Bruker AM-500
spectrometer. Chemical shifts were expressed in parts per million (ppm)
and referenced to CDCl3, CD3OD, or DMSO-d6. The general syntheses
of (S)-2-amino-7-oxoheptanoic acid 5 and (S)-(E)-2-amino-7-oxohept-
5-enenoic acid 8 are as follows.
(S)-(E)-1-tert-Butyl-[bis-(2-tert-butoxycarbonyl)amino]-7-hydroxy-
hept-5-eneoate, 6. To a stirred solution of 2 (1.4 g, mmol) in dry ether
(10 mL) was added dropwise DIBAL-H (7.9 mL, 1.0 M in hexane,
7.9 mmol) at 0 °C. The reaction mixture was stirred for 1 h. It was
quenched with H2O (0.56 mL) and allowed to warm to room
temperature. The mixture was stirred for 30 min, dried over MgSO4,
and filtered through a pad of Celite. The solvent was removed on a
rotary evaporator, and the crude residue was purified by silica gel
1
chromatography to yield 6 (1.0 g, 77% yield). H NMR (CDCl3): δ
5.67-5.58 (m, 2H), 4.75-4.72 (m, 1H), 4.03 (dd, J ) 5.5, 2.0 Hz,
2H), 2.30-2.08 (m, 3H), 1.97-1.90 (m, 1H), 1.47 (s, 18H), 1.41 (s,
13
9H). C NMR (CDCl3): δ 169.8, 152.5, 132.3, 130.1, 82.7, 81.2,
63.8, 58.3, 29.1, 27.3. MS (ES+) (m/z): [M+] calcd for (C21H37O7N)
415.26; found, 415.54.
(S)-(E)-1-tert-Butyl-[bis-(2-tert-butoxycarbonyl)amino]-7-oxo-
hept-5-enoate, 7. A solution of 6 (0.59 g, 1.43 mmol) in CH2Cl2 (20
mL) was added to Dess-Martin periodinane (3.6 mL, 1.71 mmol, 15
wt % in CH2Cl2) at 0 °C. Stirring was continued at 0 °C for 15 min,
followed by warming to ambient temperature. The reaction was
monitored by TLC. After 1 h the mixture was diluted with diethyl ether
(30 mL), poured into 40 mL of ice-cold saturated NaHCO3/saturated
Na2S2O3 (1:1), and shaken for 5 min. The organic phase was separated
and washed with saturated NaHCO3, H2O, and saturated NaCl, dried
over Na2SO4, and concentrated in vacuo. The residue was purified by
(S)-(E)-1-tert-Butyl-[bis-(2-tert-butoxycarbonyl)amino]-7-methyl-
hept-5-eneoate, 2. To a stirred solution of (S)-1-tert-butyl-[bis(2-tert-
butoxycarbonyl)amino]-5-oxopentanoate 1 (2.8 g, 7.23 mmol) in
benzene (25 mL) was added commercially available methyl (triphen-
ylphosphoranylidene)acetate (4.8 g, 14.46 mmol) at 0 °C. The mixture
was stirred until TLC showed the complete consumption of the starting
material. The solvent was removed on a rotary evaporator, and the crude
residue was purified by silica gel chromatography to yield 2 (2.2 g,
1
chromatography to afford 7 (0.348 g, 60% yield). H NMR (CDCl3):
1
79% yield). H NMR (CDCl3): δ 6.95-6.89 (m, 1H), 5.81 (d, J )
δ 9.50 (d, 1H), 6.81-6.57 (m, 1H), 6.12-6.07 (m, 1H), 4.73-4.69
(m, 1H), 2.63-2.24 (m, 3H), 2.05-1.99 (m, 1H), 1.47 (s, 18H), 1.41
(s, 9H). MS (ES+) (m/z): [M + Na] calcd for (C21H35O7N+Na) 436.22;
found 436.60, isotope 437.60.
15.6 Hz, 1H), 4.75-4.59 (m, 1H), 3.69 (s, 3H), 2.27-2.21 (m, 3H),
2.21-1.96 (m, 1H), 1.45 (s, 18H), 1.41 (s, 9H). 13C NMR (CDCl3): δ
169.4, 166.9, 152.4, 148.1, 121.6, 82.9, 81.4, 58.3, 51.4, 29.1, 28.0,
27.93, 27.8. HRMS (EI) (m/z): [M+] calcd for C22H37O8N, 443.2417;
found, 443.2430.
(S)-(E)-2-Amino-7-oxo-hept-5-enoic Acid, 8. To a stirred solution
of 7 (0.425 g, 1.03 mmol) in CH2Cl2 (8 mL) was slowly added TFA (8
mL) at 0 °C. The reaction mixture was stirred for 15 min and then
allowed to equilibrate at room temperature until the end of the reaction.
The solvent was removed on a rotary evaporator, and the crude residue
was purified by silica gel chromatography to yield 8 (0.24 g, 87% yield).
1H NMR (CD3OD): δ 9.01 (s, 1H), 6.94-6.91 (m, 1H), 5.91-5.88
(d, J ) 16 Hz, 1H), 3.97-3.95 (m, 1H), 2.50-1.90 (m, 4H). MS (ES+)
(m/z): [M+] calcd for (C7H11O3N) 157.07, found 157.86.
Kinetic Assays. Arginase inhibition by AOH was evaluated using
a modified version of the fixed-point radioactive assay developed by
Ru¨egg and Russell.21 Assay mixtures contained 500 µL of CHES-
NaOH (100 mM, pH 9.5), 100 µL of 1 mM MnCl2, 10 µL of 0.05 µCi
of [14C-guanido]-L-arginine, 90 µL of water, 100 µL of 100 mM
unlabeled L-arginine, then taking 40 µL of the assay mixtures and
5 µL of varying concentrations of AOH in a 50 µL volume per
centrifuge tube. Reactions were started by the addition of 5 µL of a 10
µg/mL wild-type arginase I solution. After 5 min, reactions were
quenched by the addition of 400 µL of “stop” solution (0.25 mM acetic
acid (pH 4.5), 7 M urea, 10 mM L-arginine, and a 1:1 (v/v) slurry of
Dowex W-X8 in water). Each reaction mixture was vortexed im-
mediately after the addition of “stop” solution, gently mixed for an
additional 5 min, and centrifuged at 6000 rpm for 10 min. A 200 µL
volume of the supernatant was removed, and 3 mL of scintillation liquid
(EcoScint) was added in preparation for liquid scintillation counting
in a Beckman counter (model LS 6000 SC). For data analysis, the plot
of Vo/V as a function of inhibitor concentration was expected to be linear
for a simple competitive inhibitor, and the slope of the line was used
to calculate Ki using the equation for competitive binding (Vo and V
were the observed velocities measured in the absence and presence of
inhibitor, respectively).
(S)-1-tert-Butyl-[bis-(2-tert-butoxycarbonyl)amino]-7-methyl-hep-
tanoate, 3. To a stirred solution of 2 (0.96 g, 2.16 mmol) in dry EtOAc
(20 mL) was added Pd/C (90 mg). The reaction was stirred under
hydrogen atmosphere until TLC showed the complete conversion of
starting material. The mixture was filtered through a pad of Celite.
The solvent was removed on a rotary evaporator, and the crude residue
was purified by silica gel chromatography to yield 3 (0.91 g, 95% yield).
1H NMR (CDCl3): δ 4.67-4.64 (m, 1H), 3.63 (s, 3H), 2.27 (t, J )
7.7 Hz, 2H), 2.05-1.98 (m, 1H), 1.86-1.78 (m, 1H), 1.68-1.52 (m,
4H), 1.46 (s, 18H), 1.40 (s, 9H). 13C NMR (CDCl3): δ 173.9, 169,8,
152.4, 82.66, 81.1, 58.7, 51.4, 33.9, 28.9, 28.0, 27.9, 26.0, 24.6. MS
(ES+) (m/z): [M+] calcd for (C22H39O8N) 445.27; found, 445.60.
(S)-1-tert-Butyl-[bis-(2-tert-butoxycarbonyl)amino]-7-oxo-hep-
tanoate, 4. To a stirred solution of 3 (0.91 g, 2.0 mmol) in dry ether
(20 mL) was added dropwise DIBAL-H (3.5 mL, 1.0 M in hexane,
3.0 mmol) to -78 °C. The reaction was stirred for 45 min. It was
quenched with H2O (0.25 mL, 7 equiv) and allowed to warm to room
temperature. The mixture was stirred for 30 min, dried over MgSO4,
and filtered through a pad of Celite. The solvent was removed on a
rotary evaporator, and the crude residue was purified by silica gel
1
chromatography to yield 4 (0.80 g, 95% yield). H NMR (CDCl3): δ
9.73 (s, 1H), 4.69-4.65 (m, 1H), 2.42-2.39 (m, 2H), 2.08-2.02 (m,
1H), 1.86-1.82 (m, 1H), 1.69-1.29 (m, 4H), 1.45 (s, 18H), 1.41 (s,
9H). 13C NMR (CDCl3): δ 202.4, 169.8, 152.5, 82.8, 81.2, 58.6, 43.7,
30.0, 28.0, 26.0, 21.8. MS (ES+) (m/z): [M+] calcd for (C21H37O7N)
415.26; found, 415.54.
(S)-2-Amino-7-oxoheptanoic Acid, 5. To a stirred solution of 4 (0.1
g, 0.24 mmol) in CH2Cl2 (1 mL) at 0 °C was slowly added
trifluoroacetic acid (TFA) (1 mL). The reaction mixture was stirred
for 15 min at 0 °C and then allowed warm to room temperature until
the end of the reaction. The solvent was removed on a rotary evaporator,
and the crude residue was purified by silica gel chromatography to
yield 5 (0.041 g, 63% yield). 1H NMR (CD3OD): δ 9.36 (s, 1H), 3.97-
3.93 (m, 1H), 2.3-1.9 (m, 2H), 1.59-1.27 (m, 6H). MS (ES+) (m/z):
[M+] calcd for (C7H13O3N) 159.08; found, 159.06.
Crystallography. For ease of protein handling, a multisite-specific
variant of rat arginase I was utilized for the preparation of the enzyme-
inhibitor complex. Crystals of this variant consistently diffracted to
higher resolution than crystals of the wild-type enzyme. The C119A/
(21) Ru¨egg, U. T.; Russell, A. S. Anal. Biochem. 1980, 102, 206-212.
9
10280 J. AM. CHEM. SOC. VOL. 126, NO. 33, 2004

Amino Acid Aldehyde Inhibitors of Arginase
A R T I C L E S
Table 1. Data Collection and Refinement Statistics
simulated annealing omit electron density map in the final stages of
refinement. It was clear at this point that AOH bound to arginase as
the hydrated aldehyde, i.e., a gem-diol. Final refinement statistics are
reported in Table 1. Figures were generated using BOBSCRIPT and
Raster3D.29,30 The atomic coordinates of the BPA-arginase I-AOH
complex have been deposited in the Protein Databank (http://
www.rcsb.org/pdb) with accession code 1T5F.
resolution (Å)
2.2
total reflections (N)
123 870
unique reflections (N)
completeness (%) (last shell)
Rmerge (last shell)b
43 716
93.0 (85.4)a
0.080 (0.327)a
41 783 (2099)a
0.222
reflections used in refinement (test set)c
c
Rcryst
c
Rfree
0.241
Results
rms deviations
bonds (Å)
angles (deg)
0.015
2.1
Amino Acid Aldehydes. The methodology for the synthesis
of R-amino acid aldehydes was based on a synthetic strategy
utilizing the commercially available L-glutamic acid derivative,
N-t-boc-L-glutamic acid-R-tert-butyl ester (Sigma), from which
(S)-1-(tert-butyl-2-[bis(tert-butoxycarbonyl)amino]-5-oxopen-
tanoate 1 was prepared as described.31 For the synthesis of (S)-
2-amino-7-oxoheptanoic acid (AOH; 5, Scheme 1), elongation
of aldehyde 1 using methyl(triphenylphosphoranylidene)acetate
under Wittig conditions32,33 provided the unsaturated ester 2 with
complete (E)-selectivity in 79% yield. The hydrogenation of
the double bond in ester 2 over Pd/C in ethyl acetate afforded
the saturated ester 3 in 95% yield. Careful reduction of 3 with
DIBAL-H (1.5 equiv, -78 °C, 45 min) afforded the aldehyde
4 in 95% yield. Subsequent deprotection with trifluoroacetic
acid yielded (S)-2-amino-7-oxoheptanoic acid (AOH) 5 in 63%
yield.
dihedrals (deg)
impropers (deg)
Rcryst
24.0
1.0
0.222
c
a Numbers in parentheses refer to the outer 0.1 Å shell of data. Rmerge
for replicate reflections, R ) Σ|Ih - <Ih>|/Σ<Ih>; Ih ) intensity measured
for reflection h; <Ih> ) average intensity for reflection h calculated from
replicate data. c Crystallographic R factor, Rcryst ) Σ||Fo| - |Fc||/Σ|Fo| for
reflections contained in the working set. Free R factor, Rfree ) Σ||Fo| -
|Fc||/Σ|Fo| for reflections contained in the test set held aside during
refinement (5% of total). |Fo| and |Fc| are the observed and calculated
structure factor amplitudes, respectively.
b
C168A/C303A/H141A/Q19C variant was prepared by site-directed
mutagenesis as described.22 The plasmid was then transformed into E.
coli BL21(DE3) cells and purified by the same procedure utilized for
the purification of recombinant wild-type arginase. Finally, the variant
was chemically modified at C19 with 3-bromopropylamine and
accordingly designated BPA-arginase I. Activity assays of BPA-arginase
I yield KM and kcat values of 0.4 mM and 0.55 s-1, respectively,
indicating 0.22% of wild-type activity.22 Since neither the binuclear
manganese cluster nor the nucleophilic metal-bridging hydroxide ion
is perturbed in BPA-arginase I, the reactivity of the metal-bridging
hydroxide ion is not compromised.
The BPA arginase I variant was crystallized using the hanging drop
vapor diffusion method, as achieved for the native rat liver enzyme.23
Briefly, a 3 µL drop of precipitant solution [50 mM bicine (pH 8.2-
8.5), 20-24% PEG 8000, 2-5 mM MnCl2] was added to a 3 µL drop
of protein solution [10-16 mg/mL BPA arginase I variant, 50 mM
bicine (pH 8.5), 100 µM MnCl2] on a silanized cover slip subsequently
inverted and sealed over a 1 mL reservoir of precipitant buffer at 4 °C.
Crystals appeared after 4 weeks and were nearly isomorphous with
those of the wild-type enzyme.4,23 Crystals belonged to space group
P32 with unit cell parameters a ) b ) 87.9 Å, c ) 105.2 Å. Data
collection statistics are recorded in Table 1. The BPA arginase I-AOH
complex was prepared by soaking enzyme crystals for 3-8 days in
26% PEG 8000, 100 mM bicine (pH 8.5), 3 mM MnCl2, and 5 mM
AOH. Prior to data collection, crystals were gradually transferred to a
cryoprotectant solution consisting of the original precipitant solution
plus 30% glycerol and then flash-cooled with liquid nitrogen.
X-ray diffraction data were collected at the Advance Light Source,
beamline 5.0.1. Intensity data integration and reduction were performed
using the HKL suite of programs.24 Initial phases were determined by
molecular replacement with AmoRe25,26 using the structure of the native
rat liver arginase trimer4 as a search probe. Iterative rounds of model
building and refinement were performed using O and CNS, respec-
tively.27,28 Strict noncrystallographic symmetry was employed at the
beginning of refinement and later relaxed to appropriately weighted
constraints as judged by Rfree. The AOH molecule was built into a
Given the high degree of (E)-stereoselectivity obtained in the
synthesis of unsaturated ester 2, this compound served as a
suitable precursor for the synthesis of (S)-(E)-2-amino-7-
oxohept-5-enoic acid (unsaturated AOH; 8, Scheme 2). Reduc-
tion of 2 with DIBAL-H under controlled conditions (1.2 equiv,
0 °C, 45 min)34 afforded the corresponding allylic alcohol 6 in
77% yield. Oxidation of 6 with Dess-Martin periodinane35
afforded the aldehyde 7 in 60% yield. Global deprotection with
trifluoroacetic acid in methylene chloride36 furnished unsaturated
AOH 8 in 87% yield.
Arginase-Inhibitor Binding Affinities. Between the two
amino acid aldehyde inhibitors studied, the highest affinity was
measured for AOH with Ki ) 60 ( 8 µM against wild-type rat
arginase I at pH 9.5. For unsaturated AOH, Ki ) 360 ( 160
µM. Accordingly, we focused our structural studies on the
arginase I-AOH complex.
Structure of the Arginase I-AOH Complex. The binding
of AOH does not trigger any major changes in the tertiary or
quaternary structure of BPA-arginase I, and the rms deviation
of 314 monomer CR atoms between the enzyme-inhibitor
complex and the native enzyme is 0.09 Å. An electron density
map is found in Figure 2a, which conclusively reveals the
binding of the tetrahedral gem-diol form, rather than the native
(28) Bru¨nger, A. T.; Adams, P. D.; Clore, G. M.; DeLano, W. L.; Gros, P.;
Grosse-Kunstleve, R. W.; Jiang, J.-S.; Kuszewski, J.; Nilges, M.; Pannu,
N. S.; Read, R. J.; Rice, L. M.; Simonson, T.; Warren, G. L. Acta
Crystallogr. 1998, D54, 905-921.
(29) Esnouf, R. M. J. Mol. Graphics 1997, 15, 132-134.
(30) Merritt, E. A.; Bacon, D. J. Methods Enzymol. 1997, 277, 505-524.
(31) Adamczyk, M.; Johnson, D. D.; Reddy, R. E. Tetrahedron: Asymmetry
1999, 10, 775-781.
(22) Cama, E.; Pethe, S.; Boucher, J.-L.; Han, S.; Emig, F. A.; Ash, D. E.;
Viola, R. E.; Mansuy, D.; Christianson, D. W. Biochemistry 2004, 43,
8987-8999.
(32) Kokotos, G.; Padro´n, J. M.; Mart´ın, T.; Gibbons, W. A.; Mart´ın, V. S. J.
Org. Chem. 1998, 63, 3741-3744.
(23) Kanyo, Z. F.; Chen, C.-Y.; Daghigh, F.; Ash, D. E.; Christianson, D. W.
J. Mol. Biol. 1992, 224, 1175-1177.
(33) Padro´n, J. M.; Kokotos, G.; Mart´ın, T.; Markidis, T.; Gibbons, W. A.;
Mart´ın, V. S. Tetrahedron: Asymmetry 1998, 9, 3381-3394.
(34) Mart´ın, R.; Islas, G.; Moyano, A.; Perica`s, M. A.; Riera, A. Tetrahedron
2001, 57, 6367-6374.
(24) Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276, 307-326.
(25) Navaza, J. Acta Crystallogr. 1994, A50, 157-163.
(26) Collaborative Computational Project, No. 4. Acta Crystallogr. 1994, D50,
760-763.
(35) Samano, V.; Robins, M. J. J. Org. Chem. 1990, 55, 5186-5188.
(36) Ellison, R. A.; Lukenbach, E. R.; Chiu, C. Tetrahedron Lett. 1975, 15,
499-502.
(27) Jones, T. A.; Zou, J.-Y.; Cowan, S. W.; Kjeldgaard, M. Acta Crystallogr.
1991, A47, 110-119.
9
J. AM. CHEM. SOC. VOL. 126, NO. 33, 2004 10281

A R T I C L E S
Shin et al.
comparable to those observed in the binding of hydrated boronic
acid inhibitors to rat arginase I and human arginase II.8,12,13
The average B factor of 48 Å2 for AOH is much higher than
that of 31 Å2 measured for all protein atoms, which likely
reflects partial occupancy inhibitor binding. If the thermal B
factors of AOH atoms are held fixed at 31 Å2 during occupancy
refinement, the AOH occupancy refines to 80%. Consistent with
partial occupancy binding of AOH, a difference electron density
map reveals a residual electron density peak (∼50% occupancy)
refined as the metal-bridging hydroxide ion of the native enzyme
(Figure 2b). It is interesting that hydroxyl group O1 of AOH,
which bridges Mn2+ and Mn2+B, is ∼0.5 Å away from the
A
position of the metal-bridging hydroxide ion in the native
enzyme, suggesting some distortion of the metal coordination
polyhedron as the metal-bridging hydroxide ion attacks the
electrophilic carbonyl group of the inhibitor.
The R-amino group of AOH donates hydrogen bonds to
D183, E186, and two solvent molecules. One of these solvent
molecules forms additional hydrogen bonds with E183, N130,
and the backbone carbonyl of G142, and the other solvent
molecule donates hydrogen bonds to the side chain and
backbone carbonyl of D181. One of the R-carboxylate oxygens
accepts a hydrogen bond from the side chain of N130 and two
solvent molecules, while the other R-carboxylate oxygen is not
observed to engage in any hydrogen-bond interactions. These
arginase-AOH interactions are comparable to those observed
for the binding of amino acid boronate12,13 and amino acid
sulfonamide10 inhibitors, except for the lack of a hydrogen-bond
interaction between the R-carboxylate group and the side chain
of S137 observed in complexes with the amino acid boronate
inhibitors.12,13 Intermolecular inhibitor interactions are illustrated
in Figure 2c.
Discussion
The X-ray crystal structure of the BPA-arginase I-AOH
complex conclusively reveals the preferential binding of AOH
as the hydrated aldehyde (gem-diol) rather than the intact
aldehyde. Accordingly, AOH is a reactive substrate analogue
that undergoes a chemical transformation to mimic the more
tightly bound tetrahedral intermediate and its flanking transition
states in the arginase mechanism. That this chemical trans-
formationsthe nucleophilic attack of solvent at the electrophilic
aldehyde carbonylsis identical to that of the first step of the
arginase mechanism (Figure 1) leads us to classify AOH as a
new type of reaction coordinate analogue inhibitor of arginase
(Figure 3).
The aldehyde moiety is sufficiently electrophilic that in
aqueous solution it exists in rapid equilibrium with a substantial
concentration of its hydrated gem-diol form. For example, the
equilibrium constant for acetaldehyde hydration in dilute D2O
solution is 1.38, indicating a mixture of 58% gem-diol and 42%
intact aldehyde.37 Accordingly, it is possible that the preformed
gem-diol form of AOH binds to the arginase I active site rather
than the enzyme itself catalyzing the hydration of the AOH
aldehyde moiety. There is precedent in the metalloenzyme
literature for either possibility. The zinc metalloenzyme carbonic
anhydrase II catalyzes the hydration of aldehyde substrates,38
so it is conceivable that arginase I similarly catalyzes the
Figure 2. Simulated annealing omit electron density maps of the arginase
I-AOH complex (The carbon skeleton of AOH is black): (a) AOH omitted
from the structure factor calculation (contoured at 6.5σ), the partial-
occupancy metal-bridging solvent molecule is not shown for clarity; (b)
metal-bridging solvent molecule omitted from the structure factor calculation
(contoured at 5.1σ), AOH is not shown for clarity. (c) Selected intermo-
lecular interactions in the arginase I-AOH complex. Hydrogen bonds are
indicated by black dashed lines, and metal coordination interactions are
indicated by magenta dashed lines. Hydroxyl group O1 bridges Mn2+A and
Mn2+ and hydroxyl group O2 donates a hydrogen bond to E277.
B
trigonal planar aldehyde form, of AOH. The Mn2+A-Mn2+
B
separation is 3.3 Å; Mn2+A is coordinated by H101 (Nδ1), D124
(Oδ1), D232 (Oδ2), D128 (Oδ1), and AOH hydroxyl group
O1 with distorted square pyramidal geometry, while Mn2+ is
B
coordinated by D124 (Oδ2), H126 (Nδ1), D232 (Oδ1), D234
(Oδ1 and Oδ2), and AOH hydroxyl group O1 with distorted
octahedral geometry. Hydroxyl group O1 asymmetrically bridges
the metal cluster with Mn2+A-O and Mn2+B-O distances of
2.0 and 2.4 Å, respectively, and donates a hydrogen bond to
D128. Hydroxyl group O2 is within hydrogen-bonding distance
to E277 Oꢀ2 (2.8 Å) and the backbone carbonyl group of A141
(3.1 Å). The Mn2+A-O2 separation is 3.2 Å, which is too long
to be considered an inner-sphere metal coordination interaction.
All interactions between the hydrated AOH aldehyde and
arginase, apart from the interactions of hydroxyl group O2, are
(37) Lewis, C. A.; Wolfenden, R. Biochemistry 1977, 16, 4886-4890.
(38) Pocker, Y.; Sarkanen, S. AdV. Enzymol. 1978, 47, 149-274.
9
10282 J. AM. CHEM. SOC. VOL. 126, NO. 33, 2004

Amino Acid Aldehyde Inhibitors of Arginase
A R T I C L E S
Figure 3. Reactive substrate analogue inhibitors of arginase, AOH and 2(S)-amino-6-boronohexanoic acid (ABH), undergo a hydration reaction presumed
to mimic the first step of arginase catalysis, the hydration of the L-arginine guanidinium group. Since these inhibitors mimic the structures and the chemistry
of the first step of catalysis, they are designated “reaction coordinate analogue” inhibitors. The X-ray crystal structure of the arginase I-AOH complex
provides the first structural evidence that active-site residue E277 can stabilize the tetrahedral intermediate with a hydrogen-bond interaction. Note the
bidentate coordination interaction of the tetrahedral intermediate in catalysis with Mn2+A, which is mimicked by the binding of ABH but not AOH. This
interaction cannot occur in the precatalytic Michaelis complex since the sp3 lone electron pair on the former guanidinium nitrogen is fully developed only
when the tetrahedral intermediate is fully formed; in the substrate, this lone electron pair is delocalized in the “Y”-shaped guanidinium π system.
hydration of the reaction coordinate analogue AOH. On the other
hand, the Ki value of leucine aldehyde against leucine ami-
nopeptidase, a binuclear metalloprotease, is insensitive to
deuterium substitution in the aldehyde functionality, suggesting
that the preformed gem-diol form of the inhibitor is selected
for binding to the enzyme.17 Regardless of the locus of the AOH
hydration reaction, that AOH binds to the arginase I active site
exclusively as the tetrahedral gem-diol reflects the preferential
binding of a transition-state-like structure. This result implies
that transition-state stabilization by the binuclear manganese
cluster makes a significant contribution to arginase catalysis.
An important new result regarding transition-state binding is
evident in the arginase I-AOH complex: AOH hydroxyl group
O2 (which does not coordinate to metal) donates a hydrogen
bond to the carboxylate group of E277 (O‚‚‚O separation )
2.8 Å). This structural feature suggests that E277 participates
in transition-state stabilization just as it participates in substrate
L-arginine binding to a deactivated bacterial arginase.6 The role
of E277 in binding the substrate and tetrahedral intermediate
was first proposed based on analysis of the structure of native
arginase I.4 Interactions with E277 in arginase I complexes with
hydrated boronic acids12,13 are too long to be considered bona
fide hydrogen bonds (O‚‚‚O separations ) 3.3-3.4 Å), possibly
indicative of electrostatic repulsion between the negatively
charged E277 carboxylate group and the negatively charged
boronate anion of the inhibitor. Such electrostatic repulsion does
not occur between E277 and the neutral gem-diol of the AOH
hydrate, so the hydrogen bond between E277 and the tetrahedral
9
J. AM. CHEM. SOC. VOL. 126, NO. 33, 2004 10283

A R T I C L E S
Shin et al.
Concluding Remarks
That arginase I-AOH affinity is in the micromolar range is
consistent with structure-activity relationships outlined by
Mansuy and colleagues for Nꢀ-hydroxy-L-lysine (Ki ) 90 µM
at pH 9.0; Ki ) 4 µM at pH 7.4), which is isosteric with the
gem-diol form of AOH less one hydroxyl group.39 Although a
crystal structure of the arginase I-Nꢀ-hydroxy-L-lysine complex
is unavailable, the inhibitor hydroxyl group likely bridges Mn2+
A
and Mn2+B in a manner similar to that of the AOH O1 hydroxyl
group, making only a monodentate metal coordination interac-
tion. An overview of the binding modes of all tetrahedral
transition-state analogue inhibitors, including amino acid sul-
fonamides,10 boronic acids,11-13,22,40 and AOH (Figure 2),
suggests that the tightest binding inhibitors make bidentate
coordination interactions with the binuclear manganese clus-
ter: one atom of the tetrahedral “warhead” bridges Mn2+A and
Figure 4. Superposition of the arginase I-AOH complex (AOH has a
black carbon skeleton) and the binding conformation of ABH (light gray
carbon skeleton with pale green boron atom) in the arginase I-ABH
complex.
gem-diol is more easily formed. Importantly, the arginase
I-AOH complex provides the first view of a neutral tetrahedral
moiety that more faithfully mimics the neutral tetrahedral
intermediate in the arginase mechanism (Figure 3).
Mn2+B, and another atom coordinates to a second site on Mn2+
A
(Figure 3). Coordination to the second site on Mn2+A therefore
plays an important role in the stabilization of transition-state
analogue binding; likewise, coordination to the second site on
Given the general similarities in the binding modes of AOH
and its boronic acid isostere, 2(S)-amino-6-boronohexanoic acid
(ABH),11 as summarized in Figure 3, it is curious as to why
AOH binds 600-fold more weakly than ABH. Affinity differ-
ences may arise from two subtle differences in the interactions
of the tetrahedral functional group of each inhibitor. First,
superposition of the AOH and ABH binding conformations in
the arginase I active site (Figure 4) shows that although hydroxyl
group O1 of each inhibitor bridges Mn2+ and Mn2+ with
Mn2+ must play an important role in the stabilization of the
A
tetrahedral transition state in catalysis. With Mn2+A and Mn2+
B
responsible for activating the bridging hydroxide nucleophile,
Mn2+A responsible for bidentate coordination by the transition
state, and Mn2+B responsible for monodentate coordination by
the transition state, the structural and functional requirement
for two metal ions in arginase I catalysis is now clarified.
A
B
Mn2+‚‚‚O separations of 2.0-2.4 Å, the gem-diol of AOH is
Acknowledgment. We thank Prof. David Ash for scientific
discussions as well as the generous gift of recombinant rat
arginase I. This work was supported by National Institutes of
Health Grant GM49758 to D.W.C. E.C. was supported in part
by the U.S. Army Medical Research and Materiel Command’s
Office of the Congressionally Directed Medical Research
Programs through a Department of Defense Breast Cancer
Research Predoctoral Fellowship.
“tilted” relative to the boronate anion of ABH such that hydroxyl
group O2 of ABH coordinates to Mn2+ (2.4 Å) but hydroxyl
A
group O2 of AOH does not coordinate to Mn2+ (3.2 Å). In
A
other words, the boronate anion of ABH makes a favorable
bidentate coordination interaction with Mn2+ via hydroxyl
A
groups O1 and O2, whereas the gem-diol of AOH makes only
a monodentate coordination interaction with Mn2+ through
A
hydroxyl group O1. Second, the O3 hydroxyl group of ABH
makes a water-mediated hydrogen-bond interaction with T246,13
but a comparable interaction cannot occur in the arginase
I-AOH complex due to the presence of the aldehyde hydrogen
atom at the corresponding position of the gem-diol (Figure 3).
JA047788W
(39) Custot, J.; Boucher, J.-L.; Vadon, S.; Guedes, C.; Dijols, S.; Delaforge,
M.; Mansuy, D. J. Biol. Inorg. Chem. 1996, 1, 73-82.
(40) Collet, S.; Carreaux, F.; Boucher, J.-L.; Pethe, S.; Lepoivre, M.; Danion-
Bougot, R.; Danion, D. J. Chem. Soc., Perkin Trans. 2000, 1, 177-182.
9
10284 J. AM. CHEM. SOC. VOL. 126, NO. 33, 2004