Synthesis and evaluation of alternative substrates for arginase

Article abstract of DOI:10.1006/bioo.2001.1228

Two novel carboxyl-containing arginase substrates, 4-guanidino-3-nitrobenzoic acid and 4-guanidino-2-nitrophenylacetic acid, have been synthesized and found to give enhanced catalysis and dramatically lower K_m values relative to 1-nitro-3-guanidinobenzene, a substrate designed for use in a chromophoric arginase assay. To more efficiently mimic the natural substrate, a series of sulfur analogs of L-arginine were synthesized and kinetically characterized. The parent compound, L-thioarginine, with the bridging guanidinium nitrogen of L-arginine replaced with sulfur, functions as efficiently as the natural substrate. The desamino analog shows extremely low turnover, while the k_cat of the descarboxy analog is only 75-fold lower than that of arginine. These results suggest that the bridging nitrogen of L-arginine is not important for either substrate binding or catalysis, while the α-carboxyl group facilitates substrate binding, and the α-amino group is necessary for efficient catalysis. Isothiourea homologs previously reported to be nitric oxide synthase inhibitors have been found to undergo a rapid non-enzymatic rearrangement to a species that is probably the true inhibitor.

Full text of DOI:10.1006/bioo.2001.1228

Bioorganic Chemistry 30, 81–94 (2002)
doi:10.1006/bioo.2001.1228, available online at http://www.idealibrary.com on
Synthesis and Evaluation of Alternative Substrates for Arginase1
Shoufa Han, Roger A. Moore, and Ronald E. Viola2
Department of Chemistry, University of Toledo, Toledo, Ohio 43606
Received August 20, 2001
Two novel carboxyl-containing arginase substrates, 4-guanidino-3-nitrobenzoic acid and 4-
guanidino-2-nitrophenylacetic acid, have been synthesized and found to give enhanced catalysis
and dramatically lower Km values relative to 1-nitro-3-guanidinobenzene, a substrate designed
for use in a chromophoric arginase assay. To more efficiently mimic the natural substrate, a
series of sulfur analogs of
L
-arginine were synthesized and kinetically characterized. The parent
-arginine replaced with
compound, -thioarginine, with the bridging guanidinium nitrogen of
L
L
sulfur, functions as efficiently as the natural substrate. The desamino analog shows extremely
low turnover, while the kcat of the descarboxy analog is only 75-fold lower than that of arginine.
These results suggest that the bridging nitrogen of L-arginine is not important for either substrate
binding or catalysis, while the ␣-carboxyl group facilitates substrate binding, and the ␣-amino
group is necessary for efficient catalysis. Isothiourea homologs previously reported to be nitric
oxide synthase inhibitors have been found to undergo a rapid non-enzymatic rearrangement to
a species that is probably the true inhibitor.
᭧ 2002 Elsevier Science (USA)
Key Words: arginase; specificity; alternative substrates; thio analogs.
INTRODUCTION
Arginase is a ubiquitous enzyme that has been found in mammals (1), reptiles (2),
plants (3), fungi (4), and bacteria (5). In mammals, arginase is found in liver (1) and
also in nonheptatic tissues including red blood cells (6), lactating mammary gland,
and kidney (1). Arginase is a trimeric manganese metalloenzyme that catalyzes the
hydrolysis of
-ornithine. In nonheptatic tissues, arginase plays a critical role in the biosynthesis
of proline (7) and polyamines (8) by regulating the availability of -ornithine. Arginase
L-arginine in the final step of the urea cycle, releasing urea and
L
L
also regulates the biosynthesis of the cell-signaling molecule nitric oxide by affecting
arginine availability, since both arginase and nitric oxide synthase share arginine as
a common substrate (9). As a consequence of the reciprocal regulatory roles of
arginase and nitric oxide synthase, arginase inhibition has therapeutic potential in
treating nitric oxide-dependent smooth muscle disorders such as erectile dysfunction
1 This work was supported by a grant from the National Science Foundation (MCB0196103).
2 To whom correspondence and reprint requests should be addressed. Fax: 419-530-1583. E-mail:
ron.viola@utoledo.edu.
81
0045-2068/02 $35.00
᭧ 2002 Elsevier Science (USA)
All rights reserved.

82
HAN, MOORE, AND VIOLA
(10). A significant effort has been directed toward the evaluation of
as possible arginase inhibitors (11–15).
L-arginine analogs
A continuous chromophoric arginase assay was recently reported using 1-nitro-3-
guanidinobenzene (NGB) as an alternative substrate (16). The kcat with NGB is 5-
orders of magnitude lower than that of
L-arginine, but the Km is almost identical.
We recently reported a more sensitive spectrophotometric arginase assay using
L
-
thioarginine as an alternative substrate (17). Thioarginine, in which the bridging
guanidinium nitrogen is replaced with sulfur, functions as efficiently as arginine.
Therefore a series of NGB and thioarginine analogs were synthesized and evaluated
to probe the interactions of selected arginine moieties with rat liver arginase.
MATERIALS AND METHODS
1
General Methods and Materials. H- and 13C-NMR spectra were obtained for all
synthetic intermediates and products on a Varian VXRS400. Mass spectra (FAB)
were obtained on Brucker Esquire-LC. Rat liver arginase, expressed and purified
from Escherichia coli, was a gift from Dr. David Ash. N-Bromosuccinimide (NBS) was
recrystallized from water and dried over P2O5 in vacuo. 2-(S)-N-(t-butyloxycarbonyl)-
glutamic acid-t-butyl ester was purchased from Sigma. S-(3-aminopropyl)isothiourea
hydrobromide and S-(2-aminoethyl)isothiourea hydrobromide were obtained from
Alexis Biochemicals. All other reagents were purchased from either Aldrich or
Lancaster. Synthetic reactions were conducted in oven-dried glassware under argon and
all solvents were distilled prior to use. Flash column chromatography was performed on
230–400 mesh silica gel from Lancaster under positive argon pressure. Thin layer
chromatography (TLC) was performed on Kodak 13181 plates with fluorescence
indicator. Compounds were visualized by iodine chamber, UV light, or a ninhydrin
dip (0.1% ninhydrin in 95% n-butanol, 4.5% water, 0.5% glacial acetic acid). Protein
concentrations were determined by the method of Bradford (18), and free thiols were
quantitated by Ellman’s method (19).
Synthesis of 4-guanidino-3-nitrobenzoic acid (NGBA). NGBA was synthesized in
three steps through the coupling of N-protected pseudothiourea with 4-amino-3-
nitrobenzoate (Scheme 1).
N,NЈ-bis-t-boc-2-methyl-2-thiopseudourea (1). A modification of the Bergeron
method (20) was used for preparation of 1 and 2. Di-t-butyldicarbonate (Boc2O) (23.2
g, 106.3 mmol) and 2-methyl-2-thiopseudourea (6.2 g, 44.5 mmol) were stirred for
5 days at rt in 100 ml of 1:1/CH2Cl2:saturated NaHCO3. Extraction with dichlorometh-
ane (100 ml) and purification by flash column chromatography (15% hexanes in
CHCl3, then CHCl3) gave 1 (10.0 g, 35.6 mmol) in 80% yield.
4-(N,NЈ-bis-t-boc-guanidino)-3-nitrobenoic acid (2). 4-amino-3-nitrobenzoic acid
(1.8 g, 10.0 mmol), silver nitrate (1.8 g, 10.6 mmol), 1 (2.5 g, 10.0 mmol), and
triethylamine (TEA) (1.5 ml, 10.8 mmol) were stirred at rt overnight in dichlorometh-
ane (250 ml). After filtration and chromatography a yellow solid (1.5 g, 4.0 mmol,
43% yield) was obtained using 10:1 hexanes:ethyl acetate. Rf ϭ 0.58 (2:1; ethyl ace-
tate:hexanes).
NGBA. A 10 mL solution of 1.0 M BCl3 in dichloromethane was introduced
dropwise at rt to a stirring solution of 2 (1.5 g, 4.0 mmol) in dichloromethane (40
ml), leading to the formation of a precipitate. The majority of BCl3 and CH2Cl2 was

ALTERNATIVE ARGINASE SUBSTRATES
83
SCHEME 1. Synthesis of NGBA.
evaporated by a stream of argon and concentrated to a yellow solid which was treated
with anhydrous methanol (10 ml). Excess methanol and the resulting B(OCH3)3 were
removed in vacuo. Chromatography using 2:1 ethyl acetate:hexanes (Rf ϭ 0.18)
1
yielded NGBA (0.7g, 2.7mmol) in 67% yield. mp: 192–194ЊC; H-NMR (DMSO):
␦ 6.99 (d, 1H, J ϭ 9.2 Hz), 7.23 (s, broad, 1H), 7.75 (s, broad, 2H), 7.85 (dd, 1H,
J ϭ 8.4, 2.0 Hz), 7.93 (s, broad, 1H), 8.55 (d, 1H, J ϭ 2.0 Hz). 13C-NMR (DMSO):
␦ 173.37, 166.93, 148.40, 134.81, 130.32, 126.50, 121.96, 119.43; Elemental analysis,
calculated (%): C: 36.87, H: 3.48, N: 21.50; found: C: 36.29, H: 3.80, N: 21.06; MS:
+
(parent ion) C8H9N4O4 ; calculated: 225.2; found: 224.8.
Synthesis of 4-guanidino-2-nitrophenylacetic acid (NGPA). NGPA was synthesized
by the oxidation of t-boc-protected 4-amino-2-nitrotoluene to the corresponding ben-
zylacetic acid, followed by coupling to protected pseudoisothiourea (Scheme 2).
4-(N,N-bis-t-boc)-amino-2-nitrotoluene (3). 4-amino-2-nitrotoluene (2.8 g, 18.4
mmol), Boc2O (9.0 g, 41.2 mmol), TEA (6.0 ml, 43.1 mmol) and 4-dimethylaminopyri-
dine (DMAP) (0.12 g, 0.98 mmol) were dissolved in acetonitrile (80 ml), stirred at
rt for 20 h, and purified by chromatography (10:1 hexanes:ethyl acetate, Rf ϭ 0.41)
producing 3 (4.3 g 12.2 mmol) in 65% yield.
(N,N-bis-t-boc)-4-amino-2-nitrophenylbromomethane (4). A mixture of 3 (4.3 g,
12.2 mmol), NBS (3.0 g, 17.0 mmol), and 2,2Ј-azobisisobutylnitrile (AIBN) (0.4 mg,
2.4 ␮mol) in CCl4 (200 ml) were heated under reflux under a 100 W bulb. After 20
h, concentration and chromatography (10:1 petroleum ether:ethyl acetate, Rf ϭ 0.56)
gave 4 as a white crystalline solid (2.0 g, 4.6 mmol, 40% yield).
4-(N,N-di-t-boc)-amino-2-nitrophenylacetonitrile (5). Potassium cyanide (0.6 g,
9.2 mmol), 4 (2.0 g, 4.6 mmol), water (10 ml), dichloromethane (100 ml), and
triethylammonium sulfate (0.3 g, 1.5 mmol) were stirred at rt overnight. Extraction

84
HAN, MOORE, AND VIOLA
SCHEME 2. Synthesis of NGPA.
with dichloromethane and chromatography (4:1 hexanes:ethyl acetate, Rf ϭ 0.11)
afforded 5 as a pale yellow solid (1.4 g, 3.7 mmol) in 80% yield.
4-amino-2-nitrophenylacetic acid (6). Deprotection of the amino group was
achieved by adding concentrated HCl (100 ml) to 5 (1.4 g, 3.7 mmol), heating under
reflux for 20 h, then concentrating in vacuo. Chromatography (1:1 hexanes:ethyl
acetate, then 3:3:1 hexanes:ethyl acetate:methanol, Rf ϭ 0.22) afforded 6 (0.50 g, 2.6
mmol) in 70% yield.
4-(N,NЈ-t-boc-guanidino)-2-nitrophenylacetic acid (7). N,NЈ-bis-t-boc-2-methyl-
2-thio-pseudourea (1.0 g, 4.0 mmol), 6 (0.5 g, 2.55 mmol), and AgNO3 (1.2 g, 7.1
mmol) were stirred in anhydrous acetonitrile (100 ml) for 12 h, filtered, and then
concentrated. After chromatography (4:1 hexanes:ethyl acetate, then 3:3:1 hex-
anes:ethyl acetate:methanol, Rf ϭ 0.45), 7 (0.6 g, 1.4 mmol) was obtained as a white
solid in 54% yield.
NGPA. Compound 7 (0.6 g, 1.4 mmol) was deprotected with 1.0 M BCl3 in
dichloromethane as described above for NGBA. NGPA (0.2 g, 0.73 mmol, 52%) was
obtained as a white hygroscopic solid after chromatography with 1:1:1 hexanes:ethyl
1
acetate:methanol, then methanol (Rf ϭ 0.18); mp: 162–164ЊC; H-NMR ␦: 3.68 (s,
2H), 4.14 (s, broad, 1H), 7.26 (d, 1H, J ϭ 8.4 Hz), 7.37 (d, 1H, J ϭ 8.4 Hz), 7.58
(s, 1H), 7.98 (broad, 4H); 13C-NMR ␦: 174.35, 156.76, 150.49, 134.27, 130.78, 129.11,
128.21, 125.52, 119.60; elemental analysis, calculated (%): C: 39.36, H: 4.04, N:
+
20.40; found (%): C: 39.10, H: 4.14, N: 19.28; MS: (parent ion) C9H11N4O4 ; calcu-
lated: 239.2, found: 239.0.
Synthesis of 2-amino-5-isothioureidovaleric acid (
L
-thioarginine).
L-Thioarginine
was synthesized by coupling pseudothiourea to protected
L
-glutamic acid (Scheme 3).
2(S)-N-t-boc-5-hydroxy-t-butylpentanoate (8). Compound 8 was prepared by modi-
fication of a reported procedure (11). To a solution of 2-(S)-N-t-boc-glutamic acid
␣-t-butyl ester (6.6 g, 21.8 mmol) at Ϫ5ЊC in anhydrous THF (600 ml) was added

ALTERNATIVE ARGINASE SUBSTRATES
85
SCHEME 3. Synthesis of L-thioarginine.
TEA (24.0 ml, 237 mmol) and ethyl chloroformate (22.5 ml, 234 mmol), the mixture
was stirred at Ϫ5ЊC for 20 min and then filtered. Water (25 ml) and then sodium
borohydride (6.0 g, 158 mmol) were added to the filtrate and stirred overnight at rt.
The supernatant was concentrated to a colorless oil and purified by column chromatog-
raphy (2:1 hexanes:ethyl acetate) to afford 8 (5.0 g, 17.3 mmol) as a colorless oil in
80% yield.
2(S)-N-t-boc-5-mesyloxy-t-butylpentanoate (9). To a solution of 8 (10.0 g, 34.6
mmol), at 0ЊC was added methanesulfonyl chloride (6.0 g, 953 mmol) and TEA (9.0
g, 81.9 mmol) in CH2Cl2 (350 ml). The reaction mixture was stirred on ice for an
additional 30 min and then at rt for 3 h. The mixture was washed successively with
4 M HCl (200 ml) and 50% saturated NaHCO3 (200 ml). The organic layer was
dried over Na2SO4, concentrated, and purified by chromatography (4:1 hexanes:ethyl
acetate, Rf ϭ 0.55), giving 9 (9.2 g, 25.0 mmol) in 73% yield.
2-(S)-N-t-boc-5-thioguanidino-t-butylpentanoate (10). Thiourea (4.0 g, 52.6 mmol)
and 9 (9.2 g, 25.0 mmol) were dissolved in acetone (250 ml) and heated overnight
under reflux. The concentrate was resuspended in CH2Cl2 (200 ml) and filtered. The
filtrate was concentrated and purified by chromatography (2:1 hexanes:ethyl acetate,
then 1:4:2 methanol:hexanes:ethyl acetate) yielding 10 (5.2 g, 11.7 mmol) in 47%
yield.
2-amino-5-isothioureidovaleric acid (
mmol) was deprotected as described above for NGBA. Chromatography (ethyl acetate,
then methanol) afforded -thioarginine (2.2 g, 9.6 mmol) as a white solid in 82%
L-thioarginine). Compound 10 (5.2 g, 11.7
L
yield. 1H-NMR (DMSO): ␦ 9.33 (s, 4H), 8.56 (s, 3H), 3.90 (t, 1H, J ϭ 6.4 Hz), 3.20
(t, 2H, J ϭ 6.4 Hz), 1.80 (m, 4H); 13C-NMR (DMSO): 170.61, 169.92, 51.31, 29.33,
28.64, 24.49; Elemental analysis, calculated (%): C: 31.23, H: 6.11, N: 18.21, S:
13.89; found (%): C: 31.09, H: 6.00, N: 17.96, S: 13.83 (these values are corrected
for a 7.6% ash content); MS: C6H14N3O2S+: calculated: 192.26, found: 192.10.
Synthesis of 5-thioguanidinovaleric acid. 5-thioguanidinovaleric acid was prepared
by modifications to the previously reported synthesis (21). Thiourea (0.6 g, 7.9 mmol)
and 5-bromovaleric acid (3.0 g, 16.6 mmol) were heated under reflux overnight in
acetone (150 ml). The white precipitate (5-thioguanidinovaleric acid) was washed
with hot acetone (150 ml) and isolated in 90% yield (1.8 g, 7.0 mmol). mp; 138–140ЊC

86
HAN, MOORE, AND VIOLA
1H-NMR (DMSO): ␦ 8.99 (broad, s, 4H), 3.14 (t, 2H, J ϭ 6.8 Hz), 1.58 (m, 4H),
2.24 (t, 2H, J ϭ 6.8 Hz); 13C-NMR (DMSO): ␦ 174.11, 169.85, 32.88, 29.82, 27.83,
23.21; MS: C6H13N2S+: calculated: 177.2; found: 177.0; Elemental analysis:
C6H13N2SBr; calculated (%): C: 28.03, H: 5.09, N: 10.89, S: 12.47; found (%): C:
28.21, H: 5.03, N: 10.69, S: 12.81.
Synthesis of S-(4-aminobutyl)isothiourea. This S-substituted isothiourea was syn-
thesized by reacting protected 4-aminobutanol with thiourea (Scheme 4).
4-N-t-boc-amino-1-butanol (11). 4-amino-1-butanol (4.4 g, 49.4 mmol), Boc2O
(10.8 g, 49.5 mmol), and TEA (9.5 ml, 68.2 mmol) were stirred in acetonitrile (250
ml) overnight at rt and then concentrated. The colorless residue was dissolved in
ethyl acetate (250 ml), washed successively with 200 ml each of saturated NaHCO3,
3.0 M HCl, and water, and finally dried over Na2SO4. A pale yellow sticky liquid
was obtained in 96% yield (9.0 g, 47.6 mmol).
4-N-t-boc-amino-1-butanoxyphenyl sulfoxide (12). To a solution of 11 (9.0 g, 47.6
mmol) in CH2Cl2 (300 ml) at 0ЊC was added benzenesulfonyl chloride (15.5 g, 87.5
mmol) and TEA (13.0 ml, 87.5 mmol) in CH2Cl2 (200 ml). This reaction was worked
up as described for the synthesis of 9. Chromatography (4:1 hexanes:ethyl acetate,
Rf ϭ 0.49; then 3:1 ethyl acetate: methanol) gave 12 (10.5 g, 31.8 mmol) in 67% yield.
S-(4-N-t-boc-aminobutyl)isothiourea (13). Thiourea (2.4 g, 31.6 mmol) and 12
(10.0 g, 30.4 mmol) were dissolved in acetone (250 ml) and heated under reflux for
30 h and then concentrated. The residue was resuspended in CH2Cl2 (200 ml) and
filtered. The filtrate was concentrated and purified by chromatography (2:1 hex-
anes:ethyl acetate, then methanol) yielding 13 in 46% yield (5.7 g, 14.1 mmol).
S-(4-aminobutyl)isothiourea. To a solution of 13 (4.3 g, 10.6 mmol) in CH2Cl2
(150 ml) at rt was added 1.0 M of BCl3 in CH2Cl2 (15 ml) dropwise. A white
precipitate formed upon the addition of BCl3. The reaction mixture was further stirred
for 2 h. The precipitate was resuspended in 2% methanol/acetonitrile (50 ml), filtered,
1
and washed with acetone to afford the product in 52% yield (1.2 g, 5.5 mmol); H-
NMR (DMSO): ␦ 9.31 (s, 4H), 8.18 (s, 3H), 3.19 (t, 2H, J ϭ 6.4 Hz), 2.75 (q, 2H,
J ϭ 5.2 Hz), 1.63 (m, 4H); 13C-NMR (DMSO): 170.02, 37.93, 29.42, 25.77, 25.65;
Elemental analysis: C5H15N3SCl2; Calculated (%): C: 27.28, H: 6.87, N: 19.09; S:
14.56; Found (%): C: 27.47, H: 6.90, N: 18.96, S: 14.45.
Rearrangement of S-(2-aminoethyl)isothiourea. S-(2-aminoethyl)isothiourea (0.05
g, 0.18 mmol) was added to 25 ml of degassed 50 mM phosphate buffer, pH 7.4.
SCHEME 4. Synthesis of S-(4-aminobutyl)isothiourea.

ALTERNATIVE ARGINASE SUBSTRATES
87
The reaction mixture was stirred at rt under argon for 15 min and then concentrated
to afford a white residue. The product was redissolved in CD3OD and characterized
1
1
by H and 13C NMR. H-NMR (CD3OD): ␦ 3.23 (t, 2H, J ϭ 6.8 Hz), 2.56 (t, 2H,
J ϭ 6.8 Hz); 13C NMR (CD3OD): ␦ 158.73, 45.99, 24.44.
Arginase assay with chromophoric products. Methanolic stock solutions of NGBA
and NGPA were prepared prior to use. Aliquots of arginase solution (24.6 ␮g) were
added to cuvettes containing 50 mM bicine, pH 9.0, 100 ␮M MnCl2, and 0.02–2.5
mM concentrations of NGBA or NGPA to a final volume of 1.0 ml. Reaction velocities
were measured spectrophotometrically for the hydrolysis of NGBA (⌬E410 ϭ 1,347
MϪ1 cmϪ1) or NGPA (⌬E380 ϭ 545 MϪ1 cmϪ1). Control assays performed in the
absence of arginase revealed no detectable hydrolysis of either NGBA or NGPA in
the time scale used.
Coupled assay for thioarginine analogs. Stock solutions of selected L-arginine
analogs in water (30 mM) and DTNB in methanol (10 mM) were prepared prior to
use. Arginase (60 ␮l, 42 ␮g) and 10 mM DTNB (20 ␮l) were added to cuvettes
containing 50 mM bicine, pH 9.0, with 0.3–6.0 mM concentrations of thioarginine
analogs to a final volume of 1.0 ml. Reaction velocities were measured by observing
the formation of 2-nitro-5-thiobenzoate (TNB) at 412 nm. Controls were run without
arginase to assess the extent of nonenzymatic hydrolysis. Previous work (17) had
shown that DTNB hydrolysis at higher pH is negligible for the duration of the assay.
For the slowest substrates, particularly 5-thioguanidinovalerate, the nonenzymatic
hydrolysis of DTNB was significant, so a fixed point assay was used for these
substrates.
Fixed-point assay for thioarginine analogs. The reaction was initiated by adding
arginase (0.41 mg) to a mixture containing 0.6 to 4.5 mM 5-thioguanidinovaleric acid
in 50 mM Hepes, pH 8.0. At defined intervals aliquots of the reaction mixture were
quenched with 1.0 M borate (pH 6.0) to terminate the reaction and adjust the pH to
7.0–7.5. Arginase was removed by ultrafiltration (Amicon) and the thiol concentration
in the filtrate was quantitated with 10 ␮l of 10 mM DTNB in methanol. The change
in absorbance was measured and product formation quantitated at 412 nm using an
extinction coefficient of 1.36 ϫ 104 MϪ1 cmϪ1 (19).
RESULTS AND DISCUSSION
Synthesis of NGB Analogs. 1-Nitro-3-guanidinobenzene (NGB) has recently been
reported as an alternative arginase substrate in a continuous spectrophotometric assay
(16). Unfortunately, NGB is a poor substrate, with a very low kcat value (0.09 minϪ1)
relative to L
-arginine (15,000 minϪ1). However, the Km of NGB (1.6 mM) is indistin-
guishable from that of arginine; an unexpected observation considering the significant
structural differences between these two substrates (Fig. 1). Arg21 of arginase was
initially implicated in substrate binding via an electrostatic interaction with the car-
boxyl group of arginine (22). More detailed structural studies on the binding of the
boronate-containing inhibitor-2-amino-6-boronohexanoic acid (10), and on the product
complex with ornithine and urea (23), have identified Asn130 and Ser137 as hydrogen-
bond donors to the substrate carboxyl group. Since NGB lacks this carboxyl group
it probably binds to arginase primarily by its guanidinium group interacting with
Glu277 similar to the way that the guanidinium group of arginine has been shown

88
HAN, MOORE, AND VIOLA
FIG. 1. Structural comparison of NGB analogs to L-arginine.
to bind (23). Several carboxyl-containing analogs of NGB have been synthesized to
probe this enzyme-substrate interaction in an attempt to design improved alternative
substrates of arginase. NGBA (Fig.1) was designed with a carboxyl group attached
to the benzene ring, while NGPA was prepared with a methylene spacer between the
benzene ring and the carboxyl group.
Both NGBA (Scheme 1) and NGPA (Scheme 2) were synthesized in good overall
yield by the coupling of a t-boc-protected pseudothiourea to the corresponding nitroani-
linic acid. Both products were obtained by deprotection with BCl3 in dichloromethane.
Although efficient removal of the protecting group was achieved, BCl3 interacts to
some extent with the unmasked guanidinium group. Thus, treatment of the crude
products with methanol was essential to completely remove BCl3 by forming trimethyl
borate, which was removed in vacuo or by washing with acetone. Without this
treatment, the residual BCl3 forms boric acid, a known arginase inhibitor, under the
assay conditions (24).
Kinetics of NGB Analogs. Enzyme-catalyzed hydrolysis of NGBA and NGPA
releases chromophoric nitroaromatic acids whose formation can be observed spectro-
photometrically. As predicted, the kcat values for both analogs are enhanced relative
to NGB, with a 40-fold increase observed for NGBA and a 60-fold increase for NGPA
(Table 1). Although considerably enhanced, these values are still more than 1000-
fold lower than that of arginine. Interestingly, the Km values of NGBA and NGPA
(7 and 10 ␮M, respectively) are significantly reduced and are now approximately
200-fold lower than those obtained for both NGB and for the physiological substrate
arginine. Incorporation of a carboxyl group thus results in a 4-order of magnitude
enhancement in kcat /Km for these substrates relative to NGB. These data support
the hypothesis that selective placement of a carboxyl group on NGB contributes to
tighter binding and enhanced catalysis by keeping the substrate in a more produc-
tive orientation.
Synthesis of thioarginine analogs. To further examine the interactions between
arginine and arginase, a series of thio analogs of arginine were synthesized and tested
for their ability to function as arginase substrates. Each analog contains a sulfur that
replaces the bridging guanidinium nitrogen. L-Thioarginine differs from L-arginine

ALTERNATIVE ARGINASE SUBSTRATES
89
TABLE 1
Kinetic Parameters of Arginase Substrates
Extinction
kcat /Km
Chromophore
measured
coefficient
Substrates
-argininea
kcat (minϪ1
)
Km (mM)
(MϪ1 sϪ1
)
(MϪ1 cmϪ1
)
L
15,000 Ϯ 1,200
0.09 Ϯ 0.02
1.4 Ϯ 0.3
1.6 Ϯ 0.2
1.8 ϫ 105 noneb
0.94 1-amino-3-
nitrobenzene
0.007 Ϯ 0.002 8.3 ϫ 103 4-amino-3-
—
1,280
NGBc
NGBA
NGPA
3.5 Ϯ 0.2
5.4 Ϯ 0.2
1,350
545
nitrobenzoate
0.010 Ϯ 0.002 9.0 ϫ 103 4-amino-2-nitro-
phenylacetate
L
-thioarginine
18,000 Ϯ 1,700
0.8 Ϯ 0.2
0.5 Ϯ 0.1
2.1 Ϯ 0.6
4.0 Ϯ 0.5
6.0 ϫ 105 2-nitro-5-
13,600
13,600
13,600
thiobenzoate
2-nitro-5-
thiobenzoate
8.3 ϫ 102 2-nitro-5-
thiobenzoate
5-thioguanidino-
valeric acid
S-(4-amino-
6.3
200 Ϯ 15
butyl)isothiourea
a Kinetic parameters determined by Cavalli et al. (31).
b Reaction measured by the release of 14C-labeled urea (32).
c Kinetic parameters determined by Baggio et al. (16).
only in the substitution of that bridging nitrogen by sulfur (Fig. 2). 5-Thioguanidinoval-
eric acid lacks the ␣-amino group and S-(4-aminobutyl)isothiourea is missing the
␣-carboxylic group. These compounds were prepared to test the role of the bridging
nitrogen, the ␣-amino, and the ␣-carboxylic group in substrate recognition. Two
isothiourea homologs, S-(3-aminopropyl)isothiourea and S-(2-aminoethyl)isothiourea,
were also examined as potential arginase substrates to evaluate the relative importance
of the separation between the functional ends of arginine in binding and catalysis.
Stability of the thioarginine analogs. While the arginase-catalyzed hydrolysis of
the NGB analogs can be followed by direct observation of the chromophoric products,
the enzyme-catalyzed hydrolyses of the selected thioarginine analogs were examined
by a coupled assay using DTNB to quantitate the release of a thiol-containing product
(17). Given the lack of specificity of this reaction it is essential that these analogs
be stable under the assay conditions. Thioarginine and 5-guanidinothiovalerate both
FIG. 2. Structural comparison of thioarginine analogs.

90
HAN, MOORE, AND VIOLA
display only minimal production of thiol in the absence of arginase (Fig. 3). However,
an appreciable nonenzymic rate of thiol release is observed in the DTNB-coupled assay
with the thioisoureas. The rate of thiol production from S-(4-aminobutyl)isothiourea is
20-fold greater than that observed with 5-thioguanidinovaleric acid (Fig. 3), while
the breakdown of the corresponding aminopropyl and aminoethyl homologs are so
rapid that the reaction is complete before a rate can be measured.
Given the similarity of these structures to thioarginine it seems unlikely that thiol
production results from hydrolysis of the thioguanidino bond. For these less stable
isothioureas there is the possibility of an internal rearrangement in which attack of
the amino group on the guanidine carbon leads to cleavage of the C–S bond and
production of a thiol-containing product (Scheme 5). The thiols generated by this
rearrangement react with DTNB in a manner that is indistinguishable from thiols
produced by enzymatic hydrolysis. The nonenzymatically produced thiol-containing
product was characterized by 1H and 13C NMR and is consistent with the rearrangment
product shown in Scheme 5. There is no NMR evidence for the direct hydrolysis
products, urea and 2-aminoethanethiol. When the rearrangement reaction is exposed
to air a second set of proton NMR peaks, consistent with oxidation of the thiol
rearrangement product, is observed. Conducting the reaction in an inert atmosphere
dramatically decreases the amount of thiol oxidation product.
Since the rearrangement of the aminoethyl and aminopropyl isothioureas would
proceed via five- or six-membered cyclic intermediates, respectively, the reaction
FIG. 3. Rates of nonenzymatic release of thiols. Samples were incubated in 50 mM Hepes, pH 8.0.
Aliquots were removed at defined times, quenched by treatment with 1.0 M borate and then quantitated by
the addition of DTNB. (•) 5-guanidinothiovaleric acid; (⅙) L-thioarginine; (᭢) S-(4-aminobutyl)isothiourea.

ALTERNATIVE ARGINASE SUBSTRATES
91
SCHEME 5. Rearrangement of S-(3-aminoethyl)isothiourea under basic conditions.
with these compounds is expected to be more facile than that with the aminobutyl
homolog which would require a seven-membered cyclic intermediate (25). The in-
creased stability of thioarginine, which could also rearrange via a seven-membered
cyclic intermediate, is due to the presence of the neighboring carboxyl group that
moderates the nucleophilicity of the attacking amino group.
S-(3-aminopropyl)isothiourea and S-(2-aminoethyl)isothiourea have been reported
to be potent inhibitors of nitric oxide synthases (26) and weak inhibitors of human
type II arginase (27). Based on the observed rearrangements with these compounds
it is likely that the resulting thiols, rather than the parent amines, are the actual
inhibitory species.
Kinetics of thioarginine analogs. Enzymatic hydrolysis of the thioarginine analogs
releases sulfhydryl-containing products that can be quantified with DTNB. Thioargin-
ine and S-(4-aminobutyl)isothiourea were examined in a continuous coupled assay,
with the enzyme concentration adjusted to compensate for the lower catalytic rate
with the latter compound (Table 1). Under these conditions the rate of the nonenzymatic
rearrangement of S-(4-aminobutyl)isothiourea is less than 10% of the arginase-cata-
lyzed hydrolysis. Although significantly more stable, attempts to use 5-thioguanidino-
valeric acid in this continuous assay were unsuccessful due to the extended durations
required, even when assayed at higher enzyme concentrations. Buffer-catalyzed
hydrolysis of DTNB during this time liberates the chromophore TNB, which interferes
with the arginase-catalyzed reaction. Instead, the fixed-point assay described above
was used to measure the enzyme-catalyzed hydrolysis of 5-thioguanidinovaleric acid.
A linear increase in absorbance was observed over at least 80 min in the full-time
course of this reaction and nonenzymatic hydrolysis under these conditions was less
than 1% of the enzymatic rate.
Role of the bridging nitrogen in L-arginine. The mechanism of arginase catalysis
has been proposed to involve formation of a tetrahedral intermediate after attack of
the guanidinium carbon by a (Mn+2)2-stabilized hydroxide ion, with subsequent expul-
sion of the bridging nitrogen to form urea and L-ornithine (22,28). Replacement of
the bridging guanidinium nitrogen of arginine with sulfur gave an alternative substrate
with kinetic properties essentially identical to those of the physiological substrate.
Thus, the bridging nitrogen in arginine does not appear to play a significant role in
either substrate binding or in catalysis. Since sulfur is typically a better leaving group
than nitrogen, collapse of the tetrahedral intermediate should proceed more rapidly
from the thio analog. The identical kinetics between thioarginine and arginine provide
supporting evidence that C–N bond cleavage is not rate-limiting step in arginase
catalysis.

92
HAN, MOORE, AND VIOLA
Role of the ␣-carboxyl and ␣-amino groups. The active site of arginase provides
a well-coordinated balance of binding interactions that take place at each end of the
substrate. Examination of a series of boronate analogs of -arginine demonstrated the
L
critical relationship between ␣-amino to boron distance and IC50 values (12). The
substrate carboxyl group makes important interactions with active site serine and
asparagine residues that contribute to substrate binding (10,23). The arginase inhibitor,
␻
N -hydroxy-
L
-arginine (NOHA) contains an N-hydroxyl group that displaces the
metal-bridged hydroxide and moves the inhibitor closer to the enzyme-bound metal
ions (23). This displacement disrupts interactions between the carboxyl end of the
inhibitor and the enzyme active site binding groups. The affinity of the descarboxy
analog of NOHA (Fig. 4) is 50-fold weaker than that of NOHA for arginase, and the
corresponding descarboxy analog of nor-NOHA is a 300-fold less potent inhibitor
(23). Despite the absence of a carboxyl moiety the alternative substrate NGB has a
comparable Km to the physiological substrate arginine. However, the addition of a
carboxyl group to NGB results in a 200-fold decrease in Km. This improved affinity
is observed both for NGBA, which contains the same number of bonds between the
bridging-nitrogen and the carboxyl group, and NGPA which is one methylene group
longer than arginine.
The Km of 5-thioguanidinovaleric acid, which lacks the ␣-amino group of thioargi-
nine, is nearly the same as that of thioarginine and arginine. However, its kcat of 0.8
minϪ1 is more than 20,000 times slower than that of thioarginine. Alternatively,
S-(4-aminobutyl)isothiourea, in which the carboxyl group is replaced by an amine
has a kcat of 200 minϪ1, 250-fold higher than that of 5-thioguanidinovaleric acid,
suggesting that the ␣-amino group plays an essential role in catalysis. Interactions
between the substrate amino group and binding groups on arginase (Asp183 and
Glu186) may place the scissile guanidinium group in a more productive orientation
relative to the metal-bridging hydroxide, resulting in more efficient catalysis.
Implications for inhibitor design. The substrate specificity of arginase is quite
broad, accepting a range of structural alterations while retaining catalytic activity.
However, the positioning and orientation between the metal-bridged hydroxide and
the guanidinium carbon of the substrate is the one essential interaction that must be
maintained. Boronate analogs such as 2-amino-6-boronohexanoic acid (11) and S-(2-
boronoethyl)-L-cysteine (29) function as transition state analog inhibitors, either by
binding of the protonated boronate to displace the metal-bridged hydroxide or binding
of the unprotonated trigonal species followed by attack of the metal-hydroxide to
␻
generate the tetrahedral intermediate (27). N -Hydroxy-nor-
L
-arginine (nor-NOHA),
a very potent inhibitor of arginase (30), also functions by displacing the metal-bridging
FIG. 4. Structures of nor-N-hydroxy-L-arginine analogs.

ALTERNATIVE ARGINASE SUBSTRATES
93
␻
hydroxide ion of arginase with its N -hydroxy group upon binding (23,30). An analog
of nor-NOHA, descarboxy-nor-NOHA (Fig. 4), is almost 20-fold more potent than
des-(␣-amino)-nor-NOHA as an arginase inhibitor (23). Binding of the amino group
of descarboxy-nor-NOHA places the N-hydroxyl moiety closer to the metal-bridging
hydroxide ion than does des-(␣-amino)-nor-NOHA, thus positioning the N-hydroxy
group to replace the bridging hydroxide ion. This observation supports the essential
role of the ␣-amino group in substrate orientation and catalysis.
ACKNOWLEDGMENTS
The authors thank Dr. David Ash (Temple University) for his generous gift of rat liver arginase and
Dr. Jun Hu and Yulin Wu (University of Akron) for helpful discussions on analog synthesis.
REFERENCES
1. Herzfeld, A., and Raper, S. M. (1976) Biochem. J. 153, 469–478.
2. Brown, G. W., Jr., and Cohen, P. P. (1960) Biochem. J. 75, 82–91.
3. Hellerman, L., and Perkins, M. E. (1935) J. Biol. Chem. 112, 175–194.
4. Davis, R. H. (1986) Microbiol. Rev. 50, 280–313.
5. Cunin, R., Glansdorff, N., Pierard, A., and Stalon, V. (1986) Microbiol. Rev. 50, 314–352.
6. Kedra-luboinska, M., Zamecka, E., and Porembska, Z. (1988) Biochem. Med. Metab. Biol. 39, 247–257.
7. Yip, M. C. M., and Knox W. E. (1972) Biochem. J. 127, 893–899.
8. Weiner, C. P., Knowles, R. G., Stegink, L. D., Dawson, J., and Moncada, S. (1996) Am. J. Obstet.
Gynecol. 174, 779–782.
9. Corraliza, I. M., Soler, G., Eichmann, K., and Modolell, M. (1995) Biochem. Biophys. Res. Commun.
206, 667–673.
10. Cox, J. D., Kim, N. N., Traish, A. W., and Christianson, D. W. (1999) Nature Struct. Biol. 6, 1043–1047.
11. Baggio, R., Elbaum, D., Kanyo, Z. F., Carroll, P. J., Cavalli, R. C., Ash, D. E., and Christianson,
D. W. (1997) J. Am. Chem. Soc. 119, 8107–8108.
12. Collet, S., Carreaux, F., Boucher, J-L., Pethe, S., Lepoivre, M., Danion-Bougot, R., and Danion, D.
(2000) J. Chem. Soc., Perkin Trans. 1, 177–182.
13. Narayanan, K., and Griffith, O. W. (1994) J. Med. Chem. 37, 885–887.
14. Daghigh, F., Fukuto, J. M., and Ash, D. E. (1994). Biochem. Biophys. Res. Commun. 202, 174–180.
15. Custot, J., Boucher, J-L., Vadon, S., Guedes, C., Dijols, S., Delaforge, M., and Mansuy, D. (1996)
J. Biol. Inorg. Chem. 1, 73–82.
16. Baggio, R., Cox, J. D., Harper, S. L., Speicher, D. W., and Christianson, D. W. (1999) Anal. Biochem.
276, 251–153.
17. Han, S., and Viola, R. E. (2001) Anal. Biochem. 295, 117–119.
18. Bradford, M. (1976) Anal. Biochem. 72, 248–253.
19. Ellman, G. L. (1959) Arch. Biochem. Biophys. 82, 70–77.
20. Bergeron, R. J., and McManis, J. S. (1987) J. Org. Chem. 53, 1700–1703.
21. Moore, E. E., and Rapala, R. T. (1947) J. Am. Chem. Soc. 69, 266.
22. Kanyo, Z. F., Scolnick, L., R., Ash, D. E., and Christianson, D. W. (1996) Nature 383, 554–557.
23. Cox, J. D., Cama, E., Colleluori, D. M., Pethe, S., Boucher, J-L., Mansuy, D., Ash, D. E., and
Christianson, D. W. (2001) Biochemistry 40, 2689–2701.
24. Pace, C. N., and Landers, R. A. (1981) Biochem. Biophys. Acta. 658, 410–412.
25. Doherty, D. G., Shapira R., and Burnett, W. T. (1957) J. Am. Chem. Soc. 79, 5567–5571.
26. Garvey, E. P., Oplinger, J. A., Tanoury, G. J., Sherman, P. A., Fowler, M., Marshall, S., Harmon,
M. F., Paith, J. E., Furfine, E. S. (1994) J. Biol. Chem. 269, 174–180.
27. Colleluori, D. M., and Ash, D. E. (2001) Biochemistry 40, 9356–9362.
28. Christianson, D. W., and Cox, J. D. (1999) Annu. Rev. Biochem. 68, 33–57.
29. Kim, N. N., Cox, J. D., Baggio, R. F., Emig, F. A., Mistry, S. K., Harper, S. L., Speicher, D. W.,
Morris, S. M., Ash, D. E., Traish, A., and Christianson, D. W. (2001) Biochemistry 40, 2678–2688.

94
HAN, MOORE, AND VIOLA
30. Custot, J., Moali, C., Brollo, M., Bocher, J. L., Delaforge M., Mansuy, D., Tenu, J. P., and Zimmermann,
J. L. (1997) J. Am. Chem. Soc. 119, 4086–4087.
31. Cavalli, R. C., Burke, C. J., Kawamoto, S., Soprano, D. R., and Ash, D. E. (1994) Biochemistry
33, 10652–10657.
32. Ruegg, U. T., and Russell, A. S. (1980) Anal. Biochem. 102, 206–212.