Substituent effects on the reaction of β-benzoylalanines with Pseudomonas fluorescens kynureninase

Article abstract of DOI:10.1021/bi100955b

Kynureninase is a pyridoxal 5′-phosphate-dependent enzyme that catalyzes the hydrolytic cleavage of l-kynurenine to give l-alanine and anthranilic acid. β-Benzoyl-l-alanine, the analogue of l-kynurenine lacking the aromatic amino group, was shown to a good substrate for kynureninase from Pseudomonas fluorescens, and the rate-determining step changes from release of the second product, l-Ala, to formation of the first product, benzoate [Gawandi, V. B., et al. (2004) Biochemistry 43, 3230-3237]. In this work, a series of aryl-substituted β-benzoyl-dl-alanines was synthesized and evaluated for substrate activity with kynureninase from P. fluorescens. Hammett analysis of k_cat and k_cat/K_m for 4-substituted β-benzoyl-dl-alanines with electron-withdrawing and electron-donating substituents is nonlinear, with a concave downward curvature. This suggests that there is a change in rate-determining step for benzoate formation with different substituents, from gem-diol formation for electron-donating substituents to C_β-C_γ bond cleavage for electron-withdrawing substituents. Rapid-scanning stopped-flow kinetic experiments demonstrated that substituents have relatively minor effects on formation of the quinonoid and 348 nm intermediates but have a much greater effect on the formation of the aldol product from reaction of benzaldehyde with the 348 nm intermediate. Since there is a kinetic isotope effect on its formation from β,β-dideuterio-β-(4-trifluoromethylbenzoyl)-dl- alanine, the 348 nm intermediate is proposed to be a vinylogous amide derived from abortive β-deprotonation of the ketimine intermediate. These results provide additional evidence for a gem-diol intermediate in the catalytic mechanism of kynureninase.

Full text of DOI:10.1021/bi100955b

Biochemistry 2010, 49, 7913–7919 7913
DOI: 10.1021/bi100955b
Substituent Effects on the Reaction of β-Benzoylalanines
with Pseudomonas fluorescens Kynureninase
Sunil Kumar,‡,§ Vijay B. Gawandi,‡, Nicholas Capito,‡ and Robert S. Phillips*,‡,§
§
‡Department of Chemistry and Department of Biochemistry and Molecular Biology, University of Georgia,
Athens, Georgia 30602. Present address: The Department of Biochemistry and Biophysics, Texas A&M University,
College Station, TX 77843-2128.
Received June 14, 2010; Revised Manuscript Received August 5, 2010
ABSTRACT: Kynureninase is a pyridoxal 50-phosphate-dependent enzyme that catalyzes the hydrolytic cleavage
of
the aromatic amino group, was shown to a good substrate for kynureninase from Pseudomonas fluorescens, and
the rate-determining step changes from release of the second product, -Ala, to formation of the first product,
L-kynurenine to give
L
-alanine and anthranilic acid. β-Benzoyl-
L
-alanine, the analogue of -kynurenine lacking
L
L
benzoate [Gawandi, V. B., et al. (2004) Biochemistry 43, 3230-3237]. In this work, a series of aryl-substituted
β-benzoyl-DL-alanines was synthesized and evaluated for substrate activity with kynureninase from P. fluorescens.
Hammett analysis of kcat and kcat/Km for 4-substituted β-benzoyl-DL-alanines with electron-withdrawing and
electron-donating substituents is nonlinear, with a concave downward curvature. This suggests that there is a
change in rate-determining step for benzoate formation with different substituents, from gem-diol formation for
electron-donating substituents to Cβ-Cγ bond cleavage for electron-withdrawing substituents. Rapid-scanning
stopped-flow kinetic experiments demonstrated that substituents have relatively minor effects on formation of the
quinonoid and 348 nm intermediates but have a much greater effect on the formation of the aldol product from
reaction of benzaldehyde with the 348 nm intermediate. Since there is a kinetic isotope effect on its formation from
β,β-dideuterio-β-(4-trifluoromethylbenzoyl)-DL-alanine, the 348 nm intermediate is proposed to be a vinylogous
amide derived from abortive β-deprotonation of the ketimine intermediate. These results provide additional
evidence for a gem-diol intermediate in the catalytic mechanism of kynureninase.
Kynureninase (kynase)1 (EC 3.7.1.3) is a pyridoxal 50-phosphate-
(PLP-) dependent enzyme that catalyzes the hydrolytic cleavage
substrate for bacterial kynase (7), with kcat and kcat/Km values of
0.7 s-1 and 8.0 ꢀ 104 M-1 s-1, respectively, compared to kcat
=
of
L-kynurenine to anthranilic acid and
L-alanine (1) (eq 1). The
16.0 s-1 and kcat/Km =6.0ꢀ 105 M-1 s-1 for
L-kynurenine. The
reaction is a key step in the catabolism of
L-tryptophan by Pseudo-
rate-determining step in the reaction of β-benzoyl-L-alanine with
monas fluorescens and some other bacteria (2). The inducible kynase
from P. fluorescens, a dimer with a subunit molecular weight of
about45000, has beenclonedandexpressedinEscherichiacoli(3).
In animals and some fungi, a similar constitutive enzyme reacts
kynase is Cβ-Cγ bond cleavage to form the first product, benzoate,
whereas the rate-determining step in the reaction of the natural
substrate,
L-kynurenine, is the release of the second product,
L-alanine (8). Furthermore, β-benzoyl-L-alanine reacts with kynur-
preferentially with 3-hydroxy-
L
-kynurenine in the catabolism of
eninase to form a transient 348 nm intermediate in stopped-flow
kinetic experiments. Addition of benzaldehyde together with
β-benzoyl-L-alanine results in decay of the 348 nm complex and for-
L-tryptophan (1). In eukaryotes, the kynurenine pathway is res-
ponsible for the de novo biosynthesis of NADþ, via the inter-
mediacy of quinolinate. However, quinolinate is also a neurotoxin,
due to its agonist effects on the N-methyl-D-asparate receptor,
mation of a new quinonoid complex that absorbs at 494 nm (7).
In the current work, we have synthesized a series of β-benzoyl-
DL-alanines substituted with electron-withdrawing and electron-
donating groups as probes of the reaction mechanism. These
compounds were examined by steady-state and stopped-flow
kinetics with P. fluorescens kynase. The results suggest that there
is a change in the rate-determining step for benzoate formation
from carbonyl hydration to Cβ-Cγ bond cleavage with electron-
donating and electron-withdrawing substituents, respectively.
and excessive levels of quinolinate have been implicated in the
etiology of a wide range of diseases such as epilepsy, stroke, and
neurological disorders, including AIDS-related dementia (4, 5).
Thus, selective inhibitors of human kynase are of interest as possi-
ble drugs for the treatment of a range of neurological disorders (6).
EXPERIMENTAL PROCEDURES
Enzyme Purification. Kynase was purified as previously
described from E. coli DH-5R cells that were transformed with
the pTZKYN plasmid containing the kyn gene cloned from
P. fluorescens (3).
It has been shown previously that β-benzoyl-L-alanine, which
lacks the aromatic amino group of the natural substrate, is a
Steady-State Kinetics. The steady-state kinetic measurements
were performed on a Cary 1E UV/vis spectrophotometer equip-
ped with a 6ꢀ6 Peltier thermoelectric cell changer, controlled by
a PC using software provided by Varian instruments. Substrate
*To whom correspondence should be addressed at the Department of
Chemistry, University of Georgia. E-mail: rsphillips@chem.uga.edu.
Phone: (706) 542-1996. Fax: (706) 542-9454.
1Abbreviations: kynase, kynureninase (EC 3.7.1.3); PLP, pyridoxal
50-phosphate; PMP, pyridoxamine 50-phosphate.
r
2010 American Chemical Society
Published on Web 08/06/2010
pubs.acs.org/Biochemistry

7914 Biochemistry, Vol. 49, No. 36, 2010
Kumar et al.
4.7 mmol) was dissolved in 1,4-dioxane (50 mL), and 6 M hydro-
chloric acid (70 mL) was added. The reaction was heated under
reflux for 8 h until no starting material was visible by TLC (silica
gel, petroleum ether/EtOAc, 1:1). The solution was then cooled
and washed with EtOAc (50 mL). The aqueous phase was concen-
trated under reduced pressure to give a brown syrup, which was
triturated with acetone to produce β-(4-methylbenzoyl)-DL-alanine
hydrochloride as an off-white crystalline solid. The hydrochlo-
ride product was then applied to a Dowex-50 cation-exchange
column and eluted with 1 M NH3 to yield the amino acid (1).
Yield 0.60 g (65%). 1H NMR (D2O þ DCl) δ (ppm) 2.34 (s, 3H),
3.73-3.74 (dd, 2H), 4.34-4.36 (t, 1H), 7.82 (d, 2H), 7.16 (d, 2H).
Deuteration of 4-Trifluoromethylbenzoylalanine. β-(4-
Trifluoromethylbenzoyl)alanine (0.0204 g) was dissolved in 5 mL
of D2O containing 10 μL of 14.2 M NaOD. A sample was removed
and examined by 1H NMR after 4 h, and the β-hydrogen
resonances were not detected. Glacial acetic acid (10 μL) was
then added to adjust the pH to 6, and the solution was left at 4 °C
for the deuterated product to crystallize. After standing over the
weekend, the fine suspension of crystals was filtered, washed with
a little cold water, and air-dried to give 0.0073 g. ESI MS: 264
(M þ 1), as predicted for the dideuterio compound.
Table 1: Spectroscopic Data for Substituted β-Benzoylalanines
substituent, X
molar extinction coeff (M-1 cm-1
)
λ
max (nm)
(1) 4-OMe
(2) 4-Me
(3) 4-Cl
2600
1660
2600
4442
2600
1946
1504
280
259
247
243
247
243
255
(4) 4-CF3
(5) 2-Cl
(6) 2-F
(7) 2-OMe
activities of all the new compounds were measured at their res-
pective absorption maxima (Table 1). All the assays were perfor-
med at 37 °C in 0.04 M potassium phosphate, pH 7.8, containing
40 μM PLP, in a total volume of 700 μL. Km, kcat, and kcat/Km
values were calculated using the FORTAN program, HYPER, of
Cleland (9).
Stopped-Flow Kinetics. Rapid-scanning stopped-flow kinetic
experiments were performed on an OLIS RSM-1000 instrument
at ambient temperature. This instrument is capable of collecting
rapid-scanning data at 1000 scans per second over the wavelength
range from 200 to 800 nm, with a dead time e2 ms. The enzyme
solutions contained 1 mg/mL P. fluorescens kynase in 0.04 M
potassium phosphate, pH 8.0. These were mixed with solutions of
1 mM compounds 1-7 in the same buffer. When benzaldehyde
was present, it was added to the substrates at 10 mM. The wave-
length region from 300 to 560 nm was used to analyze the data by
global fitting using the Global Works program (10) provided by
OLIS, Inc.
RESULTS
Chemistry. All of the substituted β-benzoyl-DL-alanines were
synthesized based on our previously published procedure (7)
(Scheme 1) used to prepared the parent compound. Compounds
1-7 were obtained starting from the R-bromination of the corres-
ponding commercially available substituted acetophenones (1a-7a)
with CuBr2 (11). In contrast to the literature (11), which states
that the R-monobrominated product is exclusively obtained, we
generally obtained a mixture of R-monobrominated (1b-7b) and
R,R-dibrominated acetophenones. However, the R-monobromi-
nated acetophenone was readily separated from the undesired
dibrominated product either by recrystallization or silica gel column
chromatography. Alkylation of the monobrominated acetophe-
none with diethylacetamidomalonate using sodium hydride as a
base afforded the diethyl benzoylmethylacetamidomalonate
(1c-7c). Attempts to alkylate diethylacetamidomalonate with
bromoacetophenones using the more common reagent, sodium
ethoxide in ethanol, gave poor yields. In some cases, a mixture of
mono- and diethyl ester products were obtained after alkylation,
along with other unknown byproducts, indicating that a partial
Krapcho decarboalkoxylation had occurred. The combined mono-
and diester products were subjected to acid hydrolysis and de-
carboxylation to furnish the substituted β-benzoyl-DL-alanine hydro-
chloride, which was then neutralized or treated with Dowex-50
cation exchange resin to give the free racemic amino acids (1-7).
The UV absorption properties of 1-7 used for the activity assays
are summarized in Table 1. Details for the synthesis of com-
pounds 2-7 are provided in the Supporting Information.
Synthesis of Substrate Analogues. β-(4-Methylbenzoyl)-
DL-alanine (1). 2-Bromo-1-(4-methylphenyl)ethanone (1b).
CuBr2 (3.19 g, 14.2 mmol) (11) was heated at reflux in 10 mL of
EtOAc with stirring. To this was added 4-methylacetophenone
(1a, 2.36 g, 11.4 mmol) in 10 mL of CHCl3. The reaction mixture
was heated at reflux for 5 h and then cooled to room temperature.
CuBr and CuBr2 residues were filtered, and the filtrate was de-
colorized with activated charcoal, filtered again through a bed of
Celite, and washed with EtOAc (4 ꢀ 50 mL). The solvent was
removed under reduced pressure to give yellow crystals. Further
purification by recrystallization gave 1.4 g (58%) of the bromi-
1
nated product 1b, mp 52-56 °C. H NMR (CDCl3) δ (ppm)
2.335 (s, 3H), 4.28 (s, 2H), 7.161 (d, 2H), 7.915 (d, 2H).
Ethyl 4-(40-Methylphenyl)-4-oxo-2-acetamido-2-ethoxy-
carbonylbutyrate (1c). Sodium hydride (1.39 g, 34.9 mmol,
2.5 equiv) (60% in mineral oil) was suspended in dry DMF (14 mL).
A solution of diethyl acetamidomalonate (4.56 g, 21 mmol,
1.5 equiv) in dry DMF was added. The solution was stirred at
0 °C under a nitrogen atmosphere for 3 h until the anion had
formed. A solution of 2-bromo-1-(4-methylphenyl)ethanone (1b,
13.9 mmol, 2.78 g) in dry DMF (10 mL) was then added, and the
solution was warmed to room temperature and stirred overnight
under nitrogen. The reaction mixture was poured into distilled
water (100 mL), acidified in an ice bath to pH 3 with 1 M hydro-
chloric acid, and extracted into diethyl ether (4 ꢀ 70 mL). The
ether extracts were dried over MgSO4, and the solvent was removed
under reduced pressure to give white crystals of ethyl 4-(40-methyl-
phenyl)-4-oxo-2-acetamido-2-ethoxycarbonylbutanoate (1c). Yield
2.3 g (45%). 1H NMR (CDCl3) δ (ppm) 1.27-1.29 (t, 6H), 1.84
(s, 3H), 2.34 (s, 3H), 3.51 (s, 2H), 4.13 (m, 4H), 7.18 (d, 2H), 7.38
(d, 2H), 7.92 (s, 1H).
Steady-State Kinetics. All of the compounds tested (1-7)
were found to be substrates for hydrolytic cleavage by P. fluorescens
kynase, with varying degrees of efficiency (Table 2). We have
shown previously that the Km for β-benzoyl-L-alanine is half that
of β-benzoyl-DL-alanine, indicating that the
D-isomer does not
interact significantly with kynureninase (7). The Hammett plot
of log(kcat) and log(kcat/Km) against the substituent constants
for the para-substituted compounds (1-4) shows a concave
downward nonlinear relationship, which suggests that there is a
change in the rate-determining step for compounds with strongly
electron-donating or electron-withdrawing substituents (Figure 1).
β-(4-Methylbenzoyl)-DL-alanine (1). Ethyl 4-(40-methyl-
phenyl)-4-oxo-2-acetamido-2-ethoxycarbonylbutanoate (1c, 1.6 g,

Article
Biochemistry, Vol. 49, No. 36, 2010 7915
Scheme 1: Synthesis of Substituted β-Benzoylalanines
the 348 nm intermediate with benzaldehyde to form a quinonoid
complex shows a much greater range of values than for its for-
mation (Table 3). 4-Me- and 4-Cl-benzoylalanine (2 and 3) show
reactivity similar to that of β-benzoylalanine (1/τ2, Table 3 and
Figure 2D), while 4-MeO- and 4-CF3-benzoylalanine (1 and 4)
show much lower reactivity (1/τ2, Table 3 and Figure 3D). 2-Cl-
and 2-F-benzoylalanine (5 and 6) react much more slowly (1/τ2,
Table 3 and Figure 5D), and 2-MeO-benzoylalanine (7) did not
show any detectable formation of the quinonoid species in the
presence of benzaldehyde for at least 30 s. Thus, the reactivity of
the 348 nm intermediates with benzaldehyde to form the 494 nm
quinonoid intermediate in the stopped-flow experiments roughly
parallels the kcat activity for the substituted benzoylalanines in
Table 2.
Table 2: Steady-State Kinetic Parameters for Substituted β-Benzoylalanines
substituent, X
kcat (s-1
)
kcat/Km ꢀ 10-3 (M-1 s-1
)
(1) 4-OMe
(2) 4-Me
(3) 4-Cl
(4) 4-CF3
(5) 2-Cl
(6) 2-F
0.38 ( 0.06
0.91 ( 0.06
1.83 ( 0.78
0.001 ( 0.0004
0.08 ( 0.006
0.06 ( 0.006
0.05 ( 0.006
0.7
1.6 ( 0.28
28 ( 0.02
18 ( 0.04
0.019 ( 0.003
12 ( 2.8
1.7 ( 0.2
1.4 ( 0.25
80
(7) 2-OMe
H
Similar slopes are seen for both kcat and kcat/Km in the electron-
donating region, with σ<0 (F=0.590 ( 0.275 and 0.455 ( 0.068,
respectively), while kcat is more sensitive (F=-4.238 ( 1.12) than
kcat/Km (F=-0.986 ( 0.031) to the effect of electron-withdrawing
groups with σ > 0. The very low activity of the CF3 derivative (4)
was confirmed by 19F NMR of reaction mixtures, which showed
a new resonance at -63.042 ppm, downfield of the resonance of 4
at -63.595 ppm, after overnight incubation. It is interesting that
all of the 2-substituted β-benzoylalanines, 2-F, 2-Cl, and 2-MeO,
are poor substrates (Table 2), even though the physiological sub-
strate, kynurenine, has a 2-amino substituent.
Pre-Steady-State Kinetics. When β-benzoylalanine is mixed
with P. fluorescens kynase in the stopped-flow spectrophoto-
meter, the enzyme spectrum changes, and a new intermediate
forms with a strong absorption peak at about 348 nm (7). Figure 2
shows the results of the reaction of β-(4-chlorobenzoyl)alanine
(3) with P. fluorescens kynase in the stopped-flow spectrophoto-
meter. The data in Figure 2A,B show a reaction very similar to
that observed previously with β-benzoylalanine, with a 348 nm
intermediate forming, concomitant with decay of the external
aldimine at 420 nm (Figure 2A,B), and an isosbestic point at
388 nm. Similar results are seen with the trifluoromethyl analogue,
4 (Figure 3). The reaction of β,β-D2-4 exhibits a kinetic isotope of
1.88 (0.14 on the formation of the 348 nm intermediate (Figure 4B).
There is also a significant equilibrium isotope effect on the forma-
tion of the 348 nm intermediate, since the steady-state absorption
of the 348 nm intermediate is reduced (Figure 4A). Compounds
1-7 all form the 348 nm intermediate, at rate constants ranging
from 2 to 20 s-1 (1/τ1, Table 3, and Figure 5), similar to the value
of 9 s-1 seen previously with β-benzoylalanine (7).
DISCUSSION
Kynase is one of a small group of enzymes that catalyze retro-
Claisen-type reactions. This reaction is proposed to proceed via a
tetrahedral intermediate on the γ-carbonyl carbon (1, 7, 8, 12-16).
We showed previously that dihydrokynurenines (12, 13, 15) and
S-phenylcysteine sulfones (14, 15) are potent competitive inhi-
bitors of P. fluorescens kynase, with Ki values in the low micro-
molar to nanomolar range. These results suggested that a gem-
diolate intermediate is involved in the reaction mechanism and
that these inhibitors are transition-state analogues. Later, we found
that β-benzoyl-
rate-determining step is formation of the first product, benzo-
ate (7), while the rate-limiting step in the reaction of -kynurenine
is release of -Ala (8). In the present work, we prepared and
L-alanine is a good substrate for kynase, but the
L
L
examined a number of ring-substituted β-benzoyl-DL-alanines to
probe the electronic effects on benzoate formation. Hammett
analysis of the steady-state kinetic data for the para-substituted
compounds (Figure 1) shows strongly downward curvature,
indicating that there is a change in rate-determining step for
benzoate formation with substitutent.
The proposed mechanism of kynase involves formation of an
external aldimine with PLP, followed by deprotonation of the
R-C and reprotonation at C-40 to form a ketimine (Scheme 2).
Addition of water to the γ-carbonyl of the ketimine forms a gem-
diolate intermediate, followed by a retro-Claisen cleavage of the
Cβ-Cγ bond (Scheme 2). The stereochemistry of the retro-aldol
reaction catalyzed by kynureninase with dihydro-L-kynurenine is
consistent with an (S)-gem-diolate intermediate (13). According
to the Dunathan hypothesis (17), the Cβ-Cγ bond of the gem-
diolate should be perpendicular to the π-system of the PMP-
ketimine for cleavage of the Cβ-Cγ bond to occur. Since there
is only one catalytic base from the pH dependence of kcat/Km (8),
The 348 nm intermediate formed with β-benzoylalanine was
found previously to decay in the presence of benzaldehyde, and
concomitantly a new quinonoid species with λmax of 494 nm
formed (7), derived from the aldol product, γ-phenylhomoserine.
In the case of compounds 1-7, the rate constant for reaction of

7916 Biochemistry, Vol. 49, No. 36, 2010
Kumar et al.
FIGURE 1: Hammett plot for reaction of substituted β-benzoylalanines. Key: circles, kcat; squares, kcat/Km. The lines were obtained by regression
analysis, with the slopes indicated in the text.
F
IGURE 2: Reaction of β-(4-chlorobenzoyl)-DL-alanine (3) with P. fluorescens kynase in the stopped-flow spectrophotometer. Reactions
contained 1 mg/mL kynase in 40 mM potassium phosphate, pH 8.0, 0.5 mM 3 (A, B, C, and D), and 5 mM benzaldehyde (C and D). (A) Reac-
tion with0.5 mM3. Curves are shown at 0.0008 s (1), 0.0308s (2), 0.0808 s (3), 0.1608 s (4), 0.3208 s (5), and 0.6408 s (6). (B) Time courses for (A) at
348 (solid line), 420 (dottedline), and 495 nm(dashedline). (C) Reactionwith0.5 mM3 and 5 mM benzaldehyde. Curvesare shown at0.0010s (1),
4.04 s (2), 12.01 s (3), 24.01 s (4), 48.01 s (5), and 96.01 s (6). (D) Time courses for (C) at 348 (solid line), 420 (dotted line), and 495 nm (dashed line).
all of the proton transfers must occur on one face of the PLP-
substrate complex, which is the bottom face of the complexes
shown in Scheme 2. This is supported by the stereochemical
experiments of Palcic et al., who found that protonation of the
enamine occurs with retention of configuration and concluded
that the Cβ-Cγ bond is oriented syn to the R-CH (18).
Electron-withdrawing groups would favor nucleophilic addi-
tion of hydroxide to the γ-carbonyl, while electron-donating
substitutents would make addition more difficult. Conversely,
the stabilizing effect of electron-withdrawing substituents on the
gem-diolate would make the Cβ-Cγ cleavage slower, while electron-
donating substituents would be expected to accelerate Cβ-Cγ
cleavage. The unsubstituted compound or those with weakly
electron-donating and electron-withdrawing groups, like p-Me
and p-Cl, balance the formation of the gem-diolate and the Cβ-Cγ
cleavage, whereas the strongly electron-withdrawing group, p-CF3,
favors the gem-diolate intermediate but reduces the rate of Cβ-Cγ
cleavage compared to β-benzoyl-L-alanine. Conversely, the strongly
electron-donating p-MeO group reduces the rate of formation of
the gem-diolate but favors the Cβ-Cγ cleavage reaction. The
stopped-flow kinetic trapping experiments with benzaldehyde
show that there is a rough correlation of the kcat values for the
β-benzoylalanines with the rates of formation of a quinonoid
species absorbing at 494 nm. Benzaldehyde reacts in an aldol-
type reaction with the enamine intermediate after benzoate
release to form γ-phenylhomoserine (13, 19). However, the very
slow 2-substituted substrates (5-7) react with benzaldehyde with
rate constants even slower than kcat
.
We found previously that β-benzoyl-L-alanine reacts with kynase
to form a new intermediate with a strong absorption peak at
about 348 nm (ε ∼ 20000 M-1 cm-1). This intermediate was not
observed in reactions with excess L-kynurenine since the strong
absorption of the substrate at 360 nm obscures this region of the
spectrum. However, in single turnover stopped-flow experiments,

Article
Biochemistry, Vol. 49, No. 36, 2010 7917
F
IGURE 3: Reaction of β-(4-trifluoromethylbenzoyl)-DL-alanine (4) with P. fluorescens kynase in the stopped-flow spectrophotometer. Reactions
contained 1 mg/mL kynase in 40 mM potassium phosphate, pH 8.0, 0.5 mM 4 (A, B, C, and D), and 5 mM benzaldehyde (C and D). (A) Reaction
with 0.5 mM 4. Curves are shown at 0.0002 s (1), 0.0202 s (2), 0.0602 s (3), 0.1202 s (4), 0.2402 s (5), and 0.4802 s (6). (B) Time courses for (A) at 348
(solid line), 420 (dotted line), and 495 nm (dashed line). (C) Reaction with 0.5 mM 4 and 5 mM benzaldehyde. Curves are shown at 0.00075 s (1),
1.00 s (2), 6.00 s (3), 20.00 s (4), 60.00 s (5), and 120.00 s (6). (D) Time courses for (C) at 348 (solid line), 420 (dotted line), and 495 nm (dashed line).
F
IGURE 4: (A) Effect of deuteration on the reaction of P. fluorescens kynase with 4-CF3-benzoylalanine (4). Key: solid line, 4-CF3-benzoylalanine
(0.5 mM); dotted line, β,β-D2-4-CF3-benzoylalanine. (B) Time courses at 348 nm. Key: solid line, 4-CF3-benzoylalanine (0.5 mM); dotted line,
β,β-D2-4-CF3-benzoylalanine.
a transient intermediate with λmax at about 350 nm was observed,
Table 3: Pre-Steady-State Kinetic Rate Constants for Substituted Benzoy-
which was interpreted at the time as evidence for a ketimine (12)
but is likely to be the same species observed in the present work.
All of the substituted β-benzoyl-DL-alanines examined in this
study also form an intermediate absorbing at about 348 nm, with
lalanines
substituent, X
1/τ1 (s-1
)
1/τ2 (s-1
)
(1) 4-OMe
(2) 4-Me
(3) 4-Cl
(4) 4-CF3
(5) 2-Cl
(6) 2-F
20 ( 3
20 ( 1
0.0174 ( 0.0015
0.71 ( 0.10
0.24 ( 0.01
0.0181 ( 0.0003
0.00285 ( 0.0009
0.0015 ( 0.0004
rate constants similar to that of β-benzoyl-L-alanine, ranging
from 2 to 20 s-1 (Figures 2A, 3A, and 4A), so the substituent
effect on its formation is relatively small. We suggested previously
that this intermediate may be the enamine product resulting from
Cβ-Cγ cleavage (7). However, since the 348 nm intermediate
shows a wide range of reactivity with benzaldehyde in the present
work (Table 3), that cannot be the case. We now propose that the
348 nm intermediate is a vinylogous amide complex formed by
abortive β-deprotonation of the ketimine intermediate (Scheme 2).
Vinylogous amides are known to have strong absorptions in
the 320-360 nm region, with molar extinction coefficients of
11.7 ( 0.1
2.57 ( 0.19 (1.37 ( 0.03)a
4.52 ( 0.18
14.4 ( 0.5
c
(7) 2-OMe
H
2.74 ( 0.11
-
9b
0.7b
aβ,β-Dideuterated 4 in parentheses. bFrom ref 7. cNot determined.
20000-30000 M-1 cm-1 (20). In fact, a compound very similar to
the proposed structure of the 348 nm intermediate in Scheme 2,

7918 Biochemistry, Vol. 49, No. 36, 2010
Kumar et al.
F
IGURE 5: Reaction of 2-fluorobenzoyl-DL-alanine (6) with P. fluorescens kynase in the stopped-flow spectrophotometer. Reactions contained
1 mg/mL kynase in 40 mM potassium phosphate, pH 8.0, 0.5 mM 6 (A, B, C, and D), and 5 mM benzaldehyde (C and D). (A) Reaction with 0.5 mM
6. Curves are shown at 0.0005 s (1), 0.0205 s (2), 0.0405 s (3), 0.0805 s (4), 0.1605 s (5), and 0.3205 s (6). (B) Time courses for (A) at 348 (solid line),
420 (dotted line), and 495 nm (dashed line). (C) Reaction with 0.5 mM 6and 5 mM benzaldehyde. Curves are shown at 0.0009 s (1), 12.00 s (2), 30.00 s (3),
60.00 s (4), 90.00 s (5), and 120.00 s (6). (D) Time courses for (C) at 348 (solid line), 420 (dotted line), and 495 nm (dashed line).
Scheme 2: Mechanism of Reaction of β-Benzoylalanines with Kynureninase
3-(N-(2-hydroxyethyl)-N-methylamino)-1-phenyl-1-propenone,
exhibits λmax = 344 nm and ε = 23300 M-1 cm-1 (20). This
assignment of the 348 nm intermediate to a vinylogous amide is
supported by the observation of a kinetic isotope effect of 1.88 on
its formation from β,β-D2-4-trifluoromethylbenzoyl-DL-alanine
(4) (Figure 4). The high absorptivity and the position of the band

Article
Biochemistry, Vol. 49, No. 36, 2010 7919
are more consistent with the trans- rather than the cis-vinylogous
amide (20). The trans geometry is also consistent with the crystal
structure of human kynureninase complexed with 3-hydroxy-
hippuric acid, which shows an extended conformation of the
ligand with a trans-amide bond (21), and also with docking experi-
ments of 3-hydroxy-L-kynurenine into the active site of human
kynureninase (22).
What is the role of this vinylogous amide in the catalytic
mechanism of kynase? Previously, we found that the 348 nm
intermediate formed from β-benzoyl-L-alanine reacted with
2. Hayaishi, O., and Stanier, R. Y. (1951) The bacterial oxidation of
tryptophan III: Enzymatic activities of cell-free extracts from bacteria
employing the aromatic pathway. J. Bacteriol. 62, 691–709.
3. Koushik, S. V., Sundaraju, B., McGraw, R. A., and Phillips, R. S.
(1997) Cloning, sequence, and expression of kynureninase from
Pseudomonas fluorescens. Arch. Biochem. Biophys. 344, 301–308.
4. Achim, C. L., Heyes, M. P., and Wiley, C. A. (1993) Quantitation of
human immunodeficiency virus, immune activation factors, and
quinolinic acid in AIDS brains. J. Clin. Invest. 91, 2769–2775.
5. Sei, S., Saito, K., Stewart, S. K., Crowely, J. S., Brouwers, P., Kleiner,
D. E., Katz, D. A., Pizzo, P. A., and Heyes, M. P. (1995) Increased
human immunodeficiency virus (HIV) type 1 DNA content and
quinolinic acid concentration in brain tissues from patients with
HIV encephalopathy. J. Infect. Dis. 172, 638–647.
benzaldehyde at a rate similar to kcat, so we concluded that it is
catalytically competent (7). In earlier pH dependence studies,
we proposed that a active site base, possibly the ε-amino group
of Lys-227, abstracts a proton from a water molecule, and the
resulting hydroxide adds to the carbonyl to form the gem-
diolate (8) (Scheme 2, path a). If the incipient hydroxide abstracts
a proton from the β-C of the PMP-ketimine, then instead a
vinylogous amide is formed (Scheme 2, path b, dashed arrows).
These β-protons should be acidic, since they are located on a
carbon between a ketone and an iminium ion. In support of this
prediction, the calculated pKa of the CH bond in the β-keto-
iminium ion derived from 2,4-pentanedione is 11.90, using the
SPARC pKa calculator (23). The absorption intensity of the
348 nm peak suggests that the equilibrium between the gem-
diolate and the vinylogous amide strongly favors the vinylogous
amide. However, the vinylogous amide does not appear to be on
the direct catalytic pathway, since it would not allow Cβ-Cγ
cleavage, but may be instead a side branch of the mechanism, in
rapid equilibrium with the PMP-ketimine. The low reactivity of
the 2-substituted benzoylalanines in steady-state kinetics and
with benzaldehyde in the stopped-flow experiments may be due
to stabilization of the 348 nm abortive complex. In this regard it is
interesting that β-(2-methoxybenzoyl)alanine is used as an inhi-
bitor of mammalian kynureninase in vivo (24). Thus, stabiliza-
tion of the vinylogous amide complex in Scheme 2 is a potential
strategy for kynureninase inhibitor design.
Conclusion. There are large electronic effects on the reaction
of P. fluorescens kynase with substituted β-benzoylalanines. The
concave downward Hammett plots indicate a change in rate-
determining step from hydration with electron-donating substi-
tutents to Cβ-Cγ bond cleavage with electron-withdrawing
substitutents. These results provide additional evidence for a
gem-diol intermediate in the catalytic mechanism of kynase. The
rapid-scanning stopped-flow data support the conclusions from
the steady-state kinetic analysis. A 348 nm intermediate formed
in the reaction of these substrates is proposed to be a vinylogous
amide not directly on the catalytic pathway.
6. Stone, T. W. (2000) Inhibitors of the kynurenine pathway. Eur. J.
Med. Chem. 35, 179–186.
7. Gawandi, V. B., Liskey, D., Lima, S., and Phillips, R. S. (2004)
Reaction of Pseudomonas fluorescens kynureninase with beta-benzoyl-
L
-alanine: Detection of a new reaction intermediate and a change in
rate-determining step. Biochemistry 43, 3230–3237.
8. Koushik, S. V., Moore, J. A., III, Sundararaju, B., and Phillips, R. S.
(1998) The catalytic mechanism of kynureninase from Pseudomonas
fluorescens: Insights from the effects of pH and isotopic substitution
on steady-state and pre-steady-state kinetics. Biochemistry 37, 1376–
1382.
9. Cleland, W. W. (1979) Statistical analysis of enzyme kinetic data.
Methods Enzymol. 63, 103–138.
10. Matheson, I. B. C., and DeSa, R. J. (1990) Robust multicomponent
analysis applied to the separation of components in a mixture of
absorbing species. Comput. Chem. 14, 49–57.
11. Kosower, E. M., Cole, W. J., Wu, G.-S., Cardy, D. E., and Meisters,
G. (1963) Halogenation with copper(II). I. Saturated ketones and
phenol. J. Org. Chem. 28, 630–633.
12. Phillips, R. S., Sundararaju, B., and Koushik, S. V. (1998) The
catalytic mechanism of kynureninase from Pseudomonas fluorescens:
Evidence for transient quinonoid and ketimine intermediates from
rapid-scanning stopped-flow spectrophotometry. Biochemistry 37,
8783–8789.
13. Phillips, R. S., and Dua, R. K. (1991) Stereochemistry and mechanism
of aldol reactions catalyzed by kynureninase. J. Am. Chem. Soc. 113,
7385–7388.
14. Dua, R. K., Taylor, E. W., and Phillips, R. S. (1993) S-Aryl-L-cysteine
S,S-dioxides: Design and evaluation of a new class of mechanism
based inhibitors of kynureninase. J. Am. Chem. Soc. 115, 1264–1270.
15. Heiss, C., Anderson, J., and Phillips, R. S. (2003) Differential effects
of bromination on substrates and inhibitors of kynureninase from
Pseudomonas fluorescens. Org. Biomol. Chem. 1, 288–295.
16. Tanizawa, K., and Soda, K. (1979) The mechanism of kynurenine
hydrolysis catalyzed by kynureninase. J. Biochem. (Tokyo) 86, 1199–
1209.
17. Dunathan, H. C. (1971) Stereochemical aspects of pyridoxal phos-
phate catalysis. Adv. Enzymol. Relat. Areas Mol. Biol. 35, 79–134.
18. Palcic, M. M., Antoun, M., Tanizawa, K., Soda, K., and Floss, H. G.
(1985) Stereochemistry of the kynureninase reaction. J. Biol. Chem.
260, 5248–5251.
19. Bild, G. S., and Morris, J. C. (1984) Detection of beta-carbanion
formation during kynurenine hydrolysis catalyzed by Pseudomonas
marginalis kynureninase. Arch. Biochem. Biophys. 235, 41–47.
20. King, L. C., and Ostrum, G. K. (1964) Vinylogous imides. 11. Ultra-
violet spectra and the application of Woodward’s rules. J. Org. Chem.
29, 3459–3461.
21. Lima, S., Kumar, S., Gawandi, V., Momany, C., and Phillips, R. S.
(2009) Crystal structure of the Homo sapiens kynureninase-3-hydro-
xyhippuric acid inhibitor complex: Insights into the molecular basis of
kynureninase substrate specificity. J. Med. Chem. 52, 389–396.
22. Lima, S., Khristoforov, R., Momany, C., and Phillips, R. S. (2007)
Crystal structure of Homo sapiens kynureninase. Biochemistry 46,
2735–2744.
SUPPORTING INFORMATION AVAILABLE
Details of the synthesis of 2-7 and the pre-steady-state kinetic
data for 1, 2, 5, and 7. This material is available free of charge via
the Internet at http://pubs.acs.org.
23. Hilal, S., Karickhoff, S. W., and Carreira, L. A. (1995) A rigorous test
for SPARC’s chemical reactivity models: Estimation of more than
4300 ionization pKa’s. Quant. Struct.-Act. Relat. 14, 348.
24. Chiarugi, A., Carpenedo, R., and Moroni, F. (1996) Kynurenine disposi-
tion in blood and brain of mice: Effects of selective inhibitors of kynu-
renine hydroxylase and of kynureninase. J. Neurochem. 67, 692–698.
REFERENCES
1. Soda, K., and Tanizawa, K. (1979) Kynureninases: Enzymological
properties and regulation mechanism. Adv. Enzymol. Relat. Areas
Mol. Biol. 49, 1–40.