Substituents effects on activity of kynureninase from Homo sapiens and Pseudomonas fluorescens

Article abstract of DOI:10.1016/j.bmc.2013.05.039

A series of substituted kynurenines (3-bromo-dl, 3-chloro-dl, 3-fluoro-dl, 3-methyl-dl, 5-bromo-l, 5-chloro-l, 3,5-dibromo-l and 5-bromo-3-chloro-dl) have been synthesized and tested for their substrate activity with human and Pseudomonas fluorescens kynureninase. All of the substituted kynurenines examined have substrate activity with both human as well as P. fluorescens kynureninase. For the human enzyme, 3- and 5-substituted kynurenines have k _cat and k_cat/K_m values higher than l-kynurenine, but less than that of the physiological substrate, 3-hydroxykynurenine. However, 3,5-dibromo- and 5-bromo-3-chlorokynurenine have k_cat and k_cat/K_m values close to that of 3-hydroxykynurenine with human kynureninase. The effects of the 3-halo substituents on the reactivity with human kynureninase may be due to electronic effects and/or halogen bonding. In contrast, for the bacterial enzyme, 3-methyl, 3-halo and 3,5-dihalokynurenines are much poorer substrates, while 3-fluoro, 5-bromo, and 5-chlorokynurenine have k_cat and k_cat/K _m values comparable to that of its physiological substrate, l-kynurenine. Thus, 5-bromo and 5-chloro-l-kynurenine are good substrates for both human as well as bacterial enzyme, indicating that both enzymes have space for substituents in the active site near C-5. The increased activity of the 5-halokynurenines may be due to van der Waals contacts or hydrophobic effects. These results may be useful in the design of potent and/or selective inhibitors of human and bacterial kynureninase.

Full text of DOI:10.1016/j.bmc.2013.05.039

Bioorganic & Medicinal Chemistry 21 (2013) 4670–4677
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Substituents effects on activity of kynureninase from Homo sapiens
and Pseudomonas fluorescens
Chandan Maitrani a, Robert S. Phillips a,b,
⇑
a Department of Chemistry, University of Georgia, Athens, GA 30602, United States
b Department of Biochemistry and Molecular Biology, University of Georgia, Athens, GA 30602, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of substituted kynurenines (3-bromo-DL, 3-chloro-DL, 3-fluoro-DL, 3-methyl-DL, 5-bromo-L,
Received 19 March 2013
Revised 2 May 2013
Accepted 10 May 2013
Available online 1 June 2013
5-chloro-L, 3,5-dibromo-L and 5-bromo-3-chloro-DL) have been synthesized and tested for their substrate
activity with human and Pseudomonas fluorescens kynureninase. All of the substituted kynurenines exam-
ined have substrate activity with both human as well as P. fluorescens kynureninase. For the human
enzyme, 3- and 5-substituted kynurenines have kcat and kcat/Km values higher than L-kynurenine, but less
than that of the physiological substrate, 3-hydroxykynurenine. However, 3,5-dibromo- and 5-bromo-3-
chlorokynurenine have kcat and kcat/Km values close to that of 3-hydroxykynurenine with human kynu-
reninase. The effects of the 3-halo substituents on the reactivity with human kynureninase may be
due to electronic effects and/or halogen bonding. In contrast, for the bacterial enzyme, 3-methyl, 3-halo
and 3,5-dihalokynurenines are much poorer substrates, while 3-fluoro, 5-bromo, and 5-chlorokynurenine
Keywords:
Pyridoxal-50-phosphate
Electronic effects
Halogen bond
have kcat and kcat/Km values comparable to that of its physiological substrate,
L-kynurenine. Thus, 5-
bromo and 5-chloro- -kynurenine are good substrates for both human as well as bacterial enzyme, indi-
L
cating that both enzymes have space for substituents in the active site near C-5. The increased activity of
the 5-halokynurenines may be due to van der Waals contacts or hydrophobic effects. These results may
be useful in the design of potent and/or selective inhibitors of human and bacterial kynureninase.
Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction
that reacts exclusively with
able activity for 3-hydroxy-
L
-kynurenine, and exhibits undetect-
L
-kynurenine.3
Kynureninase0 [EC 3.7.1.3] is a pyridoxal-50-phosphate (PLP)
dependent enzyme that catalyzes the hydrolytic cleavage of
urenine to yield anthranilic acid and -alanine (Eq. 1). This is a key
enzyme in the kynurenine pathway of tryptophan catabolism in
both bacteria and animals, and catalyzes the unique b, -cleavage
of aryl substituted -keto-
-amino acids.1 It has been found that
the mammalian kynureninase cleaves 3-hydroxy- -kynurenine
more rapidly than -kynurenine while the bacterial kynureninase
from Pseudomonas fluorescens (PfKynase) cleaves -kynurenine
more rapidly than 3-hydroxy-
-kynurenine.1 We showed that this
L-kyn-
L
ð1Þ
c
c
a
L
These two distinct types of kynureninase have also been shown
to differ in their regulation and their response to PLP. The Pseudo-
monas enzyme is induced by growing cells in the presence of
L
L
L
L
-tryptophan as carbon and nitrogen source,4 and for this reason
difference in substrate specificity is due to differences in hydrogen
bonding and steric interactions in the active sites of human and P.
fluorescens enzymes.2,3 Mutation of three active site residues
(H104W, S332G, and N333T) of human kynureninase (HsKynase)
to the corresponding residues in PfKynase gave a mutant enzyme
has been called ‘inducible kynureninase’ in the literature. Further-
more, it has been found that the inducible enzyme is reversibly
inactivated by
reaction to give pyridoxamine-50-phosphate and pyruvate from
-alanine5. However, the enzyme activity is restored either by addi-
L-alanine, resulting in an abortive transamination
L
tion of PLP or pyruvic acid; in the latter case there is a reverse
transamination between pyridoxamine-50-phosphate and the
added pyruvate to give back PLP and alanine. On the other hand,
Abbreviations: HsKynase, Homo sapiens kynureninase, EC 3.7.1.3; PfKynase,
Pseudomonas fluorescens kynureninase, EC 3.7.1.3.
the eukaryotic form of kynureninase is not induced by
L-trypto-
⇑
phan, so it has been referred to as ‘constitutive kynureninase0.1 This
Corresponding author. Tel.: +1 706 542 1996; fax: +1 706 542 9454.
E-mail address: plp@uga.edu (R.S. Phillips).
0968-0896/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmc.2013.05.039

C. Maitrani, R. S. Phillips / Bioorg. Med. Chem. 21 (2013) 4670–4677
4671
enzyme does not undergo abortive transamination in the presence
of -alanine or other amino acids. However, the rate of the hydro-
substrate, L-kynurenine, or 3-hydroxy-DL-kynurenine, in 30 mM
L
potassium phosphate buffer pH 8, containing 40 lM of pyridoxal-
lytic cleavage reduces with time, indicating that the product, 3-
50-phosphate at 37 °C, and the absorbance decrease was followed
hydroxyanthranilate, is a product inhibitor.6
for 10 min.
The kynurenine pathway has been shown to be elevated in a
number of neurodegenerative human diseases, including Hun-
tington’s Disease,7,8 HIV-related dementia,9 Parkinson’s Disease10
and Alzheimer’s Disease,11 resulting in accumulation of neuro-
toxic metabolites such as quinolinic acid. Furthermore, the
R188Q allele of kynureninase has been linked to spontaneous
hypertension in the Han Chinese population.12 In addition, the
kynurenine pathway is elevated in response to bacterial and viral
infections, presumably to starve the pathogen of tryptophan, as a
2.4. Kinetics measurements
The scanning kinetics measurements to determine the absor-
bance change for the 3-substituted substrate analogs was done
using the human enzyme, since 3-hydroxykynurenine is the nat-
ural substrate for the human enzyme.1 The scanning kinetics
measurements to determine the absorbance change for the 5-
substituted substrate analogs were done using the Pseudomonas
fluorescens enzyme since 5-substituted kynurenines have been
shown previously to be good substrates for the P. fluorescens en-
zyme.18 Depending on the reactivity of the substrate, repetitive
scans were performed at intervals ranging from 5 to 30 min until
no further changes were observed. The Km, Vmax, and Vmax/Km val-
ues were determined by fitting of the initial rate data to Equation
2 using the compiled FORTRAN program of Cleland, HYPERO.19
The fitted values of Vmax, and Vmax/Km were divided by the respec-
tive extinction coefficients in Table 1 and by the enzyme concen-
tration to give the kcat and kcat/Km values. The concentration of
the enzymes was determined using the A280 = 1.43 for a 1 mg/
mL solution.20 The molecular weights used to calculate the molar
enzyme concentrations were 46.5 kDa for PfKynase16 and 57 kDa
for HsKynase.2
result of induction by c
-interferon.13 However, Chlamydia psittaci
has been shown to circumvent this response by expression of
kynureninase in place of anthranilate synthase in the Trp oper-
on.14 Finally, kynureninase plays a role in the biosynthesis of
PQS in Pseudomonas aeruginosa, and this signaling molecule is
essential for pathogenicity.15 Thus, both human and bacterial
kynureninases emerge as potential drug targets. In the present
work, we have compared the reaction of recombinant PfKynase
and HsKynase with a range of ring-substituted kynurenines. The
results show that both the enzymes respond similarly to substitu-
tions at C-5, but differently to substitutions at C-3. These results
may be useful for the design of potent and selective inhibitors of
human and/or bacterial kynureninase.
2. Experimental methods
2.1. Materials
v
¼ Vmax½Sꢂ=ðKm þ ½SꢂÞ
ð2Þ
2.5. Synthesis of kynurenine analogues
L-Kynurenine was obtained from Sigma–Aldrich, and 3-hydro-
2.5.1. Ethyl 2-acetamido-2-carboethoxy-5-oxopentanoate 2-
chlorophenylhydrazone (3a)
xy-DL-kynurenine was a product of USBiochemicals. Common buf-
fers and solvents were obtained from Fisher Scientific.
To a suspension of 9.71 g of diethyl acetamidomalonate in
20 ml benzene was added 97 mg MeONa, with stirring. After
5 min, the suspension was cooled in an ice-water bath, and
3.6 ml of acrolein was added dropwise in about 20–25 min while
maintaining the temperature below 5 °C. After completion of addi-
tion, the reaction was warmed to ambient temperature and stirred
for about 2 h, when a clear pale yellow solution resulted. At the end
of 2 h of stirring, 2.7 ml of AcOH was added, followed by a solution
of 7 g of 2-chlorophenylhydrazine in 14 ml benzene, resulting in a
clear orange solution, which was warmed to 55–60 °C, for about
30 min and the heat was removed. The reaction was left stirring
for 2.5 days at ambient temperature, then it was concentrated in
vacuo, to give a reddish brown oil. Yield, 14 g (72%). 1H NMR:
MeOH-d4, d 7.40 (1H, d, J = 8.9 Hz), 7.29 (1H, t, J = 5.4 Hz), 7.19
(1H, d, J = 8.0 Hz), 7.13 (1H, t, J = 7.8 Hz), 6.69 (1H, t, J = 8.0 Hz),
4.20 (4H, q, J = 7.2 Hz), 2.56 (1H, t, J = 7.3 Hz), 2.22 (1H, m), 2.02
2.2. Instrumentation
1H NMR, 13C NMR, and 19F NMR spectra were recorded on a Var-
ian 400 MHz instrument operating at 400, 100, and 384 MHz,
respectively. HPLC measurements were carried out on a Thermo
Spectrasystem P 2000 pump with a C18 reverse phase column con-
nected to a Thermo UV 6000 diode array detector, at a flow rate of
1 ml/min and controlled by a Dell PC using Chromquest software.
The solvent program was 0.1 M acetic acid/20% MeOH for the first
10 min, followed by a gradient to 70% MeOH over the next 10 min.
GC–MS of the intermediate compounds was done on a Shimadzu
GC-2010 instrument with a 2010QS detector. The steady state ki-
netic measurements and scanning kinetics measurements were
performed on a Varian Cary 1E UV/Visible spectrophotometer
equipped with a Peltier 6 ꢀ 6 thermoelectric cell block for temper-
ature control. The UV instrument was controlled by a PC using soft-
ware provided by Varian Instruments.
Table 1
kmax and
2.3. Enzyme assay
De used for substituted kynurenines
Compound
kmax (nm)
D
e
(Mꢁ1 cmꢁ1
)
Anthranilate
product kmax (nm)
PfKynase was purified from Escherichia coli DH5a cells contain-
ing plasmid pTZKyn, as described previously.16 Recombinant
HsKynase was purified from E. coli BL21(DE3) cells containing the
human gene in plasmid pET100.2 PfKynase activity was measured
360
4500
310
L-Kynurenine (1)
3-Hydroxy (2)
3-Chloro (3)
3-Fluoro (4)
3-Methyl (5)
3-Bromo (6)
5-Bromo (7)
5-Chloro (8)
3,5-Dibromo (9)
5-Bromo-3-chloro (10)
370
365
360
362
360
370
370
378
379
4500
4650
3050
3440
4500
4000
4330
4480
3850
320
323
308
316
323
323
323
330
336
from the decrease in absorbance at 360 nm (De )
= 4500 Mꢁ1 cmꢁ1
upon conversion of kynurenine to anthranilic acid.17 Similarly,
the HsKynase activity was measured from the decrease in absor-
bance at 370 nm (
droxy-DL-kynurenine to 3-hydroxyanthranilic acid.2 The reaction
mixtures for these measurements contained 100 of the
De
= 4500 Mꢁ1 cmꢁ1) upon conversion of 3-hy-
l
M

4672
C. Maitrani, R. S. Phillips / Bioorg. Med. Chem. 21 (2013) 4670–4677
(3H, s), 1.20 (6H, t, J = 7.0 Hz); 13C NMR: MeOH-d4, d 13.4, 21.6,
26.9, 62.5, 66.5, 114, 116.5, 119.3, 127.7, 129.1, 129.2, 141.9,
142.4, 168, 171.2. MS (EI): 142 (100), 144 (32).
124.3, 130.6, 142.1, 170.4, 173.1, 205.3. MS (EI): 154 (100), 156
(32), 252 (90), 254 (30), 384 (21, M+), 386 (7).
2.5.6. Ethyl 2-acetamido-2-carboethoxy-5-oxo-5-(3-fluoro-2-
formamidophenyl)pentanoate (4c)
2.5.2. Ethyl 2-acetamido-2-carboethoxy-5-oxopentanoate 2-
methylphenylhydrazone (5a)
Ozonolysis of 4b (7.8 g), prepared as described previously,21 in
64 ml MeOH with workup as described for 3b gave 3.5 g (45%) of
the product as a brown oil. 1H NMR: MeOH-d4, d 7.62 (1H, d,
J = 7.7 Hz), 7.12 (1H, dd, J = 7.9, 4.0 Hz), 7.05 (1H, m), 6.35 (1H, br
s), 4.27 (4H, q, J = 7 Hz), 4.24 (2H, s), 2.01 (3H, s), 1.25 (6H, t,
J = 7 Hz); 13C NMR: MeOH-d4, d 13.8, 24.2, 36.5, 66.5, 72.5, 119.3,
121.4, 123.2, 127.6, 138.4, 162.5, 169.5, 172.2, 204.1 19F NMR:
MeOH-d4, d ꢁ137.6 (dd, J = 11.7, 4 Hz). MS (EI): 138 (100), 236
(60), 368 (10, M+).
Using the procedure above, 16 g (69%) of a reddish brown oil
was obtained from 7.5 g of 2-methylphenylhydrazine. 1H NMR:
MeOH-d4, d 7.31 (1H, d, J = 8.5 Hz), 7.23 (1H, t, J = 4.9 Hz), 7.05
(1H, t, J = 7.7 Hz), 6.98 (1H, d, J = 8.0 Hz), 6.67 (1H, t, J = 7.7 Hz),
4.21 (4H, q, J = 7.0 Hz), 2.54 (2H, t, J = 7.7 Hz), 2.23 (2H, m), 2.02
(3H, s), 1.98(3H, s), 1.2 (6H, t, J = 7.0 Hz); 13C NMR: MeOH-d4, d
13.4, 16.6, 21.5, 26.8, 30.1, 62.5, 66.6, 112.5, 118.9, 120.8, 126.7,
130.2, 140.7, 143.9, 168, 171.3.
2.5.3. Ethyl 7-chloroindole-3-(3-acetamido-3-
carboethoxy)butanoate (3b)
2.5.7. Ethyl 2-acetamido-2-carboethoxy-5-oxo-5-(2-
formamido-3-methylphenyl)pentanoate (5c)
Compound 3a (14 g) was heated in 85 ml 10% aqueous H2SO4
for about 2 h with vigorous stirring, when a dark brown mixture
resulted. The reaction was cooled to 55–60 °C, 100 ml EtOAc was
added, and the mixture was stirred for about 10 min to dissolve
the semisolid product completely. A solution of 21 g NaCl in
50 ml H2O was added, and stirring was continued at ambient tem-
perature for about 10 min. The organic layer was separated, and
the aqueous layer was extracted with 75 ml EtOAc. The combined
organic layers were washed once with 75 ml of brine, then dried
over anhydrous Na2SO4. The solvent was removed in vacuo to give
a brown semisolid. Yield = 11.0 g (82%) 1H NMR: CDCl3, d 8.38 (1H,
br s), 7.39 (1H, d, J = 7.6 Hz)), 7.18 (1H, d, J = 7.6 Hz), 7.02 (1H, t,
J = 8 Hz), 6.95 (1H, d, J = 2.4 Hz), 6.62 (1H, br s), 4.24 (4H, m),
3.84 (2H, s), 1.99 (3H, s), 1.28 (6H, t, J = 7.2 Hz); 13C NMR:
MeOH-d4 d 13.3, 19.9, 28.2, 62.5, 68.1, 109.4, 116.7, 117.1, 119.7,
120.8, 125.1, 130.2, 133.4, 167.8, 171.5. MS (EI): 164 (100), 166
(25), 321 (52), 323 (19), 380 (10, M+), 382 (3).
Ozonolysis of 5b (11 g) in 110 ml MeOH as described for 3a
gave 6.1 g (55%) of the product as a semisolid. Recrystallization
from 2-propanol gave 4.5 g of the product as a pale yellow solid,
mp 183–185 °C. 1H NMR: CDCl3, d 9.23 (1H, s), 8.30 (1H, s), 7.71
(1H, d, J = 7 Hz), 7.46 (1H, d, J = 8 Hz), 7.28 (1H, t, J = 7.6 Hz), 4.28
(2H, s), 4.26 (4H, q, J = 7 Hz), 2.28 (3H, s), 1.96 (3H, s), 1.25 (6H,
t, J = 7 Hz); 13C NMR: CDCl3, d 13.9, 19.5, 22.9, 43.9, 62.9, 63.9,
125.5, 126.2, 128.0, 128.5, 130.3, 136.2, 158.9, 167.2, 169.7,
200.1. MS (EI): 162 (100), 232 (20), 259 (18), 392 (8, M+).
2.5.8. 3-Chloro-DL-kynurenine (3)
Compound 3c (4.0 g) in 40 ml of 6 N HCl was refluxed for 4 h,
then cooled to ambient temperature, and concentrated in vacuo.
The resulting semisolid was dissolved in 20 ml H2O, treated with
charcoal at ambient temperature for 15 min, and filtered through
Celite. The filtrate was brought to pH 6.5 using 2 N NaOH, when a
solid precipitated. The solid 3-chloro-DL-kynurenine was collected
by filtration, washed with about 5 ml H2O, and allowed to air dry
overnight. Yield, 1.2 g (48%), mp 216–218 °C. 1H NMR: (1% DCl/
D2O) d 7.7 (1H, d, J = 8 Hz), 7.6 (1H, d, J = 8 Hz), 6.7 (1H, t,
J = 8 Hz), 4.2 (1H, t, J = 5 Hz), 3.7 (2H, m); 13C NMR: (1% DCl/
D2O) d 43.2, 55.4, 119.1, 121.8, 123.3, 125.2, 131.4, 143.6, 172.1,
204.2.
2.5.4. Ethyl 7-methylindole-3-(3-acetamido-3-
carboethoxy)butanoate (5b)
Compound 5a (16 g) was reacted in 96 ml of 10% aqueous H2SO4
and worked up as above for 3a. Yield = 11 g (72%) of a brown semi-
solid. 1H NMR: MeOH-d4, d 7.22 (1H, d, J = 7.8 Hz), 6.97 (1H, s), 6.91
(1H, d, J = 7.7 Hz), 6.87 (1H, t, J = 6.4 Hz), 4.18 (4H, m), 3.76 (2H, s),
2.45 (3H, s), 1.95 (3H, s), 1.22 (6H, t, J = 7.1 Hz); 13C NMR: MeOH-
d4, d 13.3, 15.9, 21.6, 28.3, 62.4, 68.2, 108.4, 115.6, 119, 120.8,
121.8, 123.7, 128.1, 135.9, 167.9, 171.3. MS (EI): 144 (100), 301
(40), 360 (10, M+).
2.5.9. 3-Fluoro-DL-kynurenine (4)
Compound 4c (3.5 g) was hydrolyzed in 32 ml of 6 N HCl and
worked up as described for 3c. Yield, 0.65 g (31%), mp 205–
210 °C. 1H NMR: (1% DCl/D2O) d 7.48 (1H, d, J = 7.2 Hz), 7.10 (2H,
m), 4.10 (1H, t, J = 5.0 Hz), 3.49 (2H, d, J = 4.7 Hz); 13C NMR: (1%
DCl/D2O) d 46.5, 53.2, 119.2, 121.6, 123.1, 128.2, 140.6, 159.2,
176.5, 204.2; 19F NMR: MeOH-d4, d ꢁ126 (dd).
2.5.5. Ethyl 2-acetamido-2-carboethoxy-5-oxo-5-(3-chloro-2-
aminophenyl)pentanoate (3c)
A solution of 3b (11 g) in 100 ml MeOH was cooled to below
ꢁ70 °C using a dry ice-acetone bath. Ozone gas was then passed
at 0.5 psig through the reaction mixture for about 90 min. The
reaction was then quenched with an aqueous solution of 44 g so-
dium bisulfite in 200 ml H2O, when a yellow suspension resulted.
The mixture was then stirred for about 10–15 min to allow it to
reach ambient temperature. The MeOH was removed in vacuo,
100 ml H2O was added, and it was extracted with two 75 ml por-
tions of EtOAc. The combined organic layers were washed with
75 ml saturated brine solution, treated with charcoal, filtered over
Celite, and dried over anhydrous Na2SO4. Evaporation of the sol-
vent gave 5.5 g (50%) of a semisolid, which after recrystallization
from 2-propanol gave 4.0 g of the product as a pale yellow solid,
mp 177–178 °C. 1H NMR: CDCl3, d 8.34 (2H, br d), 7.62 (2H, d,
J = 7.6 Hz), 7.27 (1H, t, J = 7.6 Hz), 7.1 (1H, s), 4.27 (4H, q,
J = 7.0 Hz), 4.23 (2H, s), 2.01 (3H, s), 1.26 (6H, t, J = 7.0 Hz); 13C
NMR: MeOH-d4, d 13.5, 20.1, 36.5, 62.2, 70.1, 116.2, 118.6, 122.1,
2.5.10. 3-Methyl-DL-kynurenine (5)
Compound 5c (6.1 g) was hydrolyzed in 54 ml of 6 N HCl and
worked up as described for 3c. Yield, 2.1 g (75%) of a yellow solid,
mp 215–217 °C. 1H NMR: (1% DCl/D2O) d 7.46 (1H, d, J = 7.6 Hz),
7.10 (1H, d, J = 7.2 Hz), 6.97 (1H, t, J = 7.6 Hz), 4.00 (1H, t,
J = 5.2 Hz), 3.43 (2H, d, J = 5.2 Hz), 1.81 (3H, s); 13C NMR: (1%
DCl/D2O) d 16.2, 39.1, 47.2, 126.6, 129.3, 130, 134.3, 138, 142.4,
170.6, 201.
2.5.11. 3-Bromo-
L-kynurenine (6)
3-Bromo-
L
-kynurenine was prepared as described previously.18
2.5.12. 5-Bromo-
5-Bromo- -kynurenine was prepared using our previously pub-
lished procedure.18
L-kynurenine (7)
L

C. Maitrani, R. S. Phillips / Bioorg. Med. Chem. 21 (2013) 4670–4677
4673
Scheme 1. Synthesis of 3-substituted kynurenines.
J = 2.4 Hz), 7.53 (1H, dd, J = 9, 2.4 Hz), 6.50 (1H, d, J = 7.6), 4.96
(1H, dd, J = 11.6, 4 Hz), 3.80 (1H, m), 3.77 (3H, s), 3.67 (1H, m),
2.24 (s, 3H), 2.05 (s, 3H) 13C NMR: CDCl3 d 22.8, 24.8, 40.5, 51.3,
54.6, 121.2, 122.3, 128.5, 129.8, 134.3, 135.6, 168.5, 169.3, 170.1,
200.3.
2.5.14. 5-Chloro-
L
-kynurenine (8)
a
5-Chloro-N ,N-diacetyl- -kynurenine methyl ester (1.6 g) was
L
refluxed in 15 ml of 6 N HCl for 4 h. After concentration, the result-
ing semisolid was taken up in 12 ml water, and treated with char-
coal at room temperature for 10 min. The solution was filtered
through Celite, and the filtrate was brought to pH 6.5 with 6 N
NaOH, when the product precipitated as a pale yellow solid. Yield,
0.70 g (61%), mp, 216–218 °C. 1H NMR: (1% DCl–D2O) d 8.05 (1H, d,
J = 2.4 Hz), 7.61 (1H, dd, J = 9, 2.6 Hz), 7.27 (1H, d, J = 8.2 Hz), 4.44
(1H, t, J = 8.7 Hz), 3.80 (2H, m) 13C NMR: (1% DCl–D2O) d 41.8,
53.6, 118.2, 120.1, 125.3, 130.5, 137.8, 150.1, 170.5, 201.8.
Figure 1. Scanning kinetics for 0.1 mM 5-bromo-L-kynurenine reaction with
HsKynase. The arrows indicate the direction of absorbance change in the reaction.
2.5.15. 3,5-Dibromo-
-Kynurenine (30.0 mg, 0.144 mmol) and NaOAc (26 mg) were
dissolved in 1 ml AcOH, and 16.4 L of Br2 was added with stirring.
L-kynurenine (9)
a
L
2.5.13. 5-Chloro-N,N -diacetyl-
L
-kynurenine methyl ester (8a)
a
l
N,N -Diacetyl-
L
-kynurenine methyl ester (2 g), prepared by
Within an hour, the orange color disappeared and a yellow precip-
itate formed. The reaction was diluted with water to about 10 ml
after 4 h, then filtered to give a light yellow solid, which was
washed with water and dried in vacuo over P2O5 to give 43.3 mg
(82%). 1H NMR: D2O/DCl, d 7.97 (1H, d, J = 2 Hz), 7.82 (1H,
J = 2 Hz), 4.40 (1H, t, J = 5 Hz), 3.75 (2H, m). MS (ESI): 364.8 (50),
366.8 (100), 368.8 (50), (M+1)+.
ozonolysis of N-acetyl-
L
-tryptophan methyl ester and acetyla-
tion,18 was dissolved in 40 ml acetic acid and stirred at room tem-
perature. A solution of 2 g N-chlorosuccinimide dissolved in 12 ml
HCl–saturated AcOH was added, and the reaction was stirred at
room temperature for 1 h. The reaction was quenched with a solu-
tion of 6 g sodium bisulfite dissolved in 24 ml water. The reaction
was then extracted with two 40 ml portions of chloroform, and the
combined organic layers were washed with 40 ml water and then
with 40 ml brine. The organic layer was dried over anhydrous so-
dium sulfate and the solvent was removed in vacuo to give a semi-
solid, which was recrystallized from 2-propanol to give the product
as pale yellow needles. Yield, 1.6 g (72%), mp, 185–187 °C. 1H NMR:
CDCl3 d 11.35 (1H, br s), 8.76 (1H, d, J = 9 Hz), 7.84 (1H, s,
2.5.16. 5-Bromo-3-chloro-DL-kynurenine (10)
3-Chloro-DL-kynurenine (3) (50.0 mg, 0.21 mmol) and NaOAc
(19 mg) were suspended in 1.5 ml AcOH, and 11.9 lL of Br2 was
added with stirring. Within an hour, the orange color disappeared
and a thick yellow precipitate formed. After 4 h, the reaction

4674
C. Maitrani, R. S. Phillips / Bioorg. Med. Chem. 21 (2013) 4670–4677
mixture was diluted with water to about 10 ml, then filtered to
give a light yellow solid, which was washed with water and dried
in vacuo over P2O5 to give 54.9 mg (82%). 1H NMR: D2O/DCl, d 7.74
(1H, d, J = 2.6 Hz), 7.52 (1H, d, J = 2.6 Hz), 4.32 (1H, t, J = 5 Hz), 3.64
(2H, m). MS (ESI): 320.8 (80), 322.8 (100), 324.8 (25), (M+1)+.
of a phenylhydrazone to introduce the side chain, similar to older
methods for synthesis of substituted tryptophans.24 We found that
increasing the aqueous sulfuric acid to 10% in the Fischer cycliza-
tion, rather than the 5% in the published procedure,24 gave higher
yields of indole product. Ozonolysis of the indole product, followed
by hydrolysis and decarboxylation, provides the racemic 3-substi-
tuted kynurenines in good yield (Scheme 1). In contrast to the 3-
position, direct halogenation of kynurenine at the 5-position, para
to the amino group, is facile, so we brominated and chlorinated
2.6. Modeling of 3,5-dibromo-L-kynurenine in HsKynase
The structure of HsKynase with bound 3-hydroxyhippuric acid
(PDB accession number 3E9K) was used to generate the model of
suitably protected
L-kynurenine, prepared by ozonolysis of the
3,5-dibromo-
The stucture of the ligand was modified using PyMOL22 to that of
3,5-dibromo- -kynurenine, by substitution of bromine at the 5-po-
sition, adding the -amino group, changing the 3-OH to a bromine,
and changing the amide nitrogen to a methylene carbon. In a dock-
ing experiment, the structure of HsKynase (PDB accession number,
L
-kynurenine bound to HsKynase (Figs. 2 and 3).
corresponding protected L-tryptophan, to obtain 5-bromo and 5-
chlorokynurenines, similar to our previous procedure.18 Our route
is more efficient than the previously published route to 5-chloro-
L
a
L
-kynurenine, which involved synthesis of the racemate and
resolution.25 3,5-Dibromo-
-kynurenine and 5-bromo-3-chloro-
DL-kynurenine were prepared by direct bromination of -kynuren-
L
L
2HZP) was used to dock 3,5-dibromo-
L
-kynurenine using Auto-
ine and 3-chloro-DL-kynurenine with 2 or 1 equiv, respectively, of
Br2 in glacial acetic acid. All of the kynurenines used in the kinetic
studies were found to be at least 95% pure by HPLC analysis (see
Supplementary data).
dock-Vina23 using PyMOL with the PyMOL Autodock-Vina plugin.
3. Results and discussion
3.2. Kinetic analysis of kynurenine analogues
3.1. Synthesis of kynurenine analogues
All of the analogues were examined as substrates for HsKynase
and PfKynase by repetitive UV/visible scanning of solutions. A
Electrophilic substitution of kynurenine selectively at the 3-po-
sition, ortho to the amino group, is problematic. Previously, we
tried to block the 5-position by iodination,18 but still obtained a
mixture of 3-bromo-5-iodo and 3,5-dibromo products after bro-
mination. Acid-catalyzed hydrolysis and protiodeiodination of the
former product gave a mixture of 3-bromokynurenine and 3,5-dib-
romokynurenine, which were separated by reversed-phase chro-
matography.18 In the present work we have developed a novel
route to 3-substituted kynurenines which starts from the readily
available ortho-substituted anilines, and uses a Fischer cyclization
representative scanning kinetic graph for the reaction of 100
lM
5-bromo- -kynurenine with HsKynase is shown in Figure 1. The
L
reaction shows a time-dependent decrease in the substrate absor-
bance at kmax of 370 nm, an increase in product absorbance at
320 nm, and clear isosbestic points at 245, 258, 280, and 338 nm,
indicating that there are only two absorbing species in the reaction,
the reactant, 5-bromo-L-kynurenine, and the product, 5-bromo-
anthranilate. The absorbance difference at 370 nm between the
Figure 2. Model of 3,5-dibromo-
between the Asn-333 -carbonyl-O and the 3-bromo substituent is 3.1 Å. Hydrogen bonds are shown in yellow, and the proposed halogen bond in cyan. The figure was
created with PyMOL.22
L-kynurenine in the active site of HsKynase. Crossed-eye stereo view showing the PLP, Asn-333, and the bound substrate. The distance
c
.

C. Maitrani, R. S. Phillips / Bioorg. Med. Chem. 21 (2013) 4670–4677
4675
Figure 3. Model of 3,5-dibromo-
L
-kynurenine in the active site of HsKynase. Crossed-eye stereo view of the active site surface with 3,5-dibromo-
L
-kynurenine shown in
spheres. The 5-bromo substituent can be seen as the red sphere projecting up out of the active site into the channel. The figure was created with PyMOL.22
initial and final spectra were used to determine the value of
De
Table 3
Substrates for P. fluorescens kynureninase
(Table 1). All of the compounds examined showed a decrease in
the kynurenine substrate kmax at 360–379 nm and an increase in
the anthranilate product kmax at 308–336 nm (Table 1). The other
Kynurenine substituent
kcat (sꢁ1
)
kcat/Km (Mꢁ1 sꢁ1
)
Relative kcat/
Km
substrates were analyzed in the same way to obtain
compound, the values of which are listed in Table 1.
De for each
None (1)a
16
1
6 ꢀ 105
1
3-Hydroxy-DL (2)a
3-Chloro-DL (3)
3-Fluoro-DL (4)
3-Methyl-DL (5)
2.5 ꢀ 103
0.042
0.018
0.15
0.037
0.0092
5.5
3.0
0.035
0.037
0.71 0.042 (1.1 0.1) ꢀ 104
In previous work, we compared the reactivity of wild-type
6.9 0.5
1.5 0.1
11
11.9 0.7
8.7 0.4
0.23 0.01
0.30 0.02
(9.0 0.8) ꢀ 104
(2.2 0.3) ꢀ 104
5.5 ꢀ 103
HsKynase and several active site mutants with L-kynurenine and
3-hydroxy-DL-kynurenine.3 We found that wild-type HsKynase
has about a 265-fold preference in kcat/Km for 3-hydroxy-DL-kynu-
renine (Table 2). In contrast, PfKynase shows a 24-fold preference
3-Bromo-
5-Bromo-
L
(6)b
(7)
L
(3.3 0.9) ꢀ 106
(1.8 0.3) ꢀ 106
(2.1 0.5) ꢀ 104
(2.2 0.5) ꢀ 104
5-Chloro-
L
(8)
3,5-Dibromo-
5-Bromo-3-chloro-DL
(10)
L
(9)
for L-kynurenine over 3-hydroxy-DL-kynurenine (Table 3). All of the
substituted kynurenines that we studied in the present work are
substrates for both HsKynase and PfKynase, with varying degrees
of efficiency. However, there are some interesting similarities
and differences in the effects of substituents and positions for
the two enzymes. Although some compounds studied were
a
From Ref. 30.
From Ref. 18, at 25 °C.
b
racemates, we found previously that the Km value for racemic b-
benzoylalanine is twice that of the -enantiomer, indicating that
the -iso-
-enantiomer does not significantly inhibit.26 That the
mers are also not substrates is supported by the results of the scan-
ning kinetics experiments described above, where for the DL
compounds the absorbance decreases are half of the initial sub-
Table 2
L
Substrates for human kynureninase
D
D
Kynurenine substituent
kcat (sꢁ1
)
kcat/Km (Mꢁ1 sꢁ1
)
Relative kcat/
Km
None (1)a
0.23
3.5
465
1
3-Hydroxy-DL (2)a
3-Chloro-DL (3)
3-Fluoro-DL (4)
3-Methyl-DL (5)
1.23 ꢀ 105
265
18
6
strate absorbance. For those substrates, the
based on the -amino acid concentration.
De was calculated
0.67 0.063 (8.2 1.3) ꢀ 103
0.23 0.022 (2.7 0.5) ꢀ 103
0.33 0.025 (1.8 0.2) ꢀ 103
L
Bulky substitutents at the 3-position of kynurenine result in
lower kcat and kcat/Km values for PfKynase (Table 3), but higher val-
ues for HsKynase (Table 2). This is consistent with the known pref-
erence for kynurenine over 3-hydroxykynurenine for PfKynase,1,3
and vice-versa for HsKynase. However, 3-F-kynurenine exhibits ki-
netic parameters similar to those of kynurenine for both PfKynase
and HsKynase. Thus, the human enzyme seems to prefer a large,
polar substituent at the 3-position, and in this case fluorine does
4
3-Bromo-
5-Bromo-
5-Chloro-
3,5-Dibromo-
5-Bromo-3-chloro-DL
L
(6)
(7)
(8)
1.9 0.3
(4.4 0.3) ꢀ 103
10
33
24
170
140
L
0.63 0.022 (1.5 0.1) ꢀ 104
0.47 0.022 (1.1 0.1) ꢀ 104
L
L
(9)
1.2 0.1
1.3 0.1
(7.9 1.3) ꢀ 104
(6.5 2.0) ꢀ 104
(10)
a
From Ref. 3.

4676
C. Maitrani, R. S. Phillips / Bioorg. Med. Chem. 21 (2013) 4670–4677
not substitute well for oxygen, despite their similar steric and elec-
tronic properties. In contrast, the catalytic efficiency of the bacte-
rial enzyme is affected the least by introducing a 3-F substituent,
and all larger 3-substituents show dramatic decreases in reactivity
(Table 3). We demonstrated previously that there is a potential ste-
ric clash between the 3-substituent of the substrate and Trp-102 in
PfKynase,3 which could explain the effect of substituents. For
PfKynase, the catalytic efficiency is better for the 3-methyl rather
than a 3-Cl or 3-Br substituent. For HsKynase, the kcat/Km value is
the least for 3-methyl than any other 3-substituent, but it is still
good inhibitors of rat kynurenine aminotransferase.25 We found
that 5-bromokynurenine was an excellent substrate for PfKynase
in our previous work.18 The values of kcat and kcat/Km found in
the present work are significantly higher than those reported pre-
viously, at least in part because of the higher temperature (37 °C vs
25 °C) used in this work. However, there is a significantly larger in-
crease in kcat/Km (20- to 30-fold) for HsKynase than for PfKynase
(3- to 5-fold) with the 5-halo derivatives. The effects of halogena-
tion at the 3- and 5-positions are at least partially additive, since
3,5-dibromo-L-kynurenine and 5-bromo-3-chloro-DL kynurenine
fourfold better than unsubstituted kynurenine. The relative kcat
/
are excellent substrates for HsKynase, with kcat/Km values higher
than any other substrate except the physiological substrate, 3-
OH-kynurenine (Table 2). The effect of the 5-halo substituent on
activity with both kynureninases is likely due to van der Waals
contacts and/or hydrophobic effects, since these substituents occu-
py the tunnel leading from the active site to the outside, as can be
seen from the model of 3,5-dibromokynurenine in the active site of
HsKynase (Fig. 3). In addition, electronic effects may contribute to
the effects of substituents on the activity with HsKynase (Fig. 4),
since there is a reasonable Hammett correlation of the log(relative
kcat/Km) for all substituents at both C-3 and C-5, with
Km values for 3-chloro-DL-kynurenine and 3-bromo- -kynurenine
L
are 17.5 and 8.6 for HsKynase, respectively (Table 2). Thus, having
a Cl or Br substituent at the 3-position of kynurenine increases the
substrate activity for HsKynase much more than smaller or nonpo-
lar substituents, such as methyl or F. In previous structural work,
we found that the 3-OH of an inhibitor, 3-hydroxyhippurate, can
form a hydrogen bond with the
c-carbonyl O of Asn-333 in the ac-
tive site of HsKynase. Previously, we had found a similar result in
3
docking 3-hydroxykynurenine into the active site of HsKynase.2
The roughly 10-fold increase in kcat/Km for HsKynase obtained by
substituting a chlorine or bromine at the 3-position of kynurenine
suggests that halogen bonding27,28 may be occurring. This is con-
sistent with a model of 3,5-dibromokynurenine in the active site
of Hskynase, generated from the 3-hydroxyhippurate structure,
q
= +1.73 0.74, with the major outlier being the physiological 3-
OH substituent, probably as a result of hydrogen bonding. Both
C-3 and C-5 substituents are meta to the reactive substrate car-
bonyl, so
is no significant correlation for the substituent effects on PfKynase
activity (
= ꢁ0.55 1.42), consistent with steric effects playing a
rm values were used in creating the plot. However, there
which shows there is 3.1 Å between the c-carbonyl O of Asn-333
and the 3-Br substituent of 3,5-dibromokynurenine (Fig. 2), an
optimal distance for halogen bonding.27 Halogen bonding involves
a donor electron-rich atom interacting with the positive end of the
C–X bond, the so-called ‘sigma hole’.28 Halogen bonding is minimal
with fluorine, and maximal with bromine and iodine.27,28 The
smaller positive effects of fluorine and methyl substituents at C-3
for kcat/Km with HsKynase may be due to van der Waals forces or
hydrophobic interactions. We note that 3-chlorokynurenine is a
relatively weak inhibitor of rat kynurenine aminotransferase.25
In contrast to the 3-position, chloro and bromo substituents at
the 5-position have stimulating effects on kcat/Km for both PfKyn-
ase and HsKynase (Tables 2 and 3). 5-Halokynurenines are also
q
predominant role at C-3, and van der Waals and/or hydrophobic ef-
fects at C-5. We found previously that 4-substituted (para to the
reactive carbonyl) b-benzoylalanines, which exhibit rate-deter-
mining formation of benzoate products, show a nonlinear Ham-
mett relationship for kcat and kcat/Km with PfKynase, indicating a
change in rate-determining step with substituents.29 We inter-
preted this as resulting from electron-withdrawing substituents
favoring formation of the gem-diol intermediate (Scheme 2), but
reducing its breakdown to the carboxylic acid product. Conversely,
for electron-donating substituents, the formation of the gem-diol is
reduced, but its breakdown is increased. In contrast, for kynuren-
Scheme 2. Mechanism of kynureninase.

C. Maitrani, R. S. Phillips / Bioorg. Med. Chem. 21 (2013) 4670–4677
4677
Supplementary data
Supplementary data (1H NMR, MS and HPLC chromatograms of
the final compounds used in the kinetics studies) associated with
this article can be found, in the online version, at http://
dx.doi.org/10.1016/j.bmc.2013.05.039.
References
1. Soda, K.; Tanizawa, K. Adv. Enzymol. Relat. Areas Mol. Biol. 1979, 49, 1.
2. Lima, S.; Khristoforov, R.; Momany, C.; Phillips, R. S. Biochemistry 2007, 46,
2735.
3. Lima, S.; Kumar, S.; Gawandi, V.; Momany, C.; Phillips, R. S. J. Med. Chem. 2009,
52, 389.
4. Hayaishi, O.; Stanier, R. Y. J. Biol. Chem. 1952, 195, 735.
5. Moriguchi, M.; Soda, K. Biochemistry 1973, 12, 2974.
6. Tanizawa, K.; Soda, K. J. Biochem. 1979, 86, 499.
7. Sathyasaikumar, K. V.; Stachowski, E. K.; Amori, L.; Guidetti, P.; Muchowski, P.
J.; Schwarcz, R. J. Neurochem. 2010, 113, 1416.
8. Forrest, C. M.; Mackay, G. M.; Stoy, N.; Spiden, S. L.; Taylor, R.; Stone, T. W.;
Darlington, L. G. J. Neurochem. 2010, 112, 112.
Figure 4. Hammett plot of log(relative kcat/Km
) for PfKynase (squares) and
9. Guillemin, G. J.; Kerr, S. J.; Brew, B. J. Neurotoxicol. Res. 2005, 7, 103.
10. Szabó, N.; Kincses, Z. T.; Toldi, J.; Vécsei, L. J. Neurol. Sci. 2011, 310, 256.
11. Zwilling, D.; Huang, S. Y.; Sathyasaikumar, K. V.; Notarangelo, F. M.; Guidetti,
P.; Wu, H. Q.; Lee, J.; Truong, J.; Andrews-Zwilling, Y.; Hsieh, E. W.; Louie, J. Y.;
Wu, T.; Scearce-Levie, K.; Patrick, C.; Adame, A.; Giorgini, F.; Moussaoui, S.;
Laue, G.; Rassoulpour, A.; Flik, G.; Huang, Y.; Muchowski, J. M.; Masliah, E.;
Schwarcz, R.; Muchowski, P. J. Cell 2011, 145, 863.
12. Zhang, Y.; Shen, J.; He, X.; Zhang, K.; Wu, S.; Xiao, B.; Zhou, X.; Phillips, R. S.;
Gao, P.; Jeunemaitre, X.; Zhu, D. Circ. Cardiovasc. Genet. 2011, 4, 687.
13. Zelante, T.; Fallarino, F.; Bistoni, F.; Puccetti, P.; Romani, L. Microbes Infect.
2009, 11, 133.
HsKynase (circles). The lines are the regression fits, with y = +1.73 ꢀ x + 0.76 (solid
line) for HsKynase and y = ꢁ0.55 ꢀ x ꢁ 0.90 (dashed line) for PfKynase.
ine with PfKynase, the rate-determining step is release of the sec-
ond product, Ala, not the first product, anthranilate.30 However, in
the present work the effects of halogenation on the reactivity of
HsKynase substrates (Fig. 4) suggests that the formation of the
gem-diol may be rate-determining.
14. Wood, H.; Roshick, C.; McClarty, G. Mol. Microbiol. 2004, 52, 903.
15. Farrow, J. M., 3rd; Pesci, E. C. J. Bacteriol. 2007, 189, 3425.
16. Koushik, S. V.; Sundararaju, B.; McGraw, R. A.; Phillips, R. S. Arch. Biochem.
Biophys. 1997, 344, 301.
4. Conclusion
17. Kishore, G. M. J. Biol. Chem. 1984, 259, 10669.
18. Heiss, C.; Anderson, J.; Phillips, R. S. Org. Biomol. Chem. 2003, 1, 288.
19. Cleland, W. W. Methods Enzymol. 1979, 63, 103.
20. Moriguchi, M.; Yamamoto, T.; Soda, K. Biochemistry 1973, 12, 2969.
21. Lee, M. S.; Phillips, R. S. Bioorg. Med. Chem. Lett. 1991, 1, 477.
22. The PyMOL Molecular Graphics System, Version 1.4.1 Schrödinger, LLC.
23. Trott, O.; Olson, A. J. J. Comput. Chem. 2010, 31, 455.
24. Warner, D. T.; Moe, O. A. J. Am. Chem. Soc. 1948, 70, 2765.
25. Varasil, M.; Della Zorre, A.; Heidempergher, F.; Pevarello, P.; Speciale, C.;
Guidetti, P.; Wells, D. R.; Schwarcz, R. Eur. J. Med. Chem. 1996, 31, 11.
26. Gawandi, V. B.; Liskey, D.; Lima, S.; Phillips, R. S. Biochemistry 2004, 43, 3230.
27. Hardegger, L. A.; Kuhn, B.; Spinnler, B.; Anselm, L.; Ecabert, R.; Stihle, M.; Gsell,
B.; Thoma, R.; Diez, J.; Benz, J.; Plancher, J.-M.; Hartmann, G.; Banner, D. W.;
Haap, W.; Diederich, F. Angew. Chem., Int. Ed. 2011, 50, 314.
28. Clark, T.; Hennemann, M.; Murray, J. S.; Politzer, P. J. Mol. Model. 2007, 13, 291.
29. Kumar, S.; Gawandi, V. B.; Capito, N.; Phillips, R. S. Biochemistry 2010, 49, 7913.
30. Koushik, S. V.; Moore, J. A., 3rd; Sundararaju, B.; Phillips, R. S. Biochemistry
1998, 37, 1376.
These results will be useful in the design of potent and specific
inhibitors of HsKynase or PfKynase that may be useful as drugs.
Since the known potent inhibitors of PfKynase and HsKynase are
proposed to be transition state analogues,18,31–33 the effects of sub-
stituents on kcat/Km are expected to correlate better with inhibitory
potency than Km.34 Thus, halogen substituents in the 5-position are
well tolerated by both enzymes, which is consistent with the X-ray
structures. In our previous work, we found that both diastereomers
of 5-bromodihydrokynurenine inhibit better than the correspond-
ing dihydrokynurenines.18 Furthermore, our result imply that in
the design of inhibitors for HsKynase, a hydroxyl group at the 3-po-
sition is not absolutely necessary for good activity, since the 3,5-di-
halo analogues have excellent substrate activity. Since kynurenine
aminotransferase is inhibited well by 5-halokynurenines, but not
3-halokynurenines, it is likely that compounds can be made that
are selective for HsKynase inhibition. The application of these re-
sults to selective inhibitor design is currently underway in our
laboratory.
31. Phillips, R. S.; Dua, R. K. J. Am. Chem. Soc. 1991, 113, 7385.
32. Dua, R. K.; Taylor, E. W.; Phillips, R. S. J. Am. Chem. Soc. 1993, 115, 1264.
33. Walsh, H. A.; Leslie, P. L.; O’Shea, K. C.; Botting, N. P. Bioorg. Med. Chem. Lett.
2002, 12, 361.
34. Bartlett, P. A.; Marlowe, C. K. Biochemistry 1983, 22, 4618.