A comparative study on the inhibition of human and bacterial kynureninase by novel bicyclic kynurenine analogues

Article abstract of DOI:10.1016/S0968-0896(00)00318-7

A series of novel bicyclic analogues of kynurenine were synthesised as inhibitors of kynureninase. The trytophan-induced bacterial enzyme from Pseudomonas fluorescens were compared to the constitutive recombinant human enzyme expressed in a baculovirus/insect cell system, with regard to their inhibition by these compounds. All the compounds studied were found to be simple competitive, reversible inhibitors of kynureninase. It was found that altering the size of the second ring of the inhibitor affected the observed K_i values for both enzymes. The addition of an oxygen atom into the second ring had little effect on binding to the bacterial enzyme but gave a more potent inhibitor of human kynureninase. Of the compounds tested, a naphthyl analogue of desaminokynurenine was found to be the most potent inhibitor for both enzymes with K_i values of 5 and 22 μM for bacterial and human enzyme respectively. This report also describes an alternative system for the expression of recombinant human kynureninase which is more convenient for expression in mammalian cells and produces a relatively greater quantity of enzyme. Copyright

Full text of DOI:10.1016/S0968-0896(00)00318-7

Bioorganic & Medicinal Chemistry 9 (2001) 983±989
A Comparative Study on the Inhibition of Human and Bacterial
Kynureninase by Novel Bicyclic Kynurenine Analogues
Deirdre H. Fitzgerald,y Karen M. Muirhead and Nigel P. Botting*
School of Chemistry, University of St. Andrews, St Andrews, Fife, KY16 9ST, UK
Received 23 June 2000;revised 16 November 2000;accepted 21 November 2000
AbstractÐA series of novel bicyclic analogues of kynurenine were synthesised as inhibitors of kynureninase. The tryptophan-
induced bacterial enzyme from Pseudomonas ¯uorescens was compared to the constitutive recombinant human enzyme expressed in
a baculovirus/insect cell system, with regard to their inhibition by these compounds. All the compounds studied were found to be
simple competitive, reversible inhibitors of kynureninase. It was found that altering the size of the second ring of the inhibitor
a€ected the observed Ki values for both enzymes. The addition of an oxygen atom into the second ring had little e€ect on binding to
the bacterial enzyme but gave a more potent inhibitor of human kynureninase. Of the compounds tested, a naphthyl analogue of
desaminokynurenine was found to be the most potent inhibitor for both enzymes with Ki values of 5 and 22 mM for bacterial and
human enzyme respectively. This report also describes an alternative system for the expression of recombinant human kynureninase
which is more convenient for expression in mammalian cells and produces a relatively greater quantity of enzyme. # 2001 Elsevier
Science Ltd. All rights reserved.
Introduction
receptors. Its excitotoxic e€ect at the NMDA receptor
may be a key factor in the aetiology of neuro-
degenerative diseases such as Huntington's disease, epi-
lepsy, AIDS-related dementia and septicaemia. The
design of potent and selective inhibitors of kynureninase
may be useful in the treatment of these disorders.
Furthermore, examination of the binding of such com-
pounds to the enzyme has assisted in probing the active
site of the enzyme and in investigating the mechanism of
the enzymatic reaction, which is still poorly understood.
To date there are relatively few reported inhibitors for
kynureninase and many lack speci®city.4,5 A number of
inhibitors have been developed that mimic the transition
state for the kynureninase catalysed reaction including
(4S) and (4R)-dihydro-l-kynurenine,6 a series of S-aryl-
l-cysteine S,S-dioxides7 and a phosphinic acid analogue
of kynurenine,8 which are all competitive inhibitors of
bacterial kynureninase with varying potency. More
recently Drysdale and Reinhard9 have also examined S-
aryl-l-cysteine S,S-dioxides as inhibitors of mammalian
kynureninase from rat liver and also observed good
inhibition.
Kynureninase [l-kynurenine hydroxylase E.C. 3.7.1.1] is
a pyridoxal 50-phosphate dependent enzyme which cat-
alyses the hydrolytic cleavage of l-kynurenine (1, RˆH)
to anthranilic acid (2, RˆH) and l-alanine (3).1 The
conversion is an important step in the tryptophan
metabolic pathway of Pseudomonas ¯uorescens and
some other bacteria. In mammals, a similar enzyme
[E.C.3.7.1.3], preferentially catalyses the conversion of
3-hydroxykynurenine (1, RˆOH) to 3-hydro-
xyanthranilic acid (2, RˆOH) and l-alanine (Scheme
1).
Further metabolism of hydroxyanthranilic acid on this
pathway,2 ultimately produces nicotinamide adenine
dinucleotide (NAD), an important co-factor for many
enzymatic conversions. De®ciency of vitamin B6 and
hence PLP, results in reduced kynureninase activity and
increased formation of xanthurenic acid, from 3-hydro-
xykynurenine transamination, which is excreted in the
urine.3 More interestingly, quinolinic acid, another
metabolite on the pathway, is an agonist of the N-
methyl-d-aspartate (NMDA) population of glutamate
In this laboratory, we have synthesised a series of
bicyclic analogues of kynurenine as inhibitors of the
enzyme. We report here, the results of our studies using
these compounds as inhibitors of both the bacterial
enzyme (from Pseudomonas ¯uorescens) and recombi-
nant human enzyme, expressed in a baculovirus/insect
*Corresponding author. Tel.: +44-334-463-856;fax: +44-334-463-
808;e-mail: npb@st-andrews.ac.uk
yCurrent address: School of Dentistry, Trinity College, Lincoln Place,
Dublin 2, Ireland.
0968-0896/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(00)00318-7

984
D. H. Fitzgerald et al. / Bioorg. Med. Chem. 9 (2001) 983±989
Scheme 1.
cell system. It is known that the bacterial and mamma-
lian enzyme di€er in their substrate-selectivity and
therefore it is possible that the potency of inhibitory
compounds is species dependent.
activity was determined ¯uorimetrically at 37 ꢀC,
according to the method of Shetty and Gaertner.13
The cDNA clone encoding human liver kynureninase,
which was a kind gift from Dr. Andrea Cesura, Ho€-
man la Roche Ltd., Basel, Switzerland, was isolated by
the method of Alberati-Giani et al.14 and cloned into
the eukaryotic expression vector pBC/CMV.15 The 1600
base pair (bp) kynureninase clone was inserted into the
Sma site of the multiple cloning site and was supplied, in
this form, to this laboratory. The Sma site is ¯anked on
either side by an ECoR1 site and a BamH1 site. The
1600 bp cDNA encoding kynureninase was isolated
from an agarose gel of a double digest, with these two
restriction enzymes.
Results
Synthesis of inhibitors
The inhibitors were all synthesised in racemic form
using the synthetic method5,10 as shown in Scheme 2,
which has also been used recently for the synthesis of 3-
hydroxykynurenine and its O-b-d-glucopyranoside.11
For example the tetralone derivative was prepared from
the aromatic ketone, tetralone (4). This was selectively
brominated on the carbon a to the carbonyl group using
copper (II) bromide in a mixture of ethyl acetate and
chloroform in 90% yield. The bromide (5) was then
used without further puri®cation and reacted with the
carbanion derived from treatment of diethyl acet-
amidomalonate with sodium hydride. This produced the
protected amino acid (6) derivative in 44% yield.
Finally the protecting groups were all removed in a sin-
gle step by heating the compound to re¯ux in 6N HCl.
The desired amino acid (7) was then isolated as the free
base and puri®ed giving 15% of the ®nal product. All
the compounds had the expected spectral data and were
shown to be pure by microanalysis.
The `Bac-to-Bac' Baculovirus Expression System (Gib-
coBRL) was then used to express kynureninase in SF9
insect cells. The expression of the kynureninase was
monitored by SDS±PAGE analysis and by measuring
enzyme activity in the resulting cell sonicates. The total
enzyme activity from a 500 mL suspension culture,
under these conditions was 1±2 mmol/min (speci®c
activity 8±15 nmol/mg/min).
The Km and Vmax values of the bacterial and human
kynureninase preparations were determined at 37 ꢀC
and pH 7.5. For bacterial kynureninase Km was found
to be 44.2Æ1.1 mM and Vmax 1.45Æ0.09 mmol/min/mg
for l-kynurenine. Values obtained for human kynur-
eninase were Km 44.2Æ1.1 mM and Vmax 1.45Æ0.09
mmol/min/mg for l-3-hydroxykynurenine. The kinetic
parameters determined for the human recombinant
enzyme, expressed in the baculovirus/insect cell system
are comparable to those determined14 for the enzyme
Enzyme isolation
The bacterial enzyme was isolated from Pseudomonas
¯uorescens using a modi®cation of the method of
Hayaishi and Stanier.12 The activity of the enzyme
Scheme 2.

D. H. Fitzgerald et al. / Bioorg. Med. Chem. 9 (2001) 983±989
985
expressed in a human embryonic kidney ®broblast cell
line, HEK-293.
reported herein has now allowed us to begin to study
this enzyme further and will be important in the design
of drugs to treat diseases involving impairment of the
kynurenine pathway.
Inhibition studies
For the determination of Ki values, a substrate range
was used, which spanned the Km value. Representative
Lineweaver±Burk plots, for a range of inhibitor con-
centrations, are shown for the interaction of the inda-
none derivative (8) with the human enzyme (Fig. 1).
This graph demonstrates the competitive nature of the
inhibition. Similar plots were obtained for each of the
compounds in the series. For each compound, the
slopes of these graphs were plotted against inhibitor
concentration, to determine the value of Ki. The Ki
values obtained for both the bacterial and human
enzymes, are shown in Table 1. All the compounds stu-
died showed reversibility of inhibition by dilution.
Enzyme activity was restored to 100% on dilution, fol-
lowing inhibition with each compound, but was not
restored when a known irreversible inhibitor, b-chloro-
l-alanine was used.4 A representative graph showing the
reversibility of the indanone derivative (8) is shown
(Fig. 2).
It was already known that bacterial and mammalian
kynureninase di€er in their substrate preference. The
bacterial enzyme is more active towards kynurenine as a
substrate, while the mammalian enzyme preferentially
hydrolyses 3-hydroxykynurenine. All four of the novel
inhibitors tested were indeed found to competitively
inhibit the kynureninase reaction for both bacterial and
human enzymes, Table 1 gives the full set of Ki values.
Note that the synthetic methods used to prepare the
inhibitors were not stereospeci®c. The bicyclic inhibitors
were produced as mixtures of diastereomers (7±9) (due
to chiral centres at C-2 and C-3), while the naphthyl
analogue (10) was a simple racemic mixture. The inten-
tion was that any very promising inhibitors would be
prepared in an enantiomerically pure form at a later
date. The 2S enantiomers might be expected to bind
most tightly as kynureninase only reacts with the 2S-
kynurenine, however the inhibitory e€ect of 2R-kynur-
enine has not been studied in detail.
The basis of the rationale for the design of the bicyclic
inhibitors was that they should be capable of under-
going the early steps of the kynureninase catalysed
reaction, as far as cleavage of the b,g-carbon bond.
However the presence of the new carbon chain linking
the alanine fragment to the 2 position of the benzene
ring means that following this cleavage step the two
products remain covalently attached together. Normally
the anthranilic acid debinds from the enzyme prior to
Discussion
Kynureninase represents an important target for drug
action in some disorders a€ecting the brain. However,
until recently the human enzyme was not available in
sucient quantities to conduct thorough mechanistic
and kinetic investigations. Therefore, most of our pre-
sent understanding of the chemical mechanism and
properties of kynureninase has come from studies of the
bacterial enzyme.16 The cloning of human kynureninase
and the ability to express the activity eciently as
Table 1. Ki values for the inhibition of induced kynureninase from
Pseudomonas ¯uorescens and recombinant human kynureninase
expressed in SF9 cells, by bicyclic kynurenine analogues. Data are the
mean and standard error of at least three separate experiments
Compound
Bacterial
Human
170 Æ 24 mM
227 Æ 47 mM
34.9 Æ 10 mM
162 Æ 31 mM
5 Æ 2 mM
45 Æ 12 mM
77 Æ 23 mM
22 Æ 6 mM
Figure 1. Inhibition of human recombinant kynureninase by (8). The
inhibition of human recombinant kynureninase by (8) with respect to
hydroxykynurenine. Data are presented as Lineweaver±Burk plots for
the following inhibitor concentrations, & 0 mM, ^ 0.05 mM, *, 0.1
mM, ~ 0.2 mM. Initial rates of reaction (v) were assayed as described
in the text . The concentration of substrate was varied between 2.5±20
mM. The graph shown is representative. The ®nal Ki values quoted are
the meanÆstandard error from three separate determinations.

986
D. H. Fitzgerald et al. / Bioorg. Med. Chem. 9 (2001) 983±989
the alanine, but in this case the anthranilic acid type
fragment cannot be released from the active site. It was
thought that this may give tight binding inhibitors
which may even show some irreversible inhibition.
ker. For the 5-membered ring a single CH2 group has
been inserted and in this case the new 5-membered ring
is close to planar. With the 6-membered ring there are
two CH2 groups and the geometry of the new ring is
such that it is puckered with one of the CH2 groups
forced out of the plane. This CH2 group may then clash
with amino acid side chains in the active site increasing
Ki. However the di€erence in binding is rather small to
be the result of a steric clash. The 6-membered ring will
also have more conformational ¯exibility and if only
one of the conformers were to ®t appropriately into the
active site, this would provide a more likely explanation
for the decreased binding.
From the results in Table 1 it can be seen that all the
compounds inhibited to some degree, with all the Ki
values in the mM range and of similar magnitude to the
Km values for the substrate. With the bicyclic inhibitors,
for both the bacterial and human enzymes, altering the
size of the new ring from a 6- to a 5-membered one,
reduced the observed Ki by ca. 5-fold. Such a structural
change would be expected to e€ect the spatial `®t' of the
compound in the active site. This could occur by chan-
ging the relative positions of the carbonyl group and
benzene ring. However, it is very likely that the posi-
tions of these two groups will be ®xed by interactions
with binding groups at the active site. The most impor-
tant factor may thus be the length and bulk of the lin-
When an oxygen is introduced into the linker in com-
pound (9), it has a similar Ki for the bacterial enzyme
when compared with the all carbon ring (6). With the
human enzyme, however, it is 3-fold more potent,
although still not as good as the 5-membered ring. This
may be a result of additional binding interactions
between the ring oxygen and groups at the active site
which usually bind the hydroxyl group of 3-hydro-
xykynurenine, the preferred substrate for the human
enzyme.
With all the bicyclic inhibitors experiments were carried
out to see whether the inhibition was irreversible, or
apparently irreversible due to tight binding of the
cleaved inhibitor. However, no evidence could be
obtained for this and the compounds all appeared to
be simple competitive reversible inhibitors.
The ®nal inhibitor is a naphthyl analogue of desamino-
kynurenine. This appears to be the most potent inhi-
bitor for both enzymes with Ki values of 5 and 22 mM.
In this case it appears that extending the benzene ring of
kynurenine to give a larger planar aromatic system
produces an increase in binding. Presumably the new
aromatic system is ®lling a hydrophobic pocket at the
enzyme active site. It also appeared that there was some,
albeit slow, substrate activity as TLC examination of
incubations of the naphthyl analogue with kynureninase
did show the presence of small amounts of alanine.17
Most previous studies on the inhibition of kynureninase
have included more polar functional groups on the
kynurenine benzene ring, such as methoxy and nitro
groups.6 However, recent work on the inhibition of rat
liver kynureninase by S-aryl-l-cysteine S,S-dioxides9
showed that the most potent derivative (11), was that
with a methyl group in the 5-position, which gave an
IC50 of 11 mM. This increase in binding would also be
consistent with the methyl group ®tting into a hydro-
phobic pocket. These results imply that a fruitful
approach may be to exploit this apparent hydrophobic
binding site to produce more potent inhibitors.
Figure 2. Bacterial kynureninase was assayed as described in Experi-
mental, using l-kynurenine as substrate. The reversibility of the inda-
none derivative (8) is shown above (a) and is compared to a known
irreversible inhibitor, b-chloro-l-alanine (b). In each case, ( ) indi-
cates assays in which the ®nal assay concentration of inhibitor, in a 3
mL cuvette, was as indicated and (&) indicates reactions in which the
enzyme was incubated for 10 min with the same concentration of
inhibitor and then diluted into the assay. Data shown are the mean
Æs.e.m. of three determinations assayed in duplicate.

D. H. Fitzgerald et al. / Bioorg. Med. Chem. 9 (2001) 983±989
987
Experimental
the Ki value determined. The enzyme/inhibitor reaction
solution was then diluted 300-fold into the assay bu€er
and the activity measured by the addition of substrate.
A control series was also set up in which enzyme was
incubated alone for 10 min and diluted into the assay
before adding inhibitor to give the diluted concentra-
tion. If inhibition was reversible, then the activity was
restored following dilution, when compared to an undi-
luted series, in which enzyme was assayed in the pre-
sence of the same concentrations of inhibitor.
Pseudomonas ¯uorescens was obtained from NCIMB,
Aberdeen. DH5a cells were obtained from Promega.
The components of the `Bac-to-Bac' Expression system
(pFastBac donor plasmid, MAX eciency DH10Bac
competent cells, CellFECTIN reagent, and protocols),
were obtained from GibcoBRL. BamH1, ECoR1, T4
DNA ligase and RNase 1, were obtained from Pro-
mega. Ampicillin, gentamycin, kanamycin, tetracycline,
X-Gal and dithiothreiton (DTT) were obtained from
Melford Laboratories. The cDNA clone encoding
human kynureninase was a kind gift from Dr. Andrea
Cesura, Ho€man la Roche Ltd., Basel, Switzerland. All
other compounds were obtained from Sigma Chemical
Co.
Preparation of bacterial enzyme. The bacterial enzyme
was isolated from Pseudomonas ¯uorescens using a
modi®cation of the method of Hayaishi and Stanier.13
The growth of Pseudomonas ¯uorescens was initiated in
`Lab-Lemco' nutrient broth, in an overnight culture at
28 ꢀC. This culture (50 mL) was then used to inoculate
500 mL of the tryptophan-rich growth medium which
was incubated for at least 8 h at 28 ꢀC. The culture was
then used to inoculate 6Â1 L ¯asks of the tryptophan-
rich medium which were incubated for at least 18 h at
28 ꢀC. Tryptophan-rich medium contained 0.3% l-
tryptophan, 0.2% peptone, 0.1% glycerin, 0.05% yeast
extract, 0.1% KH2PO4, 0.2% K2HPO4 and 0.01%
MgSO4, pH 7.0.
Assay of kynureninase. Enzyme activity was determined
¯uorimetrically at 37 ꢀC, according to the method of
Shetty and Gaertner.13 The rate of formation of product
was measured at wavelengths of excitation and emission
of 310 nm and 417 nm (anthranilic acid) and 330 nm
and 410 nm (hydroxyanthranilic acid) using a Perkin
Elmer luminescence spectrometer (MODEL LS50B)
connected to a GRANT circulating water bath. The
reaction mixture contained 20 mM pyridoxal-5-phos-
phate (PLP), 100 mM Tris±HCl bu€er or 40 mM
potassium phosphate bu€er, pH 7.5 and an appropriate
amount of enzyme in a ®nal volume of 3 mL. The reac-
tion was initiated by the addition of l-kynurenine or
dl-hydroxykynurenine. Preliminary assays were per-
formed to con®rm the linear relationship between
product formation and both time and protein con-
centration under these conditions. The amount of pro-
duct formed was measured by reference to standard
curves of ¯uorescence intensity against anthranilate/
hydroxyanthranilate concentration.
The total enzyme activity obtained from a 6 L culture
was about 40 mmol/min. Kynureninase was puri®ed
from the sonicated fraction using modi®cations of pre-
viously described methods.18À20 The bu€er used
throughout the puri®cation was 20 mM potassium
phosphate, pH 7.5 containing 100 mM PLP, 0.01%
mercaptoethanol and 1 mM EDTA.
Following sonication, a protamine sulphate precipita-
tion step and a 20±60% ammonium sulphate step, the
dialysed enzyme was puri®ed by chromatography on a
Fast-Flow DEAE-Sepharose 6B (Pharmacia) column,
using a linear gradient of KCl (0±0.5 M) in 20 mM
potassium phosphate bu€er, pH 7.5 and a volume of
500 mL. PLP was omitted from the bu€ers during this
step. Following concentration of the active fractions to
< 10 mL (amicon ultra®ltration unit-YM10 mem-
brane), the dialysed fraction was chromatographed on a
Sephadex G-100 (Pharmacia) column. The active frac-
tions were pooled and injected onto a mono-Q HR 5/5
FPLC column and the enzyme was eluted using a linear
gradient of KCl (0±0.5 M) in 20 mM potassium phos-
phate bu€er, pH 7.5. This protocol results in a 50±80-
fold puri®cation in a yield of 30±60% and the ®nal spe-
ci®c activity of the preparation varied between 1.2±2.2
mmol/mg/min.
Kinetic parameters were calculated using the Mac-
CurveFit computer program (version 1.2) using non-
linear regression analysis of the experimental data.
Inhibitor studies. Preliminary studies involved screening
compounds for their ability to reduce the rate of reac-
tion at 1 mM inhibitor. Ki values were then determined
by assaying the activity of the enzyme as above, but in
the presence of appropriate concentrations of inhibitor
and at a range of substrate concentrations spanning the
Km value. For inhibitor studies, the reaction was initi-
ated with enzyme rather than substrate.
The values of 1/V were plotted against the reciprocal
concentration of substrate, for each of the inhibitor
concentrations studied. Intersection of the plots on the
X-axis indicates competitive inhibition. A graph of the
slopes of these plots versus inhibitor concentration
gives a straight line with an intercept on the X axis
equal to ÀKi
Cloning and expression of recombinant human kynureni-
nase in baculovirus/insect cell system. The cDNA clone
encoding human liver kynureninase was isolated by the
method of Alberati-Giani et al.14 and cloned into the
eukaryotic expression vector pBC/CMV.15 The 1600
base pair (bp) kynureninase clone was inserted into the
Sma site of the multiple cloning site and was supplied, in
this form, to this laboratory. The Sma site is ¯anked on
either side by an ECoR1 site and a BamH1 site. The
1600 bp cDNA encoding kynureninase was isolated
Reversibility studies. Reversibility of inhibition was
examined by dilution techniques. Enzyme, at 300-fold
its normal assay concentration was incubated for 10 min
with inhibitor, at a range of concentrations, based on

988
D. H. Fitzgerald et al. / Bioorg. Med. Chem. 9 (2001) 983±989
from an agarose gel of a double digest, with these two
restriction enzymes.
Diethyl 2-(acetylamino)-2-(1-oxo-1,2,3,4-tetrahydro-2-
naphthalenyl) malonate (6). Sodium hydride (1.39 g,
34.9 mmol, 2.5 equiv) 60% in mineral oil was suspended
in dry DMF (4 mL) and a solution of diethyl acet-
amidomalonate (4.565 g, 21 mmol, 1.5 equiv) in dry
DMF (14 mL) 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-3,4-dihydro-1-(2H)-
naphthalenone (3.148 g, 13.9 mmol) in dry DMF (10
mL) was added and the solution warmed to room tem-
perature and stirred overnight under nitrogen. The
mixture was poured into distilled water (100 mL), acid-
i®ed to pH 3 with 1M hydrochloric acid in an ice-bath
and extracted into diethyl ether (4Â70 mL). The organic
phases were washed with brine (2Â50 mL), dried
(MgSO4) and the solvent removed under reduced pres-
sure to give an orange oil, which was puri®ed by column
chromatography (silica;diethyl ether) (1.86 g, 44%);
(Acc. mass found 361.151867 calcd. for C19H23NO6
361.152538); dH (300 MHz, CDCl3) 1.23 (6H, t, J 7.2,
CH3CH2CO2), 1.99 (3H, s, CH3CO), 2.83±3.26 (4H, m,
3-CH2 & 4-CH2), 3.93 (1H, dd, J1 13.9, J2 3.75, 2-CH),
4.19±4.30 (4H, m, CH3CH2CO2), 6.86 (1H, br s, NH),
7.26±7.31 (2H, m, 6-CH & 5-CH), 7.47 (1H, t, J 7.7, 7-
CH), 7.93 (1H, d, J 7.7, 8-CH); dC (73.76 MHz, CDCl3)
13.7 (CH3CH2CO2), 23.1 (CH3CO), 26.7 (4-CH2), 29.7
(3-CH2), 55.9 (CH3CH2CO2), 63.0 (2-CH), 66.1 (a-C),
126.7 (7-C), 127.4 (5-C), 128.9 (8-C), 132.3 (6-C), 134.0 (8a-
C), 144.7 (4a-C), 168.6 (CH3CH2CO2), 169.8 (CH3CO),
198.1 (CˆO); m/z (EI) 361 ([M+], 19%), 316 (10, [M-
CH3CH2O]+), 288 (16, [M-CO2CH2CH3]+), 246 (100, [M-
AcHNCCO2H]+), 217 (64, [M-AcHNCCO2Et+H]+),
171 (70, [C11H9NO]+), 145 (55, [C10H9O]+), 129 (30,
[AcHNCCO2CH2]+), 115 (24, [AcHNCCO2H]+).
The `Bac-to-Bac' Baculovirus Expression System (Gib-
coBRL) was used to express kynureninase in SF9 insect
cells. The methods used were those described in the
manufacturer's manual. Kynureninase cDNA was
cloned into pFastBac donor plasmid, which was then
transformed into DH10Bac cells for transposition to the
bacmid. SF9 cells were then transfected with the
recombinant bacmid DNA To determine the optimal
conditions for expression of kynureninase, preliminary
studies were conducted in SF9 cells grown on 6-well
plates, where both the duration of infection and the
concentration of baculovirus were varied Expression of the
kynureninase was monitored by SDS±PAGE analysis
and by measuring enzyme activity in the resulting cell
sonicates. Subsequently, cells were grown in suspension
cultures in 1L-stirred bottles, containing 500 mL total
volume and the duration of baculovirus infection was
96 h.
The harvested cells were resuspended in 20 mM Tris±
HCl, pH 7.5, containing 0.25 M sucrose, 100 mM PLP,
1mM dithiothreitol (DTT), 0.5 mM EDTA, 0.5 mM
phenylmethylsulphonyl ¯uoride (PMSF), 1 mg/mL leu-
peptin, 1 mg/mL pepstatin and 2 mg/mL aprotonin. The
suspension was sonicated (10Â30 s bursts) and cen-
trifuged at 16,400 g. The supernatant was decanted, ali-
ꢀ
quoted and stored at À20 C until required for assay.
The total enzyme activity from a 500 mL suspension
culture, under these conditions was 1±2 mmol/min (spe-
ci®c activity 8±15 nmol/mg/min).
Synthesis of inhibitors
Amino-(1-oxo-1,2,3,4-tetrahydro-2-naphthalenyl)-acetic
acid (7). Diethyl 2-(acetylamino)-2-(1-oxo-1,2,3,4-tetra-
hydro-2-naphthalenyl)-malonate (1.63 g, 4.5 mmol) was
dissolved in 1,4-dioxane (50 mL) and 6M hydrochloric
acid (70 mL) added. The reaction was heated under
re¯ux for 8 h until no starting material was visible by
TLC (silica;pet. ether: ethyl acetate;1:1). The solution
was then cooled and washed with ethyl acetate (50 mL).
The aqueous phase was concentrated under reduced
pressure to give a brown liquid which was triturated
with acetone to produce a grey solid. The solid was
taken up in isopropyl alcohol (25 mL) and propylene
oxide (10 mL) added, the solution was stirred overnight
and the solvent removed under reduced pressure to give
the product as an o€-white solid. The product was
obtained as a mixture of diastereomers in a 3:1 ratio and
spectral data is given for the major isomer (0.147 g,
15%);mp 186±192 ꢀC;(Found: C, 65.90;H, 6.09;N,
6.31. Calcd. for C12H13NO3: C, 65.74;H, 5.97;N,
6.38%); dH (200 MHz, D2O) 2.04 (2H, t, J 6.25, 4-CH2),
3.01 (2H, d, J 6.25, 3-CH2), 3.13- 3.38 (1H, m, 2-CH),
4.25- 4.37 (1H, m, a-CH), 7.29 (2H, m, 7-CH & 5-CH),
7.52 (1H, t, J 6.25, 6-CH), 7.88 (1H, d, J 6.25, 8-CH); dC
(50.31 MHz, D2O) 27.8 (4-CH2), 30.1 (3-CH2), 64.2 (2-
CH), 66.5 (a-C), 126.5 (7-C), 127.1 (5-C), 128.5 (8-C),
132.5 (6-C), 133.7 (8a-C), 144.2 (4a-C), 173.4 (CO2H),
197.0 (C=O); m/z (CI) 220 ([M+H], 25%), 203 (100,
[MH-OH]+), 147 (22, [C10H10O]+).
Full experimental details are given below for the reac-
tion sequence in Scheme 2, followed by full spectro-
scopic data for the other three inhibitors synthesised
using similar methods.
2-Bromo-3,4-dihydro-1-(2H)-naphthalenone (5). Cupric
bromide (6.126 g, 27.4 mmol, 2 equiv) was heated at
re¯ux in ethyl acetate (10 mL) with stirring. To this was
added 3,4-dihydro-1-(2H)-naphthalenone (2.091 g, 14.3
mmol) in chloroform (10 mL). The reaction was heated
at re¯ux for a further 5 h and then cooled. Copper
bromide and cupric bromide residues were ®ltered o€,
the ®ltrate decolourised with activated charcoal and
®ltered through a bed of Celite and washed with
ethyl acetate (4Â50 mL). The solvent was removed
under reduced pressure to give an orange oil which
was dried further under vacuum (2.9 g, 90%); dH (200
MHz, CDCl3) 2.49 (2H, t, J3,4 4, 4-CH2), 2.93 (1H, dt,
J2,3 17.1, J3,4 4, HA, 3-CH2), 3.26±3.41 (1H, m, HB,
3-CH2), 4.74 (1H, s, CHBr), 7.29 (1H, d, J 7.7, 5-CH),
7.38 (1H, t, J 7.7, 7-CH), 7.54 (1H, t, J 7.7, 6-CH), 8.10
(1H, d, J 7.7, 8-CH); dC (50.31 MHz, CDCl3) 26.6
(4-CH2), 32.5 (3-CH2), 51.1 (2-CH), 112.8 (7-C), 127.7
(5-C), 129.2 (8-C), 134.7 (6-C), 137.5 (8a-C), 143.2
(4a-C), 191.1 (CˆO); m/z (CI) 227 & 225 ([M+H],
98% and 100%), 147 (83, [C10H10O]+), 118 (12,
[C(O)ArCH3+H]+).

D. H. Fitzgerald et al. / Bioorg. Med. Chem. 9 (2001) 983±989
989
Amino-(1-oxo-2,3-dihydro-1H-inden-2-yl)-acetic acid (8).
Acknowledgements
The product was obtained as a mixture of diastereomers
in a 3:2 ratio and spectral data is given for the major
isomer;mp 152±158 ꢀC (decomp.);(Acc. mass found
206.082462 calcd. for C11H12NO3, 206.081718); dH (200
MHz, D2O) 3.1 (1H, dd, J1 14, J2 4.5, 2-CH), 3.77±3.93
(1H, d, J 4.5, a-CH), 4.14±4.3 (2H, m, 3-CH2), 7.31 (2H,
m, 4- and 6-CH), 7.51 (1H, t, J4,5=J5,6 7.5, 5-CH), 7.67
(1H, d, J6,7 7.5, 7-CH); dC (50.31 MHz, D2O) 44.1 (3-
CH2), 52.4(2-CH), 58.5 (a-CH), 126.7 (7-C), 129.7 (4-
C), 130.6 (5-C), 131.6 (7a-C), 138.9 (6-C), 155.3 (3a-C),
175.1 (CO2H), 190.5 (CˆO); m/z (CI) 206 ([M+H],
15%), 191 (76, [M+H-NH]+), 151 (45, [M+H-
NH2CHCO]+), 76 (99, [M+H-Ar]+).
The authors wish to thank Dr. Andrea Cesura for the
gift of the cDNA clone encoding human kynureninase.
We also wish to thank the BBSRC for ®nancial support
for DHF (a research grant) and KMM (a quota
studentship).
References
1. Soda, K.;Tanizawa, K. Biochem. J. 1979, 86, 1199.
2. Botting, N. P. Chem. Soc. Rev. 1995, 24, 401.
3. Yeh, J.-K.;Brown, R. R. J. Nutrition 1977, 107, 261.
4. Kishore, G. M. J. Biol. Chem. 1984, 259, 10669.
5. Pellicciari, R.;Natalini, B.;Constantino, G.;Mahmoud,
M. R.;Mattoli, L.;Sadeghpour, B. M.;Moroni, F.;Chiarugi,
A.;Carpenedo, R. Journal of Medicinal Chemistry 1994, 37,
647.
6. Phillips, R. S.;Dua, R. K. J. Am. Chem. Soc. 1991, 113,
7385.
7. Dua, R. K.;Taylor, E. W.;Phillips, R. J. Am. Chem. Soc.
1993, 115, 1264.
Amino-(4-oxo-3,4-dihydro-2H-chromen-3-yl)-acetic acid
(9). The product was obtained as a mixture of diaster-
eomers in a 3:1 ratio and spectral data is given for the
major isomer;mp 138±140 ꢀC [decomp.];(Acc. mass
found
244.059856
calcd.
for
C11H11NO4Na,
244.058578); dH (300 MHz, DCl/D2O) 3.37 (1H, dt, J2,3
9.3, J3,9 2.6, 3-CH), 4.01 (1H, d, J3,9 2.6, a-CH), 4.31
(2H, d, J2,3 9.3, 2-CH2), 6.60 (1H, d, J7,8 7.9, 8-CH),
6.66 (1H, t, J5,6 7.9, 5-CH), 7.17 (1H, dt, J7,8 7.9, J6,7
1.7, 7-CH), 7.35 (1H, dd, J5,67.9, J6,7 1.7, 6-CH); dC
(300 MHz, DCl/D2O) 43.9 (3-CH), 47.8 (a-CH), 66.3
(2-CH2), 115.8 (8-C), 120.0 (6-C), 124.7 (5-C), 135.7
(7-C), 159.5 (4a-C), 167.7 (1a-C), 171.7 (CO2H), 191.6
(CˆO); m/z (CI) 244 ([M+H+Na], 19%), 205 (100,
[M-NH2]).
8. Ross, F. C.;Botting, N. P.;Leeson, P. D. Bioorg. Med.
Chem. Lett. 1996, 6, 2643.
9. Drysdale, M. J.;Reinhard, J. F. Bioorg. Med. Chem. Lett.
1998, 8, 133.
10. Greenstein, J. P.;Winitz, M. Chemistry of the Amino
Acids;John Wiley and Sons Inc. 1961, p 711±712.
11. Manthey, M. K.;Jamie, J. F.;Truscott, R. J. W. J. Org.
Chem. 1999, 64, 3930.
12. Hayaishi, O.;Stanier, R. Y. J. Biol. Chem. 1952, 195, 735.
13. Shetty, A. S.;Gaertner, F. H. J. Bacteriol. 1973, 113,
1127.
14. Alberati-Giani, D.;Buchli, R.;Malherbe, M.;Broger, C.;
Lang, G.;Kohler, C.;Lahm, H. W.;Cesura, A. M.
Biochem. 1996, 239, 460.
15. Bertocci, B.;Miggiano, V.;Da Prada, M.;Dembic, Z.;
Lahm, H. W.;Malherbe, P. Proc. Natl. Acad. Sci. U.S.A.
1991, 88, 1416.
16. Koushik, S. V.;Moore, J. A.;Sundararaju, B.;Phillips, R.
S. Biochemistry 1998, 37, 1376.
17. Ross, F. C.;Botting, N. P. unpublished results.
18. Moriguchi, M.;Yamamoto, T.;Soda, K. Biochem. Bio-
phys. Res. Commun. 1971, 44, 752.
2-Amino-4-oxo-4-naphthyl-butyric acid (10). Mp 201 ꢀC;
(Found: C, 69.00;H, 5.55;N, 5.78. Calcd. for
C14H13NO3: C, 69.12;H, 5.39;N, 5.76%); umax (nujol)/
cmÀ1 1690 (CˆO), 1650 (CˆO acid), 1580 (aromatic);
m, a-CH), 7.65 (2H, m, 60H, 70H), 8.05 (4H, m, 30H,
40H, 50H, 80H), 8.68 (1H, s, 10H); dC (74.76 MHz;
C2H3O2H) 39.42 (b-C), 50.01 (a-C), 124.44 (30-C),
128.34 (70-C), 128.96 (50-C), 129.89 (40-C), 130.33
(60-C), 130.95 (80-C), 131.87 (10-C), 134.08 (100-C),
134.31 (90-C), 137.62 (20-C), 171.45 (CO2H), 197.82
(CˆO); m/z (FAB) 266 ([M+Na]+, 25%), 244 (65,
[M+H]+).
Eur. J.
d
H (200 MHz;CD 3OD), 3.95 (2H, m, 3-CH2), 4.50 (1H,
19. Ishikawa, T.;Okuno, E.;Kawai, J.;Kido, R. Comp. Bio-
chem. Phys. B. 1989, 93, 107.
20. Takeuchi, F.;Otsuka, H.;Shibata, Y. J. Biochem. 1980,
88, 997.