Purification and biochemical characterization of some
of the properties of recombinant human kynureninase
Harold A. Walsh and Nigel P. Botting
School of Chemistry, University of St Andrews, St Andrews, Fife, Scotland, UK
Recombinant human kynureninase (
L
-kynurenine hydrol-
ase, EC 3.7.1.3) was purified to homogeneity (60-fold) from
Spodoptera frugiperda (Sf9) cells infected with baculovirus
containing the kynureninase gene. The purification protocol
comprised ammonium sulfate precipitation and several
chromatographic steps, including DEAE–Sepharose
CL-6B, hydroxyapatite, strong anionic and cationic sepa-
rations. The purity of the enzyme was determined by SDS/
PAGE, and the molecular mass verified by MALDI-TOF
MS. The monomeric molecular mass of 52.4 kDa deter-
mined was > 99.99% of the predicted molecular mass. A
UV absorption spectrum of the holoenzyme resulted in a
peak at 432 nm. The optimum pH was 8.25 and the enzyme
displayed a strong dependence on the ionic strength of the
buffer for optimum activity. This cloned enzyme was highly
specific for 3-hydroxykynurenine (K
m
¼ 3.0 l
M
±0.10)
and was inhibited by
L
-kynurenine (K
i
¼ 20 l

tion into the effects of various synthetic and endogenous
inhibitors. Achieving these objectives could provide an
avenue for pharmacological modulation of the synthesis of
the N-methyl
D
-aspartate receptor agonist and the excito-
toxin, quinolinic acid, in addition to elevating the levels of
the neuroprotective kynurenate [1]. Quinolinic acid has been
implicated as an aetiological factor in a range of neurode-
generative diseases which include epilepsy, Huntington’s
disease, AIDS-related dementia, and septicaemia, where it is
released as part of the inflammatory response to injury [2,3].
Kynureninase is one of the enzymes involved in the
tryptophan metabolic pathway. It is a pyridoxal 5¢-phos-
phate (PLP)-dependent enzyme which catalyses the b,c-
hydrolytic cleavage of the amino acids
L
-kynurenine (1)and
L
-3-hydroxykynurenine (2)togive
L
-alanine (3) and either
anthranilic acid (4) or 3-hydroxyanthranilic acid (5)
(Scheme 1) [4]. This pathway is crucial in the biosynthesis
of nicotinamide nucleotides [5] and also gives rise to other
pathophysiologically important compounds such as picolinic
acid, an enhancer of nitric oxide synthase expression [6].
Kynureninase has been purified and characterized from a
number of different sources, such as bacteria, vertebrates
and fungi, but very little is known about the human form.

for concentration and buffer equilibration of active enzyme
fractions were purchased from Millipore (UK).
Protein expression
A cDNA clone encoding human liver kynureninase was a
gift from Dr Andrea Cesura, Hoffman la Roche. The
cDNA was isolated by the method of
1
Alberati-Giani et al.
Correspondence to N. P. Botting, School of Chemistry,
University of St Andrews, St Andrews, Fife KY16 9ST,
Scotland, UK. Fax: + 44 1334 463808, Tel.: + 44 1334 463856,
E-mail:
Abbreviations: PLP, pyridoxal 5¢-phosphate.
Enzyme: kynureninase (
L
-kynurenine hydrolase, EC 3.7.1.3).
7
(Received 27 November 2001, revised 25 February 2002, accepted
25 February 2002)
Eur. J. Biochem. 269, 2069–2074 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.02854.x
[10] and cloned into the ÔBac-to-BacÕ Baculovirus Expres-
sion System (Gibco-BRL) which was used to express
kynureninase in Spodoptera frugiperda (Sf9) insect cells
[11]. Sf9 cells were grown at 28 °C in TC-100 suspension
cultures of 10 L containing a minimal amount of fetal calf
serum (1–2%) until 10 mL of growing cells resulted in a
confluent layer when placed in a small Petri dish. Infection
of the Sf9 cells and expression of kynureninase proceeded in
TC-100 in the absence of any fetal calf serum. The infection
was allowed to proceed for 96 ± 12 h depending on the

)1
apro-
tinin plus 1 lgÆmL
)1
pepstatin and leupeptin, and sonicated
on ice. (All subsequent buffers contained the protease
inhibitors, PLP, dithiothreitol and EGTA at the above
concentrations.) The resultant fraction was then centrifuged
at 12 000 g for 20 min. This procedure was repeated up to
four times with retention of the supernatant. Both supern-
atants were shown to contain all the activity. The superna-
tant was brought to 20% (NH
4
)
2
SO
4
saturation centrifuged
at 10 000 g for 15 min and the pellet discarded. Then the
(NH
4
)
2
SO
4
was increased to 80% to precipitate kynuren-
inase and centrifugation carried out as above. This pellet
was redissolved in 3 mL buffer A, equilibrated with 50 mL
20 m
M

was eluted at  160 m
M
.
Active fractions were again pooled, equilibrated in
20 m
M
Tricine/NaOH, pH 8.8, concentrated to 4.0 mL
and applied to a strong anion-exchange (Macro-Prep strong
S support) column (1.5 · 30 cm) previously equilibrated
with 20 m
M
Tricine/NaOH at pH 8.8 buffer at a flow rate
of 2 mLÆmin
)1
. Bound enzyme was washed with 3 column
vol. equilibration buffer followed by stepwise elution with
NaCl (10–500 m
M
) in column equilibration buffer at a flow
rate of 3.0 mLÆmin
)1
.Theenzymewaselutedat60m
M
NaCl. A Macro-Prep strong Q support column
(1.5 · 30 cm) was equilibrated with 20 m
ML
-His/HCl
buffer at pH 6.0, and the pooled active fractions from the
anion-exchange step were equilibrated (20 m
ML

and was pooled, concentrated, and equilibrated in assay
buffer (10 m
M
Tris/HCl, pH 7.9), divided into aliquots, and
stored at )80 °C until further use. The various purification
steps were followed by SDS/PAGE (12% gels) [12]; where a
Table 1. Fractional purification of recombinant human kynureninase from the supernatant of virus-infected insect (Sf9) cells. Specific details are
outlined in the text. All activity assays were performed with 3-hydroxykynurenine as substrate and at saturating PLP.
Step
Total
protein
(mg)
Total
activity
(nmolÆmin
)1
)
Specific
activity
(nmolÆmin
)1
Æmg
)1
)
Fold
purification
%
Yield
80% (NH
4

and 310 nm and 417 nm respectively for anthranilate, by
the method of Shetty & Gaertner [7]. A Perkin–Elmer
luminescence spectrometer (model LS50B) connected to a
Grant circulating water bath was used for this purpose. The
final reaction volume was 3.0 mL consisting of 25 nmol
PLP (saturating), 10 m
M
potassium phosphate buffer at
pH 7.9, substrate 3-hydroxykynurenine, and an appropriate
volume of enzyme. Enzyme was always added last for all
reactions including the inhibition studies. The amount of
product formed was determined with reference to a
standard curve of fluorescence intensity against 3-hydroxy-
anthranilate concentration. The kinetic assays were per-
formed using both crude and pure (> 95%) forms of the
enzyme. Reproducibility of the experimental findings was
confirmed with enzyme from different batches and varying
degrees of purity in addition to replicates from within the
same batch, and kinetic analyses showed no significant
difference between the various extracts. A progress curve
was constructed to confirm the linear relationship between
product formation, protein concentration, and time. Lin-
earity of the enzymatic reaction was determined over 5 min.
To achieve temperature equilibration (37 °C), the assay
mixture was incubated for at least 5 min before initiation of
the reaction. Graphs were plotted using the
CRICKETGRAPH
and GraphPad
PRISM
3 software packages, and the kinetic

There was, however, a fair amount of tailing between the
two bands which was probably due to the continuous
association and disassociation of the respective subunits.
Owing to the asynchronized viral infection cycle, lysis of a
proportion of the transformed insect cells occurred, as was
observed microscopically. This resulted in the presence of
exogenous active kynureninase in the tissue culture medium.
Hence an 80% (NH
4
)
2
SO
4
precipitation was performed on
the supernatant obtained from harvesting the whole cells.
This fraction was purified separately, and the overall yield
was significantly lower than the whole cell fraction but
sufficient to warrant purification. The total pooled (super-
natant + whole cells) enzyme activity from 10 L culture
medium was  14 lmolÆmin
)1
with a specific activity of
164 nmolÆmin
)1
Æmg
)1
(see Table 1 for fractional purifica-
tion of the supernatant). The purified enzyme was shown to
be purified to homogeneity (Fig. 1) by SDS/PAGE on a
12% gel with subsequent Coomassie Brilliant Blue staining.

reninase (30 lg) at 52.4 kDa in the presence of PLP. The gel was run as
described by Laemmli [11]. Standards were Sigma prestained SDS
molecular mass markers (SDS-7B) in sample buffer containing 4%
SDS and 10% 2-mercaptoethanol.
Ó FEBS 2002 Recombinant human kynureninase (Eur. J. Biochem. 269) 2071
enzyme is subjected to substrate regulation (graph not
shown) and responds to both substrate and inhibitors in a
sigmoidal fashion (Fig. 2). In contrast to previous reports,
no substrate activity could be detected with
L
-kynurenine,
using either a fluorimetric or UV spectroscopic assay.
However,
L
-kynurenine was found to be a competitive
inhibitor at low substrate concentrations (K
i
¼ 20 l
M
)
and non-competitive at higher levels of substrate
(K
i
¢ ¼ 55 l
M
) (Fig. 3).
D
-Kynurenine was also found to
inhibit the enzyme (data not shown, K
i

-3,7-dihydroxydesaminokynurenine (100 n
M
)
were all distinctly sigmoidal, as was the percentage inhibition
graph obtained with
L
-kynurenine (Fig. 4). A reciprocal plot
of the data acquired for the
D
,
L
-3,7-dihydroxydesaminoky-
nurenine-inhibited enzyme clearly reveals a highly
co-operative enzyme throughout the whole substrate range,
with negative co-operativity at low concentrations, which
becomes positive as the substrate levels are increased (data
not shown). Similar results were obtained by Hill analysis.
DISCUSSION
The results describe the first purification of human recom-
binant kynureninase to homogeneity. The protein was fully
Fig. 2. Kynureninase activity as a function of 3-hydroxykynurenine ([s])
in the absence (,) and presence [160 n
M
(h) and 5 l
M
(n)3,7-
dihydroxydesaminokynurenine. Runin10l
M
Tris/HCl buffer
(pH 7.9). Data are mean values of three replicate experiments, and the

m
¼ 3.0 l
M
, specific activity of 164 nmolÆmin
)1
Æ(mg pro-
tein)
)1
and n ¼ 3. Concentrations of
L
-kynurenine in l
M
were 0 (j),
16 (n), 32 (.), 64 (e), 128 (d) and 256 (h).
2072 H. A. Walsh and N. P. Botting (Eur. J. Biochem. 269) Ó FEBS 2002
characterized by electrophoresis (Fig. 1)
3
,MALDI-TOFMS
and UV absorption spectroscopy, and the data are consis-
tent with previous reports [10,11] on the protein. The kinetic
characterization revealed that the human recombinant
kynureninase is specific for 3-hydroxykynurenine, with a
K
m
of 3.0 ± 0.1 l
M
.ThisK
m
value is much lower than
previously reported in this [11] and other laboratories [6] for

value of 1.67 l
M
for rat hepatic kynurenin-
ase. It was also not possible to show any activity towards
L
-kynurenine, and, at the previously reported K
m
values of
400 l
M
or more, there was significant inhibition of the
enzyme. At a concentration of 250 l
ML
-kynurenine in the
presence of 625 n
MD
,
L
-3-hydroxykynurenine, there was
nearly 80% inhibition (Fig. 4). This is a major difference of
human kynureninase from other mammalian enzymes, such
as rat hepatic kynureninase, and may imply that previous
reports of weak activity with
L
-kynurenine in crude cell
homogenates may be the result of additional adventitious
enzyme activity. Certainly the preference for 3-hydroxyky-
nurenine must be taken into account in inhibitor design. Rat
hepatic kynureninase, on the other hand, showed activity
towards

max
and increased K
m
. The shape of the plot is consistent with
binding of the inhibitor to both the free enzyme (E) and the
enzyme–substrate complex (ES) [15] and hence it can be
inferred that an additional ligand-binding site must be
present on the human enzyme.
When the lines intersect above the x-axis then K
i
< K
i
¢,
and when the lines intersect below the x-axis then K
i
> K
0
i
(in both instances the lines have to intersect to the left of the
y-axis). The data obtained for
L
-kynurenine gave
K
i
¼ 20 l
M
and K
i
¢ ¼ 55 l
M

obtained should provide invaluable knowledge on the
active site and also pave the way for co-crystallization of
enzyme–substrate and/or enzyme–inhibitor complexes.
These should allow further mechanistic investigation of
the catalytic reaction and hence facilitate subsequent
design and synthesis of effective inhibitors in an attempt
to combat the deleterious effects of the many serious
neurodegenerative disorders.
ACKNOWLEDGEMENTS
A fellowship to H. A. W. from the Wellcome Trust provided the funds
for this study. Dr C. H. Botting is acknowledged for helpful
discussions, performing the MALDI-TOF MS, and valuable compu-
ting assistance.
REFERENCES
1. Baran, H., Cairns, N. & Lubec, B. (1996) Increased kynurenic acid
levels and decreased brain kynurenine aminotransferase I in
patients with Downs syndrome. Life Sci. 58, 1891–1895.
2. Botting, N.P. (1993) Chemistry and neurochemistry of the
kynurenine pathway of tryptophan metabolism. Chem. Soc. Rev.
45, 309–315.
3. Stone, T.W. (2000) Development and therapeutic potential of
kynurenic acid and kynurenine derivatives for neuroprotection.
Tresnds Pharmacol. Sci. Rev. 21, 149–154.
4. Takeuchi, F., Otsuka, H. & Shibata, Y. (1980) Purification and
properties of kynureninase from rat liver. J. Biochem. (Tokyo) 88,
987–994.
Scheme 3.
Ó FEBS 2002 Recombinant human kynureninase (Eur. J. Biochem. 269) 2073
5. Nishizuka, Y. & Hayaishi, O. (1963) Studies on the biosynthesis of
nicotinamide adenine dinucleotide. J. Biochem. (Tokyo) 238,

14. Dixon, M. & Webb, E.C. (1964) The Enzymes, 2nd edn, pp. 116–
145. Academic Press, New York.
15. Kishore, G.M. (1984) Mechanism-based inactivation of bacterial
kynureninase by b-substituted amino acids. J. Biol. Chem. 259,
259–264.
16. Walsh, H.A., Leslie, P.L., O’Shea, K. & Botting, N.P. (2002)
2-Amino-4-[3¢-hydroxyphenyl]-4-hydroxybutanoic acid; a potent
inhibitor of rat and recombinant human kynureninase. Biorg.
Med. Chem. Lett. 12, 361–363.
5
17. Schott, H H. & Krause, U. (1979) Purification and charac-
terization of 3-hydroxykynureinase from yeast. Z Physiol. Chem.
360, 481–488.
6
18. Soda, K. & Tanizawa, K. (1979) Kynureninases: enzymological
properties and regulation mechanism. Adv. Enzymol. Relat. Areas
Mol. Biol. 49, 1–40.
19. Chiarugi, A., Carpanedo, R. & Moroni, F. (1996) Kynurenine
disposition in blood and brain of mice: effects of selective
inhibitors of kynurenine hydroxylase and of kynureninase.
J. Neurochem. 67, 692–698.
2074 H. A. Walsh and N. P. Botting (Eur. J. Biochem. 269) Ó FEBS 2002