CrystalstructuresofAedesaegypti kynurenine
aminotransferase
Qian Han
1,
*, Yi Gui Gao
1,2,
*, Howard Robinson
3
, Haizhen Ding
1
, Scott Wilson
2
and Jianyong Li
1
1 Department of Pathobiology, University of Illinois, Urbana, IL, USA
2 School of Chemical Sciences, University of Illinois, Urbana, IL, USA
3 Biology Department, Brookhaven National Laboratory, Upton, NY, USA
Aedes aegyptikynurenineaminotransferase (AeKAT)
is a multi-function aminotransferase [1]. Its mamma-
lian homolog, KAT-I can catalyze several amino
acids and many biologically relevant keto acids [2,3],
and is identical to glutamine transaminase K and
also a cysteine S-conjugate b-lyase [4–8]. KAT-I is
present in the brain [9,10], and also in the kidney
and liver [11,12], which indicates the important role
of KAT-I in the bioactivation of environmental pol-
lutants that contribute to liver- and kidney-associated
carcinogenesis [2]. Although the kidney and liver
show much greater KAT activity than the brain, the
emphasis of KAT research has been almost exclu-
sively on enzymes in the CNS, paralleling the investi-
gation into the pivitol role of kynurenic acid
(KYNA) therein.
Keywords
aminotransferase; crystal structure;
kynurenic acid; kynurenine
aminotransferase; mosquito
Correspondence
J. Li, Department of Pathobiology,
University of Illinois, 2001 South Lincoln
Avenue, Urbana, IL 61802, USA
Fax: +217 2447421
Tel: +217 244–3913
E-mail: jli2@uiuc.edu
*Qian Han and Yi Gui Gao contributed
equally to this work.
Note
The atomic coordinates and structure fac-
tors (PDB codes 1YIZ and 1YIY) have been
deposited in the Protein Data Bank,
Research Collaboratory for Structural Bioin-
formatics, Rutgers University, New Bruns-
wick, NJ, USA (http://www.rcsb.org).
(Received 2 February 2005, accepted 7
March 2005)
doi:10.1111/j.1742-4658.2005.04643.x
Aedes aegyptikynurenineaminotransferase (AeKAT) catalyzes the irrevers-
ible transamination ofkynurenine to kynurenic acid, the natural antagonist
of NMDA and 7-nicotinic acetycholine receptors. Here, we report the crys-
tal structure of AeKAT in its PMP and PLP forms at 1.90 and 1.55 A
˚
,
respectively. The structure was solved by a combination of single-wave-
length anomalous dispersion and molecular replacement approaches. The
initial search model in the molecular replacement method was built with
the result of single-wavelength anomalous dispersion data from the Br-
AeKAT crystal in combination with homology modeling. The solved struc-
ture shows that the enzyme is a homodimer, and that the two subunits are
stabilized by a number of hydrogen bonds, salts bridges, and hydrophobic
interactions. Each subunit is divided into an N-terminal arm and small and
large domains. Based on its folding, the enzyme belongs to the prototypical
fold type, aminotransferase subgroup I. The three-dimensional structure
shows a strictly conserved ‘PLP-phosphate binding cup’ featuring PLP-
dependent enzymes. The interaction between Cys284 (A) and Cys284 (B) is
unique in AeKAT, which might explain the cysteine effect of AeKAT
activity. Further mutation experiments of this residue are needed to eventu-
ally understand the mechanism of the enzyme modulation by cysteine.
Abbreviations
AeKAT, Aedesaegyptikynurenine aminotransferase; KAT, kynurenine aminotransferase; KYNA, kynurenic acid; MAD, multiwavelength
anomalous dispersion; PLP, pyridoxal 5-phosphate; PMP, pyridoxamine 5-phosphate; rmsd, root mean square deviation; SAD, single
wavelength anomalous dispersion.
2198 FEBS Journal 272 (2005) 2198–2206 ª 2005 FEBS
KYNA is a metabolite in the tryptophan metabolic
pathway. In mammals, it is synthesized by irreversible
transamination ofkynurenine by KATs. KYNA repre-
sents the only known endogenous antagonist of the
excitatory action of ionotropic excitatory amino acids,
showing the highest affinities for the glycine modula-
tory site of the NMDA subtype of glutamate receptor
[13–15] and the a7-nicotinic acetylcholine receptor [16–
18]. In mammals, it protects the CNS from oversti-
mulation by excitatory cytotoxins [19,20]. The inhibi-
tory actions of KYNA at excitatory amino acid KAI
receptors underlie its neuroprotective [19–22] and
anticonvulsant effects [23,24]. Fluctuations in endo-
genous brain KYNA levels significantly influence neur-
onal excitation and vulnerability to excitotoxic attack
[25–28]. In addition, KYNA is also involved in main-
taining physiological arterial blood pressure [29–33].
The role of activity-dependent synaptic plasticity in
learning and memory is a central issue in neuroscience.
Much of the relevant experimental work concerns the
possible role of long-term potentiation in learning.
Most forms of long-term potentiation are glutamater-
gic and the most prominent form is induced following
activation of the NMDA receptor [34]. KYNA, as the
only natural antagonist of NMDA, may, therefore, be
involved in the processes of memory and learning in
the CNS. Savvateeva et al. [35] demonstrated that the
mutant cardinal fly (3-hydroxykynurenine excess, local
KYNA level might be affected) shows a decline in
learning and memory, which implies a possible role for
KYNA in the formation of long-term potentiation.
Direct evidence is, however, still missing. The physiolo-
gical importance of KYNA has attracted a consider-
able amount of attention towards understanding the
molecular regulation of KYNA production in living
organisms. KATs have become the target enzymes
when studying modulation of the KYNA level in a
number of pathological conditions in animals.
A. aegypti KAT (AeKAT) shares 45–50% sequence
identity with mammalian or human KAT-Is [36].
Functional characterization of its recombinant protein,
expressed in a baculovirus ⁄ insect cell-expression sys-
tem, showed that the protein is active to kynurenine
[36]. The protein showed high activity towards many
biologically relevant keto acids. Interestingly, most
keto acids showed substrate inhibition at relatively
high concentrations. Cysteine had an intriguing effect
on the enzyme activity towards kynurenine, inducing
enhancement at relative low concentrations and inhibi-
tion at higher concentrations [1]. Moreover, AeKAT is
mainly expressed in adult heads, indicating its major
function in the CNS [36]. A bacterial homolog and
human KAT-I have been systematically characterized
using their respective recombinant enzymes [3,37], and
both three-dimensional structures have recently been
solved [38,39]. The biochemical comparison of the three
enzymes has been discussed previously [3]. To under-
stand the catalytic mechanism and structural basis
underlying these biochemical differences, it is essential
that the three-dimensional structure of AeKAT is deter-
mined and a comparative study with mammalian KATs
is carried out. Here, we provide data that describe the
crystal structure of AeKAT as obtained using macro-
molecular crystallography.
Results and Discussion
Crystallization, single wavelength anomalous
dispersion modeling, and homology modeling
Single wavelength anomalous dispersion (SAD) diffrac-
tion data for a Br-AeKAT (PMP form) derivative were
collected at cryogenic temperature at X12C at the
National Synchrotron Radiation Source in Brookha-
ven National Laboratory (BNL) (k ¼ 0.97 A
˚
). The
Br-AeKAT (PMP form) crystal has an orthorhombic
unit cell with parameters of a ¼ 55.34 A
˚
, b ¼ 95.32 A
˚
,
c ¼ 167.67 A
˚
, and diffracts to 1.90 A
˚
resolution. The
space group has been determined as P2
1
2
1
2
1
from an
auto-index using gadds program. Two molecules of
AeKAT are in an asymmetric unit, based on calcula-
tion of the Matthews coefficient [40]. SAD diffraction
data of the Br-AeKAT (native ⁄ PLP form) derivative
were collected in the same manner (k ¼ 1.1 A
˚
). The
PLP form crystal has an orthorhombic unit cell with
parameters of a ¼ 55.28 A
˚
, b ¼ 94.98 A
˚
, c ¼ 167.60 A
˚
and diffracts to 1.55 A
˚
resolution (Table 1).
An initial atomic model with 520 residues of two
molecules was obtained based on the SAD data of Br-
AeKAT crystals with a resolution at 1.90 A
˚
(Fig. 1A).
Because the SAD model has only 60% residues
assigned, we turned to a molecular replacement method
using homology modeling. The first model was built by
homology modeling using the initial SAD model and
two homology structures (PDB codes: 1gck and 1v2d)
as template structures. Briefly, SAD coordinates were
assigned to target residues of AeKAT, the structures of
other residues were built based on two search models.
Loop areas were highly optimized to target protein Ae-
KAT using the program insight ii (Accelrys). In total
100 models were built, 10 of which were used in the ini-
tial refinement tests based on the procheck [41] results.
Only one model (Fig. 1B) was used in further refine-
ment using shelx-97, o and x-plor. This strategy
enabled us to solve the three-dimensional structure of
AeKAT. Figure 1 shows the initial SAD model (A), a
Q. Han et al. Mosquito kynurenineaminotransferase structure
FEBS Journal 272 (2005) 2198–2206 ª 2005 FEBS 2199
homology model (B) and the final refined structure of
the PMP form (C) of AeKAT.
Overall structure
The PMP form of AeKAT was solved at a resolution of
1.90 A
˚
. The three-dimensional structure of the AeKAT
PLP form was solved using its PMP structure as an ini-
tial refinement model by shelx-97 and x-plor at a reso-
lution of 10 to 1.55 A
˚
. A homodimer is present in the
asymmetric unit. Excellent electron density allowed the
modeling of 418 of 429 residues, 2 PLPs and 441 sol-
vent molecules in the PLP form structure. Both forms
have the same structure except that there is no bond
formed between the cofactor and Lys255 in the PMP
form. The stereochemistry of the model was assessed
using procheck [41]. In both the PLP and PMP forms,
87% of the residues were in the most favored regions of
the Ramachandran plot. Although Tyr286 (A) and (B)
in both forms fall within a disallowed region of the
Ramachandran plot of the solved structures, the excel-
lent electron density allowed us to unambiguously
assign the observed conformation.
The protein architecture revealed by AeKAT con-
sists of the prototypical fold of aminotransferases sub-
group I [42,43], characterized by an N-terminal arm,
and a small and a large domain (Fig. 2). The N-ter-
minal arm consists of a random coiled stretch made up
of residues 12–26, the small domain (residues 27–52
and 310–429) folds into a five-stranded parallel and
antiparallel b sheet surrounded by four a helices. The
large domain (residues 53–309) adopts an a ⁄ b structure
that resembles the Rossmann fold, which shows con-
served a ⁄ b topology, in which a sharply twisted seven-
stranded b-sheet inner core is nested into a conserved
array of nine a helices which are contributed by both
the interior and external of the molecule.
As observed in other subgroup I aminotransferases,
the functional unit of AeKAT consists of a homodimer
with subunits related by a dyad axis to its two active
sites located at the domain interface in each subunit,
and at the subunit interface in the dimer (Fig. 2).
Many hydrogen bonds are formed between two sub-
units, i.e. Lys17–Gln119 (2.63, 2.68 A
˚
), Lys17–Val122
(2.69, 2.89 A
˚
), Asn71–Trp262 (2.89, 2.95 A
˚
), Asn71–
Gly261 (2.87, 2.95 A
˚
), Trp72–Thr260 (2.52, 3 A
˚
),
Fig. 1. Line ribbon representation of initial
SAD model, homology model and final
refined model. (A) Initial SAD model. (B)
Homology model. (C) Final refined model of
the PMP form. The molecules are viewed
down the molecular twofold axis. The two
subunits are colored blue and green,
respectively.
Table 1. Crystal parameters, data collection and refinement statis-
tics of AeKAT.
Data collection PLP form PMP form
Space group P2
1
2
1
2
1
P2
1
2
1
2
1
Unit cell (a, b, c) 55.28, 94.98,
167.60
55.34, 95.32,
167.67
Resolution (A
˚
) 1.55 1.90
Observation reflections 924 984 716,820
Unique reflections 128 458 68,715
R
merge
(%) 5.7 (29.1) 6.3 (19.6)
Redundancy 7.2 (6.7) 10.4 (9.8)
Completeness (%) 99.3 (99.8) 97.1 (90.5)
Refinement
R-all (%) 26.13 22.37
R-observed (%) 25.42 21.85
R-work (%) 25.4 21.8
R-free (%) 27.9 26.4
RMS bond lengths (A
˚
) 0.010 0.007
RMS bond angles (°) 2.41 1.72
All-atom RMS fit for
the two chains (A
˚
)
1.88 2.08
Ca-only RMS fit
for the two chains (A
˚
)
1.49 1.65
No. of protein atoms 2 · 3315 2 · 3315
No. of PLP ⁄ PMP atoms 2 · 15 2 · 16
No. of heavy atoms (Br-) 5 5
No. of solvent molecules (water) 441 440
Average B factor main chain (A
˚
2
) 18.43 18.26
Average B side chain (A
˚
2
) 23.24 23.78
Average B over all (A
˚
2
) 20.82 20.99
Average B factor PLP ⁄ PMP (A
˚
2
) 15.57 12.41
Average B factor
solvet (water) (A
˚
2
)
27.33 26.97
Mosquito kynurenineaminotransferase structure Q. Han et al.
2200 FEBS Journal 272 (2005) 2198–2206 ª 2005 FEBS
Trp72–Ser258 (2.84, 2.95 A
˚
), Tyr73–Thr250 (2.87,
3.9 A
˚
), Tyr111–Gln282 (2.79, 2.95 A
˚
), two Tyr115
(2.57 A
˚
), and Gly261–Thr290 (2.78, 2.85 A
˚
) of the
opposite subunits in PLP form. Several hydrophobic
interactions participate in the stabilization of the
homodimer, in particular, the interactions between
Phe46 and Leu69 (3.57, 4.39 A
˚
) and the two Val108
(4 A
˚
) of the opposite subunits. Moreover, the N-ter-
minal arms also contribute towards the stability of
the AeKAT dimer. Two salt bridges are established
between Asp112 and Lys6 (3.96, 4.15 A
˚
) and Asp112
and Arg7 (3.48, 3.55 A
˚
) of the opposite subunits. In
AeKAT structures, Cys284 (A) and Cys284 (B) with
distances of 3.46 A
˚
(PLP form) and 3.57 A
˚
(PMP
form) form a thiol–thiolate hydrogen bond, which is
unique in AeKAT (Fig. 5B).
The AeKAT active site
Similar to human KAT-I, AeKAT possesses two active
sites located around the local dyad axis. Each active
site contains one PLP molecule and is hosted in a deep
cleft at the domain interface made up of residues from
both subunits (Fig. 2). Both PLP and PMP can be
identified with confidence, with clear electronic density
maps of the structures (Fig. 3). Each PLP cofactor sits
within a binding pocket defined by two regions contri-
buted by residues from the large domains of both sub-
units (Fig. 2). The bottom of the PLP-binding pocket
is entirely defined by residues from the large domain
of the corresponding monomer. With the exception of
Lys255 and Gly110, all these residues are at, or close
to, the edge of the inner core b sheet, pointing toward
the domain interface and facing the re-face of the PLP
ring. Distinct arrays of residues form the lateral walls
Fig. 2. Stereo ribbon representation of the
AeKAT molecule. The small and large doma-
ins of one subunit are given in green and
blue, respectively. Both N-terminal arms are
grey, and other subunit is red. The cleft hos-
ting the enzyme active site can be seen at
the domain interface where the PLP cofac-
tor is shown as a ball-and-stick. The C-termi-
nus and N-terminus are indicated as C-ter
and N-ter, respectively.
Fig. 3. Diagrams of 2F
o
– F
c
electron density maps for the active
sites of the PMP and PLP forms. The map contoured at 2.0 sigma
is calculated using data between 10.0 and 1.90 A
˚
and 10.0 and
1.55 A
˚
resolution for the PMP and PLP forms of AeKAT, respect-
ively. (A) PLP form; (B) PMP form.
Q. Han et al. Mosquito kynurenineaminotransferase structure
FEBS Journal 272 (2005) 2198–2206 ª 2005 FEBS 2201
of the PLP-binding pocket. In particular, the stretch
43–46, together with the side chains of Arg405,
Asn189 and Asn193 of the large domain, make up one
pocket wall. The opposite wall consists of residues
Tyr111 and Tyr138 of the same monomer and Phe286,
His287 and Tyr73 of the other subunit.
In the PLP form, AeKAT carries the PLP molecule
covalently bound in the active site by Schiff base link-
age to the catalytic Lys255 (Figs 3B and 4A), resulting
in the formation of an internal aldimine bond (C4a ¼
Nz) with an angle of 105.5° at the pyridine ring of
PLP. Several other residues contact the PLP molecule
and participate in its recognition and binding. The
phosphate group of the cofactor is engaged in a num-
ber of interactions with residues making up the strictly
conserved ‘PLP-phosphate binding cup’, featuring
PLP-dependent enzymes [44]. In particular, its OP1
oxygen forms a set of hydrogen bonds with Ser252
(2.97 A
˚
), Ala110 (with its backbone nitrogen atom at
2.78 A
˚
) and the solvent molecule W120 at a distance
of 2.82 A
˚
. PLP OP2 forms a hydrogen bond with
Tyr73 (B) (in other subunit, 2.55 A
˚
) and the solvent
molecule W20 (2.73 A
˚
). Finally, the PLP OP3 atom
interacts with Lys263 (at 2.73 A
˚
from the Nz atom)
and with the backbone nitrogen atom of Tyr111 (at
2.87 A
˚
). The PLP phenolic oxygen is held in place by
interactions with Tyr224 (at 2.56 A
˚
from its OH atom)
and Asn193 (at a distance of 2.62 A
˚
with its OD1
atom). Moreover, the N1 atom of the PLP pyridine
moiety forms hydrogen bonds with the carboxylic oxy-
gen atoms of Asp221 (at 2.66 A
˚
). Several hydrophobic
interactions further stabilize PLP. In particular,
Phe135 and Val223 surround the anthranilic moiety
of the cofactor, on its si- and re-face, respectively
(Fig. 4A). In the PMP form, there is no internal aldi-
mine bond between Lys255 and PMP and the distance
between PMP amine and Lys255 Nz is 3.45 A
˚
(Fig. 3B). The interactions in the active site are similar
to the PLP form (Figs 3A and 4B).
It is interesting to find a thiol–thiolate hydrogen
bond formed between two subunits in AeKAT, which
might be a target for understanding AeKAT regula-
tion. Changes in the sulfur oxidation state of cysteine
residues influence the activity of many proteins [45,46].
Reversible disulfide bond formation and the associated
conformation changes are likely to play an important
role in cellular redox regulation. In particular, disulfide
bond formation between distant cysteines may be an
effective mechanism for the induction of conformatio-
nal changes that lead to switches in protein activity
[47–49]. In human mitochondrial branched chain
aminotransferase, the redox-active dithiol ⁄ disulfide
Cys315-Xaa-Xaa-Cys318 center has been proposed in
the regulation of enzyme activity. Cys315 appears to
be the sensor for redox regulation of the enzyme activ-
ity, whereas Cys318 acts as the ‘resolving cysteine’,
allowing for reversible formation of a disulfide bond
[50,51]. AeKAT has a similar sequence, Cys284-Xaa-
Xaa-Xaa-Cys288 (Fig. 5A), but there is no thiol–thio-
Fig. 5. Putative cysteine regulation site of AeKAT. (A) Partial align-
ment result of human KAT-I and AeKAT. (B) Diagrams of 2F
o
) F
c
electron density maps for Cys284 (A) and Cys284 (B) and surround-
ing residues. The map was contoured at 2.0 sigma.
Fig. 4. Schematic diagram showing active site interactions in AeKAT.
Hydrogen bonds are shown by dotted lines. Phe135 and Val223
sandwich the pyridine ring of PLP. (A) PLP form; (B) PMP form.
Mosquito kynurenineaminotransferase structure Q. Han et al.
2202 FEBS Journal 272 (2005) 2198–2206 ª 2005 FEBS
late hydrogen bond between the two cysteine residues
(9.4 A
˚
) in the AeKAT structure. To understand the
mode of enzyme regulation by cysteine, we comparat-
ively analyzed human KAT-I and AeKAT. First,
AeKAT activity can be stimulated by cysteine [1],
whereas the human homolog is not positively affected
by cysteine [3]. Second, sequence alignment showed
that AeKAT had two additional cysteine residues com-
pared with human KAT-I, as indicated in Fig. 5A
(green background), the equivalent of Cys284 in
AeKAT is Ser276 in the human enzyme. Third, there
are no thiol–thiolate hydrogen bonds or disulfide
bonds in the human KAT-I structure [39], whereas in
the AeKAT structure, Cys284 is located in the large
domain of the homodimer, the structure shows evi-
dence of a thiol–thiolate hydrogen bond between
Cys284 (A) and Cys284 (B) (Fig. 5B), which is close to
the active center. Under oxidizing conditions, these
cysteine residues in AeKAT can reasonably form a
disulfide bond because of the short distance between
the sulfur atoms (3.46 A
˚
in the PLP form and 3.57 A
˚
in the PMP form), requiring a decrease of only
1.5–1.6 A
˚
. Thus, residue Cys284 is most likely the cys-
teine-regulation target of AeKAT. Further mutation
experiments of this residue along with biochemical and
structural analyses are needed to eventually understand
the mechanism of the enzyme modulation by cysteine.
Experimental procedures
Expression and purification of recombinant
AeKAT
AeKAT lacking the N-terminal mitochondrial leader
sequence (amino acids 1–48) was expressed in a baculo-
virus ⁄ insect cell protein expression system, and purified by
DEAE Sepharose, phenyl Sepharose, hydroxyapatite, and
native PAGE separation (PMP form) or gel filtration
(PLP ⁄ native form) [36]. To begin with, we did not pay
attention to the enzyme forms, and obtained only the PMP
form of AeKAT, because the running buffer for native
PAGE has 192 mm glycine. Later, when we used gel filtra-
tion as the last step of purification, we obtained the native
form of AeKAT. The proposed reaction for aminotrans-
ferases is shown in Scheme 1. The purity of the protein was
assessed by SDS ⁄ PAGE analysis. Protein concentration
was determined by a BioRad protein assay kit using bovine
serum albumin as a standard. The purified recombinant
AeKAT was concentrated to 10 mgÆmL
)1
protein in 5 mm
phosphate buffer, pH 7.5 using a Centricon YM-30 concen-
trator (Millipore, Billerica, MA, USA).
AeKAT crystallization
Initial crystallization screening was performed using Hamp-
ton Research Crystal Screens (Hampton Research, Laguna
Niguel, CA, USA) with sitting-drop and hanging-drop
vapor diffusion methods with the volume of reservoir solu-
tion at 500 lL and the drop volume at 5 lL, containing
2.5 lL of protein sample and 2.5 lL of reservoir solution.
AeKAT crystals were obtained in a solution containing
10 mgÆmL
)1
of protein, 30% (w ⁄ v) PEG 1000, and 0.1 m of
Tris ⁄ HCl at pH 8.5. Refinement of preliminary crystalliza-
tion conditions resulted in the growth of quality crystals in
a solution containing 5 mgÆmL
)1
protein, 15% (w ⁄ v)
PEG 1000, and 0.1 m Tris ⁄ HCl at pH 8.5. Single crystals
for suitable X-ray analysis appeared in 4 days and grew to
maximum sizes of 0.5 · 0.3 · 0.2 mm
3
in 3 weeks at 4 °C.
Data collection and processing
To solve the AeKAT structure through multiwavelength
anomalous diffusion (MAD) [52] and single-wavelength
anomalous diffusion (SAD) [53], we first tried l-seleno-
methionine-labeled AeKAT, but failed to obtain quality
diffraction data from its crystals. Subsequently, the
Br-AeKAT derivative was generated by soaking AeKAT
crystals in 1 m NaBr for 5–10 s followed by transferr of
the NaBr-AeKAT crystals to a cryoprotectant solution
containing mother liquid [1 m NaBr, and 25% (v ⁄ v) glyc-
erin in Tris ⁄ HCl pH 8.5] for 30–60 s [54]. For X-ray ana-
lyses, oscillation diffraction images of Br-AeKAT were
obtained using a Bruker General Area Detector Diffrac-
tion System (Madison, WI, USA) equipped with a four-
circle diffractometer and a HiStar multiwire area detector.
Individual AeKAT crystals were frozen using 10%
sucrose plus 30% PEG 400 as a cryoprotectant solution
in order to prevent the appearance of ice diffraction dur-
ing data collection at cryogenic temperatures. Diffraction
data for Br-AeKAT crystals were collected at the Brook-
haven National Synchrotron Light Source X12C beamline
(wavelengths k ¼ 0.97 A
˚
). Single crystals were exposed to
a cold nitrogen stream during data collection using a
MAR research 165 mm CCD detector. The PEG 400,
sucrose, or glycerin provided sufficient cryoprotection dur-
ing data collection. All data were auto-indexed and integ-
rated using hkl software [55], scaling and merging of
diffraction data was performed using scalepack [56]. The
parameters ofcrystal and data collection are listed in
Table 1.
Q. Han et al. Mosquito kynurenineaminotransferase structure
FEBS Journal 272 (2005) 2198–2206 ª 2005 FEBS 2203
Structure determination
Initial phases for the PMP crystal form were obtained from
a SAD experiment. The structure of the PMP form was
determined by the molecular replacement method with the
homology module in insight ii (Accelrys) using the partial
SAD model and two search models, Thermus thermophylus
aspartate aminotransferase (Protein Data Bank code 1gck)
[57] and glutamine aminotransferase (Protein Data Bank
code 1v2d) [38] as template structures. The program amore
[58] was used to calculate both cross-rotation and transla-
tion functions in the 10–3.0 A
˚
resolution range. The initial
model was subjected to iterative cycles of crystallographic
refinement with the programs x-plor [59], shelx-97 [60]
and with graphic sessions for model building using the pro-
gram o [61]. A random sample containing 1000 reflections
was set apart to calculate the free R-factor [62]. Solvent
molecules were manually added at positions with density
> 1.5 sigma in the 2F
o
) F
c
map, considering only peaks
engaged in at least one hydrogen bond with a protein atom
or a solvent atom. The procedure converged to an R-factor
and free R-factor of 0.218 and 0.264, respectively, with
ideal geometry. Residues of the two subunits in AeKAT
are numbered 12 (A) to 429 (A) and 12 (B) to 429 (B),
respectively. The results of refinement are summarized in
Table 1.
Acknowledgements
We thank Dr John M. Sanders, Department of Chem-
istry, University of Illinois at Urbana-Champaign, for
his help in our homology model building using Insight
II. This work was supported by Grant AI 44399 from
the National Institutes of Health and the work was
carried out in part at the National Synchrotron Light
Source, Brookhaven National Laboratory.
References
1 Han Q & Li J (2004) Cysteine and keto acids modu-
late mosquito kynurenineaminotransferase catalyzed
kynurenic acid production. FEBS Lett 577, 381–
385.
2 Cooper AJ (2004) The role of glutamine transaminase
K (GTK) in sulfur and alpha-keto acid metabolism in
the brain, and in the possible bioactivation of neuro-
toxicants. Neurochem Int 44, 557–577.
3 Han Q, Li J & Li J (2004) pH dependence, substrate
specificity and inhibition of human kynurenine amino-
transferase I. Eur J Biochem 271, 4804–4814.
4 Perry SJ, Schofield MA, MacFarlane M, Lock EA,
King LJ, Gibson GG & Goldfarb PS (1993) Isolation
and expression of a cDNA coding for rat kidney cytoso-
lic cysteine conjugate beta-lyase. Mol Pharmacol 43,
660–665.
5 Mosca M, Cozzi L, Breton J, Speciale C, Okuno E,
Schwarcz R & Benatti L (1994) Molecular cloning of
rat kynurenine aminotransferase: identity with gluta-
mine transaminase K. FEBS Lett 353, 21–24.
6 Alberati-Giani D, Malherbe P, Kohler C, Lang G, Kie-
fer V, Lahm HW & Cesura AM (1995) Cloning and
characterization of a soluble kynurenine aminotransfer-
ase from rat brain: identity with kidney cysteine conju-
gate beta-lyase. J Neurochem 64, 1448–1455.
7 Schmidt W, Guidetti P, Okuno E & Schwarcz R (1993)
Characterization of human brain kynurenine amino-
transferases using [
3
H]kynurenine as a substrate. Neuro-
science 55, 177–184.
8 Baran H, Okuno E, Kido R & Schwarcz R (1994)
Purification and characterization ofkynurenine amino-
transferase I from human brain. J Neurochem 62,
730–738.
9 Cooper AJ, Abraham DG, Gelbard AS, Lai JC &
Petito CK (1993) High activities of glutamine transami-
nase K (dichlorovinylcysteine beta-lyase) and omega-
amidase in the choroid plexus of rat brain. J Neurochem
61, 1731–1741.
10 Okuno EF, Ishikawa T, Tsujimoto M, Nakamura M,
Schwarcz R & Kido R (1990) Purification and charac-
terization of kynurenine-pyruvate aminotransferase
from rat kidney and brain. Brain Res 534, 37–44.
11 Cooper AJ, Nieves E, Rosenspire KC, Filc-DeRicco S,
Gelbard AS & Brusilow SW (1988) Short-term meta-
bolic fate of
13
N-labeled glutamate, alanine, and gluta-
mine (amide) in rat liver. J Biol Chem 263, 12268–
12273.
12 Cooper AJ & Meister A (1974) Isolation and properties
of a new glutamine transaminase from rat kidney. J Biol
Chem 249 , 2554–2561.
13 Leeson PD & Iversen LL (1994) The glycine site on the
NMDA receptor: structure–activity relationships and
therapeutic potential. J Med Chem 37, 4053–4067.
14 Perkins MN & Stone TW (1982) An iontophoretic
investigation of the actions of convulsant kynurenines
and their interaction with the endogenous excitant qui-
nolinic acid. Brain Res 247, 184–187.
15 Birch PJ, Grossman CJ & Hayes AG (1988) Kynurenic
acid antagonises responses to NMDA via an action at
the strychnine-insensitive glycine receptor. Eur J Phar-
macol 154 , 85–87.
16 Pereira EF, Hilmas C, Santos MD, Alkondon M, Mae-
licke A & Albuquerque EX (2002) Unconventional
ligands and modulators of nicotinic receptors. J Neuro-
biol 53, 479–500.
17 Hilmas C, Pereira EF, Alkondon M, Rassoulpour A,
Schwarcz R & Albuquerque EX (2001) The brain meta-
bolite kynurenic acid inhibits alpha7 nicotinic receptor
activity and increases non-alpha7 nicotinic receptor
expression: physiopathological implications. J Neurosci
21, 7463–7473.
Mosquito kynurenineaminotransferase structure Q. Han et al.
2204 FEBS Journal 272 (2005) 2198–2206 ª 2005 FEBS
18 Alkondon M, Pereira EF, Yu P, Arruda EZ, Almeida
LE, Guidetti P, Fawcett WP, Sapko MT, Randall WR,
Schwarcz R et al. (2004) Targeted deletion of the kynur-
enine aminotransferase II gene reveals a critical role of
endogenous kynurenic acid in the regulation of synaptic
transmission via alpha7 nicotinic receptors in the hippo-
campus. J Neurosci 24, 4635–4648.
19 Stone TW (2001) Kynurenic acid antagonists and
kynurenine pathway inhibitors. Expert Opin Invest
Drugs 10, 633–645.
20 Stone TW (2001) Kynurenines in the CNS: from endo-
genous obscurity to therapeutic importance. Prog
Neurobiol 64, 185–218.
21 Moroni F (1999) Tryptophan metabolism and brain
function: focus on kynurenine and other indole metabo-
lites. Eur J Pharmacol 375, 87–100.
22 Schwarcz R (1993) Metabolism and function of brain
kynurenines. Biochem Soc Trans 21, 77–82.
23 Foster AC & Fagg GE (1984) Acidic amino acid bind-
ing sites in mammalian neuronal membranes: their char-
acteristics and relationship to synaptic receptors. Brain
Res 319, 103–164.
24 Scharfman HE, Goodman JH & Schwarcz R (2000)
Electrophysiological effects of exogenous and endogen-
ous kynurenic acid in the rat brain: studies in vivo and
in vitro. Amino Acids 19, 283–297.
25 Carpenedo R, Chiarugi A, Russi P, Lombardi G, Carla
V, Pellicciari R, Mattoli L & Moroni F (1994) Inhibi-
tors ofkynurenine hydroxylase and kynureninase
increase cerebral formation of kynurenate and have
sedative and anticonvulsant activities. Neuroscience 61,
237–243.
26 Cozzi A, Carpenedo R & Moroni F (1999) Kynurenine
hydroxylase inhibitors reduce ischemic brain damage:
studies with (m-nitrobenzoyl)-alanine (mNBA) and 3,4-
dimethoxy-[-N-4-(nitrophenyl) thiazol-2yl]-benzenesul-
fonamide (Ro 61–8048) in models of focal or global
brain ischemia. J Cereb Blood Flow Metab 19, 771–
777.
27 Erhardt S, Blennow K, Nordin C, Skogh E, Lindstrom
LH & Engberg G (2001) Kynurenic acid levels are ele-
vated in the cerebrospinal fluid of patients with schizo-
phrenia. Neurosci Lett 313, 96–98.
28 Battaglia G, Rassoulpour A, Wu HQ, Hodgkins PS,
Kiss C, Nicoletti F & Schwarcz R (2000) Some metabo-
tropic glutamate receptor ligands reduce kynurenate
synthesis in rats by intracellular inhibition of kynure-
nine aminotransferase II. J Neurochem 75, 2051–2060.
29 Kiely JM & Gordon FJ (1994) Role of rostral ventrolat-
eral medulla in centrally mediated pressor responses.
Am J Physiol 267, H1549–H1556.
30 Ito S, Komatsu K, Tsukamoto K & Sved AF (2000)
Excitatory amino acids in the rostral ventrolateral
medulla support blood pressure in spontaneously hyper-
tensive rats. Hypertension 35, 413–417.
31 Kapoor V, Kapoor R & Chalmers J (1994) Kynurenic
acid, an endogenous glutamate antagonist, in SHR and
WKY rats: possible role in central blood pressure regu-
lation. Clin Exp Pharmacol Physiol 21, 891–896.
32 Kapoor V, Thuruthyil SJ & Human B (1998) Reduced
kynurenine aminotransferase-I activity in SHR rats may
be due to lack of KAT-Ib activity. Neuroreport 9, 1431–
1434.
33 Kwok JB, Kapoor R, Gotoda T, Iwamoto Y, Iizuka Y,
Yamada N, Isaacs KE, Kushwaha VV, Church WB,
Schofield PR et al. (2002) A missense mutation in
kynurenine aminotransferase-1 in spontaneously hyper-
tensive rats. J Biol Chem 277, 35779–35782.
34 Martin SJ, Grimwood PD & Morris RG (2000) Synap-
tic plasticity and memory: an evaluation of the hypo-
thesis. Annu Rev Neurosci 23, 649–711.
35 Savvateeva E, Popov A, Kamyshev N, Bragina J, Hei-
senberg M, Senitz D, Kornhuber J & Riederer P (2000)
Age-dependent memory loss, synaptic pathology and
altered brain plasticity in the Drosophila mutant cardi-
nal accumulating 3-hydroxykynurenine. J Neural
Transm 107, 581–601.
36 Fang J, Han Q & Li J (2002) Isolation, characterization,
and functional expression ofkynurenine aminotransfer-
ase cDNA from the yellow fever mosquito, Aedes
aegypti (1). Insect Biochem Mol Biol 32, 943–950.
37 Hosono A, Mizuguchi H, Hayashi H, Goto M, Miyaha-
ra I, Hirotsu K & Kagamiyama H (2003) Glutamine:
phenylpyruvate aminotransferase from an extremely
thermophilic bacterium, Thermus thermophilus HB8.
J Biochem (Tokyo) 134, 843–851.
38 Goto M, Omi R, Miyahara I, Hosono A, Mizuguchi H,
Hayashi H, Kagamiyama H & Hirotsu K (2004) Crystal
structures of glutamine: phenylpyruvate aminotransfer-
ase from Thermus thermophilus HB8: induced fit and
substrate recognition. J Biol Chem 279, 16518–16525.
39 Rossi F, Han Q, Li J, Li J & Rizzi M (2004) Crystal
structure of human kynurenineaminotransferase I.
J Biol Chem 279 , 50214–50220.
40 Matthews BW (1968) Solvent content of protein crys-
tals. J Mol Biol 33, 491–497.
41 Laskowski RA, Macarthur MW, Moss DS & Thornton
JM (1993) procheck – a program to check the stereo-
chemical quality of protein structures. J Appl Crystal-
logr 26, 283–291.
42 Jansonius JN (1998) Structure, evolution and action of
vitamin B-6-dependent enzymes. Curr Opin Struct Biol
8, 759–769.
43 Nakai T, Okada K, Akutsu S, Miyahara I, Kawaguchi
S, Kato R, Kuramitsu S & Hirotsu K (1999) Structure
of Thermus thermophilus HB8 aspartate aminotransfer-
ase and its complex with maleate. Biochemistry 38,
2413–2424.
44 Denesyuk AI, Denessiouk KA, Korpela T & Johnson
MS (2002) Functional attributes of the phosphate group
Q. Han et al. Mosquito kynurenineaminotransferase structure
FEBS Journal 272 (2005) 2198–2206 ª 2005 FEBS 2205
binding cup of pyridoxal phosphate-dependent enzymes.
J Mol Biol 316, 155–172.
45 Giles NM, Watts AB, Giles GI, Fry FH, Littlechild JA
& Jacob C (2003) Metal and redox modulation of
cysteine protein function. Chem Biol 10, 677–693.
46 Xu D, Rovira II & Finkel T (2002) Oxidants painting
the cysteine chapel: redox regulation of PTPs. Dev Cell
2, 251–252.
47 Lee C, Lee SM, Mukhopadhyay P, Kim SJ, Lee SC &
Ahn WS., YuMH, Storz G & Ryu SE (2004) Redox
regulation of OxyR requires specific disulfide bond for-
mation involving a rapid kinetic reaction path. Nat
Struct Mol Biol 11, 1179–1185.
48 Georgiou G (2002) How to flip the (redox) switch. Cell
111, 607–610.
49 Choi H, Kim S, Mukhopadhyay P, Cho S, Woo J, Storz
G & Ryu S (2001) Structural basis of the redox switch in
the OxyR transcription factor. Cell 105, 103–113.
50 Conway ME, Poole LB & Hutson SM (2004) Roles for
cysteine residues in the regulatory CXXC motif of
human mitochondrial branched chain aminotransferase
enzyme. Biochemistry 43, 7356–7364.
51 Conway ME, Yennawar N, Wallin R, Poole LB & Hut-
son SM (2002) Identification of a peroxide-sensitive
redox switch at the CXXC motif in the human mito-
chondrial branched chain aminotransferase. Biochemis-
try 41, 9070–9078.
52 Hendrickson WA, Horton JR & LeMaster DM (1990)
Selenomethionyl proteins produced for analysis by mul-
tiwavelength anomalous diffraction (MAD): a vehicle
for direct determination of three-dimensional structure.
EMBO J 9, 1665–1672.
53 Dauter Z, Dauter M & Dodson E (2002) Jolly SAD.
Acta Crystallogr D Biol Crystallogr 58, 494–506.
54 Dauter Z, Dauter M & Rajashankar KR (2000) Novel
approach to phasing proteins: derivatization by short
cryo-soaking with halides. Acta Crystallogr D Biol Crys-
tallogr 56, 232–237.
55 Otwinowski Z & Minor W (1997) Processing of X-ray
diffraction data collected in oscillation mode. In Meth-
ods in Enzymology (Carter CW & Sweet RM, eds), pp.
307–326, Academic Press, London.
56 Minor W (1993)
XDISPLAYF Program. Purdue University,
West Lafayette, IN.
57 Ura H, Nakai T, Kawaguchi SI, Miyahara I, Hirotsu K
& Kuramitsu S (2001) Substrate recognition mechanism
of thermophilic dual-substrate enzyme. J Biochem
(Tokyo) 130, 89–98.
58 Navaza J (2001) Implementation of molecular replace-
ment in amore. Acta Crystallogr D Biol Crystallogr 57,
1367–1372.
59 Brunger AT (1992)
X-PLOR. Yale University Press, New
Haven, CT.
60 Sheldrick GM & Schneider TR (1997) High resolution
refinement. Methods Enzymol 277, 319–343.
61 Jones TA, Zou JY, Cowan SW & Kjeldgaard M (1991)
Improved methods for building protein models in elec-
tron density maps and the location of errors in these
models. Acta Crystallogr A 47, 110–119.
62 Brunger AT (1992) Free R-value – a novel statistical
quantity for assessing the accuracy of crystal-structures.
Nature 355, 472–475.
Mosquito kynurenineaminotransferase structure Q. Han et al.
2206 FEBS Journal 272 (2005) 2198–2206 ª 2005 FEBS
. Crystal structures of Aedes aegypti kynurenine aminotransferase Qian Han 1, *, Yi Gui Gao 1,2, *, Howard Robinson 3 , Haizhen Ding 1 , Scott Wilson 2 and Jianyong Li 1 1 Department of Pathobiology,. 2005) doi:10.1111/j.1742-4658.2005.04643.x Aedes aegypti kynurenine aminotransferase (AeKAT) catalyzes the irrevers- ible transamination of kynurenine to kynurenic acid, the natural antagonist of NMDA and 7-nicotinic. University of Illinois, Urbana, IL, USA 2 School of Chemical Sciences, University of Illinois, Urbana, IL, USA 3 Biology Department, Brookhaven National Laboratory, Upton, NY, USA Aedes aegypti kynurenine