Tài liệu Báo cáo khoa học: Crystal structures of the regulatory subunit of Thr-sensitive aspartate kinase fromThermus thermophilus pdf

Crystal structures of the regulatory subunit of
Thr-sensitive aspartate kinase from Thermus thermophilus
Ayako Yoshida
1
, Takeo Tomita
1
, Hidetoshi Kono
2
, Shinya Fushinobu
3
, Tomohisa Kuzuyama
1
and
Makoto Nishiyama
1,4
1 Biotechnology Research Center, The University of Tokyo, Japan
2 Computational Biology Group, Quantum Beam Science Directorate, Japan Atomic Energy Agency, Kyoto, Japan
3 Department of Biotechnology, The University of Tokyo, Japan
4 RIKEN SPring-8 Center, Hyogo, Japan
Aspartate kinase (AK; EC 2.7.2.4) is an enzyme that
catalyzes the first committed step, the phosphorylation
of the c-carboxyl group of aspartate, of the biosynthetic
pathway of the aspartic acid group amino acids Lys,
Thr, Ile, and Met, in microorganisms and plants. AK
is classified into two groups according to subunit orga-
nization: homo-oligomer or heterotetramer. AK from
Thermus thermophilus (TtAK), AK from C. glutami-
cum (CgAK) and AKII from Bacillus subtilis (BsAKII)
are heterotetramers containing equimolar amounts of
a-subunits and b-subunits [1–3], whereas AKIII from
Escherichia coli (EcAKIII), AKI from Arabidopsis

thaliana and AK from Methanococcus jannaschii
(MjAK) are homo-oligomers of identical subunits
[4–6]. AK of the a
2
b
2
type is encoded by in-frame
overlapping genes, so that the amino acid sequence of
the b-subunit is identical to about 160 amino acids of
the C-terminus of the a-subunit. As seen in other
enzymes involved in the first step in amino acid bio-
synthesis, AK is regulated through feedback inhibition
Keywords
ACT domain; allosteric regulation; crystal
structure; thermostability; threonine
biosynthesis
Correspondence
M. Nishiyama, Biotechnology Research
Center, The University of Tokyo, 1-1-1
Yayoi, Bunkyo-ku, Tokyo 113-8657, Japan
Fax: +81 3 5841 8030
Tel: +81 3 5841 3074
E-mail:
(Received 3 November 2008, revised 10
March 2009, Accepted 31 March 2009)
doi:10.1111/j.1742-4658.2009.07030.x
Crystal structures of the regulatory subunit of Thr-sensitive aspartate
kinase (AK; EC 2.7.2.4) from Thermus thermophilus (TtAKb) were deter-
mined at 2.15 A
˚

in the Thr-bound form (TtAKb-Thr) and at 2.98 A
˚
in the
Thr-free form (TtAKb-free). Although both forms are crystallized as
dimers, the contact surface area of the dimer interface in TtAKb-free
(3200 A
˚
2
) is smaller than that of TtAKb-Thr (3890 A
˚
2
). Sedimentation
equilibrium analyzed by ultracentrifugation revealed that TtAKb is present
in equilibrium between a monomer and dimer, and that Thr binding shifts
the equilibrium to dimer formation. In the absence of Thr, an outward
shift of b-strands near the Thr-binding site (site 1) and a concomitant loss
of the electron density of the loop region between b3 and b4 near the Thr-
binding site are observed. The mechanism of regulation by Thr is discussed
on the basis of the crystal structures. TtAKb has higher thermostability
than the regulatory subunit of Corynebacterium glutamicum AK, with a dif-
ference in denaturation temperature (T
m
)of40°C. Comparison of the
crystal structures of TtAKb and the regulatory subunit of C. glutamicum
AK showed that the well-packed hydrophobic core and high Pro content
in loops contribute to the high thermostability of TtAKb.
Abbreviations
AK, aspartate kinase; BsAKII, aspartate kinase II from Bacillus subtilis; CgAK, aspartate kinase from Corynebacterium glutamicum; CgAKb,
regulatory subunit of aspartate kinase from Corynebacterium glutamicum; DSC, differential scanning calorimetry; EcAKIII, aspartate kinase III
from Escherichia coli; MAD, multiwavelength anomalous diffraction; MjAK, aspartate kinase from Methanococcus jannaschii; TtAK, aspartate

kinase from Thermus thermophilus; TtAKb, regulatory subunit of aspartate kinase from Thermus thermophilus; TtAKb-free, Thr-free
regulatory subunit of aspartate kinase from Thermus thermophilus; TtAKb-Thr, Thr-bound regulatory subunit of aspartate kinase from
Thermus thermophilus.
3124 FEBS Journal 276 (2009) 3124–3136 ª 2009 The Authors Journal compilation ª 2009 FEBS
by end-products. CgAK is regulated through concerted
inhibition by Lys and Thr [7], whereas TtAK, which is
involved in the biosynthesis of Thr and Met but
not of Lys, because T. thermophilus synthesizes Lys
through a-aminoadipate as an intermediate [8,9], is
Thr-sensitive [1]. In a
2
b
2
-type AK, the N-terminal
regions of the a-subunits serves as catalytic domains,
and the C-terminal regions of the a-subunits and the
b-subunits act as regulatory domains [10,11].
The regulatory domains of AK contain conserved
motifs named ACT domains that are found among
many allosteric enzymes involved in amino acid and
purine biosynthesis [12,13]. The motif has a babbab
fold, and serves as a small molecule-binding domain
for allosteric regulation. Several crystal structures
have been determined for enzymes containing ACT
domains; however, the mode of association between
ACT domains is quite different among the enzymes,
as summarized by Grant [14]. For example, the
archetypical ACT domain association with two side-
by-side domains, each from a different chain, is
found in 3-phosphoglycerate dehydrogenase [15]. In

threonine deaminase, two ACT domains in a single
peptide are arranged side-by-side to form an Ile ⁄ Val-
binding unit [16,17]. For the ACT domain in AK,
two types of association modes are seen, as reviewed
by Curien et al. [18]: one is found in homo-oligo-
meric AKs [4,6,19] and the other in a
2
b
2
-type CgAK
[20]. In these ACT domains of AK proteins, there
are common structural features: (a) two ACT
domains are arranged in the C-terminal portion of a
single polypeptide; (b) the ACT1 domain is inserted
into the ACT2 domain; and (c) two ACT domains,
each from a different chain, interact to form an effec-
tor-binding unit. The effector-binding unit of the
ACT domain in the b-subunit of CgAK (CgAKb)is
organized differently from those of homo-oligomeric
AKs. In CgAKb, ACT1 and ACT2 from different
chains associate side-by-side to form an eight-
stranded b-sheet, and two eight-stranded b-sheets face
each other perpendicularly. In CgAKb, both eight-
stranded b-sheets are involved in effector binding. On
the other hand, in homo-oligomeric AKs, two ACT1
domains from different chains associate with each
other to form an eight-stranded b-sheet, and two
ACT2 domains from different chains form an addi-
tional eight-stranded b-sheet, although these two
eight-stranded b-sheets are also arranged perpendicu-

larly and face-to-face, as in CgAKb. In homo-
oligomeric AKs, only one of the eight-stranded
b-sheets is involved in effector binding. Determination
of the crystal structure of the regulatory subunit of
TtAK (TtAKb) would provide information not only
on the catalytic mechanism but also on the structural
features common to a
2
b
2
-type AKs.
As T. thermophilus is an extremely thermophilic bac-
terium, proteins produced by T. thermophilus have
high thermostability. Previously, we found that chime-
ric AK, named BTT, which is composed of a catalytic
domain from BsAKII and regulatory domains (a
regulatory domain in the a-subunit, and a b-subunit
with the same sequence as the regulatory domain)
from TtAK, improved thermostability as much as
wild-type TtAK [11]. This result indicated that the
regulatory domain of TtAK contributes not only to
catalytic regulation, but also the thermostability of
TtAK. Comparison of the crystal structures of TtAKb
and CgAKb was expected to elucidate the mechanism
of the elevated thermostability of TtAKb.
In this article, we describe the crystal structures of
TtAKb in two forms, Thr-bound and Thr-free, and
discuss the regulatory mechanism of Thr and the struc-
tural features responsible for the high thermostability
of TtAKb.

Results and Discussion
Model quality
The crystal structure of the Thr-bound form of TtAKb
(TtAKb-Thr) was determined at 2.15 A
˚
resolution,
using multiwavelength anomalous diffraction (MAD)
phases derived from selenomethionine (SeMet)-substi-
tuted TtAKb. TtAKb-Thr is a dimer containing two
Thr molecules (Fig. 1A), acetate molecules, which are
derived from the crystallization buffer, and 153 water
molecules in an asymmetric unit. The electron densities
of the N-terminal (residues 1–4 in chains A and B)
and C-terminal (residues 158–161 in chains A and B)
sections of the structure are not seen on the map,
probably owing to disorder of these regions. The over-
all geometry of the model according to the procheck
program [21] is of good quality, with 95.4% of the res-
idues in the most favored regions and 4.6% in allowed
regions of the Ramachandran plot.
The crystal structure of the Thr-free form of TtAKb
(TtAKb-free) was determined at 2.98 A
˚
resolution by
molecular replacement using the structure of TtAKb-
Thr as a search model. The TtAKb-free crystal
contains three dimers (Fig. 2A), each composed of AB,
CD and EF chains, and 79 water molecules in an
asymmetric unit. The electron densities of the N-termi-
nal (residues 1–3 in chain A, residues 1–4 in chains B

and C, and residues 1–5 in chains E–G) and C-termi-
nal (residues 158–161 in chains A–F, and residues
159–161 in chain B) portions of the structure are not
A. Yoshida et al. Crystal structures of regulatory subunit of Thr-sensitive aspartate kinase
FEBS Journal 276 (2009) 3124–3136 ª 2009 The Authors Journal compilation ª 2009 FEBS 3125
seen on the map. The electron densities of the sections
(residues 54–56 in chains A and D, residues 53–59 in
chain B, residues 55–56 in chain C, residues 56–57 in
chain E, and residues 53–56 in chain F) of the loop
between b3 and b4 are not observed in every chain.
The overall geometry of the model is good, with
89.0% of the residues in the most favored regions,
10.0% in allowed regions, 0.7% in generously allowed
regions and 0.3% in disallowed regions from pro-
check. Table 1 summarizes the refinement statistics.
Overall structure
The crystal structure of TtAKb-Thr was determined as
a homodimer (Fig. 1A). As TtAK is a heterotetramer
with an a
2
b
2
configuration, where the b-subunit is
identical to the C-terminal portion of the a-subunit, as
in CgAK, the dimeric structure revealed in this study
represents the structure of the regulatory region of an
ab-heterodimer. The rmsd is 0.50 A
˚
between two
monomers in the asymmetric unit. Structural differ-

ences between monomers are found in regions 84–88,
94–95, and 102–104 (Fig. 1B). A single chain of
TtAKb contains two ACT domains, ACT1 (N-termi-
nal domain) and ACT2 (C-terminal domain) domains.
The ACT domain organization of TtAKb-Thr is
similar to that of CgAKb [20] but not to those of
homo-oligomeric AKs. ACT1 and an ACT2, each
from different chains, are arranged side-by-side to
form an effector (Thr)-binding unit, an eight-stranded
antiparallel b-sheet with four a-helices on one side. We
assume that this characteristic dimer organization of
the regulatory domain of AK is a feature limited to
a
2
b
2
-type AKs, because, in homo-oligomeric AKs, two
equivalent ACT domains from different chains are jux-
taposed to form a structural unit, and two structural
units, each composed of two equivalent ACT
domains, are not equivalent to each other. Owing to
the difference in the ACT domain arrangement,
TtAKb binds two Thr molecules per dimer at two sites
(site 1), each in an equivalent effector-binding unit
(Fig. 1A), whereas in homo-oligomeric AKs, a single
Fig. 1. Overall structure of TtAKb-Thr. (A)
Overall structure of TtAKb-Thr. The A chain
and the B chain are shown in purple and
green, respectively. Thr molecules are
shown as an orange stick model. Both ACT

domains forming an effector-binding unit of
the front of the dimer are indicated. Site 1
and site 2 of the effector-binding unit on the
front are indicated by solid and dotted cir-
cles. (B) Superposition of two monomers in
a dimer of TtAKb-Thr. ACT domains are
shown by dotted circles. Regions showing
structural differences between monomers
are indicated by solid circles.
Fig. 2. Overall structure of TtAKb-free. (A)
Three TtAKb-free dimers in the asymmetric
unit. A chain, magenta; B chain, yellow; C
chain, cyan; D chain, green; E chain, brown;
F chain, blue. (B) EF chain dimer.
Crystal structures of regulatory subunit of Thr-sensitive aspartate kinase A. Yoshida et al.
3126 FEBS Journal 276 (2009) 3124–3136 ª 2009 The Authors Journal compilation ª 2009 FEBS
effector-binding unit binds two effectors. The other
possible effector-binding sites (site 2) in the structural
unit are vacant in the TtAKb-Thr structure.
The structure of TtAKb-free is also determined to
be a homodimer (Fig. 2), although TtAKb was eluted
in volumes close to that of a molecular mass of mono-
mer in the absence of Thr in gel filtration chromatog-
raphy, as described below. The TtAKb-free crystal
contains three dimers in the asymmetric unit, and rmsd
values for Ca among the three dimers are 0.58 A
˚
between AB and CD dimers, 0.63 A
˚
between CD and

EF dimers, and 0.58 A
˚
between AB and EF dimers.
Moreover, the rmsd values of Ca between the mono-
mers in the dimers are 0.55 A
˚
between the A and B
chains, 0.96 A
˚
between the C and D chains, and
0.51 A
˚
between the E and F chains. The main differ-
ence between TtAKb-free and TtAKb-Thr is that the
residues in the loop region between b3 and b4 near the
Thr-binding site are disordered in TtAKb-free, as
described later.
Thr-binding site
In TtAK b-Thr, the electron density of one Thr mole-
cule is observed at site 1 between ACT1 and ACT2,
each from different chains (Fig. 3A,B). The structure
of TtAKb-Thr is quite similar to that of Thr-bound
CgAKb (rmsd value of Ca is 1.73 A
˚
). Bound Thr mol-
ecules are stabilized by ionic bonds (Asp26-Od2 for the
amino group), hydrogen bonds (Gln50-Oe1 and
Ile126*-O for the side chain hydroxyl group; Asn125*-
Od1 and Ile126*-O for the amino group; Ile30-N,
Ile126-N and Asn125*-Od1 for the carboxyl group;

asterisks denote residues from another chain), and
Table 1. Data collection and refinement statistics.
SeMet-TtAKb-Thr Native
Peak Edge Remote TtAKb-Thr TtAKb-free
Data collection
X-ray source PF-NW12 PF-NW12 PF-NW12 PF-NW12 PF-NW12
Wavelength (A
˚
) 0.9792 0.9794 0.9630 1.000 1.000
Space group P4
3
32 P4
3
32 P4
3
32 P4
3
32 P3
1
Resolution (A
˚
)
a
2.40 (2.49–2.40) 2.40 (2.49–2.40) 2.40 (2.49–2.40) 2.15 (2.19–2.15) 2.98 (3.09–2.98)
Reflections (total ⁄ unique) 364 163 ⁄ 19 507 365 251 ⁄ 19 539 364 474 ⁄ 19 539 578 149 ⁄ 27 145 133 032 ⁄ 22 878
R
sym
b
(%) 9.5 (46.5) 9.5 (46.3) 9.7 (51.5) 6.7 (33.2) 9.2 (37.1)
I ⁄ r(I) 30.8 (6.0) 30.2 (5.8) 29.5 (5.5) 59.9 (11.1) 21.0 (3.2)

Completeness (%) 100.0 (100.0) 100.0 (100.0) 100.0 (100.0) 100.0 (100.0) 99.8 (100.0)
Phasing
Number of Se sites 8
FOM
c
0.40
Refinement
Resolution (A
˚
) 47.2–2.15 46.5–2.98
R-factor
d
(work ⁄ test) (%) 18.8 ⁄ 22.6 24.4 ⁄ 26.6
Number of atoms 2405 6545
Protein atoms 2228 6466
Thr molecules 2
Acetate molecules 2
Water molecules 153 79
Average B-factor
Protein atoms 32.9 56.3
Thr 23.3
Water 34.8 44.4
rmsd values
Bond length (A
˚
) 0.009 0.010
Bond angle (°) 1.50 1.40
Ramachandran plot
e
Most favored (%) 95.4 89.0

Additionally favored (%) 4.6 10.0
Generously allowed (%) 0 0.7
Disallowed (%) 0 0.3
a
Values in parentheses are data of the highest-resolution shell.
b
R
sym
¼ RjI
i
À <I>j=R<I>.
c
Figure of merit (FOM) was calculated with the
SOLVE program.
d
R - factor ¼ R
hkl
F
o
jjÀF
c
jjjj=R
hkl
F
o
jj.
e
Calculated using PROCHECK.
A. Yoshida et al. Crystal structures of regulatory subunit of Thr-sensitive aspartate kinase
FEBS Journal 276 (2009) 3124–3136 ª 2009 The Authors Journal compilation ª 2009 FEBS 3127

hydrophobic interactions (Ile24, Ile30 and Met62 for
the side chain methyl group). Two water molecules
present near the two oxygen atoms contribute to a
hydrogen bond network between the carboxyl group
of Thr, Gly29-N, Ala31-N, Ala32-N, and Phe116-O
from another chain. Most of the residues and water
molecules recognizing the bound Thr in TtAKb are
conserved in CgAKb. As seen in CgAKb, the carboxyl
group of Thr is located near the N-terminal section of
helix a1, suggesting that the positive charge of the
N-terminal helix dipole facilitates recognition of the
carboxyl group. Importantly, Thr is bound between
two chains and is not exposed to the solvent, suggest-
ing that bound Thr plays an important role in stabiliz-
ing the dimeric structure, as in CgAKb.
Monomer–dimer equilibrium
In CgAKb, Thr binding induces the dimerization of
CgAKb [20]. Thr is bound at an effector-binding unit
formed between two chains in TtAKb in a manner
almost identical to that in CgAKb, suggesting that Thr
binding plays a role in stabilizing the dimeric form of
TtAKb. To examine the effect of Thr on dimerization,
we analyzed the oligomeric state of TtAKb in the pres-
ence or absence of Thr, using two different methods:
Fig. 3. Thr-binding site. (A) 2F
o
)F
c
map of bound Thr molecule and two water molecules. The contour level of the map is 1.0r. (B) Thr-bind-
ing site in TtAKb-Thr. Residues in purple are in the A chain, and residues in green are in the B chain. (C) Vacant Thr-binding site in TtAKb-

free. Residues in blue are in the E chain, and residues in orange are in the F chain. (D) Structure-based sequence alignment of TtAKb and
CgAKb. Alignment was performed with
CLUSTALW [42], and alignment with secondary structures of TtAKb and CgAKb was performed with
ESPRIPT [43]. Regions for ACT1 and ACT2 are shown by solid and broken divergent arrows, respectively.
Crystal structures of regulatory subunit of Thr-sensitive aspartate kinase A. Yoshida et al.
3128 FEBS Journal 276 (2009) 3124–3136 ª 2009 The Authors Journal compilation ª 2009 FEBS
sedimentation equilibrium by analytical ultracentri-
fugation, and gel filtration chromatography.
In gel filtration, TtAKb was eluted in a volume
corresponding to a molecular mass of 21.7 kDa in the
absence of Thr (Fig. 4A). The molecular mass, esti-
mated by gel filtration, is a little larger than the calcu-
lated mass of a monomer (17.7 kDa). When gel
filtration was performed in the presence of 5 mm Thr,
elution profiles gave an estimated molecular mass of
30.9 kDa.
To further confirm that Thr affects monomer–dimer
equilibrium even in TtAKb, we also analyzed the sub-
unit arrangement by sedimentation equilibrium. The
data fitted well with the monomer–dimer equilibrium,
with equilibrium constants of 5.6 · 10
)3
m
)1
(goodness
of fit = 4.0 · 10
)4
) and 9.1 · 10
)4
m

)1
(goodness of
fit = 2.3 · 10
)4
) in the presence and absence of 5 mm
Thr, respectively. According to the constants, TtAKb
at 1 mgÆmL
)1
is mostly (91%) present as a monomer
in the absence of Thr, whereas at the same protein
concentration, 31% of TtAKb is present in a dimeric
form in solution containing Thr. At 5 mgÆmL
)1
which
is the protein concentration used for crystallization,
58% and 27% are present as dimers in the presence
and absence of Thr, respectively. Thus, TtAKb is in
monomer–dimer equilibrium, which is displaced by
Thr and ⁄ or protein concentrations; therefore, dimer
formation in the crystal structure in the absence of
Thr can be explained by the high concentration of
TtAKb-free under crystallization conditions.
CgAK is easily dissociated into a-subunits and b-sub-
units during purification without Thr. On the other
hand, TtAK is purified in the a
2
b
2
form even without
Thr. This observation suggests that binding affinity

between a-subunits and b-subunits is stronger in TtAK
than in CgAK; however, even TtAK showed sharp
and broad elution profiles in gel filtration in the pres-
ence and absence of Thr, respectively. SDS⁄ PAGE of
the fractions in gel filtration showed that a-subunits
and b-subunits are eluted in the same volumes from
the column in the presence of Thr, whereas b-subunits
are eluted from the column later than a-subunits in the
absence of Thr (Fig. 4B–E). From these results, we
conclude that b-subunits can interact with the regula-
tory domains of a-subunits even without Thr, but the
interaction is tighter in the presence of Thr.
Fig. 4. Stabilization of oligomer formation of
TtAK and TtAKb by Thr. (A) Elution profiles
of TtAKb in the presence and absence of
5m
M Thr. The solid line with circles and the
dotted line with squares indicate profiles in
the presence and absence of 5 m
M Thr,
respectively. Elution volumes for BSA
(67 kDa), chymotrypsinogen A (43 kDa),
ovalbumin (25 kDa) and ribonuclease A
(13 kDa) are indicated by a, b, c, and d,
respectively. (B, C) SDS ⁄ PAGE of each frac-
tion by gel filtration for TtAK in the presence
(B) and absence (C) of 5 m
M Thr. Elution
profiles of TtAK were quantitated by
IMAGEJ

[44]. (D) Densitometric calibration of TtAK
subunits of SDS ⁄ PAGE in (B). The a-sub-
units and b-subunits are indicated by a solid
line with circles and a dotted line with
squares, respectively. (E) Densitometric cali-
bration of TtAK subunits of SDS ⁄ PAGE in
(C). The a-subunits and b-subunits are indi-
cated by a solid line with circles and a
dotted line with squares, respectively.
A. Yoshida et al. Crystal structures of regulatory subunit of Thr-sensitive aspartate kinase
FEBS Journal 276 (2009) 3124–3136 ª 2009 The Authors Journal compilation ª 2009 FEBS 3129
Conformational change of TtAKb upon Thr
binding and its implications
Unexpectedly, the structural difference between
TtAKb-Thr and TtAKb-free is not so large: the rmsd
for Ca between two structures is about 1.5 A
˚
. This
contrasts with Lys-sensitive EcAKIII, which shows a
larger conformational change of the regulatory domain
dimer upon Lys binding, resulting in the displacement
of several residues responsible for catalytic function in
the catalytic domain. The most distinct difference
between the structures is that the electron density of
the b3–b4 loop around the Thr-binding site is missing
in TtAKb-free, and that b-strands surrounding the
Thr-binding site show outward shifts in the absence of
Thr, with 12° rotation of the ACT2 domain from the
fixed ACT1 domain (Figs 3B,C and 5D). The regula-
tory domain dimer of CgAK inhibited in a concerted

manner by both Lys and Thr binds two Thr molecules
at site 1, like TtAKb-Thr, and easily dissociates into
monomers in the absence of Thr [20]. Therefore, a sim-
ilar conformational change is expected for these two
enzymes, depending on the presence or absence of Thr.
In CgAK, mutations of the residues in the b3–b4 loop
close to site 1 induced resistance to Lys or a Lys ana-
log, S-2-aminoethyl-l-cysteine [20]. As Lys is bound to
a vacant effector-binding site (site 2) in the effector-
binding unit composed of the ACT1 and ACT2
Fig. 5. Comparison of a single effector-binding unit between TtAKb-Thr and TtAKb-free. (A) Superposition of the effector-binding units of
TtAKb-Thr and TtAKb-free. ACT1 displays the Ca models of the B chain (residues 15–93) from TtAKb-Thr and the F chain (residues 15–93)
from TtAKb-free, and ACT2 shows Ca models of the A chain (residues 5–14 and 94–157) from TtAKb-Thr and the E chain (residues 6–14
and 94–157) from TtAKb-free. TtAKb-Thr and TtAKb-free are in blue and red, respectively. The Thr molecule is shown as a stick model. The
loop between b4 and a2 corresponding to the latch loop in EcAKIII is shown as a dotted oval. (B) Movement of Ca atoms caused by Thr
binding mapped on the effector-binding unit of TtAKb-Thr. Cyan, < 1 A
˚
; green, < 2 A
˚
; yellow, < 3 A
˚
; orange, < 4 A
˚
; red, > 4 A
˚
. Regions
showing larger movement are marked as A–E. The loop between b4 and a2 corresponding to the latch loop in EcAKIII is shown as a dotted
oval. (C) Ca distance between TtAKb-Thr and TtAKb-free. Blue indicates the distance between the A chain from TtAKb-Thr and the E chain
from TtAKb-free, and red indicates the distance between the B chain from TtAKb-Thr and the F chain from TtAKb-free. The regions shown
in A–E are: A, 42–45; B, 46–50; C, 51–58; D, 102–110; E, 131–134. (D) Domain motion in TtAKb caused by Thr binding. The structures of

TtAKb-Thr and TtAKb-free are shown in blue and pink. Domain motion was analyzed by
DYNDOM [45]. A broken line indicates hinge axis for
movement.
Crystal structures of regulatory subunit of Thr-sensitive aspartate kinase A. Yoshida et al.
3130 FEBS Journal 276 (2009) 3124–3136 ª 2009 The Authors Journal compilation ª 2009 FEBS
domains (Yoshida et al., unpublished result), the direct
function of the loop in catalytic control is unexpected
in CgAK. In TtAKb, accompanied by an outward
shift of b-strands, especially b2–b4 from ACT1 near
site 1, a significant shift is also found around site 2
(Fig. 5A–C). This result may suggest that two Thr
molecules bound at site 1 induce a conformational
change in site 2, thereby facilitating the binding of
additional Thr molecules at site 2. In addition, TtAKb
has a Pro-Gly sequence in the N-terminal section
(positions 28 and 29) of helix a1, which contribute to
the recognition of the carboxyl group of bound Thr by
a putative helix dipole (Fig. 3B). Interestingly, the
dihedral angle of Gly29 changes upon Thr binding (for
example, / = 89.85°, w = )13.64° in the A chain of
TtAKb-Thr and / = 103.16°, w = )11.53° in the E
chain of TtAKb-free). In CgAKb, which can bind the
Thr molecule at site 1, the Pro-Gly sequence is con-
served at the same position (positions 27 and 28)
(Fig. 3D). In addition to Pro27-Gly28, CgAKb has a
similar Pro-Gly sequence at positions 109–110 in the
N-terminal portion of helix a3 forming site 2
(Fig. 3D). We also found a similar change in the dihe-
dral angle of Gly110 upon binding of Lys at site 2 of
CgAKb (details will be published elsewhere). These

observations suggest that the Pro-Gly motif functions
as a hinge to facilitate conformational change upon
effector binding in a
2
b
2
-type AKs. In TtAKb, on the
other hand, the corresponding section (positions 109
and 110) has a Pro-Glu sequence (Fig. 3D). Although
both dihedral angles shown by Gly29 in the presence
and absence of Thr are within the permissible range,
an allowed or generously allowed region on the Rama-
chandran plot, for Glu in general, such a marked
change upon effector binding would not be expected
for Glu. At present, we cannot judge whether the sec-
ond Thr is bound to site 2 for catalytic control of
TtAK. In order to further clarify the regulatory mech-
anism of TtAK by Thr, the crystal structure of full-
length TtAK in the a
2
b
2
form is obviously required.
It should be noted that, on comparison of the amino
acid sequences of CgAK and TtAK, CgAK had an
extra 11 residues at the C-terminus, forming b9, con-
sisting of a b-sheet with a b1-strand at the N-terminus
(Fig. 5D). As TtAK, which is only inhibited by Thr,
does not have this extra b-strand, the b-strand may be
involved in a process of concerted inhibition by Thr

and Lys in CgAK.
Comparison with other AKs
Recently, the crystal structures of Thr-sensitive MjAK
have been determined in three forms [22]: (a) complex
with magnesium adenosine 5¢-(b,c-imido)triphosphate
and Asp; (b) complex with Asp; and (c) complex with
Thr. MjAK has a homotetrameric structure, and
shows high overall structural similarity to the
inhibitory complex of EcAKIII bound to Lys.
Although EcAKIII binds the effector, Lys, at the
binding unit formed between ACT1 domains from
different chains, MjAK binds Thr at the binding sites
formed between ACT2 domains from different chains.
In EcAKIII, transition from the R-state to the
T-state, accompanied by rotational rearrangements to
form a tetramer, occurs through large movement of a
latch loop from the regulatory domains [4]. In MjAK,
however, the loop corresponding to the latch of
EcAKIII is shortened, and shows no conformational
change upon Thr binding. Instead, Thr binding
rotates the regulatory domain away from the kinase
domain. Accompanied by the rotation of the regula-
tory domain, other loops from the catalytic domains
are displaced to orient the residues important for
cofactor and Asp binding in unfavorable positions.
Thus, in spite of their structural similarity, the
regulatory mechanism is different between these
homo-oligomeric AKs. In TtAKb, the loop, b4–a2,
corresponding to the latch in EcAKIII, is short and
shows no structural rearrangement upon Thr binding

(Fig. 5A,B), similar to Thr-sensitive MjAK. In MjAK,
Thr binding causes the entire regulatory domain to
rotate 6.5 away from the fixed kinase domain,
resulting in opening of the catalytic site. In this case,
the entire domain moves as a rigid body with no
significant change in the interaction between ACT
domains [22]. In contrast, Thr binding causes the
ACT2 domain to rotate by 12° from the fixed ACT1
domain (Fig. 5D), which is presumed to be the
motion that closes the active site of TtAK. Thus, the
direction of domain motion is different between TtAK
and MjAK, suggesting a different inhibitory mecha-
nism in TtAK.
Isothermal titration calorimetry suggests that MjAK
has not only Thr-binding sites with high affinity in the
regulatory domain, but also five weak Thr-binding
sites per dimer, which may include a Thr bound near
the Asp-binding site and those bound on the protein
surface nonspecifically [22]. As we have not yet
determined the crystal structure of TtAK in the a
2
b
2
form, we do not know whether TtAK also contains
weak Thr-binding sites. However, it should be noted
that TtAK has a K
i
value of less than 10 lm for Thr,
which is markedly lower than that of MjAK (0.3 mm)
[19]. The low K

i
value of TtAK may indicate that AK
activity is controlled through the high-affinity Thr site
present in the regulatory domain in TtAK.
A. Yoshida et al. Crystal structures of regulatory subunit of Thr-sensitive aspartate kinase
FEBS Journal 276 (2009) 3124–3136 ª 2009 The Authors Journal compilation ª 2009 FEBS 3131
Potential factors involved in the high
thermostability of TtAKb
In our previous study, we found that chimeric AK,
BTT, which is composed of the catalytic domain of
BsAKII and the regulatory domain and b-subunit of
TtAK, had thermostability as high as that of wild-type
TtAK, suggesting that the regulatory domain of TtAK
is also responsible for the thermal stability of TtAK
[11].
TtAKb and CgAKb show 36% sequence identity,
and Thr-bound crystal structures of these proteins are
very similar. To understand the mechanism of
enhanced stability of TtAKb, we examined the dena-
turation of TtAKb and CgAKb by differential scan-
ning calorimetry (DSC) in the presence and absence,
respectively, of their inhibitors. TtAKb has a denatur-
ation temperature approximately 40 °C higher than
that of CgAKb (Table 2). Both CgAKb and TtAKb
are more stable at 4.3–4.4 °C in the presence of Thr.
Considering that TtAKb and, putatively, CgAKb are
in equilibrium between monomers and dimers, and
bound Thr shifts the equilibrium towards dimer forma-
tion, this observation indicates that the small increase
in stability results from a shift of the equilibrium to

dimer formation caused by Thr. Similar protein stabil-
ization via oligomer formation has been shown for a
thermostable homoisocitrate dehydrogenase [23]. In
the crystal, the contact surface area in the dimer inter-
face is larger in TtAKb-Thr (3890 A
˚
2
) than in TtAKb-
free (3200 A
˚
2
), indicating that Thr binding tightens the
interaction of the two chains.
Many factors are involved in protein stability, such
as hydrophobic interactions [24], hydrogen and ionic
bonds [25], cavity volume [26], and other entropic fac-
tors [27]. Proteins are generally stabilized by a combi-
nation of these factors [28]. Among them, an increased
number of hydrogen bonds (ionic interactions) and
better internal packing are reported to be the most
important protein-stabilizing factors [29,30]. To under-
stand the difference in thermostability between CgAKb
and TtAKb, we compared the crystal structures of the
two proteins. When the numbers of ionic bonds and
hydrogen bonds are compared, unexpectedly, both
numbers are larger in CgAKb than in TtAKb
(Table 3). In contrast, when cavity volumes were calcu-
lated from the crystal structure of TtAKb and CgAKb,
the volume of TtAKb was smaller than that of CgAKb
(Table 4), suggesting that TtAKb is more tightly

packed than CgAKb. With regard to the amino acid
composition, TtAKb has a higher ratio of hydrophobic
residues than CgAKb. It is also remarkable that
TtAKb contains more proline residues than CgAKb
(Table 5). Considering that most Pro residues are
located at the N-termini or C-termini of loops in
TtAKb (Fig. 6A), the flexibility of the loop conforma-
tion of TtAKb is likely to be suppressed in the dena-
tured state. We therefore suggest that smaller loss of
entropy upon folding contributes to the stabilization
of TtAKb.
We next calculated changes in Gibbs free energy
from the native to the denatured state, which were esti-
mated on the basis of the solvent-accessible surface
area (Table 4). The difference in changes in Gibbs free
energy between TtAKb and CgAKb was 22 kcalÆ
mol
)1
, indicating that TtAKb is more stable than
CgAKb. A more detailed examination showed that the
difference in the solvent-accessible surface area per
hydrophobic amino acid residue between the native
and denatured states was significantly larger in
TtAKb-Thr than in Thr-bound CgAKb, whereas that
of hydrophilic residues did not change, suggesting a
contribution of internal hydrophobic residues to the
stability of TtAKb. In fact, hydrophobicity inside the
molecule was apparently higher in TtAKb-Thr than in
CgAKb (Fig. 6B). From these results, we conclude
that better internal packing, ensured by tight

Table 2. Denaturation temperatures of TtAKb and CgAKb.
T
m
(°C)
No additive 5 m
M Thr
CgAKb 50.9 55.2
TtAKb 91.7 96.1
Table 3. Thermostabilization factors. Numbers of hydrogen bonds
and ionic bonds. Data in parentheses are number of bonds
between subunits.
TtAKb CgAKb
Hydrogen bonds 237 (12) 262 (11)
Ionic bonds 22 (2) 78 (16)
<3A
˚
3 (0) 14 (2)
<4A
˚
8 (1) 27 (5)
<5A
˚
11 (1) 37 (9)
Table 4. Thermostabilization factors. Accessible surface area and
cavity. Data in parentheses are values per amino acid residue.
TtAKb CgAKbD(Tt)Cg)
Difference in monomer ASA value (D)N)
Hydrophobic (A
˚
2

) 21 313 (70) 22 291 (68) )978 (2)
Hydrophilic (A
˚
2
) 6389 (21) 6709 (21) )320 (0)
DG 283 261 22
Cavity volume
(probe 1.4 A
˚
)(A
˚
3
)
41.4 110.1 )68.7
Crystal structures of regulatory subunit of Thr-sensitive aspartate kinase A. Yoshida et al.
3132 FEBS Journal 276 (2009) 3124–3136 ª 2009 The Authors Journal compilation ª 2009 FEBS
hydrophobic interactions in the interior of protein, and
the richness of Pro residues mainly contribute to the
stabilization of TtAK b. This information might be use-
ful for the generation of more thermostable CgAK
variants for industrial use.
Experimental procedures
Enzyme production and crystallization
Gene cloning and the production, purification and crystalli-
zation of TtAKb-Thr were performed as previously
described [31]. TtAKb-free was crystallized by the hanging
drop, vapor diffusion method. Crystals appeared in 0.1 m
sodium acetate (pH 5.0) and 1.2–2.0 m NaCl.
Data collection
The collection of TtAKb-Thr data and MAD data

collection for SeMet-substituted TtAKb-Thr has been
previously reported [20,31]. Before data collection for
TtAKb-free, a crystal was soaked briefly in cryoprotec-
tant solution of 25% (v ⁄ v) glycerol in reservoir solution,
flash-cooled in a nitrogen gas stream at 95 K, and stored
in liquid nitrogen. Diffraction data were collected with a
CCD camera on the beamline NW12 of the Photon
Factory AR [High Energy Accelerator Research Organi-
zation (KEK), Tsukuba, Japan]. Data on TtAKb-free
crystals were recorded at 2.98 A
˚
resolution. Diffraction
data were indexed, integrated and scaled using the
hkl2000 program suite [32].
Structure determination and refinement
The structure of TtAKb-Thr was determined by the MAD
phasing method. The detailed structure determination and
refinement of TtAKb-Thr were as previously described [20].
TtAKb-free crystals have three dimers per asymmetric unit,
and belong to the space group P3
1
, with unit cell parame-
ters of a = b = 107.2 A
˚
, c = 87.22 A
˚
, a = b =90°, and
c = 120°. The structure determination for TtAKb-free by
molecular replacement was performed by molrep [33] in
the ccp4 program suite [34], using the model of TtAKb-

Thr. Subsequent refinement was conducted using the
program cns1.1 [35], and model correction in the electron
density map was carried out with the xtalview program
suite [36]. Figures were prepared using xfit in the xtal-
view program suite and pymol [DeLano WL, The PyMOL
Molecular Graphics System (2002) at ol.
org]. The atomic coordinates and structure factors deter-
mined in this study have been deposited in the Protein Data
Bank (accession numbers 2dt9 and 2zho).
Determination of quaternary structure
The subunit organization of TtAKb was analyzed by analyt-
ical ultracentrifugation and gel filtration chromatography.
Fig. 6. Factors important for thermostabilization of TtAKb. (A) Pro
residues in TtAKb monomer. (B) Pro residues in CgAKb monomer.
(C, D) Cross-sectional views of TtAKb-Thr dimer and Thr-bound
CgAKb dimer, respectively, drawn by
UCSF CHIMERA [46]. The surface
of the molecules is shown in cyan, and inner hydrophobic residues
are shown in pink.
Table 5. Thermostabilization factors. Comparison of the amino acid
composition of TtAKb and CgAKb.
TtAKb CgAKb
Residues (%) Residues (%)
Hydrophobic 102 63.4 88 51.2
Gly 12 7.45 14 8.14
Ala 28 17.4 18 10.5
Val 15 9.32 19 11.1
Leu 11 6.83 14 8.14
Ile 16 9.94 10 5.81
Met 6 3.73 5 2.92

Phe 5 3.11 4 2.33
Trp 0 0 1 0.58
Pro 9 5.59 3 1.74
Neutral 21 13.0 34 19.8
Ser 6 3.73 10 5.81
Thr 5 3.11 11 6.40
Asn 1 0.62 6 3.49
Gln 9 5.59 6 3.49
Cys 0 0 1 0.58
Hydrophilic 38 23.6 50 29.1
Asp 9 5.59 14 8.14
Glu 13 8.07 15 8.72
Lys 7 4.35 8 4.65
His 3 1.86 2 1.16
Arg 5 3.11 9 5.23
Tyr 1 0.62 2 1.16
A. Yoshida et al. Crystal structures of regulatory subunit of Thr-sensitive aspartate kinase
FEBS Journal 276 (2009) 3124–3136 ª 2009 The Authors Journal compilation ª 2009 FEBS 3133
Sedimentation equilibrium experiments were performed
with a Beckman Optiman XL-A analytical ultracentrifuge
equipped with An-50Ti rotor at 4 °C, with scanning
absorbance at 280 nm. TtAKb at three different concentra-
tions (A
280 nm
of 0.125, 0.25, and 0.5) in buffer A with or
without 5 mm Thr were loaded into a six-channel
centerpiece. Samples were centrifuged at 9757, 20 644
and 35 562 g. Nine curves in total were used for data
fitting. The subunit organization of TtAKb was analyzed
by nonlinear least square analysis using the software

package origin 2.8, provided by Beckman, with the partial
specific volume 0.747 cm
3
Æg
)1
estimated by sednterp soft-
ware [37]. Goodness of fit was also determined by global
analysis in origin 2.8. Gel filtration was performed using a
HiLoad 26 ⁄ 60 Superdex 75 column on an FLPC system
(GE-Healthcare Japan, Tokyo, Japan) at 4 °C. The column
was equilibrated with buffer A (20 mm Tris ⁄ HCl pH 7.5,
150 mm NaCl) or buffer A supplemented with 5 mm Thr.
Protein samples of 5 mg were loaded and eluted at a flow
rate of 2.5 mLÆmin
)1
. Samples of TtAK eluted from the gel
filtration column were collected every 2.5 mL for the
elution volume between 125 and 225 mL. Gel filtration for
TtAK was performed using a HiLoad 26 ⁄ 60 Superdex 200
column under the same elution conditions as described
above.
DSC measurement
Proteins TtAKb and CgAKb were purified by Ni
2+
affinity
with Ni
2+
–nitrilotriacetic acid resin (Novagen, Madison,
WI, USA) and subsequent gel filtration chromatography
with HiLoad 26 ⁄ 60 Superdex 75 equilibrated with buffer B

(20 mm sodium ⁄ potassium phosphate, pH 7.5). The pro-
teins were adjusted to 0.1–0.2 mgÆmL
)1
, and Lys and Thr
were added to the samples arbitrarily at 5 mm. DSC was
carried out by increasing the temperature from 20 to
105 °C with a scanning rate of 1 °C per min, using a
VP-DSC Micro Calorimeter (Microcal, Northampton, MA,
USA). All samples were filtered with a 0.22 lm pore mem-
brane and degassed before measurements.
Calculations of solvent-accessible surface area
and cavity volumes of proteins
To understand the thermostabilization mechanism of
TtAKb, we investigated the factors contributing to the sta-
bilization of two proteins, TtAKb-Thr and Thr-bound
CgAKb, using crystal structures. The cavity volumes of the
proteins were calculated by the voidoo program [38]. The
hbplus program [39] was used to analyze hydrogen bonds
in the crystal structures. To estimate the changes in Gibbs
free energy (DG) of TtAKb and CgAKb, we calculated the
solvent-accessible surface area of two native and denatured
protein forms using the asc program [40]. The denatured
structures were assumed to be in the extended form, and
were constructed using the program insight II (Accelrys
K.K., Tokyo, Japan). The coefficient in the calculation of
Gibbs free energy changes based on the solvent surface area
was as used by Eisenberg and McLachlan [41].
Acknowledgements
This work was supported in part by a grant-in-aid for
scientific research from the Ministry of Education,

Culture, Sports, Science and Technology Japan and by
the Noda Institute for Scientific Research. We thank
the staff of the Photon Factory for their assistance
with data collection. This work was approved by the
Photon Factory Program Advisory Committee (Pro-
posal no. 2005G268, 2007G531). We are also grateful
to H. Fukushima and S. Watabe (University of Tokyo)
for their assistance with analysis of protein thermosta-
bility by DSC.
References
1 Nishiyama M, Kukimoto M, Beppu T & Horinouchi S
(1995) An operon encoding aspartokinase and purine
phosphoribosyltransferase in Thermus flavus. Microbiol-
ogy 141, 1211–1219.
2 Kalinowski J, Bachmann B, Thierbach G & Pu
¨
hler A
(1990) Aspartokinase genes lysCa and lysCb overlap
and are adjacent to the aspartate b-semialdehyde dehy-
drogenase gene asd in Corynebacterium glutamicum.
Mol Gen Genet 224, 317–324.
3 Kalinowski J, Cremer J, Bachmann B, Eggeling L,
Sahm H & Pu
¨
hler A (1991) Genetic and biochemical
analysis of the aspartokinase from Corynebacterium glu-
tamicum. Mol Microbiol 5, 1197–1204.
4 Kotaka M, Ren J, Lockyer M, Hawkins AR & Stam-
mers DK (2006) Structures of R- and T-state Escheri-
chia coli aspartokinase III. Mechanisms of the allosteric

transition and inhibition by lysine. J Biol Chem 281,
31544–31552.
5 Rognes SE, Lea PJ & Miflin BJ (1980) S-adenosylme-
thionine – a novel regulator of aspartate kinase. Nature
287, 357–359.
6 Mas-Droux C, Curien G, Robert-Genthon M, Lauren-
cin M, Ferrer JL & Dumas R (2006) A novel organiza-
tion of ACT domains in allosteric enzymes revealed by
the crystal structure of Arabidopsis aspartate kinase.
Plant Cell 18, 1681–1692.
7 Shiio I & Miyajima R (1969) Concerted inhibition
and its reversal by end products of aspartate kinase
in Brevibacterium flavum. J Biochem (Tokyo) 65, 849–
859.
8 Kobashi N, Nishiyama M & Tanokura M (1999)
Aspartate kinase-independent lysine synthesis in an
extremely thermophilic bacterium, Thermus thermophi-
Crystal structures of regulatory subunit of Thr-sensitive aspartate kinase A. Yoshida et al.
3134 FEBS Journal 276 (2009) 3124–3136 ª 2009 The Authors Journal compilation ª 2009 FEBS
lus: lysine is synthesized via a-aminoadipic acid, not via
diaminopimelic acid. J Bacteriol 181, 1713–1718.
9 Nishida H, Nishiyama M, Kobashi N, Kosuge T,
Hoshino T & Yamane H (1999) A prokaryotic gene
cluster involved in synthesis of lysine through the amino
adipate pathway: a key to the evolution of amino acid
biosynthesis. Genome Res 9, 1175–1183.
10 Kobashi N, Nishiyama M & Tanokura M (1999)
Kinetic and mutation analyses of aspartate kinase from
Thermus flavus. J Biosci Bioeng 87, 739–745.
11 Kato C, Kurihara T, Kobashi N, Yamane H & Nishiy-

ama M (2004) Conversion of feedback regulation in
aspartate kinase by domain exchange. Biochem Biophys
Res Commun 316, 802–808.
12 Aravind L & Koonin EV (1999) Gleaning non-trivial
structural, functional and evolutionary information
about proteins by iterative database searches. J Mol
Biol 287, 1023–1040.
13 Chipman DM & Shaanan B (2001) The ACT domain
family. Curr Opin Struct Biol 11, 694–700.
14 Grant GA (2006) The ACT domain: a small molecule
binding domain and its role as a common regulatory
element. J Biol Chem 281, 33825–33829.
15 Schuller DJ, Grant GA & Banaszak LJ (1995) The allo-
steric ligand site in the Vmax-type cooperative enzyme
phosphoglycerate dehydrogenase. Nat Struct Biol 2,
69–76.
16 Gallagher DT, Gilliland GL, Xiao G, Zondlo J, Fisher
KE, Chinchilla D & Eisenstein E (1998) Structure and
control of pyridoxal phosphate dependent allosteric
threonine deaminase. Structure 6, 465–475.
17 Wessel PM, Graciet E, Douce R & Dumas R (2000)
Evidence for two distinct effector-binding sites in threo-
nine deaminase by site-directed mutagenesis, kinetic,
and binding experiments. Biochemistry 39, 15136–15143.
18 Curien G, Biou V, Mas-Droux C, Robert-Genthon M,
Ferrer JL & Dumas R (2008) Amino acid biosynthesis:
new architectures in allosteric enzymes. Plant Physiol
Biochem 46, 325–339.
19 Faehnle CR, Liu X, Pavlovsky A & Viola RE (2006)
The initial step in the archaeal aspartate biosynthetic

pathway catalyzed by a monofunctional aspartokinase.
Acta Crystallogr F 62, 962–966.
20 Yoshida A, Tomita T, Kurihara T, Fushinobu S,
Kuzuyama T & Nishiyama M (2007) Structural insight
into concerted inhibition of a
2
b
2
-type aspartate kinase
from Corynebacterium glutamicum. J Mol Biol 368,
521–536.
21 Laskowski RA, MacArthur MW, Moss DS & Thornton
JM (1993) PROCHECK: a program to check the ste-
reochemical quality of protein structures. J Appl Crys-
tallogr 26, 283–291.
22 Liu X, Pavlovsky AG & Viola RE (2008) The structural
basis for allosteric inhibition of a threonine-sensitive
aspartokinase. J Biol Chem 283, 16216–16225.
23 Miyazaki J, Asada K, Fushinobu S, Kuzuyama T &
Nishiyama M (2005) Crystal structure of tetrameric
homoisocitrate dehydrogenase from an extreme thermo-
phile, Thermus thermophilus: involvement of hydropho-
bic dimer–dimer interaction in extremely high
thermotolerance. J Bacteriol 187, 6779–6788.
24 Kauzmann W (1959) Some factors in the interpretation
of protein denaturation. Adv Protein Chem 14, 1–63.
25 Perutz MF & Raidt H (1975) Stereochemical basis of
heat stability in bacterial ferredoxins and in haemoglo-
bin A2. Nature 255, 256–259.
26 Kellis JT Jr, Nyberg K, Sali D & Fersht AR (1988)

Contribution of hydrophobic interactions to protein sta-
bility. Nature 333, 784–786.
27 Matthews BW, Nicholson H & Becktel WJ (1987)
Enhanced protein thermostability from site-directed
mutations that decrease the entropy of unfolding. Proc
Natl Acad Sci USA 84, 6663–6667.
28 Petsko GA (2001) Structural basis of thermostability
in hyperthermophilic proteins, or ‘there’s more than
one way to skin a cat’. Methods Enzymol 334, 469–
478.
29 Querol E, Perez-Pons JA & Mozo-Villarias A (1996)
Analysis of protein conformational characteristics
related to thermostability. Protein Eng 9, 265–271.
30 Vogt G & Argos P (1997) Protein thermal stability:
hydrogen bonds or internal packing? Fold Des 2, S40–
S46.
31 Yoshida A, Tomita T, Kuzuyama T & Nishiyama M
(2007) Purification, crystallization and preliminary
X-ray analysis of regulatory subunit of aspartate kinase
from Thermus thermophilus. Acta Crystallogr F 63, 96–
98.
32 Otwinowski Z & Minor W (1997) Processing of X-ray
diffraction data collected in oscillation mode. Methods
Enzymol 276, 307–326.
33 Vagin A & Teplyakov A (1997) MOLREP: an auto-
mated program for molecular replacement. J Appl Crys-
tallogr 30, 1022–1025.
34 Collaborative Computational Project N (1994) The
CCP4 suite: programs for protein crystallography. Acta
Crystallogr D 50, 760–763.

35 Brunger AT, Adams PD, Clore GM, DeLano WL,
Gros P, Grosse-Kunstleve RW, Jiang JS, Kuszewski J,
Nilges M, Pannu NS et al. (1998) Crystallography &
NMR system: a new software suite for macromolecular
structure determination. Acta Crystallogr D 54
, 905–
921.
36 McRee DE (1999) XtalView ⁄ Xfit – a versatile program
for manipulating atomic coordinates and electron den-
sity. J Struct Biol 125, 156–165.
37 Laue TM, Shah BD, Ridgeway TM & Pelletier SL
(1992) Computer-aided interpretation of analytical sedi-
mentation data for proteins. In Analytical Ultracentrifu-
gation in Biochemistry and Polymer Science (Harding S,
A. Yoshida et al. Crystal structures of regulatory subunit of Thr-sensitive aspartate kinase
FEBS Journal 276 (2009) 3124–3136 ª 2009 The Authors Journal compilation ª 2009 FEBS 3135
Rowe A & Horton L, eds), pp. 99–125. Royal Society
of Chemistry, Cambridge, UK.
38 Kleywegt GJ & Jones TA (1994) Detection, delineation,
measurement and display of cavities in macromolecular
structures. Acta Crystallogr D 50, 178–185.
39 McDonald IK & Thornton JM (1994) Satisfying hydro-
gen bonding potential in proteins. J Mol Biol 238, 777–
793.
40 Eisenhaber F & Argos P (1993) Improved strategy in
analytic surface calculation for molecular systems: han-
dling of singularities and computational efficiency.
J Comput Chem 14, 1272–1280.
41 Eisenberg D & McLachlan AD (1986) Solvation energy
in protein folding and binding. Nature 319, 199–203.

42 Thompson JD, Higgins DG & Gibson TJ (1994)
CLUSTALW: improving the sensitivity of progressive
multiple sequence alignment through sequence weight-
ing, position-specific gap penalties and weight matrix
choice. Nucleic Acids Res 22, 4673–4680.
43 Gouet P, Courcelle E, Stuart DI & Metoz F (1999)
ESPript: analysis of multiple sequence alignments in
PostScript. Bioinformatics 15, 305–308.
44 Abramoff MD, Magelhaes PJ & Ram SJ (2004) Image
processing with ImageJ. Biophotonics Int 11, 36–42.
45 Hayward S & Lee RA (2002) Improvements in the anal-
ysis of domain motions in proteins from conformational
change: DynDom version 1.50. J Mol Graph Model 21,
181–183.
46 Pettersen EF, Goddard TD, Huang CC, Couch GS,
Greenblatt DM, Meng EC & Ferrin TE (2004) UCSF
Chimera – a visualization system for exploratory
research and analysis. J Comput Chem 25, 1605–1612.
Crystal structures of regulatory subunit of Thr-sensitive aspartate kinase A. Yoshida et al.
3136 FEBS Journal 276 (2009) 3124–3136 ª 2009 The Authors Journal compilation ª 2009 FEBS