Báo cáo khoa học: Crystal structure of salt-tolerant glutaminase from Micrococcus luteus K-3 in the presence and absence of its product L-glutamate and its activator Tris pdf

Crystal structure of salt-tolerant glutaminase from
Micrococcus luteus K-3 in the presence and absence of its
product
L-glutamate and its activator Tris
Kazuaki Yoshimune
1
, Yasuo Shirakihara
2
, Mamoru Wakayama
3
and Isao Yumoto
1
1 Research Institute of Genome-based Biofactory, National Institute of Advanced Industrial Science and Technology (AIST), Sapporo,
Hokkaido, Japan
2 Structural Biology Center, National Institute of Genetics, Mishima, Shizuoka, Japan
3 Department of Biotechnology, Faculty of Life Science, Ritsumeikan University, Kusatsu, Shiga, Japan
Introduction
Glutaminase (EC 3.5.1.2) hydrolyzes l-glutamine to
produce l-glutamate and ammonia. It is highly specific
for l-glutamine and is distinct from asparaginase (EC
3.5.1.1), which hydrolyzes both glutamine and aspara-
gine [1]. Glutaminase is found in many organisms,
including mammals, fungi, yeasts and bacteria [1–5].
Glutaminase from Micrococcus luteus K-3 (Mglu) and
AoGlS glutaminase from Aspergillus oryzae RIB40 are
salt-tolerant [6,7]. Glutaminase (YbaS) from Escheri-
chia coli exhibits low b-lactamase activity [5], and
Keywords
crystal structure;
L-glutamate;
Micrococcus luteus K-3; salt-tolerant

glutaminase; Tris
Correspondence
K. Yoshimune, Research Institute of
Genome-based Biofactory, National Institute
of Advanced Industrial Science and
Technology (AIST), Tsukisamu-Higashi,
Toyohira-ku, Sapporo, Hokkaido 062-8517,
Japan
Fax: +81 11 857 8980
Tel: +81 11 857 8444
E-mail:
Note
Coordinates and structure factors have been
deposited in the Protein Data Bank under
accession codes 3ih8 (N), 3ih9 (T), 3iha (G),
and 3ihb (TG)
(Received 5 September 2009, revised 1
November 2009, accepted 26 November
2009)
doi:10.1111/j.1742-4658.2009.07523.x
Glutaminase from Micrococcus luteus K-3 [Micrococcus glutaminase
(Mglu); 456 amino acid residues (aa); 48 kDa] is a salt-tolerant enzyme. Our
previous study determined the structure of its major 42-kDa fragment. Here,
using new crystallization conditions, we determined the structures of the
intact enzyme in the presence and absence of its product l-glutamate and its
activator Tris, which activates the enzyme by sixfold. With the exception of
a ‘lid’ part (26-29 aa) and a few other short stretches, the structures were all
very similar over the entire polypeptide chain. However, the presence of the
ligands significantly reduced the length of the disordered regions: 41 aa in
the unliganded structure (N), 21 aa for l-glutamate (G), 8 aa for Tris (T)

and 6 aa for both l-glutamate and Tris (TG). l-Glutamate was identified in
both the G and TG structures, whereas Tris was only identified in the TG
structure. Comparison of the glutamate-binding site between Mglu and salt-
labile glutaminase (YbgJ) from Bacillus subtilis showed significantly smaller
structural changes of the protein part in Mglu. A comparison of the sub-
strate-binding pocket of Mglu, which is highly specific for l-glutamine, with
that of Erwinia carotovora asparaginase, which has substrates other than
l-glutamine, shows that Mglu has a larger substrate-binding pocket that
prevents the binding of l-asparagine with proper interactions.
Structured digital abstract
l
MINT-7305730: Mglu (uniprotkb:Q4U1A6) and Mglu (uniprotkb:Q4U1A6) bind (MI:0407)
by x-ray crystallography (
MI:0114)
Abbreviations
aa, amino acid residues; F, major fragment of Mglu; G, presence of
L-glutamate; Mglu, Micrococcus glutaminase; PDB, protein data bank;
N, absence of additives; T, presence of Tris; TG, presence of Tris +
L-glutamate.
738 FEBS Journal 277 (2010) 738–748 ª 2009 The Authors Journal compilation ª 2009 FEBS
glutaminases from E. coli [5,8], Bacillus subtilis [5], pigs
[9] and humans [2] exhibit allosteric behavior for l-glu-
tamine with positive cooperativity [2,8,9]. Moreover,
the crystal structures of a major fragment of Mglu
[protein data bank (PDB) 3if5] [10], and glutaminases
from E. coli (PDB 1u60) [5], B. subtilis (PDB 1mki,
3brm, 2osu) [5], Geobacillus kaustophilus (PDB 2pby)
and human kidney (PDB 3czd) have been determined.
The structure of glutaminase from B. subtilis (YbgJ,
Bacillus glutaminase) with covalently bound 5-oxo-

l-norleucine (PDB 3brm) reveals its Ser74 to be the
catalytic nucleophile [5]. On the basis of their crystal
structures and the conserved amino acid residues (aa)
of their active sites, these glutaminases are classified
into a large group of bacterial penicillin-binding
proteins and b-lactamases [5,10].
Most enzymes from nonhalophilic microorganisms
are inhibited in high salt concentrations [1,11]. Mglu is
a salt-tolerant enzyme with about 40% residual activ-
ity, even in 2.7 m NaCl [6,12]. Salt-tolerant enzymes
are distinct from halophilic enzymes that are inactive
in the absence of salt [13]. Although the structures of
the fragments of Mglu [10] and algal carbonic anhydr-
ase [14] have been determined, the mechanism of salt
tolerance is not very well understood. Mglu is com-
posed of N-terminal (1-305 aa) and C-terminal (315-
456 aa) domains, with a peptide linker (306-314 aa).
We previously reported the structure of a major frag-
ment (referred to as F) of Mglu. The fragment (F)
consists of the conserved N-terminal domain that con-
tains active-site residues and part of the C-terminal
domain. The structure of the N-terminal domain is
conserved among species (E. coli, B. subtilis, G. kausto-
philus and human kidney) with superposition rmsd val-
ues ranging from 1.2 to 1.5 A
˚
(when superimposed
onto Mglu). The C-terminal domain is not present in
other glutaminases, such as those from B. subtilis and
E. coli [5,10]; therefore, it was hypothesized that the

C-terminal domain might be responsible for salt
tolerance [10].
Here, we report the crystal structure of the intact glu-
taminase under four different conditions: in the absence
of the additives (referred to as N); in the presence of
Tris (referred to as T); in the presence of l-glutamate
(referred to as G); and in the presence of both Tris and
l-glutamate (referred to as TG). Furthermore, the li-
ganded (TG) and unliganded (T) structures of Mglu
were compared with those of salt-labile Bacillus gluta-
minase to gain insight into the mechanism of salt toler-
ance. The liganded (TG) structure was also compared
with that of Erwinia carotovora asparaginase, which
shows a broader substrate specificity, to investigate the
mechanism of the narrow substrate specificity.
Results and Discussion
Determination of overall structure
In a previous study [10], we determined the F struc-
ture containing the truncated region of the C-terminal
domain; however, we failed to determine the structure
of the intact glutaminase because the crystals exhib-
ited heavy twinning. In the present study, we found
new crystallization conditions that employed poly(eth-
ylene glycol) 4000 as a precipitant (previously sodium
citrate was used), and these crystallization conditions
led us to determine nearly the full-length structure of
Mglu in the presence and absence of l-glutamate and
Tris (N, T, G and TG). l-Glutamate was identified in
the G and TG structures, and Tris was identified only
in the TG structure. The crystallographic statistics are

summarized in Table 1. There were two molecules in
the asymmetric unit of space group C2, and a physio-
logical dimer [10] was generated by a twofold axis of
crystallographic symmetry. The two independent
physiological dimers shared a similar structure. The
physiological dimer of the TG structure is shown in
Fig. 1. The monomer structures were very similar
among the four different forms. However, small, but
significant, differences were found to exist among
those structures and these will be described in the
subsequent sections. Figure 2 shows the C-terminal
domain in the F, N, G, T and TG structures to dem-
onstrate the structural similarity and varying degree
of disordered regions that are present. The N struc-
ture was found to have the largest number of disor-
dered residues (354-376, 395-401 and 446-456 aa;
41 aa in total). Bound l-glutamate reduced the num-
ber of disordered residues (353-359, 397-402 and 449-
456 aa; 21 aa in total), and Tris dramatically reduced
this number even further (449-456 aa; eight aa in
total). In the presence of both Tris and l-glutamate,
the number of disordered residues was minimal (451-
456 aa; six aa in total).
A search for structural homologues of the C-termi-
nal domain using the DALI [15] databases identified
SpoIIAA (anti-anti-sigma factor; PDB 1til, chain D;
14% sequence identity). No structural homologue for
any part (354-403 aa) of the C-terminal domain was
found by the search. SpoIIAA [16] (6-37 and 43-
81 aa) could be superimposed on parts of the C-ter-

minal domain (319-350 and 404-442 aa) with rmsd
values of 1.39 A
˚
. SpoIIAA was found to bind to Spo-
IIAB, thereby antagonizing the anti-r
F
function of
SpoIIAB; the structure of SpoIIAA, in complex with
SpoIIAB, contained 11 interface residues (21, 23-25,
K. Yoshimune et al. Crystal structure of salt-tolerant glutaminase
FEBS Journal 277 (2010) 738–748 ª 2009 The Authors Journal compilation ª 2009 FEBS 739
31, 58, 59, 61, 67, 89 and 90 aa). The structurally
homologous residues in Mglu (334, 336-338, 344, 419,
420, 422 and 428 aa; nine aa in total) were located
near the interface of the physiological dimer. The C-
terminal domain may contain structural features
involved in protein–protein interactions to form the
physiological dimers.
Active site
Figure 3 shows l-glutamate in the G and TG struc-
tures. The electron density of l-glutamate became
more pronounced upon the addition of Tris. The
location and orientation of l-glutamate were deter-
mined on the basis of the lower electron density of
Table 1. Crystallographic statistics and refinement parameters.
X-ray source
NT G TG
BL-6A AR-NW12A AR-NW12A BL-6A
Data collection
Resolution (A

˚
) 2.3 2.5 2.6 2.4
Mean I ⁄ r (I)
a
10.9 (2.2) 11.0 (2.1) 9.1 (2.8) 8.7 (1.8)
No. of reflections
Measure 189 811 149 207 125 111 147 220
Unique 52 671 40 451 35 920 46 819
Completeness (%)
a
100 (99.9) 98.4 (98.0) 98.2 (99.6) 99.5 (98.6)
R
merge
(%)
a,b
Overall 11.1 12.4 11.9 13.1
Highest-resolution shell 54.6 49.2 35.6 41.2
Lowest-resolution shell 6.1 8.2 8.7 5.7
Multiplicity
a
3.6 (3.6) 3.7 (3.7) 3.5 (3.5) 3.1 (2.8)
Unit-cell constants
a(A
˚
) 118.08 117.70 117.45 118.36
b(A
˚
) 142.25 141.39 141.62 141.20
c(A
˚

) 74.25 75.23 75.15 76.12
a (°) 90909090
b (°) 104.1 104.5 104.0 105.4
c (°) 90909090
B value from Wilson plot (A
˚
2
) 42.1 52.8 58.0 41.0
Molecular replacement and refinement
Initial model F TG N N
Resolution range (A
˚
) 19.9–2.3 19.9–2.5 19.9–2.6 19.9–2.4
No. of reflections 52 579 40 379 35 873 46 741
R factor for 95% of data
c
0.247 0.217 0.271 0.222
Free R factor for 5% of data 0.291 0.258 0.309 0.268
No. of atoms
Protein 6 118 6 664 6 505 6 679
Water 433 288 191 511
Glutamate 0 0 20 20
Tris 0 0 0 16
Disordered amino acid residues rmsd 86 16 39 14
from ideality
Bond length (A
˚
) 0.006 0.008 0.008 0.007
Bond angle (°) 1.30 1.38 1.43 1.37
Average B value (A

˚
2
) 40.8 48.9 70.4 40.9
Ramachandran analysis
Favoured (%) 89.1 88.3 85.7 86.2
Allowed (%) 10.2 10.8 13.2 12.8
Generally allowed (%) 0.7 0.9 1.1 0.9
Disallowed (%) 0 0 0 0
Error in coordinate by Luzzati (A
˚
) 0.32 0.36 0.51 0.34
a
Numbers in parentheses refer to data for the highest-resolution shell.
b
R
merge
= RhRi|I(h)-I(h)i| ⁄ RhRiI(h)I, where h is a unique reflection index, I(h)I is the intensity of symmetry-related reflections and I(h) is the
mean intensity.
c
R factor=Rh||F
o
|
h
| ⁄ R
h
|F
o
|
h
, where h is a unique reflection index.

Crystal structure of salt-tolerant glutaminase K. Yoshimune et al.
740 FEBS Journal 277 (2010) 738–748 ª 2009 The Authors Journal compilation ª 2009 FEBS
the a-amino group, and on the higher electron den-
sity of the a-carboxyl group, of l-glutamate. Figure 3
shows the seven residues (Gln63, Ser64, Asn114,
Glu160, Asn167, Cys195 and Val261) that interact
with l-glutamate, and also shows Lys67, which is
probably hydrogen bonded (2.9 A
˚
) to Ser64. These
Fig. 1. The TG structure in the physiological dimeric form. One monomer is shown in cyan and the other is shown in yellow. L-Glutamate
(magenta) and Tris (red) are shown as cylinders. The ‘lid’ (26-29 aa) is shown in magenta.
Fig. 2. C-terminal domains in the F, N, G, T and TG structures. The backbone atoms of the C-terminal domain (1-305 aa) in the F (black), G
(yellow), T (cyan) and TG (magenta) structures are superimposed on that in the N structure (blue). The newly determined part of the C-termi-
nal domain in the T and TG structures suggests a revised F structure (PDB 3if5). There are disordered regions in the F (354-403 and 449-
456 aa), N (354-376, 395-401 and 446-456 aa), G (353-359, 397-402 and 449-456 aa), T (449-456 aa) and TG (451-456 aa) structures.
Fig. 3. Active site of Mglu. L-Glutamate and
active-site residues in the G (yellow) and TG
(magenta) structures are represented as cyl-
inders. F
o
-F
c
electron densities in the maps
of the G (yellow) and TG (magenta) struc-
tures are contoured at +2.0 and +3.0 sigma
levels, respectively. Putative interactions
(atom distance less than 3.9 A
˚
) are shown

by dotted lines.
K. Yoshimune et al. Crystal structure of salt-tolerant glutaminase
FEBS Journal 277 (2010) 738–748 ª 2009 The Authors Journal compilation ª 2009 FEBS 741
residues are conserved among glutaminases from
B. subtilis (YbgJ) [5], E. coli (YbaS) [5], G. kaustophi-
lus (PDB 2pby) and humans (PDB 3czd), except for
Gln63, for which a glutamate residue is substituted in
YbaS. The C-a locations and side-chain orientations
of the seven conserved residues are similar among
these glutaminases, suggesting that these residues have
similar functions. On the basis of the proposed cata-
lytic mechanism of Bacillus glutaminase, Ser64 is
likely to be the putative catalytic nucleophile, and
Lys67 may function as general base to assist Ser64.
Tyr27 (Tyr37 in YbgJ) is hypothesized to interact
with the a-amino group of l-glutamine, and Tyr191
(Tyr201 in YbgJ) and Tyr243 (Tyr253 in YbgJ) are
proposed to participate in the proton-transfer reaction
during the catalysis. In addition to the seven con-
served residues, these residues surround l-glutamate
in the TG structure and they are also structurally
conserved among these glutaminases.
Effect of
L-glutamate on the structure
Figure 4A shows the effect of l-glutamate on the struc-
ture of Mglu. A comparison between the T and TG
structures shows clearly that a single region (26-29 aa)
changes upon l-glutamate binding. We refer to this seg-
ment as the ‘lid’, which is located at the surface of gluta-
minase and near its active site (Fig. 1), because it

appears to enclose the bound l-glutamate in the G and
TG forms. In the unliganded form, the lid is placed fur-
ther from the active site, opening up the active site. This
open–close motion, caused by the presence of l-gluta-
mate, is small, as shown in Fig. 5; however, it is signifi-
cant, as judged from the plot shown in Fig. 4A.
The N and G structures were also compared to eval-
uate the l-glutamate effect. Figure 4A shows the peak
caused by the lid motion, described above, and two
additional peaks (100-105, 303-309 aa). However, those
two peaks were considered to be insignificant as they
also appear in the N ⁄ T comparison shown in Fig. 4B.
It seems that the N structure is structurally unique
among all the structures at 100-105 aa and 303-309 aa.
This is supported by the fact that these two peaks are
not observed in the F⁄ GorF⁄ T comparisons, in
which the F form has no bound ligand (data not
shown). It was difficult to determine what brought
about the different conformations of the regions at
100-105 and 303-309 aa in the N form, and we investi-
gated the following points. First, these areas are not
involved in any contacts with other molecules in crys-
tals in all the crystal forms. Second, the overall contact
area of the N structure (3 890 A
˚
2
) is significantly smal-
ler than those of the T (4 583 A
˚
2

), G (4 435 A
˚
2
) and
TG (4 559 A
˚
2
) structures. This is apparently because
Fig. 4. Effect of L-glutamate and Tris on
backbone atoms. Backbone atoms of the
overall amino acid residues of the T and TG
(T ⁄ TG), N and G (N ⁄ G), G and TG (G ⁄ TG),
and N and T (N ⁄ T) structures are superim-
posed. Effects of (A)
L-glutamate (T ⁄ TG and
N ⁄ G) and (B) Tris (G ⁄ TG and N ⁄ T) on the
structure are shown. Circles and numbers
indicate residues mentioned in the text.
Crystal structure of salt-tolerant glutaminase K. Yoshimune et al.
742 FEBS Journal 277 (2010) 738–748 ª 2009 The Authors Journal compilation ª 2009 FEBS
of the more extensive disordered regions in the C-ter-
minal domain of the N structure.
Salt-tolerant mechanism
A comparison between the structures of Mglu and
salt-labile glutaminase was expected to elucidate the
structural features of salt-tolerant enzymes. Bacillus
glutaminase is a structural homologue of Mglu, and
thus the enzyme is an appropriate counterpart. As the
salt tolerance of Bacillus glutaminase had not yet been
determined, the gene was cloned and the overproduced

enzyme was purified. The purified Bacillus glutaminase
exhibited no activity in 1.3 m NaCl at its optimum pH
8.0 (data not shown). Figure 5 shows the active sites
of Mglu and Bacillus glutaminase in the presence and
absence of the ligands. The l-glutamate of Mglu is
located near the seven conserved amino acid residues
(shown in Fig. 3), the lid part (26-29 aa) and the three
conserved tyrosine residues (Tyr27, Tyr191, and
Tyr243). Homologous residues of the three conserved
tyrosine residues (Tyr37, Tyr201, and Tyr253) in Bacil-
lus glutaminase have been hypothesized to be impor-
tant for the catalytic activity of the enzyme [5]. The
location of l-glutamate in the TG structure is similar
to that of 5-oxo-l-norleucine, which covalently binds
to Bacillus glutaminase. Table 2 shows the displace-
ment, upon ligand binding, of the amino acid residues
that are shown in Fig. 5. The displacement of these
residues in Mglu is much smaller than those in Bacillus
glutaminase. In addition to the conserved amino acid
residues, the displacement of the overall amino acid
residues in Mglu is also small compared with those of
Bacillus glutaminase. Figure 6 shows a comparison of
the displacement of all the backbone atoms of Mglu
and Bacillus glutaminase induced by ligand binding.
Mglu shows no significant conformational change
(average, 0.25 A
˚
; SD, 0.15 A
˚
) except for the lid motion

of residues 26-29 (average, 1.14 A
˚
; SD, 0.38 A
˚
), which
Fig. 5. Active sites of Mglu and Bacillus glutaminase in the presence and absence of ligands. Backbone atoms of the seven conserved
amino acid residues and the three conserved tyrosine residues in Mglu and Bacillus glutaminase are superimposed. T (cyan) and TG
(magenta) structures, and structures of Bacillus glutaminase in the presence (BD, grey) and absence (B, white) of 5-oxo-
L-norleucine are
shown in different colors. Atoms of ligands are represented by a ball-and-stick model. Putative interactions (atom distance less than 3.9 A
˚
)
are shown by dotted lines. Residues in the T structure (residue number and type) and in the BD structure (residue number) are labeled.
Residues 36-39 in the B structure (homologous residue of 26-29 in Mglu) are disordered.
Table 2. Displacement of the backbone atoms of Mglu and
Bacillus glutaminase induced by ligand binding. Displacement val-
ues for backbone atoms of the nonliganded and liganded structures
are shown. The average and SD of these conserved residues (nine
aa in total) are 0.16 ± 0.07 (Mglu) and 0.51 ± 0.34 (Bacillus gluta-
minase).
Mglu
a
Bacillus glutaminase
b
Residue rmsd (A
˚
) Residue rmsd (A
˚
)
The seven conserved residues that interact with

L-glutamate
Gln63 0.10 Gln73 0.30
Ser64 0.15 Ser74 0.29
Asn114 0.14 Asn126 0.57
Glu160 0.11 Glu170 0.57
Asn167 0.14 Asn177 0.94
Cys195 0.14 Cys205 0.30
Val261 0.17 Val271 0.25
Average ± SD 0.14 ± 0.02 0.46 ± 0.25
The three conserved tyrosine residues
Tyr27
c
1.43 Tyr37 -
d
Tyr191 0.16 Tyr201 0.20
Tyr243 0.33 Tyr253 1.16
Average ± SD 0.25 ± 0.12 0.68 ± 0.68
a
Backbone atoms of the N-terminal domains (1-305 aa) in the T
and TG structures are aligned.
b
Backbone atoms of Bacillus gluta-
minase are superimposed on those of the liganded enzyme.
c
The
value of Tyr27 is excluded from the calculation of the average and
SD.
d
Tyr37 is disordered in the nonliganded structure of Bacillus
glutaminase.

K. Yoshimune et al. Crystal structure of salt-tolerant glutaminase
FEBS Journal 277 (2010) 738–748 ª 2009 The Authors Journal compilation ª 2009 FEBS 743
is described above. By contrast, Bacillus glutaminase
shows relatively large displacement (average, 0.42 A
˚
;
SD, 0.37 A
˚
). The small conformational change is a
possible explanation for the salt-tolerant glutaminase.
As inactivation of nonhalophilic enzymes by high salt
concentrations is usually caused by a decrease in flexi-
bility [17], an enzyme reaction with a small conforma-
tional change might be favourable under high salt
conditions. However, this concept may be specific to
glutaminase because such a small motion (average
0.24 A
˚
; SD 0.11 A
˚
) as a result of ligand binding is also
observed in b-lactamase Toho-1 (PDB 1bza and 1iyq)
from E. coli.
Substrate specificity
Glutaminase is strictly specific to l-glutamine [1] and
distinct from asparaginase whose substrates are both
glutamine and asparagine [18–23]. To gain structural
insights regarding substrate specificity, the TG struc-
ture was compared with the structure of E. carotovora
asparaginase in complex with l-glutamate (PDB 2hln)

[18]. Mglu binds l-glutamate in an extended form
(Fig. 7A) with a contact area of 288 A
˚
2
. Asparaginase
binds l-glutamate in a folded form (Fig. 7B) in an
apparently much smaller pocket with a contact area of
233 A
˚
2
. These findings suggest that the substrate-bind-
ing pocket of Mglu is too large to bind l-asparagine
with proper interactions.
Effect of Tris on the structure
Tris was found to activate Mglu by approximately
sixfold at pH 7.5 (data not shown). This activation is
specific for Tris, and Tris analogues, such as Bistris or
tricine, induce no such effect. Similar specific activa-
tions have also been observed in other proteins [24–
30]. As shown in Fig. 8, Tris is identified only in the
TG structure. Residues Arg223 and Ser227 interact
with Tris. Tris binding in the TG structure seems to be
coupled to the protein local structure in the region
encompassing aa 222-228, which is distinct from the
corresponding structures in the N, T and G forms, in
which Tris is not identified.
In contrast to l-glutamate, Tris binding induces a
lesser degree of structural changes, as shown in the
G ⁄ TG superposition plot in Fig. 4B. The N ⁄ T super-
position is not taken into account for the reasons

described in the previous section. The N-terminal resi-
dues (Met1-His3)that give a single peak in the G ⁄ TG
plot interact with the C-terminal domain residues
(Asp345, Leu347, Thr352, Gly428 and Asp435) in the
physiological dimer in all the structures. Therefore, the
observed structural difference in the N-terminal resi-
dues must have arisen from subtle differences in the
interactions across the domain interface. All the resi-
dues involved here seem to be unrelated to the Tris-
binding site residues (including Arg223 and Ser227),
either directly or indirectly. Similarly, the 33 ordered
residues in the T structure (that are disordered in the
N form) are not related to the Tris-binding site resi-
dues in any way. Thus, the effect of Tris binding on
the glutaminase structure remains unclear.
Catalytic mechanism of Mglu
The similar spatial arrangement of the probable active-
site residues in Mglu and Bacillus glutaminase suggests
that the catalytic mechanism of Mglu is similar to the
proposed catalytic mechanism of Bacillus glutaminase
[5]. The structure of Bacillus glutaminase that cova-
lently binds 5-oxo-l-norleucine (PDB 3brm) mimics
that of the acyl-enzyme intermediate, while the Mglu
TG structure may represent a stage just prior to prod-
Fig. 6. Displacement of backbone atoms for all amino acid residues
of Mglu and Bacillus glutaminases induced by ligand binding.
Displacements of the conserved amino acid residues in Table 2 are
indicated by circles. (A) Backbone atoms of all amino acid residues
in the T structure are superimposed on those in the TG structure.
The arrow indicates the Tyr27 residue. The mean and SD of the val-

ues are 0.25 A
˚
and 0.15 A
˚
, respectively. (B) The backbone atoms
of liganded and unliganded structures of Bacillus glutaminase are
superimposed. Displacement of 55 aa for Bacillus glutaminase
(327 aa) is not shown because of the disordered regions in the
structures. The mean and SD of the values are 0.42 A
˚
and 0.37 A
˚
,
respectively.
Crystal structure of salt-tolerant glutaminase K. Yoshimune et al.
744 FEBS Journal 277 (2010) 738–748 ª 2009 The Authors Journal compilation ª 2009 FEBS
uct release. The acyl-enzyme intermediate may be
formed by the nucleophilic attack of Ser64 (Ser74 in
YbgJ) on the glutamine, which is assisted by the gen-
eral base Lys67 (Lys77 in YbgJ) and Tyr243 (Tyr253
in YbgJ) as a mediator of proton transfer. Deacylation
may occur via nucleophilic attack by water, which is
assisted by the general base Tyr191 (Tyr201 in YbgJ).
In this study, the lid part containing Tyr27 (Tyr37
in YbgJ) was present in both ligand-free and l-gluta-
mate-bound states. The positions of the lid were dis-
tinct from that in the acyl-enzyme intermediate state
(disordered in the ligand-free state in YbgJ). In the
absence of the ligand, the lid part is separated from
the active site, which leads to a more open active site.

Upon the binding of ligand in the acyl-enzyme inter-
mediate state, the lid part is located very close to the
active site, allowing interaction with the covalently
bound 5-oxo-l-norleucine. Before the release of l-glu-
tamate, the lid may be in the intermediate position
found in the Mglu TG structure. By contrast, b-lac-
tamase has no amino acid residue equivalent to the lid
part, and shows no significant conformational change
upon the binding of ligand [31]. The smaller size of
l-glutamine than the substrates for b-lactamase could
be one of the reasons for the conformational changes
of the lid part.
Experimental procedures
Enzyme assay
Glutaminase activity was assayed by determining the for-
mation of l-glutamate by l-glutamate dehydrogenase, as
previously described [12]. The activity was assayed with
30 mml-glutamine and 50 mm potassium phosphate buffer
(pH 7.5) at 30 °C for 10 min, unless stated otherwise. The
pH values of Tris and l-glutamate were adjusted to the pH
of the reaction mixture before addition to the mixture. One
Fig. 8. Tris in the TG structure. The F
o
-F
c
electron density in the map of the TG structure is contoured at the +2.4 sigma level. Putative
interactions (atom distance less than 3.9 A
˚
) are shown by dotted lines. The backbone atoms of all amino acid residues in the N (blue),
G (yellow) and T (cyan) structures are superimposed on those in the TG (magenta) structure.

A
B
Fig. 7. L-Glutamate in complex with Mglu
and asparaginase of Erwinia carotovora.
L-
Glutamate (cylinder) on the surfaces of the
structures of TG (A) and asparaginase of
E. carotovora (B) are shown. Putative inter-
actions (atom distance less than 3.9 A
˚
) are
shown by dotted lines.
K. Yoshimune et al. Crystal structure of salt-tolerant glutaminase
FEBS Journal 277 (2010) 738–748 ª 2009 The Authors Journal compilation ª 2009 FEBS 745
unit of glutaminase was defined as the amount of enzyme
that catalyzed the formation of 1 lmol of l-glutamate per
minute. The protein concentration of the cell solution was
estimated using the bicinchoninic acid (BCA) Protein Assay
Reagent kit (Pierce Biotechnology, Rockford, IL, USA)
with BSA as the standard.
Structural determination
The recombinant Mglu was expressed and purified as
described previously [32]. The crystals of Mglu (N) were
grown at 20 °C using the hanging-drop vapor-diffusion
method with 0.75 mL of reservoir solution [90 mm HEPES
buffer (pH 7.5), 180 mm sodium acetate and 27% (w ⁄ v)
poly(ethylene glycol) 4000] and 6 lL of protein solution
[10 mgÆmL
)1
of protein, 50 mm HEPES buffer (pH 7.5),

100 mm sodium acetate and 15% (w ⁄ v) poly(ethylene gly-
col) 4000]. The protein solutions that additionally contained
0.3 m Tris (T), 0.2 ml-glutamate (G), and 0.3 m Tris + 0.2
ml-glutamate (TG) were prepared at pH values of 7.0, 7.5
and 7.2, respectively. After treatment with paratone, the
crystals were flash-frozen using liquid nitrogen.
Diffraction data were collected at beamlines at the Pho-
ton Factory, Tsukuba, Japan. All data sets were collected
at 95 K. All images were indexed and integrated using the
program MOSFLM [33], and the data sets were phased
with molecular replacement using the program AMoRe [34]
in the CCP4 program package [35]. The F structure (PDB
3if5) [10] was used as an initial phasing model for the N
structure. The G and TG structures were solved using the
N structure as a search model, and the TG structure was
used in phasing of the T structure. The structures were
refined using the program CNS [36] with manual rebuilding
using the program O [37]. In CNS, the structures were
refined using a combination of rigid-body, conjugate gradi-
ent minimization refinement and B-factor refinement. The
two molecules in the asymmetric units were refined without
noncrystallographic symmetry restraints, which typically
increased the free R values by 0.1%. Although the R and
free R values of the four structures were in the range of the
values of the published structures, those of the N and G
structures were slightly poor [38,39]. For the N structure,
the extensive disordered regions in its structure might be
responsible for these values. For the G structure, the rela-
tively poor quality of the diffraction data sets was indicated
by the slightly worse statistics of the data sets compared

with the other structures, particularly R
merge
in the lowest-
resolution shell and the high B values from a Wilson plot
(Table 1). In spite of these unfavourable values, the main
chain and the side chain were clearly identified in the 2F
o
-
F
c
electron density map, and the final difference Fourier
maps did not contain any significant peaks. To confirm the
validity of the structures, each of the four structures was
fitted to each of the four diffraction data sets using rigid-
body refinement. All the four structures had minimum
R and free R values (each difference was in the range of
1–5%), when the structure was fitted to the cognate diffrac-
tion data sets. The programs PROCHECK [40] and
SFCHECK [41] in the CCP4 package were used for stereo-
chemistry analysis of all models and for calculating the
rmsd as well as the average error by the Luzzati plot.
The structures were superimposed on each other using
the program SUPERPOSE [42] in the CCP4 package.
All the figures illustrating these structures were prepared
using the program CCP4mg [43,44]. The contact areas were
calculated using AREAIMOL [45] in the CCP4 package
with the use of the 1.4 A
˚
probe. The coordinates have been
deposited in the PDB under accession codes 3ih8 (N), 3ih9

(T), 3iha (G) and 3ihb (TG).
Gene cloning and purification of Bacillus
glutaminase
The Bacillus glutaminase gene was amplified using the chro-
mosomal DNA of B. subtilis 168 (ATCC23857). The PCR
product was digested with NdeI and SalI and ligated into the
corresponding sites of pET21a (Novagen, San Diego, CA,
USA). The Bacillus glutaminase gene in the resulting plasmid
pYbgJ was confirmed to have the correct sequence. The E.
coli BL21 (DE3) strain was transformed with pYbgJ, cul-
tured at 37 °C for 12 h in Luria–Bertani (LB) medium sup-
plemented with ampicillin (50 lgÆmL
)1
) and 1 mm isopropyl
thio-b-d-galactoside, and then sonicated in 10 mm Tris ⁄ HCl
(pH 7.5). Bacillus glutaminase was purified using Q-sepha-
rose (Tosoh, Tokyo, Japan) and hydroxyapatite (Wako,
Osaka, Japan). The homogeneity of the final preparation of
Bacillus glutaminase was confirmed by SDS ⁄ PAGE.
Acknowledgements
Part of this study was supported by The Salt Science
Research Foundation (no. 0720) and the NIG Cooper-
ative Research Program (A50).
References
1 Nandakumar R, Yoshimune K, Wakayama M &
Moriguchi M (2003) Microbial glutaminase: biochemis-
try, molecular approaches and applications in the food
industry. J Mol Catal B: Enzym 23, 87–100.
2 Campos-Sandoval JA, Lopez de la Oliva AR, Lobo C,
Segura JA, Mates JM, Alonso FJ & Marquez J (2007)

Expression of functional human glutaminase in baculo-
virus system: affinity purification, kinetic and molecular
characterization. Int J Biochem Cell Biol 39, 765–773.
3 Wakayama M, Yamagata T, Kamekura A, Boontim N,
Yano S, Tachiki T, Yoshimune K & Moriguchi M
(2005) Characterization of salt-tolerant glutaminase
from Stenotrophomonas maltophila NYW-81 and its
Crystal structure of salt-tolerant glutaminase K. Yoshimune et al.
746 FEBS Journal 277 (2010) 738–748 ª 2009 The Authors Journal compilation ª 2009 FEBS
application in Japanese soy sauce fermentation. J Ind
Microbiol Biotechnol 32, 383–390.
4 Masuo N, Ito K, Yoshimune K, Hoshino M, Matsushi-
ma K, Koyama Y & Moriguchi M (2004) Molecular
cloning, overexpression, and purification of Micrococcus
luteus K-3-type glutaminase from Aspergillus oryzae
RIB40. Protein Expr Purif 38, 272–278.
5 Brown G, Singer A, Proudfoot M, Skarina T, Kim Y,
Chang C, Dementieva I, Kuznetsova E, Gonzalez CF,
Joachimiak A et al. (2008) Functional and structural
characterization of four glutaminases from Escherichia
coli and Bacillus subtilis. Biochemistry 47, 5724–5735.
6 Nandakumar R, Wakayama M, Nagano Y, Kawamura
T, Sakai K & Moriguchi M (1999) Overexpression of
salt-tolerant glutaminase from Micrococcus luteus K-3
in Escherichia coli and its purification. Protein Expr
Purif 15, 155–161.
7 Masuo N, Yoshimune K, Ito K, Matsushima K,
Koyama Y & Moriguchi M (2005) Micrococcus luteus
K-3-type glutaminase from Aspergillus oryzae RIB40 is
salt-tolerant. J Biosci Bioeng 100, 576–578.

8 Hartman SC & Stochai EM (1973) Glutaminase A of
Escherichia coli. Subunit structure and cooperative
behavior. J Biol Chem 248, 8511–8517.
9 Nimmo GA & Tipton KF (1981) Kinetic comparisons
between soluble and membrane-bound glutaminase
preparations from pig brain. Eur J Biochem 117, 57–64.
10 Yoshimune K, Shirakihara Y, Shiratori A, Wakayama
M, Chantawannakul P & Moriguchi M (2006) Crystal
structure of a major fragment of the salt-tolerant gluta-
minase from Micrococcus luteus K-3. Biochem Biophys
Res Commun 346, 1118–1124.
11 Madern D, Ebel C & Zaccai G (2000) Halophilic adap-
tation of enzymes. Extremophiles 4, 91–98.
12 Yoshimune K, Yamashita R, Masuo N, Wakayama M &
Moriguchi M (2004) Digestion by serine proteases
enhances salt tolerance of glutaminase in the marine
bacterium Micrococcus luteus K-3. Extremophiles 8, 441–
446.
13 Fukuchi S, Yoshimune K, Wakayama M, Moriguchi M
& Nishikawa K (2003) Unique amino acid composition of
proteins in halophilic bacteria. J Mol Biol 327, 347–357.
14 Premkumar L, Greenblatt HM, Bageshwar UK,
Savchenko T, Gokhman I, Sussman JL & Zamir A
(2005) Three-dimensional structure of a halotolerant
algal carbonic anhydrase predicts halotolerance of a
mammalian homolog. Proc Natl Acad Sci U S A 102,
7493–7498.
15 Holm L & Sander C (1996) Mapping the protein uni-
verse. Science 273, 595–603.
16 Masuda S, Katsuhiko KS, Wang S, Anders Olson C,

Donigian J, Leon F, Darst SA & Campbell EA (2004)
Crystal structures of the ADP and ATP bound forms of
the Bacillus anti-sigma factor SpoIIAB in complex with
the anti-anti-sigma SpoIIAA. J Mol Biol 340, 941–956.
17 Mevarech M, Frolow F & Gloss LM (2000) Halophilic
enzymes: proteins with a grain of salt. Biophys Chem
86, 155–164.
18 Kravchenko OV, Kislitsin YA, Popov AN, Nikonov SV
& Kuranova IP (2008) Three-dimensional structures of
l-asparaginase from Erwinia carotovora complexed with
aspartate and glutamate. Acta Crystallogr D Biol
Crystallogr 64, 248–256.
19 Aghaiypour K, Wlodawer A & Lubkowski J (2001)
Structural basis for the activity and substrate specificity
of Erwinia chrysanthemi l-asparaginase. Biochemistry
40, 5655–5664.
20 Derst C, Henseling J & Rohm KH (2000) Engineering the
substrate specificity of Escherichia coli asparaginase. II.
Selective reduction of glutaminase activity by amino acid
replacements at position 248. Protein Sci 9, 2009–2017.
21 Roberts J (1976) Purification and properties of a highly
potent antitumor glutaminase-asparaginase from
Pseudomonas 7Z. J Biol Chem 251, 2119–2123.
22 Holcenberg JS & Teller DC (1976) Physical properties
of antitumor glutaminase-asparaginase from Pseudomo-
nas 7A. J Biol Chem 251, 5375–5380.
23 Howard JB & Carpenter FH (1972) l-Asparaginase
from Erwinia carotovora. Substrate specificity and enzy-
matic properties. J Biol Chem 247, 1020–1023.
24 Nillius D, Jaenicke E & Decker H (2008) Switch

between tyrosinase and catecholoxidase activity of scor-
pion hemocyanin by allosteric effectors. FEBS Lett 582,
749–754.
25 Stemer R, Bardehele K, Paul R & Decker H (1994)
Tris: an allosteric effector of tarantula haemocyanin.
FEBS Lett 339, 37–39.
26 Nakano M & Tauchi H (1986) Difference in activation
by Tris(hydroxymethyl)aminomethane of Ca, Mg-AT-
Pase activity between young and old rat skeletal mus-
cles. Mech Ageing Dev 36, 287–294.
27 Irvin RT, MacAlister TJ & Costerton JW (1981)
Tris(hydroxymethyl)aminomethane buffer modification
of Escherichia coli outer membrane permeability. J Bac-
teriol 145, 1397–1403.
28 Goldman SJ & Slakey LL (1981) The effect of tris(hy-
droxymethyl) aminomethane on membrane bound
5¢nucleotidase from swine aortic smooth muscle. Bio-
chim Biophys Acta 658, 169–173.
29 Voss JG (1967) Effects of organic cations on the gram-
negative cell wall and their bacterial activity with ethyl-
enediaminetetraacetate and surface active agents. J Gen
Microbiol 48, 391–400.
30 Jensen-Holm J (1970) Tris-(hydroxymethyl) aminometh-
ane as an activator of acetylcholinesterase. Acta Phar-
macol Toxicol 28, 55.
31 Doucet N & Pelletier JN (2007) Simulated annealing
exploration of an active-site tyrosine in TEM-1 beta-lac-
tamase suggests the existence of alternate conforma-
tions. Proteins 69, 340–348.
K. Yoshimune et al. Crystal structure of salt-tolerant glutaminase

FEBS Journal 277 (2010) 738–748 ª 2009 The Authors Journal compilation ª 2009 FEBS 747
32 Chantawannakul P, Yoshimune K, Shirakihara Y,
Shiratori A, Wakayama M & Moriguchi M (2003) Crys-
tallization and preliminary X-ray crystallographic stud-
ies of salt-tolerant glutaminase from Micrococcus luteus
K-3. Acta Crystallogr D Biol Crystallogr 59, 566–568.
33 Leslie AGW (1992) Recent changes to the MOSFLM
package for processing film and image plate data. Joint
CCP4 and ESRF-EESF-EACMB Newsletter on Protein
Crystallography, No.26 SERC Daresbury Laboratory,
Warrington, UK.
34 Navaza J (1994) AMoRe: an automated package for
molecular replacement. Acta Crystallogr A 50,
157–163.
35 Collaborative Computational Project Number 4
(1994) The CCP4 Suite: programs for protein crystal-
lography. Acta Crystallogr D Biol Crystallogr 50,
760–763.
36 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 Biol
Crystallogr 54, 905–921.
37 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.
38 Kleywegt GJ & Jones TA (2002) Homo crystallographi-
cus – Quo vadis? Structure 10, 465–472.

39 Urzhumtseva L, Afonine PV, Adams PD &
Urzhumtsev A (2009) Crystallographic model quality
at a glance. Acta Crystallogr D Biol Crystallogr 65,
297–300.
40 Laskowski RA, MacArthur MW, Moss DS & Thornton
JM (1993) PROCHECK: a program to check the
stereochemical quality of protein structures. J Appl
Crystallogr 26, 283–291.
41 Vaguine AA, Richelle J & Wodak SJ (1999)
SFCHECK: a unified set of procedure for evaluating
the quality of macromolecular structure-factor data and
their agreement with atomic model. Acta Crystallogr D
Biol Crystallogr 55, 191–205.
42 Krissinel E & Henrick K (2004) Secondary-structure
matching (SSM), a new tool for fast protein structure
alignment in three dimensions. Acta Crystallogr D Biol
Crystallogr 60, 2256–2268.
43 Potterton L, McNicholas S, Krissinel E, Gruber J,
Cowtan K, Emsley P, Murshudov GN, Cohen S,
Perrakis A & Noble M (2004) Developments in the
CCP4 molecular-graphics project. Acta Crystallogr D
Biol Crystallogr 60, 2288–2294.
44 Potterton E, McNicholas S, Krissinel E, Cowtan K &
Noble M (2002) The CCP4 molecular-graphics pro-
ject. Acta Crystallogr D Biol Crystallogr 58, 1955–
1957.
45 Lee B & Richards FM (1971) The interpretation of
protein structures: estimation of static accessibility.
J Mol Biol 55, 379–400.
Crystal structure of salt-tolerant glutaminase K. Yoshimune et al.

748 FEBS Journal 277 (2010) 738–748 ª 2009 The Authors Journal compilation ª 2009 FEBS