Crystal structure and functional analysis of the glutaminyl cyclase from Xanthomonas campestris

Article abstract of DOI:10.1016/j.jmb.2010.06.012

Glutaminyl cyclases (QCs) (EC 2.3.2.5) catalyze the formation of pyroglutamate (pGlu) at the N-terminus of many proteins and peptides, a critical step for the maturation of these bioactive molecules. Proteins having QC activity have been identified in animals and plants, but not in bacteria. Here, we report the first bacterial QC from the plant pathogen Xanthomonas campestris (Xc). The crystal structure of the enzyme was solved and refined to 1.44-A resolution. The structure shows a β-propeller and exhibits a scaffold similar to that of papaya QC (pQC), but with some sequence deletions and conformational changes. In contrast to the pQC structure, the active site of XcQC has a wider substrate-binding pocket, but its accessibility is modulated by a protruding loop acting as a flap. Enzyme activity analyses showed that the wild-type XcQC possesses only 3% QC activity compared to that of pQC. Superposition of those two structures revealed that an active-site glutamine residue in pQC is substituted by a glutamate (Glu⁴⁵) in XcQC, although position 45 is a glutamine in most bacterial QC sequences. The E45Q mutation increased the QC activity by an order of magnitude, but the mutation E45A led to a drop in the enzyme activity, indicating the critical catalytic role of this residue. Further mutagenesis studies support the catalytic role of Glu⁸⁹ as proposed previously and confirm the importance of several conserved amino acids around the substrate-binding pocket. XcQC was shown to be weakly resistant to guanidine hydrochloride, extreme pH, and heat denaturations, in contrast to the extremely high stability of pQC, despite their similar scaffold. On the basis of structure comparison, the low stability of XcQC may be attributed to the absence of both a disulfide linkage and some hydrogen bonds in the closure of β-propeller structure. These results significantly improve our understanding of the catalytic mechanism and extreme stability of type I QCs, which will be useful in further applications of QC enzymes.

Full text of DOI:10.1016/j.jmb.2010.06.012

doi:10.1016/j.jmb.2010.06.012
J. Mol. Biol. (2010) 401, 374–388
Available online at www.sciencedirect.com
Crystal Structure and Functional Analysis of the
Glutaminyl Cyclase from Xanthomonas campestris
1
,2
2
2,3
2
Wei-Lin Huang , Yu-Ruei Wang , Tzu-Ping Ko , Cho-Yun Chia ,
2
,3
1,2,3
Kai-Fa Huang ⁎ and Andrew H.-J. Wang
⁎
1Graduate Institute of
Biochemical Science, College of
Life Science, National Taiwan
University, Taipei 106, Taiwan
2Institute of Biological
Chemistry, Academia Sinica,
Taipei 115, Taiwan
3Core Facility for Protein
Production and X-ray
Glutaminyl cyclases (QCs) (EC 2.3.2.5) catalyze the formation of pyroglu-
tamate (pGlu) at the N-terminus of many proteins and peptides, a critical
step for the maturation of these bioactive molecules. Proteins having QC
activity have been identified in animals and plants, but not in bacteria. Here,
we report the first bacterial QC from the plant pathogen Xanthomonas
campestris (Xc). The crystal structure of the enzyme was solved and refined
to 1.44-Å resolution. The structure shows a five-bladed β-propeller and
exhibits a scaffold similar to that of papaya QC (pQC), but with some
sequence deletions and conformational changes. In contrast to the pQC
structure, the active site of XcQC has a wider substrate-binding pocket, but
its accessibility is modulated by a protruding loop acting as a flap. Enzyme
activity analyses showed that the wild-type XcQC possesses only 3% QC
activity compared to that of pQC. Superposition of those two structures
revealed that an active-site glutamine residue in pQC is substituted by a
Structural Analysis, Academia
Sinica, Taipei 115, Taiwan
Received 26 March 2010;
received in revised form
45
glutamate (Glu ) in XcQC, although position 45 is a glutamine in most
3
1 May 2010;
bacterial QC sequences. The E45Q mutation increased the QC activity by an
order of magnitude, but the mutation E45A led to a drop in the enzyme
activity, indicating the critical catalytic role of this residue. Further
accepted 5 June 2010
Available online
1
5 June 2010
89
mutagenesis studies support the catalytic role of Glu as proposed
previously and confirm the importance of several conserved amino acids
around the substrate-binding pocket. XcQC was shown to be weakly
resistant to guanidine hydrochloride, extreme pH, and heat denaturations,
in contrast to the extremely high stability of pQC, despite their similar
scaffold. On the basis of structure comparison, the low stability of XcQC
may be attributed to the absence of both a disulfide linkage and some
hydrogen bonds in the closure of β-propeller structure. These results
significantly improve our understanding of the catalytic mechanism and
extreme stability of type I QCs, which will be useful in further applications
of QC enzymes.
©
2010 Elsevier Ltd. All rights reserved.
Keywords: glutaminyl cyclase; Xanthomonas campestris; pyroglutamate;
β-propeller
Edited by G. Schulz
Introduction
Glutaminyl cyclases (QCs) (EC 2.3.2.5) catalyze
the intramolecular cyclization of the protein N-
*
Corresponding authors. E-mail addresses:
terminal glutaminyl or glutamyl residue to form
pyroglutamate (5-oxoproline; pGlu) with the con-
comitant release of an ammonia or a water molecule,
respectively. Two types of QCs have been defined
thus far. Type I QCs were found in plants and in
huangkf@gate.sinica.edu.tw; ahjwang@gate.sinica.edu.tw.
Present address: Y.-R. Wang, Graduate Program in
Biomolecular Structure and Design, University of
Washington, Seattle, WA 98195, USA.
Abbreviations used: QC, glutaminyl cyclase; pGlu,
pyroglutamate; hQC, human QC; Gln-βNA, L-glutaminyl
1
several pathogenic bacteria and human parasites,
while type II QCs were mainly identified in the
neuroendocrine tissues of mammals. Up to now,
2
–7
2
-naphthylamide.
0022-2836/$ - see front matter © 2010 Elsevier Ltd. All rights reserved.

Crystal Structure and Functional Analysis of XcQC
375
human QC (hQC) is the most extensively studied
type II QC, which is known to be important in the
maturation of numerous neuropeptides and cyto-
Results
3
,4,8,9
Protein expression and purification
kines in their secretory pathways.
A number of
studies reported that hQC is involved in many
human pathological processes, such as amyloidotic
diseases, rheumatoid arthritis, osteoporosis, and
By searching the National Center for Biotechnol-
ogy Information (NCBI) protein database† for pQC-
homologous proteins, we found a number of
bacterial proteins whose sequences share 30–40%
identities to the sequence of pQC but their biological
activities remain unclear (see Fig. 1). We chose the
homologue (37.5% identity) from the plant pathogen
X. campestris as a target to study the bacterial type I
1
0–18
malignant melanoma.
In contrast, the physio-
logical role of type I QCs in plants, bacteria, and
parasites still remains to be elucidated.
Papaya QC (pQC) is the best-known type I QC.
This enzyme was first discovered in the latex of the
1
9
tropical plant Carica papaya. It was believed that
pQC belongs to the wound-induced protein family,
which also includes the plant defensive enzymes in
1
267
QC. The DNA coding for Met -Gln
of the
putative X. campestris QC (XcQC) was amplified
by PCR, and the recombinant protein was over-
expressed in E. coli cells and then purified to near
homogeneity (see Supplementary Figs. 1a and b).
The typical production yield was 3–5 mg per liter
culture. N-terminal sequence analysis of the purified
2
0
the papaya latex. Previous biochemical studies
demonstrated that pQC is a monomeric glycoprotein
of 266 amino acids and is highly resistant to
proteolytic (e.g., bovine trypsin and chymotrypsin
and papaya cysteine proteinases), chemical (e.g., 7 M
guanidium hydrochloride), and thermal (up to
2
2
protein revealed that Arg is the first residue,
1
1
,21,22
indicating that the N-terminal 21 residues (Met -
9
5 °C) denaturations.
The enzyme exhibited
2
1
Ala ) encoded by the expression construct were
cleaved out during expression in the E. coli cells. CD
spectra analysis showed a typical curve of high β-
sheet content for the recombinant XcQC (Supple-
lementary Fig. 1d), confirming that it is correctly
folded. QC activity analysis showed that XcQC has
only ∼13% activity compared to that of the hQC
activity (Tables 1 and 2). On the basis of the results
catalytic activity over a broad pH range, that is,
pH 3.5–11.0. In contrast, hQC is quite susceptible to
chemical and thermal denaturation, although it is
also a monomeric glycoprotein of 329 amino acids.
Schilling et al. have shown that hQC was significant-
2
3
2
4
2
3
ly unstable above pH 8.5 and below pH 6.0. X-ray
crystallographic analyses revealed that pQC is a
hatbox-shaped molecule, consisting of a five-bladed
3
0
1
,25
published by Schilling et al., which showed that
hQC possesses ∼22% QC activity relative to that of
pQC (Table 2), the activity of XcQC is estimated to
be only ∼3% of the pQC activity. In addition, as
shown in Table 2, weak substrate inhibitions were
observed during activity assays of the wild-type and
mutant XcQCs, probably owing to the nonproduc-
tive binding of the substrate Gln-βNA (L-glutaminyl
β-propeller traversed by a central channel,
whereas hQC adopts an α/β topology and
shares a conserved scaffold with the bacterial
2
6
aminopeptidases. Notably, recent structural and
mutagenesis studies have uncovered an unusual and
catalytically critical hydrogen-bond network neigh-
boring the zinc-binding center of the hQC active
2
7
site. This finding allowed us to propose an essential
proton-transfer process during the catalysis of type II
QCs. However, the catalytic mechanism of type I
QCs still remains to be further characterized.
2-naphthylamide) to XcQC.
Structure determination
Up to now, the production of sufficient quantities
of QCs is still a challenge. Previously, Huang et al.
reported a relatively high-yield (∼10 mg per liter of
Because we failed to obtain suitable crystals using
the wild-type XcQC despite extensive crystallization
attempts, we have tried to use some XcQC mutants
to screen for crystals. CD spectra analysis indicated
that these mutants have no observable conforma-
tional change when compared to the spectra of wild-
type enzyme. Finally, we used crystals of the mutant
E89A to perform the data collection and structure
determination experiments. The crystal structure of
the enzyme was solved by the molecular replace-
ment method and refined to 1.44-Å resolution with
R-factor/Rfree values being on an acceptable scale
2
8
culture) production of hQC in Escherichia coli cells.
However, considering the industrial production of
bioactive peptides and proteins, which require the
pGlu residue at their N-terminus, the high stability
of pQC makes the enzyme more attractive for
industrial uses than hQC. Currently, pQC protein
samples are obtained primarily by purification of
2
2
the latex of C. papaya, because it was reported that
the expression of recombinant pQC in E. coli cells is
very difficult, and the expression in insect cells gave
2
9
(
see Table 3 and Supplementary Fig. 2). The refined
very poor yields.
24
model of XcQC comprises 236 residues (Pro -
Here, we describe the efficient production in E. coli
cells of a type I QC from the plant pathogenic
bacterium Xanthomonas campestris. We also report
the 1.44-Å-resolution crystal structure of this bacte-
rial QC. On the basis of the successful expression
and crystal structure determination of this enzyme,
we further investigate the catalytic mechanism and
extreme stability of type I QCs.
2
59
His ), with 10 residues located at the N- and C-
terminal regions being excluded from the model due
to their high flexibility. Ramachandran plot analysis
showed that 99% of the residues were in the core
† http://blast.ncbi.nlm.nih.gov/

376
Crystal Structure and Functional Analysis of XcQC
Fig. 1. Sequence alignment of several representative bacterial QCs and pQC. The secondary structural elements of β-
strands and 310-helix according to our refined structure are also illustrated. The residues that are identical in more than 7
out of the 11 sequences are shaded and the completely conserved residues are further painted in cyan. Residue numbering
refers to that of XcQC. The five blades in the XcQC structure are indicated above the alignment. The stars in magenta mark
the residues in the active site of XcQC. The blue diamonds mark the calcium ion-coordinated residues. The double circles
in red mark the additional cysteine residues in papaya and other plant QCs. The green stars mark the hydrophobic
residues that contribute to the contacts between the five blades in the structures of XcQC and pQC. The sequence
identities listed were calculated by the software BioEdit. The GenBank accession numbers of these sequences are as
follows: NP_637572 (X. campestris), ABJ88842 (Solibacter usitatus Ellin6076), AAK32972 (Caulobacter crescentus CB15),
ZP_01305395 (Sphingomonas sp. SKA58 ), YP_002590309 (Brevundimonas sp. BAL3), YP_163612 (Zymomonas mobilis subsp.
mobilis ZM4), AAP58621 (uncultured Acidobacteria bacterium), ZP_04038429 (Meiothermus ruber DSM 1279), ZP_04772590
(
Allochromatium vinosum DSM 180), CAI78568 (uncultured candidate division OP8 bacterium), and AAC27745 (C. papaya).
9
5
31
region, with Arg being in the disallowed region.
This may be due to its location in the second position
of a β-hairpin, which is usually occupied by a
glycine residue. Glycines in a protein have less
restricted main-chain bond angles, and they usually
play an important structural role. Occasionally these

Crystal Structure and Functional Analysis of XcQC
377
Table 1. Enzyme activity assay
polypeptide chain progresses outward successively
through each strand from the central channel to
complete a four-stranded blade. The five blades are
then successively joined together in a counterclock-
wise direction, establishing the disk architecture.
Notably, in each blade, the loops between the first
two inner strands (i.e., β3–β4, β7–β8, β11–β12,
β16–β17, and β20–β21) and between the first two
outer strands (i.e., β5–β6, β9–β10, β13–β14, and
β18–β19) are positioned on the same surface of the
disk and surround the wide exit side of the central
channel. Conversely, the loops connecting the
second and third strands (i.e., β4–β5, β8–β9,
β12–β13, β17–β18, and β21–β22) and the loops
connecting two adjacent blades between the outer-
most strand and the innermost one of the next blade
Activity
Relative
μmol·min ·mg−1
−1
(
)
activity
hQC
3.1±0.2
(4.3±0.7)×10
4.7±0.1
(2.4±0.2)×10
(0.5±2.7)×10
(2.7±1.3)×10
(1.6±1.0)×10
(1.3±1.2)×10
725.40
100.00
1079.68
5.46
−
1
XcQC
XcQC E45Q
XcQC E89A
XcQC E45A
XcQC F43A
XcQC F87A
XcQC W103A
−
−
−
−
−
2
2
2
2
2
1.12
6.37
3.87
3.05
Results are shown as means±SEM from three independent
experiments using the synthetic substrate Gln-βNA.
are taken over by other amino acids with special
conformations.
(
i.e., β2–β3, β6–β7, β10–β11, β15–β16, and
β19–β20) are located on the other surface and
surround the narrow entry side of the channel.
Such a configuration makes the molecular environ-
ments at both entrances to the central channel very
different. An additional β-strand (β15) was identi-
fied at the interface of blades III and IV, which packs
against the extended part of β14 (Fig. 2a). Moreover,
a short 3 -helix structure was identified in the loop
Overall structure
The crystal structure of XcQC revealed a five-
bladed β-propeller, which is shaped like a thick disk
crossed by a central channel (see Fig. 2). This disk is
characterized by a diameter of ∼40 Å and a height of
∼
25 Å. The disk scaffold consists of five blades
1
0
(
Fig. 2A), each of which is composed of four
region connecting blades IV and V.
β-strands connected by loops and progressing in a
twisted and antiparallel manner. The five blades are
arranged radially around the central channel and
the first inner strand of each blade packs against the
other blades through their respective innermost
strands, thus contributing force for tethering these
five blades together (Fig. 2A). The five inner strands
The closure of the β-propeller scaffold is achieved
by β1 and β2, because they act as a double
molecular velcro contacting β19 of blade IV and
β22 of blade V, respectively, before the polypeptide
chain enters into blade I (Fig. 2a). The velcro is
strengthened by two main-chain hydrogen bonds
between β1 and β19 and nine between β2 and β22.
(
β3, β7, β11, β16, and β20) are parallel with one
another and to the central channel. Based on the
orientation of the inner strands, the openings to the
central channel at both disk surfaces can be
distinguished, one being the “entry,” where the
strands begin, and the other being the “exit,” where
they end (Fig. 2b). The entry is very narrow (∼4.5–
Table 3. Data collection and refinement statistics
Parameters
XcQC E89A
Data collection
Unit cell parameters (Å)
Space group
a=b=95.33, c=65.09
I4
5
Å in diameter) and the exit is more open (∼7.5 Å in
Resolution range (Å)
Total no. of reflections
Unique reflections
Redundancy
30.0–1.44 (1.49–1.44)
332,561 (22,915)
52,509 (5208)
6.3 (4.4)
diameter), thus characterizing a conical shape of the
central channel, which is due to the curved nature of
the inner β-strands.
The four β-strands in each blade are linked by an
up-and-down connectivity, except for the outermost
strands β1 of blade IV and β2 of blade V. The
Completeness (%)
Average I/σ(I)
99.9 (100.0)
55.1 (3.2)
3.4 (45.4)
R
merge (%)
Refinement
No. of reflections [N0 σ(F)], working/test
49,664/2669
0.174
Table 2. Kinetic parameters for the XcQC and E45Q
mutant
R
R
working
free (5.1% data)
0.209
r.m.s.d. bond distance (Å)
r.m.s.d. bond angle (°)
Average B value (Å )/no. of atoms
0.022
1.953
K
cat/K
m
cat (S−1
(mM ·S−1)
−1
a
K
m
(μM)
K
)
K
i
(mM)
2
XcQC
XcQC E45Q 39.0±12.6
121.2±34.2
1.4±0.3
5.8±0.9
5.4
12.6±4.5
1.54±0.78
All nonhydrogen atoms
Protein
24.15/2154
22.07/1882
20.79/1
166.1±60.1 5.06±3.27
b
hQCc
58.0
70.0
38.0
92.6
294.3
1352.6
N3.5
1.21
1.20
Calcium
hQC
20.6
Water
38.61/271
c
pQC
51.4
Ramachandran plot (excluding prolines and glycines)
Residues in most favored regions
Residues in additional allowed regions
Residues in generously allowed regions
Residues in disallowed regions
169 (84.1%)
All data were obtained using the substrate Gln-βNA.
30 (14.9%)
1 (0.5%)
1 (0.5%)
a
Inhibition constant for the detected substrate inhibition.
Data taken from Ref. 27, which were assayed using the same
b
fluorescence system as that for XcQC.
c
Data taken from Ref. 30, which were assayed using a different
Values in parentheses in data collection correspond to the highest-
resolution shell.
system.

378
Crystal Structure and Functional Analysis of XcQC
Fig. 2. Overall structure of XcQC. (a) A ribbon diagram of the XcQC structure. The calcium ion located at the center of
the β-propeller is shown as a gray ball. The roman numbers I–V indicate the five blades with color codes identical to those
illustrated in Fig. 1. (b) A view that is perpendicular to the central channel. The entry and exit sides are based on the
direction of the innermost β-strands. (c) Surface charge potential of the XcQC structure. The residues that surround the
catalytic center are indicated. The charge potentials were calculated by using the software PyMOL (DeLano Scientific,
http://www.pymol.org) with the values ranging from −62.262 to 26.262, colored from red to blue. Note that the active-
site pocket has a negatively charged surface.
9
1
131
A metal ion binding site
atoms from Ile and Leu . The ligand coordina-
tion of the calcium ion adopts an octahedral
geometry, with the three carbonyl oxygen atoms
and the water molecule lying in a plane, and the two
carboxylic oxygen atoms occupying the apical
positions both above and below the plane. The
distances between the calcium ion and the six
ligands range from 2.30 to 2.52 Å (see Fig. 3a),
close to the average calcium–oxygen distance
The experimental electron-density map clearly
showed a heavy atom (Fig. 3a), halfway in the
central channel (see Fig. 2a). Atomic absorption
analysis of the recombinant XcQC indicated that this
heavy atom is a calcium ion. This calcium ion is
coordinated by five protein oxygen atoms (Fig. 3a),
that is, three backbone carbonyl oxygen atoms from
4
7
176
230
32
Leu (β3, blade I), Leu (β16, blade IV), and Ile
β3, blade V) and two side-chain carboxylic oxygen
reported in octahedral coordination spheres and
consistent with the result from our atomic absorp-
tion analysis.
By superimposition of the calcium and zinc ion
binding sites in XcQC and pQC (see Fig. 3b), we
(
1
75
177
atoms from Glu (β16, blade IV) and Glu (β16,
blade IV). The sixth ligand is a water molecule,
which is bound by two backbone carbonyl oxygen

Crystal Structure and Functional Analysis of XcQC
379
Fig. 3. The calcium ion binding site in the XcQC structure. (a) The F −F simulated annealing omit density of the
o
c
calcium ion. The density map, colored in yellow and contoured at the 3.0σ level, is overlaid with the refined models of
calcium ion and its coordinated residues. The height of the calcium peak density is around 30σ. The coordination bonds
are drawn with distances indicated in angstroms. (b) Superimposition of the calcium/zinc ion binding sites in the
structures of XcQC and pQC. Color codes: XcQC bound with calcium ion (gray), pQC (hexagonal crystal form) bound
with calcium ion (yellow), pQC (orthorhombic crystal form) bound with zinc ion (green). Both figures were prepared in
stereo view.
found that the calcium/zinc-coordinated residues in
both QCs have almost identical orientations, except
conserved scaffold (Fig. 4). The root-mean-square
deviations, when comparing the XcQC structure
1
75
25
for Glu
in XcQC. The distinct orientation of
with the structures of the hexagonal
and
1
75
1
Glu is due to a slight movement in the backbone
of the XcQC residues 174–176 compared to the
corresponding region in pQC. As a consequence,
orthorhombic crystal forms of pQC, are 1.247 and
1.369 Å, respectively. Some notable differences
between XcQC and pQC structures are described
below. (i) The pQC structure contains an additional
disulfide bridge connecting the C-terminal α-helix
with the second outer strand (β16) of blade IV. This
disulfide bond was believed to contribute to the
polypeptide chain closure within the pQC structure.
However, this disulfide bond was not found in
XcQC structure and other bacterial QC sequences
(see Fig. 1). (ii) In the structure of XcQC, the
polypeptide chain after the short 3 -helix forms a
1
75
the side chain of Glu
slightly swings away from
the calcium ion compared to that of pQC, which
causes its two carboxylic oxygen atoms to be unable
to coordinate to the calcium ion simultaneously.
This backbone movement in XcQC leads to its two
calcium-coordinated glutamate residues to be in
monodentate oxygen coordination. Moreover,
mutations on these two glutamate residues, that is,
E175A and E177A, have resulted in failure to
produce the soluble enzymes (data not shown),
which strongly suggests a structural role of the
calcium ion in XcQC.
1
0
loop folding over the central channel and is close to
the catalytic site (see Fig. 4a). By contrast, the
corresponding loop in two known pQC structures
flips outward from the central channel. The differ-
ence in loop conformations leads to the quite
different active-site structures between these two
QCs (compare panels c and d in Fig. 4). The
protruding loop in XcQC generates two main-chain
Comparison with the pQC structure
Superimposition of XcQC structure with the
two published pQC structures revealed a highly

380
Crystal Structure and Functional Analysis of XcQC
2
26
246
hydrogen bonds to the residues Val
and Trp
cyclization reaction. The bottom of the active site
in XcQC is restricted by the three tryptophan indole
rings (Fig. 6), with diameters of 12–14 Å. Notably, as
shown in Fig. 4b, the orientations of the side chains
(
see Fig. 5), which were not observed in the pQC
structures. (iii) The closure in the pQC structure is
achieved by the first N-terminal strand (β1), in
addition to its C-terminal disulfide bridge. This β1,
acting as a double velcro, contacts the outermost
strand (β17) of blade IV via 3 or 4 main-chain
hydrogen bonds and attaches itself to the last C-
terminal strand (β20) in blade V via 10 or 11 main-
chain hydrogen bonds (Table 4). By contrast, the
closure in the XcQC structure is achieved by β1 and
β2, which interact with blades IV and V, respec-
tively, through 2 and 9 hydrogen bonds. (iv) As
mentioned above, the XcQC structure contains a
calcium ion held in position by six oxygen atoms.
A similar calcium ion binding site was identified in
the orthorhombic-form structure of pQC, but the
calcium ion is ligated by seven oxygen atoms (see
Fig. 3B). By contrast, in the hexagonal structure of
pQC, a zinc ion occupied the calcium ion binding
site (Fig. 3B). Because the zinc-bound form of pQC
was extracted from the crude papaya latex and the
calcium-bound one was produced in insect cell
expression systems, it is possible that the metal ion
binding site of type I QCs is naturally occupied by a
zinc ion but prefers to bind with a calcium ion if the
enzymes are obtained as a recombinant protein. (v)
Two N-glycosylation sites have been identified in
the pQC sequence (see Fig. 1). They are believed to
be critical in the maintenance of the pQC structure
because no soluble pQC could be obtained with the
E. coli expression system and no expression was
detected with the insect cell system when the
4
3
87
188
of Phe , Phe , and Trp
(numbering refers to
XcQC) are significantly different between XcQC and
pQC, resulting in the different dimensions of the
hydrophobic pocket. This finding suggests that the
substrate specificities of the two enzymes are
different.
As reported for the structures of pQCs, four
residues in the active site are involved in the binding
of a Tris molecule or the N-terminus of a symmetry-
related molecule, suggesting that these residues are
important in the substrate binding and catalytic
process. By superimposition, the four corresponding
4
5
89
residues in the XcQC structure are Glu , Glu ,
1
74
244
Asn , and Lys
(Fig. 4b). Based on sequence
4
5
alignment (Fig. 1), Glu of XcQC is substituted by a
glutamine residue in pQC and in most bacterial
QCs. Because the recombinant XcQC exhibited a
relatively low QC activity compared to that of pQC,
we have determined the activity of the XcQC E45Q
mutant. Surprisingly, this mutant showed an ∼10-
fold increase in QC activity compared to that of the
wild-type XcQC (Tables 1 and 2). However, the
activity of the E45A mutant was decreased by 2
orders of magnitude (Table 1), indicating the
4
5
45
importance of Glu or Gln in the catalytic process.
On the other hand, as proposed for the pQC
8
9
structure, Glu may act as a general base and acid
to assist the transfer of a proton from the α-amino
group of substrate N-terminus to the leaving amino
(or hydroxyl) group. The mutant E89A of XcQC was
shown to exhibit ∼5% activity compared to that of
the wild-type enzyme (Table 1), totally agreeing
with the hypothesis for pQC structures.
2
9
glycosylation sites were mutated to alanine. The
published pQC structures had identified clear
electron density protruding from the glycosylation
sites (Fig. 4A). By contrast, bacterial proteins are
usually nonglycosylated, including the bacterial
2
5
3
3
QCs (Fig. 1).
Tests for resistance to chemical, acid, and
thermal denaturation
The active-site structure
Because pQC has been demonstrated to possess
high stability in extreme conditions, and XcQC and
pQC share a conserved scaffold, we further ana-
lyzed the resistance of XcQC toward the chemical,
acid, and thermal denaturation. After a 24-h incuba-
tion in 1 M guanidine hydrochloride, we found that
the recombinant XcQC became aggregated. Further
CD spectral analysis showed that the scaffold of the
enzyme was disrupted in 2 M guanidine hydrochlo-
ride (data not shown), sharply contrasting with the
stability of pQC, which was stable even in 4 M
On the basis of the proposed active-site structure
of pQC, the XcQC active site is located at the entry
side of the central channel (see Fig. 2b). This putative
active site is quite shallow and surface-located,
possessing a negatively charged surface potential
(
Fig. 2c). The active site is lined by the nine residues
43 45 87 89 103 129 174
Phe , Glu , Phe , Glu , Trp , Trp , Asn
,
1
88
244
Trp , and Lys , which are highly conserved in
bacterial QCs and pQCs (see Figs. 1 and 4b), except
4
5
for Glu . The three tryptophan and the two
phenylalanine residues locate at the edge of the
active site and form a hydrophobic pocket (see Figs.
2
2
guanidine hydrochloride. Moreover, the enzyme
activity of XcQC at the pH range 3.5–9 was analyzed
and showed that this enzyme possessed relatively
high activities between pH 5 and 7.5, but exhibited
significantly decreased activities above pH 8 and
below pH 5 (Fig. 7). Such a result is also different
from that of pQC, which was found to possess high
2
c and 6), reminiscent of the active-site structure of
hQC, which has also a similar pocket formed by two
2
6
tryptophan and two phenylalanine residues. This
pocket in hQC was shown to accommodate a
portion of substrate or inhibitor upon its binding
to the enzyme. As shown in Table 1, mutations on
3
0
activities between pH 4 and 10. In addition, the
activity of XcQC at temperatures of 20–75 °C was
analyzed, and the results showed that this enzyme
was only stable below 40 °C and exhibited
4
3
87
103
Phe , Phe , and Trp
of XcQC decreased
the enzyme activity significantly, implicating the
importance of the hydrophobic pocket in the

Crystal Structure and Functional Analysis of XcQC
381
Fig. 4 (legend on next page)

382
Crystal Structure and Functional Analysis of XcQC
2
15
225
Fig. 5. The protruding loop that may modulate the accessibility of the XcQC active site. The loop Asp –Asp (cyan)
2
26
246
folds over the active-site pocket (yellow) and forms two main-chain hydrogen bonds to the residues Val and Trp
magenta). In addition, six hydrogen bonds within the loop stabilize the loop conformation. The hydrogen bond distances
are indicated in angstroms. The figure was prepared in stereo view.
(
4
3
87
remarkable instability above 50 °C (Fig. 7), also
different from that of pQC, which was stable even at
and two phenylalanine (i.e., Phe and Phe )
residues located at the edge of the active site.
Notably, this pocket is significantly wider than
that in pQC due to the outward orientations of the
2
2
9
0 °C.
4
3
87
188
side chains of Phe , Phe , and Trp (see Fig. 4b).
This finding suggests that XcQC favors substrates
with a more bulky residue at their second position
(see Fig. 8a for substrate-binding model). Mutations
on three of the residues surrounding the hydropho-
bic pocket reduced the enzyme activity significantly
(see Table 1), implicating the importance of the
hydrophobic pocket in the cyclization reaction.
Moreover, by comparing the entire active-site
pockets of XcQC and pQC (Fig. 4c and d), it is
obvious that these two active sites have different
solvent accessibilities, primarily owing to the dis-
tinct conformation of the protruding loop in XcQC
structure. Two structural observations support the
view that the protruding loop is not due to
crystallographic artifact. (i) The protruding loop in
the XcQC structure has only very limited contacts
with its symmetry-related XcQC molecules (see
Supplementary Fig. 3). (ii) In the two published
pQC structures, the regions that correspond to the
XcQC protruding loop have nearly identical
Discussion
Enzymes exhibiting QC activity have been iden-
tified in animals (e.g., human, bovine, mouse, and
Drosophila) and plants (e.g., papaya, potato, and a
7
,19,34–37
small flowering plant),
but no report has
described a protein with QC activity from bacteria
thus far. Here we present the cloning and expression
of the first bacterial QC from the plant pathogen X.
campestris. We also describe the high-resolution
crystal structure of the enzyme and its relevant
structure–function relationships.
Previous enzyme kinetic studies had shown that
hQC and pQC are highly active toward substrates
with a large hydrophobic residue at their second
3
0
position. These data may reflect the fact that the
active-site structures of hQC and pQC contain a
1
,25,26
hydrophobic pocket.
A similar pocket was also
defined in the structure of XcQC, being lined by
1
03
129
188
three tryptophan (i.e., Trp , Trp , and Trp
)
Fig. 4. Structural comparison of XcQC and pQC. (a) Superimposition of the XcQC and pQC structures. XcQC is
colored in blue, the hexagonal crystal form of pQC (PDB code 2IWA) in red, and the orthorhombic crystal form of pQC
(
2FAW) in yellow. The loop modulating the active-site accessibility in XCQC structure is further colored in cyan. The gray
ball depicts the calcium ion in XcQC. The sugar molecules in both pQC structures are also shown. The figure was
prepared in stereo view. (b) Superimposition of the residues around the catalytic center in XcQC (gray), the hexagonal
4
3
87
form of pQC (green), and the orthorhombic form of pQC (magenta). Note the outward orientations of Phe , Phe , and
188
Trp in XcQC compared to pQCs. The figure was prepared in stereo view. (c and d) Surface representations of the XcQC
and pQC (hexagonal crystal form) structures. The active-site pocket in both structures are colored red. The protruding
loop that folds over the active-site pocket in XcQC structure is in cyan, and its corresponding region in the pQC structure
is painted with the same color. The green surfaces indicate the regions close to the C-terminus in both structures.

Crystal Structure and Functional Analysis of XcQC
383
Table 4. Numbers of backbone hydrogen bonds in the
structure of XcQC and pQC
different crystal form, which showed that the N-
terminus of pQC specifically inserts into the C-
terminal entrance of the central channel of a
symmetry-related molecule, likely mimicking the
β-Sheet Helix
Velcro structure
XcQC
101
93
1
2 (β1–β19)+9 (β2–β22)
3 (β1–β17)+10 (β1–β20)
25
N-terminus of a bound substrate. These findings
pQC (hexagonal
crystal form)
pQC (orthorhombic
crystal form)
13
allow us to map the residues that participate in the
substrate binding and catalysis. On the basis of these
structures, a glutamate residue in the active site was
proposed to act as a general base and acid to assist
the intramolecular cyclization reaction. The
corresponding glutamate residue in the present
95
10
4 (β1–β17)+11 (β1–β20)
8
9
conformations (see Fig. 4a), that is, flipped outward
from the central channel, although the two pQC
crystal structures belong to different crystal forms. It
is notable that the two published pQC structures
have found small molecules or symmetry-related
pQC molecule bound in their active site, but no
bound materials could be identified in the active site
of XcQC.
In the study of hQC structure, a zinc ion found in
the active site was proved to be crucial to the
catalysis of hQC. In contrast, either a calcium or a
zinc ion was identified in the structures of pQC, but
slightly distant from the putative catalytic center,
suggesting that the metal ion does not participate in
the catalysis of pQC. In the present structure, this
metal ion binding site is occupied by a calcium ion.
Our mutagenesis studies on some of the calcium-
coordinated residues also supported the structural
role of the calcium ion in XcQC.
structure is Glu . Mutation of this glutamate to
alanine resulted in a significant drop in QC activity
(Table 1), totally consistent with the previous
hypothesis.
On the other hand, the oxyanion intermediate
generated during catalysis of type II QCs was
believed to be stabilized by an active-site zinc
2
3,27
ion.
In the pQC structures, three residues (Gln,
Asn, and Lys) were proposed to be candidates that
1
,25
play a role similar to that of the zinc ion.
Among
the three candidates, the glutamine residue is
4
5
substituted by a glutamate (Glu ) in XcQC, whose
QC activity was only ∼3% of that of pQC.
4
5
Surprisingly, mutating this Glu to glutamine
elevated the QC activity by ∼10-fold and increased
its K
value by ∼4-fold (see Tables 1 and 2),
cat
whereas mutating it to alanine resulted in an ∼90-
fold decrease of the QC activity. These data support
the likeliness that, in general, the oxyanion interme-
diate during catalysis of type I QCs is stabilized, in
part, by a glutamine residue (see Fig. 8b). When the
glutamine is substituted by a negatively charged
glutamate, partial destabilization of the oxyanion
intermediate (also negatively charged) may occur
during catalysis due to charge–charge repulsion,
The crystal structure of pQC published by
Wintjens et al. showed either a Tris molecule or an
acetate ion bound in the C-terminal entrance of the
1
central tunnel. By further inhibition analyses, they
1
concluded that Tris is a weak inhibitor of pQC.
Later, Guevara et al. reported a pQC structure in a
Fig. 6. The active-site structure of XcQC. The hydrogen bond network in the active site is shown. Note that the three
tryptophan and the two phenylalanine residues surround the putative catalytic center, forming a hydrophobic pocket.
8
9
89
Because the present structure comes from crystals of the E89A mutant, the position Ala should be Glu in the wild-type
XcQC.

384
Crystal Structure and Functional Analysis of XcQC
Fig. 7. Effects of different pH environments and temperatures on the QC activity of XcQC. Blue: QC activity was
measured in different buffers as follows: acetate (pH 3.5, 4.0, 4.5, and 5.0), Mes (pH 5.5, 6.0, and 6.5), Mops (pH 7.0 and
7
.5), and Tris (pH 8.0, 8.5, and 9.0). Results are shown as mean±SEM from three independent experiments. Red:
evaluation of QC activity for XcQC preheated at different temperatures for 10 min. Results are also shown as mean±SEM
from three assays.
accounting for a lower activity. This hypothesis is
reminiscent of the transition-state stabilization in the
oxyanion hole of papain. Ménard et al. proposed that
XcQC may be attributed to the absence of both a
disulfide linkage and some hydrogen bonds in the
closure structure.
1
9
Gln in papain plays a role in stabilizing the
transition-state oxyanion during catalysis of the
To date, the biological function and natural
substrates of QCs in bacteria are still elusive. Some
time ago, Cook and Russell proposed that QC might
exist in some bacteria strains, such as Streptococcus
bovis, and the enzyme activity in this bacterium
might be related to the increase in membrane
3
8
enzyme. Their experiments showed that muta-
tions of this glutamine residue to glutamate or
alanine caused the significant decreases in enzyme
3
9
activities, which are totally consistent with our
results described in the present study. On the other
hand, the greatly decreased activity of the E45A
4
0
potential and the formation of intracellular ATPs.
This hypothesis is still waiting for more experimen-
tal evidence. Notably, in addition to X. campestris,
QC-like sequences were also identified in other
pathogenic bacteria, such as Corynebacterium
4
5
mutant of XcQC also reflected the role of Glu or
4
5
Gln in substrate–intermediate binding.
Because the XcQC and pQC structures share a
conserved scaffold but display remarkably different
stabilities, the information from the XcQC structure
gave useful hints in the understanding of the
mechanism of the extreme stabilities of pQC. From
the previous studies of pQC structure, some
structural features were believed to contribute to
the stability of the enzyme: (i) the bound metal ion,
4
1
diphtheriae, which causes diphtheria in humans.
This raises a possibility that bacterial QCs act as a
virulence factor that affects the metabolisms of
bacteria-invaded animals and plants. In this regard,
natural substrates of bacterial QCs might exist in the
hosts rather than in the bacteria themselves.
In summary, we report the first bacterial QC from
the plant pathogen X. campestris and its high-
resolution crystal structure. We also describe the
results from further mutagenesis, enzyme kinetics,
and structural comparison studies. These data
significantly improve our understanding of the
mechanism of catalysis and extreme stabilities of
type I QCs, which will be useful in further
applications of QC enzymes.
(
(
ii) the double velcro closure at the last two blades,
iii) a large number of tightly packed hydrophobic
contacts between the blades, (iv) the extensive
hydrogen bond networks, and (v) a disulfide
linkage connecting the C-terminal domain and the
blade IV. In regard to the XcQC structure, the metal
ion and the double velcro closure have also been
identified, but several hydrogen bonds in the closure
structure are missing. The hydrophobic residues
that contribute to the interactions between blades
are also highly conserved between pQC and XcQC
Materials and Methods
(
see Fig. 1). In addition, the numbers of hydrogen
bonds that participate in regular secondary struc-
tures in both QCs are close to each other (see Table 4
and Supplementary Fig. 4). On the other hand, by
contrast, no disulfide linkage could be found in the
XcQC structure. In this regard, the low stability of
Protein expression and purification
The gene coding for XcQC was amplified from the
genomic DNA of the plant pathogen X. campestris pv.
campestris str. ATCC33913 by PCR using the primers 5′-

Crystal Structure and Functional Analysis of XcQC
385
Fig. 8. Proposed substrate binding and catalysis mechanism of XcQC. (a) Model of a substrate (Gln-Phe-Ala-Ala-Ala)
bound to the active site of XcQC. The model was generated on the basis of the pQC structure published by Guevara
et al.,27 which showed that the N-terminus of pQC specifically inserted into the active site of a symmetry-related pQC
molecule, presumably mimicking the N-terminus of a bound substrate. The substrate molecule is colored in pink and the
three important catalytic residues in yellow. The three tryptophan residues responsible for accommodating the phenyl
group of the bound substrate are shown in blue. The figure was prepared in stereo view. (b) Proposed reaction mechanism
of type I QCs. In general, but not for XcQC, there is a glutamine residue instead of a glutamate residue in position 45 and
this has been included in the figure.
GTCGAGAAGCTTCCATCGAAGGTCGTATGC-
C A C G T C T C G T C C C T G C C C T G C - 3 ′ a n d 5 ′ -
GTGGTGCTCGAGCTGCGCGTGCTTGCCGGCCGCGG-
CGTG-3′. The purified PCR product was inserted into the
pET-21a expression vector (Novagen) via HindIII and
XhoI cloning sites. This construct contains a hexahistidine
tag adjacent to the C-terminal end of the XcQC coding
region. The constructed expression vector was trans-
formed into E. coli BL21 (DE3) cells (Novagen). The
bacteria were grown in Terrific Broth containing ampi-
cillin (200 μg/mL) at 37 °C until the cell density reached an
OD600 of 0.8–0.9 . The cultures were induced with 0.25 mM
isopropyl β-D-1-thiogalactopyranoside (IPTG) for 16–20 h
at 20 °C. The cells were harvested by centrifugation at
pooled and dialyzed against the dialysis buffer (150 mM
sodium chloride and 10% glycerol in 50 mM phosphate,
pH 7.0) at 4° C overnight. A further purification step using
a second Ni-NTA column was then carried out. The
column was preequilibrated with the dialysis buffer to
which 10 mM imidazole was added, and the QCs were
eluted by a linear gradient of 10–250 mM imidazole. The
eluted protein samples were concentrated and desalted by
a desalting column (HiPrep™ 26/10) (Amersham Phar-
macia) equilibrated with the desalting buffer (10% glycerol
in 50 mM Tris–HCl, pH 7.0). The final purified XcQCs had
a purity of higher than 95% as judged by SDS-polyacryl-
amide gel electrophoresis (see Supplementary Fig. 1a).
For site-directed mutagenesis studies, the expression
vectors for mutant XcQCs were constructed on the basis of
the protocol of the QuikChange® Site-Directed Mutagen-
esis kit (Stratagene). The mutants were expressed and
purified in the same manner as that of the wild-type
enzyme.
6000 rpm for 30 min followed by freezing at −80 °C.
Frozen bacterial pellets were resuspended in the lysis
buffer (300 mM sodium chloride and 10 mM imidazole in
5
0 mM phosphate, pH 7.5), and then the cells were lysed
with a French press. The cell lysate was clarified by
centrifugation at 28,000 rpm for 1 h, and the supernatant
was loaded onto a Ni-NTA column (Amersham Pharma-
cia, Uppsala, Sweden) preequilibrated with the lysis
buffer. The column was washed with the washing buffer
Enzyme activity assay
(
300 mM sodium chloride, 20 mM imidazole in 50 mM
The activities of wild-type and mutant XcQCs were
evaluated at 25 °C using the fluorescent substrate Gln-
βNA. The 100 -μl reaction mixtures contained 300 μM
phosphate, pH 7), and then the bound QCs were eluted by
a linear gradient of 20–400 mM imidazole. The elutes were

386
Crystal Structure and Functional Analysis of XcQC
fluorogenic substrate, ∼0.2 U of human pyroglutamyl
aminopeptidase I (1 U is defined as the amount hydrolyz-
ing 1 μmol of pGlu-βNA per minute under the same
condition), and an appropriately diluted aliquot of wild-
type or mutant XcQCs in 50 mM Tris–HCl and 10%
glycerol (pH 7.0). The excitation and emission wavelengths
were set at 320 and 410 nm, respectively. The reactions
were initiated by the addition of QC. Enzymatic activity
was determined by the amount of released βNA, calculat-
ed using a standard curve for βNA under the same assay
condition. The measurements were made using an LS-55
fluorescent spectrometer (PerkinElmer, Waltham, MA).
Crystallization and X-ray data collection
XcQC was concentrated to 8 mg/mL with the centrifugal
filter Ultracel® 5 K (Millipore, Billerica, MA). The crystal-
lization screening was achieved with screening kits from
Hampton Research (Laguna Niguel, CA), Emerald BioSys-
tems (Bainbridge Island, WA), and Molecular Dimension
(
Apopka, FL). Because no suitable crystals could be found
for the wild-type XcQC after screening ∼1000 crystalliza-
tion conditions, we used the low-activity mutant E89A
instead. The purified E89A of XcQC in 50 mM Tris–HCl and
1
0% glycerol (pH 7.0) was mixed with an equal volume of
the reservoir solution (50 mM imidazole and 0.8 M sodium
citrate, pH 8.68) and then crystallized at 25 °C by the
hanging-drop vapor-diffusion method. Crystals with
dimensions reaching 0.2 mm×0.2 mm×0.3 mm appeared
in 2 weeks (see Supplementary Fig. 1c). A 1.95-Å X-ray
diffraction data set was collected at the beamline 13B1 of the
National Synchrotron Radiation Research Center (Hsinchu,
Taiwan). A second 1.44-Å data set was collected at the
beamline 5A of KEK Photon Factory (Tsukuba, Japan).
Before being mounted on the X-ray machine, crystals were
briefly soaked in reservoir solutions containing 10% (v/v)
glycerol as a cryoprotectant. All diffraction data were
Enzyme kinetic assay
The kinetic constants were determined at pH 7.0 at 25 °C
using the substrate Gln-βNA. Enzyme concentrations
were used on the basis of the known relative activities of
XcQC and its mutant. The reaction was initiated by adding
QC to the 100 -μl reaction mixture. The initial rate was
measured with less than 10% substrate depletion for the
first 2–10 min. Because substrate inhibition was observed,
the kinetic parameters K , V, and K were evaluated by
m i
fitting Eq. (1) to initial velocity data by nonlinear regression
using KaleidaGraph software:
43
processed and scaled with the HKL2000 package. The
data-collection statistics are listed in Table 3. The space
group of the crystals is I4, with typical unit cells of
a=b=95 Å and c=65 Å and a solvent content of 52.4%, in
which an asymmetric unit comprises one XcQC molecule.
ꢀ
ꢁ
2
v0 = Vmax½Sꢀ= Km + ½Sꢀ + ½Sꢀ = Ki
ð1Þ
where v is the initial velocity, V
the limiting rate, [S] the
0
max
substrate concentration, Km the Michaelis constant, and Ki
Structure determination and refinement
the inhibition constant. Correlation coefficients better than
0
.989 were obtained throughout all fittings. Substrate
concentrations used were in the range 0.025–1.5 mM. The
cat values were calculated from Vmax/[E] according to the
enzyme concentrations determined from UV absorption in
With the 1.95-Å X-ray data set, the crystal structure of
XcQC was solved by the molecular replacement method
using the published pQC structure [Protein Data Bank
K
80 nm (ɛ=51,350 M− cm ) in the presence of 6.0 M
1
−1
1
2
(PDB) code 2FAW] as the search model and using the
4
4
guanidine hydrochloride. Therefore, the Kcat values
reported in this study represent minimal estimates on the
assumption that the enzyme active-site concentration is
equivalent to its concentration as protein.
software CNS. Only one clear solution for the rotation
and translation functions was found, consistent with one
XcQC molecule estimated in the asymmetric unit. The
residues that differ between XcQC and pQC were replaced
4
5
using O, and manual adjustments were made using the
same software. The well-ordered water molecules and one
Evaluation of enzyme activities at various pH values
and temperatures
2+
46
Ca ion were located using XtalView. Subsequently,
iterative cycles of computational refinements were per-
formed with the 1.44-Å X-ray data set and the software
The enzyme activities at various pH values were
evaluated using a variety of buffer systems, including
47
packages CNS and Refmac5. Both the R-factor and Rfree
values were used to monitor the progress of the refine-
5
0 mM acetate (pH 3.5, 4.0, 4.5, and 5.0), 50 mM Mes
pH 5.5, 6.0, and 6.5), 50 mM Mops (pH 7.0 and 7.5), and
0 mM Tris (pH 8.0, 8.5, and 9.0). The buffers were added
ment. The R-factor/Rfree of the refinement reached 0.198/
(
0
.231 after the CNS refinement and 0.174/0.209 by
5
Refmac5. The parameters for ideal protein geometry of
with 10% glycerol and titrated to the desired pH values
using HCl or NaOH. All measurements were performed
by the same fluorescence assay method as described
above.
For activity measurements at varying temperatures, the
enzyme was preincubated at various temperatures for
48
Engh and Huber were used during the refinements, and
the stereochemical quality of the refined structure was
49
checked with PROCHECK. The molecular figures were
produced with PyMOL (DeLano Scientific)‡.
1
0 min and then cooled on ice. The enzyme activities were
Modeling of a substrate bound in the active site of
XcQC
evaluated by the same fluorescence method.
The initial substrate-binding model was generated on
the basis of the hexagonal-form crystal structure of pQC
Atomic absorption analysis
(
PDB code 2IWA), which showed that the active site of the
The metal ion (Cu, Zn, Ca, Fe, Mn, and Cd) content of
the recombinant XcQC was analyzed with an atomic
absorption spectrophotometer (Hitachi, Tokyo, Japan)
using A-type graphite cuvettes (Hitachi) as described
enzyme was bound with the N-terminus of a symmetry-
related pQC molecule. Furthermore, because previous
42
previously. The experiment was carried out in triplicate,
and the results showed that only the absorption for
calcium ion was statistically significant.
‡ http://www.pymol.org

Crystal Structure and Functional Analysis of XcQC
387
structural studies on pQC and the mutagenesis study
enzymes in rat hypothalamus and an immortalized
hypothalamic neuronal cell line. Neuroendocrinology,
62, 166–177.
8
9
described in this report have suggested that Glu of XcQC
plays a role as a general base and acid, we have put the
8
9
side chain of Glu in proximity to the N-terminal α-amino
group of the bound substrate. The resulting model was
then optimized by energy minimization with the software
CNS.44 The final model was verified with the software
6. Sykes, P. A., Watson, S. J., Temple, J. S. & Bateman,
R. C., Jr. (1999). Evidence for tissue-specific forms of
glutaminyl cyclase. FEBS Lett. 455, 159–161.
7. Fischer, W. H. & Spiess, J. (1987). Identification of a
mammalian glutaminyl cyclase converting glutaminyl
into pyroglutamyl peptides. Proc. Natl Acad. Sci. USA,
84, 3628–3632.
. Vale, W., Spiess, J., Rivier, C. & Rivier, J. (1981).
Characterization of a 41-residue ovine hypothalamic
peptide that stimulates secretion of corticotropin and
beta-endorphin. Science, 213, 1394–1397.
50
Profiles-3D, which showed a score of 93.4, with the
expected high and low scores of 109.9 and 49.5,
respectively.
8
PDB accession numbers
Coordinates and structure factors have been deposited
in the PDB with accession number 3MBR.
9
. Song, I., Chuang, C. Z. & Bateman, R. C., Jr. (1994).
Molecular cloning, sequence analysis and expression
of human pituitary glutaminyl cyclase. J. Mol.
Endocrinol. 13, 77–86.
1
1
0. Batliwalla, F. M., Baechler, E. C., Xiao, X., Li, W.,
Balasubramanian, S., Khalili, H. et al. (2005). Periph-
eral blood gene expression profiling in rheumatoid
arthritis. Genes Immun. 6, 388–397.
Acknowledgements
1. Cynis, H., Schilling, S., Bodnar, M., Hoffmann, T.,
Heiser, U., Saido, T. C. & Demuth, H. U. (2006).
Inhibition of glutaminyl cyclase alters pyroglutamate
formation in mammalian cells. Biochim. Biophys. Acta,
We thank Dr. Wen-Yih Jeng at the Institute of
Biological Chemistry, Academia Sinica (Taipei,
Taiwan) for assistance in X-ray data collection.
We are grateful to Prof. Shan-Ho Chou at the
Institute of Biochemistry of National Chung-Hsing
University (Taichung, Taiwan) for kindly provid-
ing the genomic DNA of X. campestris. We are also
grateful for the access to the synchrotron radiation
beamline 13B1 (National Synchrotron Radiation
Research Center, Taiwan) and beamline 5A (Pho-
ton Factory, Japan). This work was supported by
Academia Sinica and Core Facility for Protein
Production and X-Ray Structural Analysis of the
National Research Program for Genomic Medicine
1
764, 1618–1625.
12. Ezura, Y., Kajita, M., Ishida, R., Yoshida, S., Yoshida,
H., Suzuki, T. et al. (2004). Association of multiple
nucleotide variations in the pituitary glutaminyl
cyclase gene (QPCT) with low radial BMD in adult
women. J. Bone Miner. Res. 19, 1296–1301.
1
1
1
3. Gillis, J. S. (2006). Microarray evidence of glutaminyl
cyclase gene expression in melanoma: implications for
tumor antigen specific immunotherapy. J. Transl. Med.
4, 27.
4. Huang, Q. Y. & Kung, A. W. (2007). The association of
common polymorphisms in the QPCT gene with bone
mineral density in the Chinese population. J. Hum.
Genet. 52, 757–762.
(NSC97-3112-B-001-035-B4).
5. Muthusamy, V., Duraisamy, S., Bradbury, C. M.,
Hobbs, C., Curley, D. P., Nelson, B. & Bosenberg, M.
Supplementary Data
(
2006). Epigenetic silencing of novel tumor sup-
pressors in malignant melanoma. Cancer Res. 66,
1187–11193.
Supplementary data associated with this article
can be found, in the online version, at doi:10.1016/
j.jmb.2010.06.012
1
1
6. Schilling, S., Appl, T., Hoffmann, T., Cynis, H., Schulz,
K., Jagla, W. et al. (2008). Inhibition of glutaminyl
cyclase prevents pGlu-Abeta formation after intracor-
tical/hippocampal microinjection in vivo/in situ.
J. Neurochem. 106, 1225–1236.
7. Schilling, S., Hoffmann, T., Manhart, S., Hoffmann, M.
& Demuth, H. U. (2004). Glutaminyl cyclases unfold
glutamyl cyclase activity under mild acid conditions.
FEBS Lett. 563, 191–196.
References
1
1
. Wintjens, R., Belrhali, H., Clantin, B., Azarkan, M.,
Bompard, C., Baeyens-Volant, D. et al. (2006). Crystal
structure of papaya glutaminyl cyclase, an archetype
for plant and bacterial glutaminyl cyclases. J. Mol. Biol.
18. Schilling, S., Zeitschel, U., Hoffmann, T., Heiser, U.,
Francke, M., Kehlen, A. et al. (2008). Glutaminyl
cyclase inhibition attenuates pyroglutamate Abeta
and Alzheimer's disease-like pathology. Nat. Med. 14,
1106–1111.
19. Messer, M. (1963). Enzymatic cyclization of L-glutamine
and L-glutaminyl peptides. Nature, 197, 1299.
20. Azarkan, M., Wintjens, R., Looze, Y. & Baeyens-
Volant, D. (2004). Detection of three wound-induced
proteins in papaya latex. Phytochemistry, 65, 525–534.
21. Oberg, K. A., Ruysschaert, J. M., Azarkan, M.,
Smolders, N., Zerhouni, S., Wintjens, R. et al. (1998).
Papaya glutamine cyclase, a plant enzyme highly
resistant to proteolysis, adopts an all-beta conforma-
tion. Eur. J. Biochem. 258, 214–222.
357, 457–470.
2
3
. Bateman, R. C., Jr. (1989). A spectrophotometric assay
for glutaminyl-peptide cyclizing enzymes. J. Neurosci.
Methods, 30, 23–28.
. Pohl, T., Zimmer, M., Mugele, K. & Spiess, J. (1991).
Primary structure and functional expression of a
glutaminyl cyclase. Proc. Natl Acad. Sci. USA, 88,
10059–10063.
4
5
. Bockers, T. M., Kreutz, M. R. & Pohl, T. (1995).
Glutaminyl-cyclase expression in the bovine/porcine
hypothalamus and pituitary. J. Neuroendocrinol. 7,
4
45–453.
. Wetsel, W. C., Liposits, Z., Seidah, N. G. & Collins, S.
1995). Expression of candidate pro-GnRH processing
(

388
Crystal Structure and Functional Analysis of XcQC
2
2. Zerhouni, S., Amrani, A., Nijs, M., Smolders, N.,
Azarkan, M., Vincentelli, J. & Looze, Y. (1998).
Purification and characterization of papaya glutamine
cyclotransferase, a plant enzyme highly resistant to
chemical, acid and thermal denaturation. Biochim.
Biophys. Acta, 1387, 275–290.
characterization of glutaminyl cyclases from Drosoph-
ila: evidence for enzyme forms with different subcel-
lular, localization. Biochemistry, 46, 10921–10930.
37. Schilling, S., Stenzel, I., von Bohlen, A., Wermann, M.,
Schulz, K., Demuth, H. U. & Wasternack, C. (2007).
Isolation and characterization of the glutaminyl
cyclases from Solanum tuberosum and Arabidopsis
thaliana: implications for physiological functions.
Biol. Chem. 388, 145–153.
38. Menard, R., Carriere, J., Laflamme, P., Plouffe, C.,
Khouri, H. E., Vernet, T. et al. (1991). Contribution of
the glutamine 19 side chain to transition-state stabi-
lization in the oxyanion hole of papain. Biochemistry,
30, 8924–8928.
39. Menard, R., Plouffe, C., Laflamme, P., Vernet, T.,
Tessier, D., Thomas, D. & Storer, A. C. (1994).
Modification of the electrostatic environment is
tolerated in the oxyanion hole of the cysteine protease
papain. Biochemistry, 34, 464–471.
40. Cook, G. M. & Russell, J. B. (1993). The glutamine
cyclotransferase reaction of Streptococcus bovis: a novel
mechanism of deriving energy from non-oxidative
and non-reductive deamination. FEMS Microbiol. Lett.
111, 263–268.
41. McCracken, A. W. & Mauney, C. U. (1971). Identifi-
cation of Corynebacterium diphtheriae by immunofluo-
rescence during a diphtheria epidemic. J. Clin. Pathol.
24, 641–644.
42. Huang, K. F., Chiou, S. H., Ko, T. P., Yuann, J. M. &
Wang, A. H. (2002). The 1.35 Å structure of
cadmium-substituted TM-3, a snake-venom metallo-
proteinase from Taiwan habu: elucidation of a
TNFalpha-converting enzyme-like active-site struc-
ture with a distorted octahedral geometry of cadmi-
um. Acta Crystallogr., Sect. D: Biol. Crystallogr. 58,
1118–1128.
43. Otwinowski, Z. & Minor, W. (1997). Processing of
X-ray diffraction data collected in oscillation mode.
In Macromolecular Crystallography, Part A (Sweet,
R. M., ed.), Macromolecular Crystallography, Part A,
276, pp. Academic, New York..
44. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano,
W. L., Gros, P., Grosse-Kunstleve, R. W. et al. (1998).
Crystallography & NMR system: a new software suite
for macromolecular structure determination. Acta
Crystallogr., Sect. D: Biol. Crystallogr. 54, 905–921.
45. Jones, T. A., Zou, J. Y., Cowan, S. W. & Kjeldgaard, M.
(1991). Improved methods for building protein
models in electron density maps and the location of
errors in these models. Acta Crystallogr. A, 47(Pt 2),
110–119.
2
3. Schilling, S., Niestroj, A. J., Rahfeld, J. U., Hoffmann,
T., Wermann, M., Zunkel, K. et al. (2003). Identifica-
tion of human glutaminyl cyclase as a metalloenzyme.
Potent inhibition by imidazole derivatives and het-
erocyclic chelators. J. Biol. Chem. 278, 49773–49779.
4. Bateman, R. C., Jr., Temple, J. S., Misquitta, S. A. &
Booth, R. E. (2001). Evidence for essential histidines in
human pituitary glutaminyl cyclase. Biochemistry, 40,
2
2
11246–11250.
5. Guevara, T., Mallorqui-Fernandez, N., Garcia-Castellanos,
R., Garcia-Pique, S., Ebert Petersen, G., Lauritzen, C.
et al. (2006). Papaya glutamine cyclotransferase shows
a singular five-fold beta-propeller architecture that
suggests a novel reaction mechanism. Biol. Chem. 387,
1479–1486.
2
2
2
2
6. Huang, K. F., Liu, Y. L., Cheng, W. J., Ko, T. P. &
Wang, A. H. (2005). Crystal structures of human
glutaminyl cyclase, an enzyme responsible for protein
N-terminal pyroglutamate formation. Proc. Natl Acad.
Sci. USA, 102, 13117–13122.
7. Huang, K. F., Wang, Y. R., Chang, E. C., Chou, T. L. &
Wang, A. H. (2008). A conserved hydrogen-bond
network in the catalytic centre of animal glutaminyl
cyclases is critical for catalysis. Biochem. J. 411,
181–190.
8. Huang, K. F., Liu, Y. L. & Wang, A. H. (2005). Cloning,
expression, characterization, and crystallization of a
glutaminyl cyclase from human bone marrow: a
single zinc metalloenzyme. Protein Expression Purif.
43, 65–72.
9. Dahl, S. W., Slaughter, C., Lauritzen, C., Bateman,
R. C., Jr., Connerton, I. & Pedersen, J. (2000). Carica
papaya glutamine cyclotransferase belongs to a novel
plant enzyme subfamily: cloning and characterization
of the recombinant enzyme. Protein Expression Purif.
20, 27–36.
3
0. Schilling, S., Manhart, S., Hoffmann, T., Ludwig,
H. H., Wasternack, C. & Demuth, H. U. (2003).
Substrate specificity of glutaminyl cyclases from
plants and animals. Biol. Chem. 384, 1583–1592.
1. Richardson, J. S. (1981). The anatomy and taxonomy
of protein structure. Adv. Protein Chem. 34, 167–339.
2. Harding, M. M. (1999). The geometry of metal-ligand
interactions relevant to proteins. Acta Crystallogr., Sect.
D: Biol. Crystallogr. 55, 1432–1443.
3
3
3
3. Benz, I. & Schmidt, M. A. (2002). Never say never
again: protein glycosylation in pathogenic bacteria.
Mol. Microbiol. 45, 267–276.
4. Busby, W. H., Jr., Quackenbush, G. E., Humm, J.,
Youngblood, W. W. & Kizer, J. S. (1987). An
enzyme(s) that converts glutaminyl-peptides into
pyroglutamyl-peptides. Presence in pituitary, brain,
adrenal medulla, and lymphocytes. J. Biol. Chem. 262,
46. McRee, D. E. (1999). XtalView/Xfit—a versatile
program for manipulating atomic coordinates and
electron density. J. Struct. Biol. 125, 156–165.
47. Murshudov, G. N., Vagin, A. A. & Dodson, E. J.
(1997). Refinement of macromolecular structures by
the maximum-likelihood method. Acta Crystallogr.,
Sect. D: Biol. Crystallogr. 53, 240–255.
48. Engh, R. A. & Huber, R. (1991). Accurate bond and
angle parameters for X-ray protein structure refine-
ment. Acta Crystallogr. A, 47, 392–400.
49. Laskowski, R. A., Moss, D. S. & Thornton, J. M. (1993).
Main-chain bond lengths and bond angles in protein
structures. J. Mol. Biol. 231, 1049–1067.
3
8532–8536.
3
5. Schilling, S., Cynis, H., von Bohlen, A., Hoffmann, T.,
Wermann, M., Heiser, U. et al. (2005). Isolation,
catalytic properties, and competitive inhibitors of the
zinc-dependent murine glutaminyl cyclase. Biochem-
istry, 44, 13415–13424.
6. Schilling, S., Lindner, C., Koch, B., Wermann, M.,
Rahfeld, J. U., von Bohlen, A. et al. (2007). Isolation and
50. Luthy, R., Bowie, J. U. & Eisenberg, D. (1992).
Assessment of protein models with three-dimensional
profiles. Nature, 356, 83–85.
3