Báo cáo khoa học: Flexibility and communication within the structure of the Mycobacterium smegmatis methionyl-tRNA synthetase Henrik Ingvarsson and Torsten Unge potx

Flexibility and communication within the structure of the
Mycobacterium smegmatis methionyl-tRNA synthetase
Henrik Ingvarsson and Torsten Unge
Department of Cell and Molecular Biology, Uppsala Biomedical Center, Uppsala University, Sweden
Introduction
With the aim of developing a new anti-tuberculosis
drug, we selected methionyl-tRNA synthetase (MetRS)
as a potential target. The aminoacyl-tRNA synthetases
(aaRSs) have been considered as promising targets as
a result of their central role in cell metabolism, the sig-
nificant sequence differences between the prokaryotic
and eukaryotic enzymes, the availability of enzyme
material and access to structural information [1–3].
Furthermore, the potential of aaRSs as drug targets
has been illustrated by the work on isoleucyl-tRNA
synthetase. The isoleucyl-tRNA synthetase-specific
inhibitor mupirocin has been shown to be active
against the Gram-positive pathogens Staphylococ-
cus aureus and Streptococcus pyogenes and is currently
used as a topical antibiotic [4].
The c harging of th e tRNA species with their corre-
sponding amino acids is catalyzed b y their cognate aaRSs,
one for each amino acid [5]. However, not all bacterial
pathogens possess all 20 aaRSs; only the eukaryotic spe-
cies and a few eubacteria have a complete set [6–8]. The
aminoacylated tRNA molecules bind to the A-site of the
ribosome for participation in protein biosynthesis [9].
The aminoacylation is a two-step reaction [2]. In the
first step, the corresponding amino acid is activated by
Keywords
adenosine; methionine; methionyl-tRNA

synthetase; MetRS;
Mycobacterium smegmatis
Correspondence
T. Unge, Department of Cell and Molecular
Biology, Uppsala Biomedical Center,
Uppsala University, Box 596, SE-751 24,
Uppsala, Sweden
Fax: +46 18 53 69 71
Tel: +46 18 471 50 62
E-mail:
Database
Structural data are available in the Protein
Data Bank/BioMagResBank databases
under the accession numbers 2X1L for
M. smegmatis MetRS:M/A and 2X1M for
M. smegmatis MetRS:M
(Received 17 May 2010, revised 15 July
2010, accepted 20 July 2010)
doi:10.1111/j.1742-4658.2010.07784.x
Two structures of monomeric methionyl-tRNA synthetase, from Mycobac-
terium smegmatis, in complex with the ligands methionine ⁄ adenosine and
methionine, were analyzed by X-ray crystallography at 2.3 A
˚
and at 2.8 A
˚
,
respectively. The structures demonstrated the flexibility of the multidomain
enzyme. A new conformation of the structure was identified in which the
connective peptide domain bound more closely to the catalytic domain
than described previously. The KMSKS(301-305) loop in our structures

was in an open and inactive conformation that differed from previous
structures by a rotation of the loop of about 90° around hinges located at
Asn297 and Val310. The binding of adenosine to the methionyl-tRNA
synthetase methionine complex caused a shift in the KMSKS domain that
brought it closer to the catalytic domain. The potential use of the
adenosine-binding site for inhibitor binding was evaluated and a potential
binding site for a specific allosteric inhibitor was identified.
Abbreviations
aaRS, aminoacyl-tRNA synthetase; CAPS, 3-cyclohexylaminopropane-1-sulfonic acid; CP, connective peptide; MetRS, methionyl-tRNA
synthetase; MetRS:M, methionyl-tRNA synthetase in complex with methionine; MetRS:M ⁄ A, methionyl-tRNA synthetase in complex with
methionine ⁄ adenosine; NCS, noncrystallographic symmetry; PDB ID, Protein Data Bank identification.
FEBS Journal 277 (2010) 3947–3962 ª 2010 The Authors Journal compilation ª 2010 FEBS 3947
the formation of an aminoacyl–adenylate complex. In
the second step the amino acid is linked to the 3¢-end
of the cognate tRNA molecule through the formation
of an ester bond.
The aaRSs have been divided into two classes based
on the presence of certain conserved motifs and also
on the structural framework supporting the active site
[10–12]. Class 1 aaRSs are characterized by the amino
acid sequence motifs HIGH and KMSKS, and by
having a catalytic domain with a classical Rossmann-
fold topology [13–16]. Class 2 aaRSs have their active-
site residues located in an antiparallel b-sheet [10].
Each class consists of about 10 aaRSs; these are fur-
ther divided into three subclasses – a, b and c – based
on their structures and sequences [17]. A modified clas-
sification, which includes structural data, has been pro-
duced [18]. MetRS belongs to class 1a, together with
isoleucyl-, leucyl-, valyl-, cysteinyl- and arginyl-tRNA

synthetases. The MetRS molecules occur either in
monomeric or dimeric forms. The dimeric forms
contain a dimerization domain appended to the C-ter-
minus [19,20]. The MetRS structure from Mycobacte-
rium smegmatis studied here is monomeric, has only
one knuckle in its connective peptide (CP) domain and
no coordinated zinc ion.
The MetRS core unit consists of four domains
(Fig. 1) [21]: the catalytic domain, the CP domain, the
KMSKS domain and an anticodon-binding domain
(anticodon domain). The catalytic domain contains the
binding pockets for the substrates methionine, ATP
and the 3¢-CCA end of tRNA
Met
.
In this study we present the unique features of the
M. smegmatis MetRS in complex with methionine
(MetRS:M; 2.8 A
˚
resolution) and methionine ⁄ adeno-
sine (MetRS:M ⁄ A, 2.3 A
˚
resolution). Initially, our
study of the mycobacterial MetRSs included the enzyme
from Mycobacterium tuberculosis as well as that from
M. smegmatis, but because of solubility problems with
the M. tuberculosis enzyme, our study analyzed only the
M. smegmatis variant. However, the sequence similarity
between M. tuberculosis MetRS and M. smegmatis
MetRS is high (74% identity) and the active-site resi-

dues are identical, except for two residues in the ATP-
binding site. We describe, in detail, the binding modes
of the methionine, the adenosine and the residues
involved. In order to evaluate the possibilities for
designing a mycobacterial-specific competitive inhibitor,
the differences in the active site relative to the human
mitochondrial variant were also described. The implica-
tions of the tight binding of methionine and adenosine
were discussed, together with the functional significance
of the structural shift of the entire KMSKS domain that
is induced by addition of the adenosine molecule.
The KMSKS domain of the class 1a MetRSs con-
tains the conserved sequence Lys-Met-Ser-Lys-Ser
[12,15]. In the M. smegmatis MetRS structure this
sequence is located in a loop from residues 297 to 310.
The structural information currently available shows
that this loop is highly flexible and adopts different
conformations (which may have different functions),
depending on the composition of the aaRS complex
(apo, adenylate or tRNA complex) and crystal-packing
interactions. Data on the structures and sequences we
used for comparisons are summarized in Table 1. The
Fig. 1. An overall view of the crystallographic model of Mycobacte-
rium smegmatis MetRS:M ⁄ A. The structure is divided into four
domains: the catalytic domain in red, the CP domain in green, the
KMSKS domain in yellow and the anticodon domain in blue. The
intervening helices a2 and a8 in the catalytic domain are colored
cyan. The linking p-helix a13 between the KMSKS domain and the
anticodon domain is purple. The two ligands (methionine and aden-
osine) bound to the active site are shown in gray.

Crystal structure of M. smegmatis MetRS H. Ingvarsson and T. Unge
3948 FEBS Journal 277 (2010) 3947–3962 ª 2010 The Authors Journal compilation ª 2010 FEBS
second lysine residue in the KMSKS sequence has pre-
viously been predicted to be involved in coordination
of the transition state [22]. Analysis of the Escherichi-
a coli glutaminyl-tRNA synthetase tRNA
Gln
complex
[E. coli GlnRS:tRNA
Gln
; Protein Data Bank identifica-
tion (PDB ID): 1QTQ] revealed that the MSK Lys
residue (amino acid 270 in E. coli GlnRS:tRNA
Gln
)
hydrogen bonds to the adenylate phosphate and
interacts via a water molecule with the 3¢-end of the
ribose of the acceptor stem [23]. The positioning of
this Lys residue is supported by the interaction of the
methonine residue with the adenylate-adenine group
and the hydrophobic residues making up the pocket
for the methionine side chain. This arrangement of the
MSK methionine residue is also present in the Aquifex
aeolicus MetRS tRNA
Met
complex (A. aeolicus Met-
RS:tRNA
Met
, PDB ID: 2CT8) [24]. In the A. aeolicus
structure, however, the tRNA acceptor stem is disor-

dered and the MSK lysine residue is not in contact
with the adenylate phosphate. The structures of the
class 1 aaRS enzymes show a small variation in struc-
ture of the KMSKS loop and that the loop is in a con-
formation that positions the MSK Lys residue close
to the catalytic site. This is shown in the structural
studies of E. coli MetRS [25], Thermus thermophilus
leucyl-tRNA synthetase [26] and Methanocaldococcus
jannaschii tyrosyl-tRNA synthetase [27].
In the M. smegmatis MetRS structures presented
here, the loop containing the KMSKS sequence is
trapped in an open and inactive conformation, which
is the result of a 90° rotation compared with previ-
ously described structures. The possible functional
significance of this conformation is discussed, as is
the new structural situation that this conformation
opens for allosteric inhibition of the enzyme. The pos-
sibilities for design of a competitive inhibitor are also
discussed.
These structures exhibit a new structural feature of
the CP domain. In one of the complexes, the MetRS
molecule is trapped in a conformation where the CP
domain is in a tighter contact with the catalytic
domain than has been observed earlier.
By comparison with the A. aeolicus MetRS:tRNA
Met
anticodon domain we identified the three conserved
residues Trp433, Arg363 and Asn359 (M. smegmatis
MetRS) that, together with four more residues, are
involved in the recognition of the anticodon triplets

CAU in initiator and elongator tRNA
Met
(tRNA
Met
f
and tRNA
Met
m
, respectively).
Results
Crystal properties
The crystals of M. smegmatis MetRS:M ⁄ A and
M. smegmatis MetRS:M belong to the space groups C2
and R32, respectively. M. smegmatis MetRS:M ⁄ A has
three molecules (A, B and C), and M. smegmatis
MetRS:M has one molecule, in the asymmetric unit.
Residues 2–512 out of 515 residues in total could be
built in the A molecule, whereas in the B and C mole-
cules, there was no electron density for residues 124–
160 and residues 123–158, respectively. Crystal packing
contacts of the A molecule stabilize the structure of
these flexible regions, which belong to the CP domain.
See Table 2 for data collection and refinement statistics.
Table 1. Sequence identities between MetRSs, including the catalytic domain, the CP domain, the KMSKS domain and the anticodon
domain from the species referred to in the Introduction, the Results and the Discussion. The numbers denote percentage identity. The PDB
IDs of the MetRSs, with their corresponding ligands, are shown.
aaRS Organism PDB ID Ligands
Sequence identity
M. smegmatis (%) M. tuberculosis (%)
MetRS M. tuberculosis Not available – 74 –

A. aeolicus 2CT8 Methionyl-sulfamoyl-adenosine,
tRNA
Met
44 43
T. thermophilus 1A8H Zn ion 39 41
T. thermophilus 2D54 Zn ion 39 41
Human (mitochondrial) Not available – 36 38
E. coli 1F4L Met, Zn ion 25 25
P. abyssi 1RQG Zn ion 28 27
GlnRS E. coli 1QTQ Glutaminyl-sulfamoyl-adenosine,
tRNA
Gln
sulfate ion
––
LeuRS T. thermophilus 1H3N Leucine, leucyl-sulfamoyl-adenosine,
sulfate ion, Zn ion
––
TyrRS M. jannaschii 1J1U tRNA
Tyr
-tyrosine, Mg ion – –
H. Ingvarsson and T. Unge Crystal structure of M. smegmatis MetRS
FEBS Journal 277 (2010) 3947–3962 ª 2010 The Authors Journal compilation ª 2010 FEBS 3949
In the crystals of both M. smegmatis MetRS:M ⁄ A
and M. smegmatis MetRS:M, the crystallization buffer
component 3-cyclohexylaminopropane-1-sulfonic acid
(CAPS), plays a crucial role for crystal formation in
that it coordinates two symmetry-related molecules in
the crystal lattice. The CAPS molecule binds to a shal-
low pocket situated in the junction between the
C-terminal end of helix a4 of the catalytic domain, the

N-terminal end of strand b4 and helix a5, and the
C-terminal end of strand b8 belonging to the CP
domain (Figs 1 and 2). Each MetRS binds one phos-
phate molecule coordinated by the main-chain amide
group of Asp189 in the N-terminal part of helix a6
from the CP domain and Arg288 positioned in the
N-terminal end of b10 from the catalytic domain.
The catalytic domain
The catalytic domain is an a ⁄ b domain which is divided
into two segments: the N-terminal part (residues
1–115) and the C-terminal part (residues 229–292)
(Fig. 1). The inner core of the domain consists of a
Table 2. Data collection, crystal parameters and refinement statistics for Mycobacterium smegmatis MetRS.
Data set
M. smegmatis MetRS:M ⁄ A
(PDB ID: 2X1L)
M. smegmatis MetRS:M
(PDB ID: 2X1M)
Data collection
Wavelength (A
˚
) 1.038 1.038
Resolution range (A
˚
) 30.0–2.3 (2.42–2.30) 20.0–2.8 (2.95–2.80)
No. of measured reflections 384352 (50671) 116080 (16911)
No. of unique reflections 93306 (13381) 15164 (2221)
Average multiplicity 4.1 (3.8) 7.7 (7.6)
Completeness (%) 97.8 (96.1) 98.7 (99.4)
Mean I ⁄ r(I) 11.8 (3.2) 17.8 (5.1)

R
merge
a
(%) 9.3 (38.1) 8.6 (36.3)
R
p.i.m
b
(%) 5.1 (21.6) 3.3 (13.9)
Crystal parameters
Solvent content (%) 61.1 54.4
Matthews coefficient, V
M
(A
˚
3
ÆDa
)1
) 3.2 2.7
No. of molecules in the asymmetric unit 3 1
Space group C2 R32:H
Unit-cell lengths (A
˚
) a = 155.9, b = 138.9, c = 123.3 a = 210.0, b = 210.0, c = 73.9
Unit-cell angles (°) a = c = 90, b = 124.8 a = b = 90, c = 120
Mosaicity (°) 0.81 0.57
Refinement statistics
Resolution range (A
˚
) 30.0–2.3 20.0–2.8
No. of reflections used in working set 88634 14405

No. of reflections used in test set 4670 756
R-factor (%) 21.8 21.0
R
free
(%) 24.9 24.9
No. of nonhydrogen atoms 12263 4120
No. of solvent water molecules 484 16
Mean B factor for protein atoms
Protein atoms (A
˚
2
) 23.8 47.0
Ligand atoms (A
˚
2
)
Methionine 15.5 28.9
Adenosine 42.2 –
CAPS
c
32.2 46.1
Dihydrogen phosphate
c
23.1 41.2
Water atoms (A
˚
2
) 24.2 31.8
Ramachandran plot outliers
d

(%) 0 0
rmsd from ideal values
e
Bond lengths (A
˚
) 0.006 0.005
Bond angles (°) 0.89 0.81
a
R
merge
= R
h
R
l
ŒI
hl
) ÆI
h
æŒ ⁄R
h
R
l
ÆI
h
æ, where I
hl
is the lth observation of reflection h.
b
R
p.i.m

= R
hkl
[1 ⁄ (N ) 1)]
1 ⁄ 2
R
i
|I
i
(hkl) ) ÆI(hkl)æŒ ⁄R
hkl
R
i
I(hkl).
c
Component of the crystallization condition.
d
PROCHECK (CCP4i program suite) – disallowed regions [37].
e
Calculated for the protein using
ideal values [50].
Crystal structure of M. smegmatis MetRS H. Ingvarsson and T. Unge
3950 FEBS Journal 277 (2010) 3947–3962 ª 2010 The Authors Journal compilation ª 2010 FEBS
b-sheet built up by five parallel b-strands: b1 (residues
4-9), b2 (residues 42–49), b3 (residues 91–94), b9
(residues 258–263) and b10 (residues 289–292). The
b-strands are interconnected by seven a-helices: a1 (res-
idues 19–39), a2 (residues 53–64), a3 (residues 66–85),
a4 (residues 97–114), a7 (residues 228–241), a8 (resi-
dues 248–255) and a9 (residues 267–282). In MetRS
structures determined previously, helix a4 has been

assigned to either the catalytic domain or the CP
domain. Here we include a4 in the catalytic domain.
Specific features for some of the class 1a synthetases
are additional helices in the Rossmann-fold structure
[28]. In M. smegmatis MetRS, two helices (a2 and a8)
are inserted. Helix a2 connects the b2 strand with helix
a3, and helix a8 connects helix a7 with strand b9
(Fig. 1). The active site is positioned at the C-terminal
edge of the sheet formed by the five parallel b-strands
and the N-terminal ends of the helices a1, a7 and a9.
The methionine-binding site
The active site in MetRS contains the binding pockets
for methionine and ATP. The methionine pocket is a
tight cavity that only exposes the carboxylate group of
the methionine ligand. The cavity, which is composed
of 12 residues, is mainly hydrophobic (Fig. 3A). The
atoms close-packing (within 3.8 A
˚
) to the sulfur atom
of methionine are Ce and Ne of His270, OH of
Tyr237, Cb of Ala10 and N of Ile11 – thus three polar
and two hydrophobic interactions. The free carboxyl-
ate oxygen atoms form hydrogen bonds to three water
molecules and the 5¢-hydroxyl of the adenosine mole-
cule. The Tyr13 aromatic ring covers the positively
charged amino group of the methionine ligand and
partly covers the carboxylate group. Residue Trp230
stacks from the side. The position of Tyr13 is
guided by a tight polar interaction to Trp230 through
close-packing of the OH and Ne atoms. The impor-

tance of Tyr13 and Trp230 residues for the methionine
binding is indicated by the fact that they are con-
served, and by the safe positioning of Trp230 through
face-to-face stacking against Phe269 on the opposite
side, and close-packing to Glu131, Tyr228 and Ile266.
The ATP-binding site
In our attempts to form different active-site complexes,
we managed to form a complex with methionine alone,
and methionine and adenosine, but not with adenosine
alone. The structure of the methionine ⁄ adenosine
complex shows that methionine binds to the ribose part
of adenosine, and that residues Ala10 and Ala12 have
close-packing interactions with both methionine and
the ribose. The ATP-binding site is a deep groove rather
than a cavity. The groove comprises 14 residues in the
catalytic domain, including the conserved HIGH(19–
22) sequence, and three residues from the KMSKS
domain (Fig. 3B). Binding of the adenosine molecule
leads to re-arrangement of the site residues. The
adenine moiety is oriented through a hydrogen bond
from N6 to O of Leu295, face-to-face stacking against
Trp294, end-to-end stacking against the two histidines
19 and 22, and close-packing interactions to Gly21.
Analysis of the catalytic-site amino acid sequences
General analysis of catalytic-site sequence conservation
has been carried out by Serre et al. and Landes et al.
[16,29]. In the present study, in order to evaluate the
catalytic site as a drug target, comparison of the
Pro498
Pro379

Tyr452
Arg495
Phe382
Wat
Wat
Arg174
Ala170
Tyr171
Asp116
Arg167
Ala114
Gly115
Fig. 2. CAPS molecules are found in the
contact points between the symmetry-
related proteins in the lattices of the
Mycobacterium smegmatis MetRS:M and
the Mycobacterium smegmatis MetRS:M ⁄ A
crystals. Green residues belong to the
catalytic domain and to the CP domain.
Residues in cyan belong to the anticodon
domain of a symmetry-related protein. The
interactions holding the CAPS in its place
are mainly of a close-packing type. The
dotted green lines represent electrostatic
or hydrogen bonds. Wat denotes water
molecule.
H. Ingvarsson and T. Unge Crystal structure of M. smegmatis MetRS
FEBS Journal 277 (2010) 3947–3962 ª 2010 The Authors Journal compilation ª 2010 FEBS 3951
M. smegmatis sequence with the human mitochondrial
(Hs-mit) and cytoplasmic sequences was performed

(Fig. 4). The human cytoplasmic sequence was not
included in the alignment due to too low sequence simi-
larity. Comparison with the Hs-mit sequence showed
high degree of identity among the methionine- and
adenosine-binding residues. The residues Ala10, Ala12,
Glu25 and Trp294, however, were M. smegmatis
specific. Trp294 is a phenylalanine in M. tuberculosis.
The KMSKS domain
The KMSKS domain (residues 293-350), which has
also been called the ‘stem contact fold domain’ and
the b-a-a-b-a topology domain, connects the catalytic
domain with the anticodon domain (Fig. 1). The sec-
ondary structure elements are: b11 (residues 295-297),
a10 (residues 311-319), a11 (residues 321-332), b12
(residues 338-340) and a12 (residues 341-350). This
domain contains the conserved sequence KMSKS(301-
305) located in a loop (comprising residues 297–310)
with a b-hairpin like structure at the tip of the domain
(Fig. 5). In our structures this loop is trapped in an
open and inactive conformation, with Lys304 unable
to reach the active site. This arrangement of the loop
is stabilized by hydrogen bonds to residues within the
KMSKS domain (Fig. 5). Analysis of the crystal-pack-
ing interactions verifies that this conformation is not
induced by contacts made to neighbouring molecules.
Comparison with the class 1a A. aeolicus Met-
RS:tRNA
Met
and T. thermophilus MetRS (apo) struc-
tures reveal that the position of the KMSKS loop in

our structures is the result of a rotation of  90°
around the hinges H1 and H2 located at Asn297 and
at Val310, respectively. In A. aeolicus MetRS and
T. thermophilus MetRS the corresponding hinge posi-
tions are Val291 ⁄Val304 and Gly292 ⁄Val306 (Figs 6–8).
The properties of the two hinges differ significantly.
A
B
Fig. 3. Stereoviews of the methionine (Met) and adenosine (Ade) binding sites. The residues in the figures are all within a 4 A
˚
distance from
the Met and Ade molecules. (A) The residues lining the Met molecule form a tight cavity with a mainly hydrophobic interior. An ion-pair inter-
action with Od2 of Asp50 and a hydrogen bond to O of Ile11 orients the amino group of the Met. The Tyr13 aromatic ring covers the posi-
tively charged amino group of the Met and partly covers the carboxylate group. Trp230 stacks from the side against the Met and is in a
close-packing polar interaction with Tyr13. (B) The orientation of the ribose ring is controlled by hydrogen bonds to main-chain atoms as well
as to side-chain atoms. The adenine base is oriented by hydrogen bonding to O of Leu295, by stacking against Trp294, by end-to-end stack-
ing with His19 and His22, and by close packing to Gly21. Residues 293–295 belong to the KMSKS domain and residues 19–22 constitute
the conserved HIGH sequence. Water molecules fill the volume between the ligands.
Crystal structure of M. smegmatis MetRS H. Ingvarsson and T. Unge
3952 FEBS Journal 277 (2010) 3947–3962 ª 2010 The Authors Journal compilation ª 2010 FEBS
Hinge H2 is located in the conserved sequence
GNV
VDP (the position of H2 is underlined) (Fig. 7).
The rotation axis is around the Ca-C bond of the Val
residue. Comparison with the A. aeolicus structure
shows that, despite the large conformational changes
accompanying the rotation at H2, the side chain of the
hinge Val residue maintains its interaction with the
conserved isoleucine of the HIGH sequence (leucine in
A. aeolicus MetRS and in T. thermophilus MetRS). At

H1, the rotation mechanism is more complex, and the
result of conformational changes of the peptide chain.
The changes include a rotation of the peptide plane
between Asn297 and Arg298. Furthermore, the
sequences are not conserved at H1, and in the T. ther-
mophilus MetRS structure a proline residue has also
been inserted.
The open conformation of the KMSKS loop exposes
the vacant binding site for the KMSKS methionine
residue (Figs 8 and 9). The binding site is a cavity
composed of the adenine group and hydrophobic
Fig. 4. Alignment of the active-site sequences from Mycobacterium smegmatis MetRS (Ms) and Hs-mit MetRS (Hs). The adenosine-binding
residues are denoted with ‘a’ and the methionine-binding residues with ‘m’.
Fig. 5. The open conformation of the KMSKS loop. The structure
of the KMSKS loop is stabilized by internal hydrogen bonds as well
as by hydrogen bonds to the KMSKS domain. In order for the loop
to adopt the closed active conformation, the bonds Asn314 in a10
to Ile306 and Asn308, Glu342 in a12 to Ser303 and Ser305, and
Tyr340 to Val309, have to be broken.
H2
H1
Fig. 6. Least-squares superimpositioning of the KMSKS domains
from Mycobacterium smegmatis (yellow), Aquifex aeolicus (magenta)
and Thermus thermophilus (gray). Hinges (H1 and H2) located at
sequence positions 297 ⁄ 310 (M. smegmatis), 291 ⁄ 304 (A. aeolicus)
and 292 ⁄ 306 (T. thermophilus) allow rotation of the KMSKS loop.
The two positions of the loop correspond to a rotation of  90°.
H. Ingvarsson and T. Unge Crystal structure of M. smegmatis MetRS
FEBS Journal 277 (2010) 3947–3962 ª 2010 The Authors Journal compilation ª 2010 FEBS 3953
residues. The hydrophobic nature or this site makes it

attractive for binding the side chain of an inhibitor.
The majority of the residues in the site are identical
between the M. smegmatis and the human enzymes,
but at two positions – residues Leu295 and Leu296 –
the sequences differ, and the human sequence has
instead a trypyophan and a threonine (Fig. 7).
The M. smegmatis MetRS:M ⁄ A structure reveals a
shift of the entire KMSKS domain relative to its posi-
tion in the M. smegmatis MetRS:M structure that
Fig. 7. Alignment of the MetRS KMSKS domain sequences from Mycobacerium smegmatis, Aquifex aeolicus, Thermus thermophilus, and
Hs-mit. The hinges H1 and H2 are indicated. H2 is located in the conserved sequence GNV
VDP (Val310 is underlined). At H1, however, the
sequences are not conserved. The LL residues of the LLR sequence are important for the stability of the hydrophobic core below the ade-
nine group.
Fig. 8. A stereoview of the conformation shift of the KMSKS loop. In the Mycobacerium smegmatis MetRS structures, the KMSKS loop is
rotated  90° compared with previous structures. Even though the loop is not rotated exactly as a rigid body, two hinges could be identified.
The M. smegmatis MetRS loop is shown in yellow, the Aquifex aeolicus MetRS loop (PDB ID: 2C8T) in magenta and the Thermus thermo-
philus MetRS loop (PDB ID: 2D54) in gray. At hinge H2, which is located at Val310 in M. smegmatis, at Val304 in A. aeolicus and at Val306
in T. thermophilus, the rotation is around the Ca-C bond at the conserved Val residue. At hinge H1, which is located at Asn297 in
M. smegmatis, at Val291 in A. aeolicus and at Gly292 in T. thermophilus, the rotation is more complex than a simple hinge rotation and the
peptide chain makes new turns compared with the closed structure, including a rotation of the peptide plane between Asn297 and Gly298.
Crystal structure of M. smegmatis MetRS H. Ingvarsson and T. Unge
3954 FEBS Journal 277 (2010) 3947–3962 ª 2010 The Authors Journal compilation ª 2010 FEBS
seems to be induced by the interactions with the ade-
nine moiety of the adenosine group. The entire
KMSKS domain is moved towards the catalytic site
through the interactions between adenine and the resi-
due His292, the KMSKS domain residues Gly293,
Trp294 and Leu295, and the HIGH sequence residues
His19 and Gly21. At the tip of the KMSKS domain

the greatest structural shifts are about 1.8 A
˚
(Fig. 10).
Structural shifts of the same size (up to 1.7 A
˚
) are also
observed in the anticodon domain.
The CP domain
The CP domain covers the active site in the catalytic
domain (Fig. 1). It consists of two subdomains: one
mainly b and one mainly a. A distinctive feature of the
mainly b subdomain is the arched antiparallel
b-strands, b4 (residues 117-127) and b8 (residues 156-
166). On the sides of the arch there is one flanking
b-strand b5 (residues 132-134) and an arched tip with
the antiparallel strands b6 (residues 139-141) and b7
(residues 147-149) (Fig. 1). The mainly a subdomain
consists of the helices a5 (residues 168-182) and a6
(residues 189-200), which pack against helix a9 and
b10 in the catalytic domain.
The mainly b subdomain of the M. smegmatis Met-
RS structure is trapped in a conformation which
differs significantly from that of the previously
determined homologus enzymes. This is illustrated in
Fig. 11, which shows a superimposition of the
M. smegmatis CP domain on the homologus CP
domains from E. coli (the rmsd is  1.4 A
˚
for 243 Ca
atoms), Pyrococcus abyssi (the rmsd is  1.3 A

˚
for 236
Ca atoms) and T. thermophilus (the rmsd is  1.2 A
˚
for 254 Ca atoms). The comparison shows that the
mainly b subdomain in the M. smegmatis CP domain
is shifted downwards towards the active site. The tip
of the subdomain located between b4 and b8 has a
unique conformation. It shows the special features of a
representative of the ‘one knuckle’ without a metal
motif of the CP domain. It consists of a concave b-
sheet, b6 and b7, flanked by two loops on both N-
and C-terminal ends of the sheet, which extends the
concave shape. The concave area faces the solution. In
order to show the unique structural features of the
‘one knuckle’ without a metal compared to with a
metal, a superimposition of the M. smegmatis
tip
structure (residues 127–155) on the T. thermophilus tip
(residues 127–152) was made (Fig. 11D). In the
T. thermophilus structure, a zinc ion is tetrahedrally
coordinated by three cysteines and one histidine.
Fig. 9. The binding site of the KMSKS methionine (Met) residue.
The open conformation of the KMSKS loop exposes the binding
site of the KMSKS loop Met residue in the closed conformation.
Through an alignment with complexes of Aquifex aeolicus
MetRS:tRNA
Met
and Escherichia coli GlnRS:tRNA
Gln

, the binding-
site residues shown in the figure were identified. The residues
Leu295 and Leu296 are tryptophan and threonine, respectively, in
the Hs-mit sequence (Fig. 7).
Fig. 10. The binding of adenosine (Ade) (green) to the My cobacterium
smegmatis MetRS:M complex causes a shift in the KMSKS domain
and in the HIGH sequence of the catalytic domain. The superim-
positions shown here are made between all Ca atoms of the
catalytic domains of the M. smegmatis MetRS structures (the rmsd
is  0.3 A
˚
for all proteins). The three proteins A, B and C of
the asymmetric unit of the M. smegmatis MetRS:M ⁄ A crystal
(purple) and the protein M. smegmatis MetRS:M (gray) are
included. Molecules A, B and C are grouped together. This super-
imposition clearly shows that the binding of Ade causes a shift in
the KMSKS domain.
H. Ingvarsson and T. Unge Crystal structure of M. smegmatis MetRS
FEBS Journal 277 (2010) 3947–3962 ª 2010 The Authors Journal compilation ª 2010 FEBS 3955
However, in the M. smegmatis structure there is a
water molecule close to the metal position. This water
molecule is coordinated through a network of hydro-
gen bonds to the carbonyl oxygen of Pro155, the
main-chain amide groups of Ile128 and Arg129.
Despite these differences in the coordination of the res-
idues in the tip, the fold of the loop is strikingly simi-
lar. The rmsd for the Ca atoms of the residues 127 to
138 was  0.5 A
˚
.

The mainly b subdomain is anchored to the catalytic
domain through an extensive hydrogen-bond network
(Fig. 12A). Three water molecules participate in this
interaction. The position and curvature of the b8 and
the b4 strands is stabilized by an ion-pair interaction
between Glu161 and Arg210, which also forms hydro-
gen bonds to Gln54 and a water molecule (Fig. 12A).
The stabilizing interactions involve hydrophobic stack-
ing interactions through the aromatic side chains of
Tyr122, Tyr126, Phe133 and Tyr228. In the B and C
molecules in the MetRS:M ⁄ A (with both methionine
and adenosine) crystal, these interactions are not pres-
ent and the tip of the mainly b subdomain is not visi-
ble (Figs 12A,B).
The anticodon domain
By comparison with the A. aeolicus MetRS:tRNA
Met
complex, the anticodon-binding domain of M. smegma-
tis MetRS was assigned to include residues 358–515
(Fig. 1) [24]. This domain is mainly a-helical. The
helices a15 (residues 383–406), a16 (residues 408–430)
and a18 (residues 441–465) form a bundle, and the heli-
ces a14 (residues 357–372), a19 (residues 470–480) and
the short 3
10
helix a20 (residues 488–493) interact with
a18 at an angle (Fig. 1). The same arrangement of
the helices is found in the structures of T. thermophilus
MetRS, Pyrococcus abyssi MetRS, A. aeolicus MetRS
and E. coli MetRS. A p-helix a13 (residues 351–357)

links the KMSKS domain with the anticodon domain.
By alignment of the anticodon-binding domains of the
M. smegmatis holo structure to the A. aeolicus Met-
RS:tRNA
Met
complex (the rmsd  1.4 A
˚
) we identified
residues Asn359, Arg363, Trp433, Ile508 and Phe509 as
primarily the most important residues involved in the
binding of the tRNA anticodon loop. Asn359 and
Arg363 are located on helix a14. In our structure,
Trp433 interacts with Arg363, but in the A. aeolicus
AB
CD
Fig. 11. Structural comparisons of CP
domains from homologus structures to
Mycobacterium smegmatis MetRS (green):
(A) Escherichia coli MetRS in blue (PDB ID:
1F4L), (B) Pyrococcus abyssi MetRS in pink
(PDB ID: 1RQG) and (C) Thermus thermo-
philus MetRS in gray (PDB ID: 2D54).
M. smegmatis MetRS was found to have a
significant shift (highlighted in the oval in
Fig. 11A) of the protruding arm in com-
parison with the other homologus
structures. (D) The tip of the protruding arm
in T. thermophilus MetRS (residues
127–152) was superimposed onto the
counterpart of M. smegmatis MetRS

(residues 127–155). The zinc atom in yellow
and the water molecule in red belong to
T. thermophilus MetRS and M. smegmatis
MetRS:M ⁄ A, respectively.
Crystal structure of M. smegmatis MetRS H. Ingvarsson and T. Unge
3956 FEBS Journal 277 (2010) 3947–3962 ª 2010 The Authors Journal compilation ª 2010 FEBS
MetRS:tRNA
Met
complex both the corresponding Trp
and Arg residues interact with the nucleic acid, and
thereby form polar and stacking interactions with
slightly different orientations of the side chains
(Fig. 13). Thus, the corresponding arginine residue (357)
in the A. aeolicus MetRS:tRNA
Met
complex is posi-
tioned between nucleotides C34 and A35 that, together
with U36, form the anticodon triplet (CAU) (Fig. 13).
There is a kink involving residues 394–396 in helix
a15 on the back of the molecule relative to the
tRNA-binding site. This kink is of importance for the
positioning of the helix a16 and the residue Asn424.
The potentially nucleotide-binding residues Asn359,
Ile508 and Phe509 all hydrogen bond to water mole-
cules. One of these water molecules occupies a pocket
for U36 in the A. aeolicus MetRS:tRNA
Met
complex.
Discussion
The multidomain aaRS molecules have the capability

to specifically recognize a relatively small substrate, the
amino acid, incorporate ATP and covalently link these
together through the formation of an anhydride bond.
To complete the tRNA-charging process, these
enzymes exclusively recognize their cognate tRNA
molecules, expose the 2¢-or3¢-hydroxyl of the 3¢-end
ribose of the tRNA to the catalytic centre, break the
anhydride bond and link the amino acid to the tRNA
by formation of an ester bond. These functions require
a relatively large molecule with specialized domains,
which in addition should be able to communicate with
each other [29–31].
MetRS is of special importance among the aaRS
enzymes, because of the fact that protein synthesis
starts with the formylated methionine in prokaryotes.
Thus, if the function of the MetRSs could be stopped
through intervention with an inhibitor, the cell would
die. The challenging problem here is specificity. Can
an enzyme with these fundamental functions possess
host-specific properties? In order to scrutinize the
M. tuberculosis MetRS structure in the search for a
potential inhibitor-binding site, we set up trials to rec-
ombinantly express and crystallize this enzyme. It was
discovered that the production of the material was
uncomplicated, but the enzyme had poor solubility
properties that impeded structural work. However, the
closely related enzyme, M. smegmatis MetRS, could
not only be readily produced, but also crystallized.
Tyr228
Ser207

Ser209
Tyr228
Arg210
Tyr122
Glu161
Gln54
Lys55
Met56
Gln54
Arg210
Tyr122
Glu161
W748
W777
W835
W758
W702
W712
Phe133
β8
β4
β8
β4
β5
β7
β6
Glu131
Lys55
Met56
A

B
Fig. 12. Anchoring of the protruding arm of
the CP domain (green) to the catalytic
domain (yellow). (A) A network of hydrogen
bonds between Gln54, Lys55 and Met56 in
the catalytic domain and Glu131, Glu161
and Arg210 in the CP domain tightly associ-
ate the two domains with each other. Three
water molecules (red) take part in this
hydrogen network. The position of Phe133
forces Arg210 to change conformation and
interact with Glu161, Gln54 and a water
molecule. The aromatic side chains of
Tyr122, Tyr126 (not included in the figure),
Phe133 and Tyr228 form a stabilizing clus-
ter. (B) In the absence of this interaction
between the two domains, Arg210 changes
its position and forms hydrogen bonds with
Ser207, Ser209 and a water molecule.
H. Ingvarsson and T. Unge Crystal structure of M. smegmatis MetRS
FEBS Journal 277 (2010) 3947–3962 ª 2010 The Authors Journal compilation ª 2010 FEBS 3957
The amino acid sequences between M. tuberculosis
and M. smegmatis show 74% identity, and among the
catalytic site residues only one residue is different: a
tryptophan in M. smegmatis is replaced with a phenyl-
alanine in M. tuberculosis. Preliminary catalytic activ-
ity experiments (not shown here), based on charging
tRNA
Met
f

with radiolabelled methionine, clearly showed
that the material was biologically active.
Our structural analysis confirms that the MetRS
molecule is flexible [29,30,32]. The mobility of the
domains reflects the structural changes required for
binding, catalysis, communication and release of the
reaction partners. A new conformation of the CP
domain was identified in our structures – in the
M. smegmatis MetRS:M structure as well as in the
M. smegmatis MetRS:M ⁄ A structure. The CP domain
adopted a conformation where it was firmly anchored
to the catalytic domain. The biological significance of
this conformation is not clear, but the structure shows
that a conformation of the enzyme exists where the CP
domain, by binding tightly to the catalytic domain,
increases the rigidity of the complex between the
enzyme and the substrate methionine. Comparisons of
the apo, methionine and adenylate complexes of the
homologus E. coli enzyme showed that the re-arrange-
ment of the active-site residues accompanying the
coordination of the methionine ligand is a prerequisite
for this compact conformation [25,29]. The presence of
two molecules (B and C) in the M. smegmatis
MetRS:M ⁄A crystals with the CP domain in the relaxed
conformation, verifies the flexibility of this domains.
The importance of the conserved KMSKS sequence
for the catalytic activity has already been confirmed in
the structure of the class 1b enzyme E. coli GlnRS
in complex with its cognate tRNA
Gln

. This complex
verifies that the lysine residue in the MSK sequence
interacts directly with the adenylate and indirectly,
through a water molecule, with the tRNA acceptor
stem [23]. The loop containing the MSK sequence is,
in the E. coli GlnRS structure, in close contact with
the catalytic domain. In all of our M. smegmatis Met-
RS structures, however, the KMSKS loop was in an
open conformation with the MSK Lys residue out of
reach of the catalytic site. This conformation of the
KMSKS loop has not been observed previously. The
X-ray structures do not throw any light on the mecha-
nism responsible for the structural change, but it is
clear that the new conformation is not induced by
AB
Fig. 13. Overall view of the anticodon domain of Mycobacterium smegmatis MetRS:M ⁄ A (blue) compared with that of Aquifex aeolicus
MetRS:tRNA
Met
(magenta, PDB ID: 2CT8). (A) All residues that interact directly with the nucleotides in the anticodon triplet CAU of tRNA
Met
are conserved in M. smegmatis MetRS:M ⁄ A (Trp433, Arg363 and Asn359). A water molecule is positioned in the uridine-binding pocket.
The electron-density maps (mF
o
) DF
c
) are contoured at 0.21 eÆA
˚
)3
[the average density (F
o

) F
c
) for 50 carbonyl oxygen in M. smegmatis
MetRS:M ⁄ A is 0.21 eÆA
˚
)3
]. Hydrogen bonds are represented by green dotted lines. (B) The structure of A. aeolicus MetRS (PDB ID: 2CT8)
in complex with the elongator tRNA
Met
. The conserved residues Trp422, Arg357 and Asn353 are positioned to allow interactions to take
place with the nucleotides in the anticodon triplets.
Crystal structure of M. smegmatis MetRS H. Ingvarsson and T. Unge
3958 FEBS Journal 277 (2010) 3947–3962 ª 2010 The Authors Journal compilation ª 2010 FEBS
crystal packing interactions. It is noteworthy that the
loop is not flexible in this new position but anchored
to the KMSKS domain by hydrogen bonds. Com-
parison of the M. smegmatis MetRS structures with
the E. coli GlnRS:tRNA
Gln
and A. aeolicus Met-
RS:tRNA
Met
complexes shows that the loop has
rotated  90° and that two hinges, H1 and H2, could
be identified. Hinge H2 (Val310) in M. smegmatis Met-
RS is located in the conserved sequence GNV
VDP
(Val310 is underlined). The side chain of Val310 is
anchored against the conserved Ile20 in the HIGH
sequence and remains so also after the rotation,

according to the comparison. The sequence conserva-
tion at H2 and the rotation mechanism might indicate
the biological relevance of the open conformation. The
properties of hinge H1 (located at Asn297) differ from
H2. First, there is no sequence conservation at this
site. Second, the rotation mechanism is complex and
involves re-arrangements of the peptide chain.
Through the rotation, positively charged residues,
and also hydrophobic residues, are exposed, and there-
fore able to assist in interaction with a tRNA
molecule. We have so far not been able to show
structurally the triggering factor that makes the loop
change from the open to the closed conformation, but
because our enzyme is active in the presence of tRNA,
it is not unlikely that it is the interaction with the
tRNA that releases the loop. We speculate that the
KMSKS loop could assist in the positioning of the
acceptor stem in the catalytic site. The exact steering
and positioning of the loop, and particularly the cata-
lytic lysine residue, could be controlled through the
positioning of the KMSKS methionine residue in the
hydrophobic pocket that also includes the adenosine as
one of the lining groups. The discovery of the flexibil-
ity of the KMSKS loop opens new possibilities
for designing mycobacterial-specific inhibitors. These
inhibitors should be designed to lock the loop in an
inactive open conformation by occupying the KMSKS
methionine-binding site. Similarly to mupirocin, the
new inhibitor should be an analog of the amino-acid
adenylate, but instead of binding to the KMSKS loop,

the compound should have a side group that could
bind to the KMSKS methionine-binding site [4,33].
Furthermore, the compound should be designed to uti-
lize the mycobacterial-specific residues in this pocket.
The structures of the M. smegmatis MetRS com-
plexes with methionine and adenine can be superim-
posed with good agreement on the structures of
corresponding complexes of homologus enzymes. Few
mycobacterial-specific residues are present in the bind-
ing site. Compared with catalytic site residues of
Hs-mit MetRS, only four of the lining residues are
unique, namely Ala10, Ala12, Glu25 and Trp294 (Phe
in M. tuberculosis). Interestingly, they are located in
the area where the methionine connects to adenosine
and on the same side of these residues. A mycobacte-
rial-specific inhibitor with capacity to bind to these
residues would have to be used in combination
with methionine, or contain a methionine-mimicking
moiety.
Comparison of the structure with only methionine in
the catalytic site to the structure with methionine and
adenosine, showed a small shift in the KMSKS
domain. The shift that brings the KMSKS domain clo-
ser to the catalytic site is induced by the interactions
between the adenine group and adjacent residues in
the KMSKS domain. The shift, which is small (1.7–
1.8 A
˚
), was also observed in the anticodon domain.
This shift may not have any biological significance

compared with the large shifts of  15° that were
observed upon binding of tRNA to A. aeolicus MetRS
[24], but the shift would bring the MSK lysine residue
closer to the catalytic site if the loop was in the closed
conformation. A similar shift, but only of  0.6 A
˚
,is
observed in the corresponding structures of E. coli
MetRS [25]. The smaller shift in this case might be
caused by crystal packing interactions between the
KMSKS domain and a dimerization domain helix.
In summary, our structures of M. smegmatis MetRS
in complex with methionine and methionine ⁄ adenosine
were used to confirm the flexible properties of the
enzyme. The CP domain was found to assist in the sta-
bilization of the methionine ⁄ adenosine complex by
forming a tight interaction with the catalytic domain.
Binding of adenosine to the methionine complex
induced a 1.8 A
˚
shift of the entire KMSKS domain in
the direction closer to the catalytic domain. A new, inac-
tive conformation of the KMSKS loop was identified
that was caused by a rotation of  90° compared with
previous structures. The new conformation exposed a
possible binding site for an allosteric inhibitor, which,
similarly to the inhibitor mupirocin, could be an analog
of the adenylated amino acid. The possibility of design-
ing a competitive inhibitor was indicated. Both binding
sites contained mycobacterial-specific residues.

Materials and methods
Cloning and expression
Details of the cloning, expression, purification and crystalli-
zation procedures have been described previously, together
with an initial analysis of the diffraction data [34].
The gene metG, coding for MetRS (MSMEG_5441), was
isolated by PCR from genomic DNA of M. smegmatis
H. Ingvarsson and T. Unge Crystal structure of M. smegmatis MetRS
FEBS Journal 277 (2010) 3947–3962 ª 2010 The Authors Journal compilation ª 2010 FEBS 3959
MC2 155. The protein sequence was provided with a 6-his-
tidine tag at the N-terminus. Cloning was performed in
E. coli Top10 cells (Invitrogen) and expression was carried
out in the E. coli strain BL21 Star (DE3) cells (Invitrogen).
Induction was performed with 100 mgÆL
)1
of isopropyl
1-thio-b-d-galactopyranoside (Sigma) at OD
600
 0.9. Cells
were harvested after 3 h and the yield was 2 g of cells per
liter of cell culture. The cells were lysed using a Constant
Cell Disruptor (Constant Systems Ltd, Daventry, North-
ants, UK), and the protein was isolated using immobilized
metal affinity chromatography. After a final size-exclusion
chromatography step, where the MetRS molecules migrated
as monomers, the material was homogenous, as judged by
SDS ⁄ PAGE.
Crystallization
Initial screening for crystallization conditions was carried
out using the Wizard I & II screen (Emerald BioSystems).

The experiment was set up at 293 K as a 96-well vapour-
diffusion experiment with sitting-drops. The droplets were
made by mixing 0.5 lL of protein solution with 0.5 lLof
reservoir solution. The protein solution contained
6mgÆmL
)1
of MetRS, 2.5 mm adenosine, 2.5 mm methio-
nine, 2.5 mm tetrasodium pyrophosphate and 50 mm
Hepes, pH 7.0. Crystals were produced with the reservoir
solution containing 1.2 m NaH
2
PO
4
, 800 mm K
2
HPO
4
,
200 mm Li
2
SO
4
and 100 mm CAPS (pH 10.5). The final pH
in the crystallization droplets was 7.0. Crystals were also
produced without adenosine.
Data collection
A cryogenic solution was made by adding 16.6% glycerol
to the crystallization buffer. The crystals were soaked in the
cryo-solution for 20 s and flash-cooled in liquid nitrogen
before being mounted in the beam. Diffraction data were

collected at beamline I911-2, at MAX-lab (Lund, Sweden),
for the two complexes M. smegmatis MetRS:M and
M. smegmatis MetRS:M ⁄ A. Data-collection statistics are
shown in Table 2. Both data sets were processed using
MOSFLM [35] and scaled using SCALA [36] in the CCP4
program suite [37].
Model building and refinement
The structure solution was made using the molecular-
replacement method with PHASER [38]. The starting
model was made from the structure of T. thermophilus
(PDB ID: 1A8H, 39% identical amino acids). The model
was modified using the program CHAINSAW [39], such
that side-chain atoms not identical with the M. smegmatis
MetRS amino acid sequence were deleted.
The models were completed during the course of manual
building in O [40] and refinement in the CCP4 program
REFMAC5 [41]. The crystallographic refinement of the
methionine ⁄ adenosine complex made use of the noncrystal-
lographic symmetry (NCS) associated with this crystal form.
Reciprocal space refinement was carried out using
REFMAC5 with tight NCS-restraints. The resulting Sig-
maA-weighted maps [42] were used in O [32] to start cyclic
NCS-averaging of the 2m|F
o
| ) D|F
c
| maps for rebuilding,
and a zero-cycle average of the m|F
o
| ) D|F

c
| map as an aid
to identify water molecules. Carbonyl oxygen profiling,
calculated using the Water_profile command of O, was used
to decide levels in the water-adding step in O. As a reference
for comparisons between electron-density maps, 50 ran-
domly selected carbonyl oxygen atoms in the model of
M. smegmatis MetRS:M ⁄A were used to calculate a 3D
profile and an average value of the difference in electron
density. This calculation was made using the command
Water_profile in O [40] and generated a peak height value of
0.21 eÆA
˚
)3
. Refinement statistics can be found in Table 2.
The presence of ligands in the active site in both
M. smegmatis MetRS:M ⁄A and M. smegmatis MetRS:M
was confirmed by mF
o
) DF
c
omit maps produced in CNS
[43] by simulated annealing. Validation of structures was
made using MOLPROBITY [44] and MOLEMAN [45].
The amino acid sequence alignments were performed using
the programs JALVIEW [46] and MAFFT [47]. All struc-
tural superimpositions onto M. smegmatis MetRS of ho-
mologus counterparts from various organisms were made
in O [40]. Pictures were prepared in PYMOL [48] and O
[40]. Rendering was performed using MOLRAY [49].

Acknowledgements
Genomic DNA of Mycobacterium smegmatis was a gift
from Dr Mary Jackson, Institut Pasteur, France. We
thank Professor T. Alwyn Jones and Professor Sherry
Mowbray for continuous support during this work, and
Professor Alwyn especially for allowing us to use his
recently developed program in O, the cyclic averaging
procedure, which was crucial for the improvement of the
initial molecular replacement maps. We thank Professor
Lars Liljas for fruitful discussions, Drs Terese Bergfors,
Annette Roos and Daniel Ericsson for reading the man-
uscript, and Dr Lena Henriksson and the staff of beam-
line I911-2 at MAX-lab Lund in Sweden for support
and advice during the collection of diffraction data. This
work was financially supported by the Foundation for
Strategic Research (SSF), the Swedish Research Council
(VR) and the European Commission program NM4TB.
References
1 Hurdle JG, O’Neill AJ & Chopra I (2005) Prospects for
aminoacyl-tRNA synthetase inhibitors as new antimicro-
bial agents. Antimicrob Agents Chemother 49, 4821–4833.
Crystal structure of M. smegmatis MetRS H. Ingvarsson and T. Unge
3960 FEBS Journal 277 (2010) 3947–3962 ª 2010 The Authors Journal compilation ª 2010 FEBS
2 Ibba M & Soll D (2000) Aminoacyl-tRNA synthesis.
Annu Rev Biochem 69, 617–650.
3 Raczniak G, Ibba M & Soll D (2001) Genomics-based
identification of targets in pathogenic bacteria for
potential therapeutic and diagnostic use. Toxicology
160, 181–189.
4 Sutherland R, Boon RJ, Griffin KE, Masters PJ,

Slocombe B & White AR (1985) Antibacterial activity
of mupirocin (pseudomonic acid), a new antibiotic for
topical use. Antimicrob Agents Chemother 27, 495–498.
5 Arnez JG & Moras D (1997) Structural and functional
considerations of the aminoacylation reaction. Trends
Biochem Sci 22, 211–216.
6 Tumbula DL, Becker HD, Chang WZ & Soll D (2000)
Domain-specific recruitment of amide amino acids for
protein synthesis. Nature 407, 106–110.
7 Sheppard K, Akochy PM, Salazar JC & Soll D (2007)
The Helicobacter pylori amidotransferase GatCAB is
equally efficient in glutamine-dependent transamidation
of Asp-tRNAAsn and Glu-tRNAGln. J Biol Chem 282,
11866–11873.
8 Slesarev AI, Mezhevaya KV, Makarova KS, Polushin
NN, Shcherbinina OV, Shakhova VV, Belova GI,
Aravind L, Natale DA, Rogozin IB et al. (2002) The
complete genome of hyperthermophile Methanopyrus
kandleri AV19 and monophyly of archaeal methano-
gens. Proc Natl Acad Sci USA 99 , 4644–4649.
9 Schmeing TM & Ramakrishnan V (2009) What recent
ribosome structures have revealed about the mechanism
of translation. Nature 461, 1234–1242.
10 Eriani G, Delarue M, Poch O, Gangloff J & Moras D
(1990) Partition of tRNA synthetases into two classes
based on mutually exclusive sets of sequence motifs.
Nature 347, 203–206.
11 Cusack S, Hartlein M & Leberman R (1991) Sequence,
structural and evolutionary relationships between class
2 aminoacyl-tRNA synthetases. Nucleic Acids Res 19,

3489–3498.
12 Burbaum JJ & Schimmel P (1991) Structural relation-
ships and the classification of aminoacyl-tRNA synthe-
tases. J Biol Chem 266 , 16965–16968.
13 Barker DG & Winter G (1982) Conserved cysteine and
histidine residues in the structures of the tyrosyl and
methionyl-tRNA synthetases. FEBS Lett 145, 191–193.
14 Webster T, Tsai H, Kula M, Mackie GA & Schimmel P
(1984) Specific sequence homology and three-dimen-
sional structure of an aminoacyl transfer RNA synthe-
tase. Science (New York, NY) 226, 1315–1317.
15 Hountondji C, Dessen P & Blanquet S (1986) Sequence
similarities among the family of aminoacyl-tRNA
synthetases. Biochimie 68, 1071–1078.
16 Landes C, Perona JJ, Brunie S, Rould MA, Zelwer C,
Steitz TA & Risler JL (1995) A structure-based multiple
sequence alignment of all class I aminoacyl-tRNA syn-
thetases. Biochimie 77, 194–203.
17 Cusack S (1995) Eleven down and nine to go. Nat
Struct Biol 2, 824–831.
18 Ribas de Pouplana L & Schimmel P (2001) Two
classes of tRNA synthetases suggested by sterically
compatible dockings on tRNA acceptor stem. Cell
104, 191–193.
19 Cassio D & Waller JP (1971) Modification of
methionyl-tRNA synthetase by proteolytic cleavage and
properties of the trypsin-modified enzyme. Eur J
Biochem ⁄ FEBS 20, 283–300.
20 Mellot P, Mechulam Y, Le Corre D, Blanquet S &
Fayat G (1989) Identification of an amino acid region

supporting specific methionyl-tRNA synthetase: tRNA
recognition. J Mol Biol 208, 429–443.
21 Deniziak MA & Barciszewski J (2001) Methionyl-tRNA
synthetase. Acta Biochim Pol 48, 337–350.
22 Fersht AR (1987) Dissection of the structure and
activity of the tyrosyl-tRNA synthetase by site-directed
mutagenesis. Biochemistry 26, 8031–8037.
23 Rould MA, Perona JJ, Soll D & Steitz TA (1989) Struc-
ture of E. coli glutaminyl-tRNA synthetase complexed
with tRNA(Gln) and ATP at 2.8 A resolution. Science
(New York, NY) 246, 1135–1142.
24 Nakanishi K, Ogiso Y, Nakama T, Fukai S & Nureki
O (2005) Structural basis for anticodon recognition by
methionyl-tRNA synthetase. Nat Struct Mol Biol 12,
931–932.
25 Crepin T, Schmitt E, Mechulam Y, Sampson PB,
Vaughan MD, Honek JF & Blanquet S (2003) Use of
analogues of methionine and methionyl adenylate to
sample conformational changes during catalysis in
Escherichia coli methionyl-tRNA synthetase. J Mol Biol
332, 59–72.
26 Cusack S, Yaremchuk A & Tukalo M (2000) The 2 A
crystal structure of leucyl-tRNA synthetase and its
complex with a leucyl-adenylate analogue. EMBO J 19,
2351–2361.
27 Kobayashi T, Nureki O, Ishitani R, Yaremchuk A,
Tukalo M, Cusack S, Sakamoto K & Yokoyama S
(2003) Structural basis for orthogonal tRNA specifici-
ties of tyrosyl-tRNA synthetases for genetic code expan-
sion. Nat Struct Biol 10, 425–432.

28 Sugiura I, Nureki O, Ugaji-Yoshikawa Y, Kuwabara S,
Shimada A, Tateno M, Lorber B, Giege R, Moras D,
Yokoyama S et al. (2000) The 2.0 A crystal structure of
Thermus thermophilus methionyl-tRNA synthetase
reveals two RNA-binding modules. Structure 8,
197–208.
29 Serre L, Verdon G, Choinowski T, Hervouet N, Risler
JL & Zelwer C (2001) How methionyl-tRNA synthe-
tase creates its amino acid recognition pocket upon
L-methionine binding. J Mol Biol 306, 863–876.
30 Ghosh A & Vishveshwara S (2008) Variations in clique
and community patterns in protein structures during
allosteric communication: investigation of dynamically
H. Ingvarsson and T. Unge Crystal structure of M. smegmatis MetRS
FEBS Journal 277 (2010) 3947–3962 ª 2010 The Authors Journal compilation ª 2010 FEBS 3961
equilibrated structures of methionyl tRNA synthetase
complexes. Biochemistry 47, 11398–11407.
31 Ghosh A & Vishveshwara S (2007) A study of commu-
nication pathways in methionyl- tRNA synthetase by
molecular dynamics simulations and structure network
analysis. Proc Natl Acad Sci USA 104, 15711–15716.
32 Schmitt E, Tanrikulu IC, Yoo TH, Panvert M, Tirrell
DA & Mechulam Y (2009) Switching from an induced-
fit to a lock-and-key mechanism in an aminoacyl-tRNA
synthetase with modified specificity. J Mol Biol 394,
843–851.
33 Nakama T, Nureki O & Yokoyama S (2001) Structural
basis for the recognition of isoleucyl-adenylate and an
antibiotic, mupirocin, by isoleucyl-tRNA synthetase.
J Biol Chem 276, 47387–47393.

34 Ingvarsson H, Jones TA & Unge T (2009) Crystalliza-
tion of Mycobacterium smegmatis methionyl-tRNA
synthetase in the presence of methionine and adenosine.
Acta Crystallogr Sect F Struct Biol Cryst Commun 65,
618–620.
35 Leslie AGW (1992) Recent changes to the MOSFLM
package for processing film and image plate data. Joint
CCP4 ⁄ ESF-EAMCB Newsletter Protein Crystallogra-
phy 26.
36 Evans PR (1993) Data reduction. Proceedings of the
CCP4 Study Weekend on Data Collection and Process-
ing, 114–122.
37 Collaborative Computational Project (Number 4) (1994)
The CCP4 Suite: programs for protein crystallography.
Acta Crystallogr Sect D 50, 760–763.
38 McCoy AJ, Grosse-Kunstleve RW, Adams PD, Winn
MD, Storoni LC & Read RJ (2007) Phaser crystallo-
graphic software. J Appl Crystallogr 40, 658–674.
39 Stein N (2008) CHAINSAW: a program for mutating
pdb files used as templates in molecular replacement.
J Appl Crystallogr 41, 641–643.
40 Jones TA, Zou JY, Cowan SW & 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.
41 Murshudov GN, Vagin AA & Dodson EJ (1997)
Refinement of macromolecular structures by the
maximum-likelihood method. Acta Crystallogr 53, 240–
255.

42 Read RJ (1986) Improved fourier coefficients for maps
using phases from partial structures with errors. Acta
Crystallogr 42, 140–149.
43 Brunger AT (2007) Version 1.2 of the Crystallography
and NMR system. Nat Protoc 2, 2728–2733.
44 Davis IW, Leaver-Fay A, Chen VB, Block JN, Kapral
GJ, Wang X, Murray LW, Arendall WB III, Snoeyink
J, Richardson JS et al. (2007) MolProbity: all-atom
contacts and structure validation for proteins and
nucleic acids. Nucleic Acids Res 35, W375–W383.
45 Kleywegt GJ (1992–2004) Uppsala University, Uppsala,
Sweden. Unpublished program.
46 Waterhouse AM, Procter JB, Martin DM, Clamp M &
Barton GJ (2009) Jalview Version 2–a multiple sequence
alignment editor and analysis workbench. Bioinformat-
ics (Oxford, England) 25, 1189–1191.
47 Katoh K, Kuma K, Toh H & Miyata T (2005) MAFFT
version 5: improvement in accuracy of multiple
sequence alignment. Nucleic Acids Res 33, 511–518.
48 DeLano WL (2008) The PyMOL Molecular Graphics
System. DeLano Scientific LLC, Palo Alto, CA, USA.
49 Harris M & Jones TA (2001) Molray – a web interface
between O and the POV-Ray ray tracer.
Acta Crystal-
logr 57, 1201–1203.
50 Engh RA & Huber R (1991) Accurate bond and angle
parameters for X-ray protein structure refinement. Acta
Crystallogr 47, 392–400.
Crystal structure of M. smegmatis MetRS H. Ingvarsson and T. Unge
3962 FEBS Journal 277 (2010) 3947–3962 ª 2010 The Authors Journal compilation ª 2010 FEBS