Báo cáo khoa học: Structural flexibility of the methanogenic-type seryl-tRNA synthetase active site and its implication for specific substrate recognition pptx

Structural flexibility of the methanogenic-type seryl-tRNA
synthetase active site and its implication for specific
substrate recognition
Silvija Bilokapic
1
, Jasmina Rokov Plavec
1
, Nenad Ban
2
and Ivana Weygand-Durasevic
1
1 Department of Chemistry, Faculty of Science, University of Zagreb, Croatia
2 Department of Biology, Institute of Molecular Biology and Biophysics, Swiss Federal Institute of Technology, Zu
¨
rich, Switzerland
The fidelity of protein synthesis depends on the correct
attachment of amino acids to the 3¢-ends of their cog-
nate tRNA species by aminoacyl-tRNA synthetases
(aaRS), a family of enzymes that bridges the informa-
tion gap between nucleic acids and proteins. Although
strategies for the specific recognition of amino acid
and tRNA are unique to each enzyme, two reaction
steps in the esterification of amino acids to their cog-
nate tRNAs are conserved among all aaRS: (a) the
activation of amino acids with ATP by formation of
aminoacyl-adenylate, and (b) the transfer of the
aminoacyl-moiety from the aminoacyl-adenylate to the
cognate tRNA substrate [1].
Sequence alignment and later structural data showed
that aaRSs constitute a family of enzymes in which the
same catalytic reaction is performed at two topologi-

cally different structural domains. Class I aaRSs are
built around a canonical dinucleotide-binding fold
(Rossmann fold) with the consensus motifs HIGH and
KMSKS, which define two regions of sequence conser-
vation for all class I aaRSs. Class II synthetases share
an antiparallel b sheet partly enclosed by helices, thus
Keywords
conformational flexibility; motif 2 loop; seryl-
tRNA synthetase; specificity of substrate
recognition; synthetase:tRNA model
Correspondence
I. Weygand-Durasevic, Department of
Chemistry, Faculty of Science, University of
Zagreb, Horvatovac 102a, 10000 Zagreb,
Croatia
Fax: +385 1 460 6401
Tel: +385 1 460 6230
E-mail:
(Received 20 February 2008, revised 21
March 2008, accepted 26 March 2008)
doi:10.1111/j.1742-4658.2008.06423.x
Seryl-tRNA synthetase (SerRS) is a class II aminoacyl-tRNA synthetase
that catalyzes serine activation and its transfer to cognate tRNA
Ser
. Previ-
ous biochemical and structural studies have revealed that bacterial- and
methanogenic-type SerRSs employ different strategies of substrate recogni-
tion. In addition to other idiosyncratic features, such as the active site zinc
ion and the unique fold of the N-terminal tRNA-binding domain, metha-
nogenic-type SerRS is, in comparison with bacterial homologues, charac-

terized by a notable shortening of the motif 2 loop. Mutational analysis of
Methanosarcina barkeri SerRS (mMbSerRS) was undertaken to identify the
active site residues that ensure the specificity of amino acid and tRNA
3¢-end recognition. Residues predicted to contribute to the amino acid spec-
ificity were selected for mutation according to the crystal structure of
mMbSerRS complexed with its cognate aminoacyl-adenylate, whereas those
involved in binding of the tRNA 3¢-end were identified and mutagenized
on the basis of modeling the mMbSerRS:tRNA complex. Although
mMbSerRSs variants with an altered serine-binding pocket (W396A,
N435A, S437A) were more sensitive to inhibition by threonine and cyste-
ine, none of the mutants was able to activate noncognate amino acids to
greater extent than the wild-type enzyme. In vitro kinetics results also
suggest that conformational changes in the motif 2 loop are required for
efficient serylation.
Abbreviations
aaRS, aminoacyl-tRNA synthetase; HTH, helix–turn–helix; mMbSerRS, archaeal (Methanosarcina barkeri) SerRS; Ni-NTA, nickel-nitrilotriacetic
acid; SerRS, seryl-tRNA synthetase; SOL, serine ordering loop.
FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS 2831
creating a fold found only in the class II synthetases
and in their paralogues. Three short conserved
sequence motifs (motifs 1, 2 and 3) are characteristic
of all class II synthetases [1].
The separation of aaRSs into two classes correlates
with the modes of tRNA binding: class I synthetases
approach the acceptor helix from the minor groove
side, whereas class II synthetases approach it from the
opposite, major groove side [2,3]. The functional differ-
ences between the two classes can also be seen in the
mode of ATP binding: the ATP molecule adopts an
extended structure in the class I active site, whereas

bound ATP in the class II catalytic core has a bent
conformation. Recent completion of representative
crystal structures for all 20 aaRSs has created a new
context to determine if synthetases of the same class
share kinetic features that parallel those based on struc-
ture [4,5]. Again, a distinct mechanistic signature has
been shown to divide the two classes of synthetase [6].
Seryl-tRNA synthetase (SerRS), which catalyzes
serylation of the corresponding set of isoaccepting
tRNAs
Ser
, belongs to the class II aaRSs [7]. The crys-
tal structure of Escherichia coli SerRS was the first
example of an aaRS that does not possess the Ross-
mann dinucleotide-binding fold [8]. The enzyme is
homodimeric, with motif 1 involved in dimer interface
contacts. As revealed by the crystal structures of two
prokaryotic seryl-tRNA synthetases from E. coli and
Thermus thermophilus, each subunit possesses a C-ter-
minal active site domain typical for class II aaRSs,
whereas the first 100 N-terminal residues form an anti-
parallel a-helical coiled-coil crucial for the selection
and binding of tRNA [7]. In the crystal structure of
T. thermophilus SerRS complexed with tRNA
Ser
, the
N-terminal helical arm is buried between the TwC arm
and the long extra arm of tRNA
Ser
[9,10]. The crystal

structure of mammalian mitochondrial SerRS from
Bos taurus has an N-terminal domain that also consists
of a long a-helical arm, and extensions in the N- and
C-termini of the enzyme ensure recognition of two
unusual mitochondrial tRNAs
Ser
[11]. In general, the
active sites of seryl-tRNA synthetases contain a very
long loop in motif 2 (the longest of all class II synthe-
tases), involved in binding ATP and the acceptor end
of tRNA [8,10,12].
With the first sequence of archaeal genome (Methan-
ococcus jannaschii) [13], it became apparent that it
encodes an atypical SerRS, later found in all methano-
genic archaea, and thus named methanogenic-type
SerRS [14,15], to be distinguished from SerRSs found
in all other organisms (bacterial-type SerRSs). The
N-terminal region of this enzyme is significantly longer
than the corresponding domain of its bacterial-type
counterparts. Accordingly, the crystal structure of
Methanosarcina barkeri SerRS (mMbSerRS) [12]
revealed two idiosyncratic features: a novel N-terminal
tRNA-binding domain and a zinc ion in the active site.
Biochemical analysis confirmed and absolute require-
ment of zinc for enzymatic activity. In addition, meth-
anogenic-type SerRS is, in comparison with bacterial
homologues, characterized by a notable shortening of
the motif 2 loop, questioning the mode of tRNA bind-
ing. Evidently, bacterial- and methanogenic-type
SerRSs have diverse modes of substrate recognition

[10,15–19]. Although the widely distributed bacterial-
type SerRS has been extensively characterized, the
unusual methanogenic-type still represents an intrigu-
ing puzzle in structure–functional and evolutionary
terms. Using mutational and kinetic analysis, we eva-
luated the importance of mMbSerRS active site
flexibility in the discriminating between substrates and
the proper positioning of the tRNA CCA-end.
Results
Structure-based design of mMbSerRS variants
The positions of particular motifs along the
mMbSerRS polypeptide are shown in Fig. 1. Each
mMbSerRS subunit consists of two domains linked by
a short flexible oligopeptide (L). The N-terminus
(residues 1–165) is a mixed a ⁄ b domain composed of a
six-stranded antiparallel b sheet capped by a bundle of
four helices (H1, H2, H3, H4). This fold is distinct
from the coiled-coil of the tRNA-binding domain in
bacterial-type SerRS, but presumably also interacts
with the tRNA extra arm (see later). The catalytic
module (residues 174–502), built upon eight anti-
parallel strands surrounded by three a helices, repeats
the general structure of the catalytic core in the bacte-
rial enzyme. The catalytic domain contains three
class II signature motifs, designated M1, M2 and M3
in Fig. 1A. M1 motifs participate at the dimer inter-
face, whereas M2 and M3 contain residues involved in
serine, ATP and tRNA 3¢-end interactions. The tetra-
coordinated Zn
2+

ion is bound to three conserved
protein ligands (Cys306, Glu355 and Cys461; marked
with asterisks in Fig. 1A) and a water molecule, which
dissociates from the zinc ion to allow coordination of
the amino group of the serine substrate (Fig. 1B).
Accordingly, biochemical analysis revealed a loss of
enzyme function as a consequence of alterations in
zinc-binding residues [12]. In this study, we aimed to
characterize the other active site residues, which
according to the crystal structure may influence the
specificity of substrate recognition. These are marked
mMbSerRS active site mutants S. Bilokapic et al.
2832 FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS
with a black lozenge in Fig. 1A. The mutations were
generated using site-specific in vitro mutagenesis of the
mMbSerRS gene, and N-terminally His-tagged vari-
ants with alanines (or valine in one case) at the alter-
ation sites were expressed and purified from E. coli.
All mutated enzymes possessed a characteristic class II
dimeric nature, as confirmed by gel-filtration chroma-
tography (not shown). CD spectra of the wild-type
and all mutants were similar, confirming that the over-
all fold is maintained (Fig. 2). In addition, selected
mutants were crystallized and the structures were
essentially identical to that of the wild-type enzyme
(not shown).
Kinetic analysis of amino acid activation
by mMbSerRS variants with an altered
serine-binding pocket
Serine binding in wild-type mMbSerRS is accompanied

by a significant localized conformational change in the
A
B
Fig. 1. Structural and functional features of mMbSerRS. (A) Idiosyncratic and general structural–functional motifs along the polypeptide chain
of one subunit. H1–H4 represent four idiosyncratic helices in the N-terminal domain, L is a linker region, M1, M2 and M3 are class II signa-
ture motifs, HTH and SOL denote idiosyncratic oligopeptide insertions into the mMbSerRS catalytic core. Zinc-ion ligands are marked by
asterisks (*). A black lozenge denotes other active site residues that are, according to a crystal or model structure, presumed to be impli-
cated in the specificity of substrate recognition. These are changed to alanine (or valine) in this study. (B) View of the active site-bound
seryl-adenylate analogue. The characteristic extended conformation of seryl-adenylate in the active site of mMbSerRS can be observed.
Residues that interact with the analogue are indicated. The active site zinc ion is in cyan.
200 210 220 230 240 250 260
–12
–10
–8
–6
–4
–2
0
2
4
6
8
[Θ]X 10
3
(deg·cm
2
· dmol
–1
)
Wavelen

g
th (nm)
Fig. 2. CD spectra of wild-type mMbSerRS
(h) and its mutants: N435A (s), S437A (n),
W396A ⁄ N435A S437A (,), R267A (X),
E338A (e), R347A (+).
S. Bilokapic et al. mMbSerRS active site mutants
FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS 2833
‘serine ordering loop’ (SOL, residues 394–410; Fig. 1),
which brings the loop into the proximity of the zinc ion
and enables direct contact between Gln400 and the
carbonyl oxygen of the serine substrate [12]. These
movements are required to position the carboxylate
oxygen for nucleophilic attack of the a-phosphate of
ATP. Thus, according to the crystal structure, the spec-
ificity of serine recognition depends on: (a) the zinc ion,
(b) the size of the active site and (c) the hydrophilic
nature of the serine-binding pocket. Despite precise
recognition of the cognate amino acid, functional
assays have shown slight, but notable, misactivation of
threonine by mMbSerRS, which does not seem to be
edited in vitro [12]. To obtain further insight into the
discrimination strategy, we designed several mMbS-
erRS mutants and tested their ability to activate serine
and misactivate noncognate amino acids. Because the
size of the active site depends predominantly on
Trp396, which is located at the bottom of the serine-
binding subsite and packs above the amino group of
the serine substrate (Fig. 1B), by substituting alanine,
we expected to produce a variant with an enlarged

active site that was capable of accommodating a larger
noncognate amino acid. However, formation of a
hydrophilic environment around Asn435 and Ser437
(which coordinate a structurally conserved water mole-
cule) may reduce the binding of amino acids with
hydrophobic side chains (e.g. threonine). Accordingly,
the mMbSerRS triple-mutant (W396A ⁄ N435A ⁄ S437A)
was expected to combine important misactivating
features, namely a larger binding site and increased
hydrophobicity.
The kinetic parameters for serine activation were
determined using the pyrophosphate-exchange reaction
for the wild-type and five mMbSerRS variants with
alterations in the amino acid-binding pocket (W396A,
N435A, S437A, the double-mutant N435Al ⁄ S437A
and the triple-mutant W396A ⁄ N435A ⁄ S437A). All
mutants showed significant reductions in catalytic effi-
ciency (k
cat
⁄ K
m
) for serine compared with wild-type
mMbSerRS (Table 1). In accordance with structural
data and the sequence conservation of Trp396
(replacement by Phe in some methanogenic-type
SerRSs is presumably functionally equivalent), muta-
tion of this residue had deleterious effects on k
cat
and
K

m
for serine in seryl-adenylate synthesis (Table 1).
Replacing Asn435 does not affect the affinity for serine
in the first step of the aminoacylation reaction as much
as mutation of Trp396. However, catalytic efficiency is
seriously hampered. The alteration S437A leads to the
K
m
value for serine being elevated by two orders of
magnitude. This residue is highly conserved in all
SerRSs, although its direct interaction with serine has
not been observed in bacterial- and methanogenic-type
SerRS structures [12,16]. CD spectrometry and X-ray
analysis showed that the observed kinetic effect
(Table 1) is not due protein misfolding (Fig. 2).
Having established the effect of alterations in
both Asn435 and Ser437 on serine recognition, we
investigated the catalytic properties of the enzyme
using double substitutions. The variant N435A ⁄ S437A
displayed kinetic parameters similar to those deter-
mined for S437A (Table 1), suggesting that the large
reduction in the affinity for serine was caused primar-
ily by the mutation S437A. Although mutations
N435A ⁄ S437A altered the chemical properties of the
amino acid binding pocket significantly, the substrate
specificity of the double mMbSerRS mutant was not
relaxed. Activation of noncognate amino acids in the
ATP–PP
i
exchange reaction was not increased in com-

parison with wild-type enzyme (not shown). This
mutant showed decreased activation efficiency towards
both cognate and noncognate substrates (serine and
threonine, respectively). Although the triple-mutant
W396A ⁄ N435A ⁄ S437A showed a 10
3
-fold decrease in
serine affinity (Table 1) its k
cat
value was only three
times lower than for the wild-type enzyme, suggesting
a compensatory effect for the mutations. It seems that
in an enlarged amino acid-binding subsite orientation
of the carboxylate group is preserved and the attack
on ATP is presumably facilitated. An electron-density
map for the mutated enzyme in complex with serine
Table 1. Kinetic parameters for serine in ATP–PP
i
exchange reaction.
mMbSerRS K
m
(lM) k
cat
(s
)1
) k
cat
⁄ K
m
(s

)1
ÆlM
)1
) k
cat
⁄ K
m
(s
)1
ÆM
)1
)(k
cat
⁄ K
m
)
rel
a
Wild-type 6.61 ± 1.325 2.06 ± 0.208 0.31143 311430.7 1
W396A 264.5 ± 12.55 0.287 ± 0.0631 0.00108 1083.932 287.32
N435A 70.28 ± 0.280 0.172 ± 0.0167 0.00245 2448.776 127.18
S437A 573.8 ± 69.50 0.729 ± 0.169 0.00127 1271.349 244.96
N435A S437A 689.5 ± 15.25 0.576 ± 0.0283 0.00084 835.388 372.79
N435A ⁄ S437A ⁄ W396A 2879 ± 303.5 1.32 ± 0.2421 0.00046 459.535 677.71
a
(k
cat
⁄ K
m
)

wt
⁄ (k
cat
⁄ K
m
)
mut
.
mMbSerRS active site mutants S. Bilokapic et al.
2834 FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS
(data not shown) revealed the ordering of 394–410
loop residues and interactions with Gln400, despite
replacing Trp396 with alanine.
The influence of mutations in the serine-binding
pocket on magnesium binding
The characteristic conformation of ATP in the class II
active site is usually stabilized by hydrogen bonds with
motif 2 and motif 3 conserved arginine residues and
three Mg
2+
ions. In addition to phosphate groups,
acidic residues from the active site and water molecules
coordinate Mg
2+
ions. Interestingly, an exception has
been observed in the structure of HisRS: two magne-
sium cations are bound in the active site and a HisRS-
specific arginine residue occupies the same position as
the third Mg
2+

and takes over its role in binding the
a-phosphate of the ATP molecule [20].
Magnesium ions play a crucial role in the aminoa-
cylation reaction because they decrease and delocalize
the negative charge of the triphosphate moiety of
ATP, which is in close proximity to the negatively
charged carboxylate group of the amino acid sub-
strate [21]. The binding of ATP and the amino acid
substrate in the active site are linked, and incorrect
binding of ATP can affect serine binding and vice
versa [22]. Asn435, together with Asp416 and Glu432,
contributes to binding of the Mg
2+
ion between the
a- and b-phosphates of ATP. Indeed, a 10-fold
increase in the K
m
value for serine has been observed
for the N435A mutant compared with the wild-type
enzyme. Therefore, we assayed the activity depen-
dence of mMbSerRS variants comprising N435A
replacement on magnesium concentration in the
ATP–PP
i
exchange reaction at saturating levels of
serine and ATP substrates. The double-mutant
N435A ⁄ S437A and the triple-mutant N435A ⁄ S437A ⁄
W396A showed maximal reaction activity at magne-
sium concentrations that were higher than for the
wild-type enzyme or W396A mutant, supporting the

role of Asn435 in magnesium binding (Fig. 3).
Kinetic analysis of aminoacylation by mMbSerRS
variants with an altered serine-binding pocket
The catalytic parametars of W396A, N435A, S437A,
the double-mutant N435Al ⁄ S437A and the triple-
mutant W396A ⁄ N435A ⁄ S437A were significantly dif-
ferent from those of the wild-type mMbSerRS in serine
activation. The kinetic properties of the active site
mutants were further investigated in the aminoacyla-
tion reaction. The production and characterization of
different tRNAs
Ser
substrates is described below. The
data reveal that alteration in the SOL (W396A) has a
more pronounced effect on adenylate synthesis than
on tRNA aminoacylation (Table 2). Also, mutations in
Asn435 and Ser437 have a less deleterious influence on
tRNA aminoacylation. This suggests that tRNA bind-
ing stabilizes the active site despite perturbations
caused by the mutations.
Although only four class I aaRSs, glutamyl-,
glutaminyl-, arginyl- and lysyl-tRNA synthetases (GluRS,
GlnRS, ArgRS and LysRS, respectively), are known to
require their cognate tRNA for amino acid activation,
a number of other synthetases (including several
class II representatives) [1], use tRNA to optimize the
amino acid binding pocket. Indeed, tRNA-mediated
amino acid recognition has been documented for yeast
SerRS [23,24]. Our results show that tRNA also con-
tributes to optimization of the active site in methano-

genic-type SerRS, which can be detected when kinetic
parameters for the mutated proteins are measured.
Selectivity of the active site towards amino
acid substrates
Based on previous biochemical and crystallographic
studies [12], we assumed that of the 20 amino acids
alanine, glycine, threonine and cysteine may potentially
bind into the active site of methanogenic-type SerRS
and cause mischarging problems. The ability of these
noncognate amino acids to impair the specific aminoa-
cylation of tRNA
Ser
with serine was measured for
wild-type and mutant mMbSerRS enzymes (Fig. 4).
Alanine and glycine were chosen because they are
smaller than serine; yet they do not have similar chem-
ical properties. Valine was used as a positive control
0
0.2
0.4
0.6
0.8
1
1.2
0 5 10 15 20 25
c(Mg)/mM
Relative reaction activity
Fig. 3. Comparison of the activity dependence of wild-type and
mMbSerRS mutants on magnesium concentration. Activity was
assayed in an ATP–PP

i
exchange reaction: wild-type (r), W396A
(
), N435A ⁄ S437A ( ), W396A ⁄ N435A ⁄ S437A (s).
S. Bilokapic et al. mMbSerRS active site mutants
FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS 2835
because it is bigger and cannot form the H-bonds that
are characteristic of serine in the active site. Threonine
and cysteine were used because of their chemical simi-
larity with serine. Binding of threonine may be particu-
larly likely because of the similarity between the
mMbSerRS and threonyl-tRNA synthetase (ThrRS)
catalytic domains, both of which contain a zinc ion.
Serylation activity of the wild-type protein was 10%
lower in the presence of cysteine, and the inhibition of
N435A ⁄ S437A and W396A ⁄ N435A ⁄ S437A mutants
was 55% and 70%, respectively (Fig. 4). Interestingly,
the noncognate amino acid was not activated in the
ATP–PP
i
assay, indicating that the inhibition of seryla-
tion was not caused by cysteinyl-tRNA
Ser
formation.
Cysteine may be a competitive inhibitor of aminoacy-
lation in the presence of tRNA or the –SH group
of cysteine may attack serylated-tRNA, resulting
in the formation of a serine–cysteine dipeptide and
deacylated tRNA, in accordance with the hypothesis
that aaRSs are capable of performing thioester-depen-

dent peptide synthesis [25,26]. It remains to be deter-
mined which of two proposed scenarios is actually
taking place in methanogenic-type SerRS.
In addition, inhibition of serylation with 6 mm
threonine was 10–20% higher with mutated mMbSerRS
variants than with wild-type enzyme. Again, the
mutants did not misactivate threonine (or other
noncognate amino acids) more efficiently than the wild-
type enzyme (data not shown). It seems that noncognate
substrates are not bound in the proper orientation,
which would allow their carboxyl group to attack the
a-phosphate of bound ATP and lead to the completion
of the first step of the aminoacylation reaction.
tRNA acceptor-stem binding by mMbSerRS
motif 2 loop residues and the design of the
mutants
Multiple sequence alignments of members of two
distinct groups of SerRS showed that the motif 2
loop sequence, which according to crystal structures
participates in acceptor-stem contact, is highly con-
served within each type, but differs between them [12].
The crystal structure of T. thermophilus SerRS in com-
plex with cognate tRNA revealed that Phe262 from
this long loop interacts with the fifth base pair of the
acceptor stem [10]. In contrast to bacterial-type SerRS,
the discriminator base (G73) and the first base pair
in the acceptor stem were shown to be important
determinants of specific tRNA
Ser
recognition in metha-

nogenic-type enzyme [15]. We assumed that the differ-
ences in acceptor-stem recognition between the two
SerRS types are due to differences in the motif 2 loop
sequence and that each motif 2 loop is capable
of performing different but specific interactions with
cognate tRNA.
Because the acceptor end of tRNA
Ser
in the T. ther-
mophilus SerRS:tRNA co-crystal structure is disor-
dered, we made a mMbSerRS:tRNA docking model
using the crystal structures of ThrRS in complex with
the cognate tRNA. Because of the similar size of the
motif 2 loop in methanogenic-type SerRS and ThrRS
and the involvement of zinc ions in the recognition of
amino acid substrates, the M. barkeri loop was homol-
Table 2. Kinetic parameters for serine in aminoacylation reaction. The tRNA used in these experiments was M. barkeri tRNA
Ser
overpro-
duced in E. coli.
mMbSerRS K
m
(lM) k
cat
(s
)1
) k
cat
⁄ K
m

(s
)1
ÆlM
)1
) k
cat
⁄ K
m
(s
)1
ÆM
)1
)(k
cat
⁄ K
m
)
rel
a
Wild-type 11.13 ± 1.310 0.473 ± 0.0323 0.04247 42470.80 1
W396A 760.2 ± 92.74 0.239 ± 0.0320 0.00032 315.049 134.80
N435A 283.1 ± 14.91 0.435 ± 0.0322 0.00154 1535.853 27.65
S437A 412.9 ± 75.04 0.396 ± 0.0297 0.00096 958.828 44.29
N435A ⁄ S437A 509.9 ± 48.49 0.361 ± 0.0113 0.00071 707.394 60.038
N435A ⁄ S437A ⁄ W396A 1191 ± 299.2 0.175 ± 0.0348 0.00015 146.600 289.70
a
(k
cat
⁄ K
m

)
wt
⁄ (k
cat
⁄ K
m
)
mut
.
0
20
40
60
80
100
120
Activity %
val
g
ly ala cys thr
Fig. 4. Inhibition of mMbSerRS with noncognate amino acids.
Activity of wild-type mMbSerRS (black bar) and different mutants
(W396A, dark gray; N435A ⁄ S437A, gray; W396A ⁄ N435A ⁄ S437A,
white) in the presence of other amino acids. The level of inhibition
was determined as a ratio of initial velocities of inhibited (6 m
M
noncognate amino acid) and uninhibited reactions.
mMbSerRS active site mutants S. Bilokapic et al.
2836 FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS
ogy modeled based on the structure of the E. coli

ThrRS loop. We observed that the positioning of the
cognate tRNA in the active site of mMbSerRS would
be facilitated upon the conformational change of the
motif 2 loop, which was the only proximal protein
region in the proposed mMbSerRS:tRNA complex
able to mediate the interactions with the tRNA accep-
tor stem. The model used to predict the amino acids
that might participate in the binding of the 3¢-end of
tRNA, which were therefore subjected to mutational
and kinetic analyses, is shown in Fig. 5.
Kinetic parameters for M. barkeri tRNA
Ser
amino-
acylation by wild-type and mutant mMbSerRS enzymes
are given in Table 3. The results are consistent with
the involvement of several motif 2 residues (Glu338,
Arg347, Gly340 and Gly341) in the serylation reaction.
Loss of the catalytic efficiency of mutant E338A is
essentially the consequence of a decreased k
cat
value
(Table 3), and is in agreement with previous studies on
aspartyl-tRNA synthetase (AspRS) [27,28] and LysRS
[29]. Residue Glu338 in mMbSerRS is presumably
functionally equivalent to Glu258 in T. thermophilus
SerRS. The crystal structure of bacterial binary SerRS
complexes reveals an interaction between Glu258 and
N6 of ATP or Ser-AMS. Upon tRNA binding, this
residue forms a hydrogen bond with N2 of G73. How-
ever, the side chain of Glu338 may adopt a different

conformation, which would facilitate interactions with
the O2¢ of C74 ribose, analogous to those observed in
the structure of the E. coli AspRS:tRNA complex.
Because G73 is a discriminator base in mMbSerRS,
specific interactions, in addition to Glu338, must con-
tribute to the importance of G73. The class-invariant
Arg347 from the motif 2 loop in mMbSerRS is posi-
tioned to interact with N1 of G73. It cannot be
excluded, however, that upon tRNA binding the orien-
tation of Arg347 side chain changes, allowing recogni-
tion of C74 as in other class II synthetases. The crystal
structure of mMbSerRS shows that Arg347 and
Glu338 also form contacts with ATP (Fig. 1B). Our
kinetic results revealed that replacement of these two
residues with alanine affects tRNA aminoacylation
(Table 3), which is consistent with the involvement of
these two residues in both steps of the aminoacylation
reaction in all class II synthetases [30]. The R347A
variant shows a greater reduction in k
cat
than the
E338A variant (8-fold for E338A vs. 175-fold for
R347A). Moreover, Arg347 also contributes slightly to
the binding of tRNA, as deduced from the twofold
decrease in K
m
(Table 3).
Strikingly, two glycines in succession, Gly340 and
Gly341, in the motif 2 loop are conserved among
methanogenic-type SerRSs. We analyzed the functional

significance of this apparent flexibility for tRNA
CCA-end binding and recognition of the G1:C72 iden-
tity determinant. In agreement with our model, the
G340V ⁄ G341A mutant showed a 410-fold diminished
Fig. 5. Model of the acceptor end of tRNA
Ser
bound in the active
site of mMbSerRS. The view is focused on the acceptor part of
tRNA in the active site of mMbSerRS. The tRNA is shown as an
orange tube with the CCA-end and G73 shown as sticks. The cata-
lytic domain of mMbSerRS in complex with Ser-AMS is shown in
blue. The structure of mMbSerRS with the proposed motif 2 loop
conformation upon tRNA binding is depicted in gray. Residues that
participate in the interaction with tRNA and that were tested are
shown as sticks. It can be seen that the side chain of Ile342 points
away from tRNA molecule. All molecular depictions were produced
using
PYMOL ().
Table 3. Kinetic parameters for M. barkeri tRNA
Ser
aminoacylation with wild-type and mutant mMbSerRS enzymes. The tRNA used in these
experiments was M. barkeri tRNA
Ser
overproduced in E. coli.
mMbSerRS K
m
(lM) k
cat
(s
)1

) k
cat
⁄ K
m
(s
)1
ÆlM
)1
) k
cat
⁄ K
m
(s
)1
ÆM
)1
) k
cat ⁄
K
m
rel
Wild-type 2.99 ± 0.31 0.454 ± 0.018 0.152076 152076.4 1
R267A 3.40 ± 0.39 0.032 ± 0.0020 0.009436 9435.79 16.12
E338A 2.17 ± 0.14 0.059 ± 0.0043 0.027221 27220.68 5.59
R347A 6.38 ± 1.12 0.0026 ± 0.00046 0.000403 403.13 377.23
G340V ⁄ G341A 6.76 ± 0.55 0.0011 ± 9.42 · 10
)5
0.000171 171.28 887.9
S. Bilokapic et al. mMbSerRS active site mutants
FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS 2837

catalytic efficiency and a twofold higher K
m
(Table 3).
Thus, the flexibility of this region might be important
in avoiding steric clashes with tRNA and allowing
hydrogen bonding of the first base pair.
The crystal structure of mMbSerRS revealed that
the unique  30-residue insertion between motifs 1
and 2 in methanogenic-type SerRSs adopts a helix–
turn–helix (HTH) fold. The HTH fold of one mono-
mer is positioned above the catalytic core of the other.
Interestingly, Arg267 from the beginning of helix 9 in
the HTH fold of one monomer is positioned towards
the active site of the other monomer and can contrib-
ute to interactions with the tRNA acceptor end. In
the proposed model, Arg267 coordinates the ribose
O2¢ of C74. Substitution of Arg267 with alanine
significantly decreased the k
cat
value for tRNA
aminoacylation (Table 3). Amino acids subjected to
alterations in G340V ⁄ G341A and R267A mutants
presumably interact with tRNA in the transition state
and contribute to structural rearrangements that result
in new contacts between the acceptor end of tRNA
and the active site of synthetase. Furthermore, Gly340
and Gly341 from the motif 2 loop also contribute to
the affinity for tRNA. Importantly, our kinetic data
indicate that Arg267 from the HTH-motif, belonging
to one monomer of the dimeric mMbSerRS, is essen-

tial for tRNA aminoacylation in the active site of the
other monomer. The crystal structures of mMbSerRS
in complex with ATP or the seryl-adenylate analogue
do not reveal any interactions between this arginine
and small substrates. Therefore, detected changes in
the kinetic parameters for this mutant can only be due
to tRNA CCA-end binding (Table 3). Thus, Arg267
from the HTH-motif seems to be essential for the
second, but not the first, step in the aminoacylation
reaction.
Wild-type enzyme and mutants were tested for their
ability to bind tRNA using a gel-shift assay. All pro-
teins were able to shift the cognate tRNA, although
some mutants form weaker complexes, especially
R347A, as shown in Fig. 6.
Variability of tRNA substrates for mMbSerRS
We encountered many difficulties in preparing suffi-
cient quantities of tRNA substrates to determine the
kinetic parameters with mMbSerRS variants. Because
of the very inefficient in vitro transcription of M. bark-
eri synthetic tRNA
Ser
genes we decided to attempt to
overproduce M. barkeri tRNAs
Ser
in E. coli. Both
tRNA
CGA
and tRNA
GGA

isoacceptors were found to
be well expressed in vivo. K
m
values for serylation were
estimated for the expressed tRNAs and for in vitro
transcribed tRNA
CGA
as a reference. The K
m
and k
cat
values for both isoacceptors were similar and compara-
ble with tRNA
Ser
obtained by in vitro transcription
(Table 4). It was thus evident that the M. barkeri
tRNAs
Ser
were processed correctly in the bacterial
host.
tRNA
GGA
acceptor made up 50–60% of the total
tRNA and was used for subsequent experiments because
it was better expressed than tRNA
CGA
, which contrib-
uted only 25–30%. Endogenous E. coli tRNA
Ser
, 3.5%

in the same preparation, was negligible in comparison
with the overexpressed M. barkeri tRNA
GGA
. Because
of the 25-fold higher k
cat
value (Table 4), E. coli
tRNA
Ser
was not expected to influence the kinetic results
obtained with in vivo produced M. barkeri tRNA
Ser
.To
confirm this, in vivo expressed M. barkeri tRNA
GGA
was purified on urea PAGE. The kinetic parameters
obtained with pure M. barkeri tRNA
GGA
and mMbS-
erRS were the same as for the ‘crude’ sample (data not
shown).
Methanococcus maripaludis tRNA
GCU
gene tran-
script was not efficiently recognized by wild-type
mMbSerRS (Table 4), which is consistent with struc-
tural differences between homologous and heter-
ologous serine-specific tRNAs in the D-loops.
Although M. barkeri, in contrast to Me. jannaschii
and Me. maripaludis, does not possess tRNA

Sec
[31],
our experiments revealed charging of Me. jannaschii
tRNA
Sec
by mMbSerRS, comparable with the recogni-
tion of tRNA
Sec
in the human and bacterial systems
(Table 4).
Fig. 6. Native PAGE analysis of complex formation between over-
expressed tRNA
Ser
and mMbSerRS mutants. (Upper) Wild-type
mMbSerRS forms a more stable complex with overexpressed
tRNA
Ser
in comparison with R347A mutant: lane 1, wild-type
mMbSerRS (1.6 l
M); lanes 2–6, wild-type and tRNA (lane 2,
0.12 l
M; lane 3, 0.3 lM; lane 4, 0.6 lM; lane 5, 1.2 lM, lane 6,
2.4 l
M); lanes 7–10, R347A and tRNA (lane 7, 0.12 lM; lane 8,
0.3 l
M; lane 9, 0.6 lM; lane 10, 1.2 lM). (Lower) Complex formation
between tRNA
Ser
(1.2 lM) and different mMbSerRS mutants
(1.6 l

M): lane 1, wild-type mMbSerRS; lane 2, N435A S437A; lane
3, W394A; lane 4, catalytic domain (does not make complex with
tRNA; unpublished results); lane 5, catalytic domain mutant; lane 6,
R347A; lane 7, G340V ⁄ G341A. The gel was stained with Coomas-
sie Brilliant Blue. Arrow denotes the position of the complex.
mMbSerRS active site mutants S. Bilokapic et al.
2838 FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS
Discussion
Switching the specificity for amino acid substrate
Although alteration of the three residues (W396A,
N435A and S437A) in the amino acid binding pocket
of mMbSerRS was expected to increase the level of
threonine misrecognition, no improvement in threonine
activation was observed with mutated enzymes. Thus,
it appears to be far easier to abolish cognate amino
acid acivation than to replace it with noncognate
activation, despite redesigning the active site in terms
of size, polarity and hydrogen-bonding capacity.
Specific amino acid recognition is therefore substan-
tially more resistant to mutation than structure-based
predictions would suggest, as observed previously for
other systems [32]. Our results suggest that engineering
methanogenic-type SerRS in an attempt to change its
substrate specificity is not straightforward and may
possibly be achieved only by multiple alterations of the
active site residues. Furthemore, contemporary metha-
nogenic-type SerRS and ThrRS may have diverged too
far to readily allow the switching of substrate specifici-
ties between the two, as seen for tryptophanyl-tRNA
synthetase (TrpRS) and tyrosyl-tRNA synthetase

(TyrRS) [32]. By contrast, substrate-specificity
switching was possible for GlnRS ⁄ GluRS, a much
more closely related pair of aaRSs, which have only
relatively recently diverged from a common ancestor
[33,34].
Flexibility of the motif 2 loop is required for tRNA
binding
The studies described here were undertaken to test the
role of the motif 2 loop in functional binding of cog-
nate tRNA. Sequence variability in this loop which
exists in different class II aaRSs clearly points to its
role in the selectivity of tRNA recognition. In order to
study binding of the tRNA acceptor end into the
active site of mMbSerRS, we superimposed the ThrRS
in complex with cognate tRNA onto the catalytic core
of methanogenic-type SerRS. The most striking feature
of the obtained model is the severe steric clash between
the acceptor end of tRNA and the motif 2 loop of
mMbSerRS, especially with residue Ile342 (Fig. 7). In
the co-crystal structures of AspRS:tRNA and
ThrRS:tRNA complexes, the conformation of motif 2
loop is closed, in comparison with the mMbSerRS
motif 2 loop. The clash seen in the mMbSerRS:tRNA
model, combined with the results of our mutational
and kinetic analyses, suggest that a conformational
change in this loop is required upon tRNA binding.
Recognition by methanogenic SerRS relies, in addi-
tion to the long extra arm, on G1:C72 in the cognate
tRNA, which is achieved by the motif 2 residues
(Figs 5 and 7). Two successive glycines in the motif 2

loop are conserved among all the methanogenic-type
SerRSs. Our biochemical experiments point to the
importance of these residues for the flexibility of the
tRNA 3¢-end binding region. The ability of the loop to
change its conformation upon tRNA binding is crucial
for correct positioning of the tRNA acceptor end in
the active site. However, whether these residues also
contribute to specific interactions with the tRNA
G1:C72 pair remains to be investigated.
Role of motif 2 and the SOL in the two steps of
the aminoacylation reaction
Strucutural [12] and kinetic data (Table 3) showed that
the SOL has a crucial role in cognate amino acid
Table 4. Kinetic parameters for wild-type mMbSerRS enzyme with various tRNA substrates in aminoacylation reaction. Mb, Mm and Mj
denote M. barkeri, M. maripaludis and Me. jannaschii, respectively.
tRNA K
m
(lM) k
cat
(s
)1
) k
cat
⁄ K
m
(s
)1
ÆlM
)1
) k

cat
⁄ K
m
(s
)1
ÆM
)1
) k
cat
⁄ K
m
rel
a
In vivo
Mb tRNA
Ser
GGA
2.99 ± 0.31 0.454 ± 0.018 0.1518395 151839.5 1
In vivo
Mb tRNA
Ser
CGA
3.07 ± 0.52 0.467 ± 0.038 0.1521173 152117.3 0.998
In vitro transcript
Mb tRNA
Ser
10.10 ± 1.67 0.439 ± 0.111 0.0434653 43465.3 3.49
In vitro transcript
Mm tRNA
Ser

3.316 ± 0.129 0.079 ± 0.027 0.0238239 2382.39 63.73
tRNA
Ser
from
total E. coli tRNA
2.34 ± 0.37 0.0175 ± 0.0015 0.0074786 7478.63 19.74
In vitro transcript
Mj tRNA
Sec
2.41 ± 0.04 0.005 ± 0.0006 0.0020746 2074.69 73.19
a
k
cat
⁄ K
m
(in vivo Mb tRNA
Ser
GGA
) ⁄ k
cat
⁄ K
m
(tRNA).
S. Bilokapic et al. mMbSerRS active site mutants
FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS 2839
binding to mMbSerRS. In addition, the motif 2 loop
participates in ATP binding and recognition of the
tRNA 3¢-end in all class II aaRSs. The motif 2 loop of
T. thermophilus SerRS is disordered in the absence of
substrate. It adopts two different conformations upon

substrate binding: an ‘A-conformation’ in the presence
of ATP or adenylate or a ‘T-conformation’ when
tRNA
Ser
is bound. In mMbSerRS, the motif 2 loop is
fully ordered in an apo-enzyme structure. However, in
the presence of ATP or seryl-adenylate the motif 2
loop shifts (Fig. 8). In addition, the side chains of the
conserved arginine residues in the motif 2 loop show a
concerted movement following substrate binding in the
active site. In the apo-enzyme, the motif 2 Arg336
occupies the adenine-binding site and upon binding of
ATP or Ser-AMS shifts to form interactions with the
a-phosphate group. Arg336 also interacts with the
carbonyl oxygen of Ser-AMS, but not with serine.
Therefore, this residue can sense the presence of both
substrates needed for the first step of the amino-
acylation reaction.
Fig. 7. Proposed model of the mMbSerRS:tRNA complex. Overall
view with ThrRS:tRNA
Thr
complex (1qf6) shown as a ribbon repre-
sentation and colored in orange. The apo-structure mMbSerRS
(2cim) is shown as a blue transparent surface. For clarity, only the
tRNA CCA-end and the first two base pairs are shown as sticks.
Two synthetases are superposed by means of their catalytic
domains. The view reveals steric clashes between the mMbSerRS
motif 2 loop and the acceptor end of tRNA (circled). The gray arrow
indicates the likely movement of the mMbSerRS N-terminal domain
upon tRNA binding. (Inset) Comparison of the motif 2 loops in

AspRS:tRNA (1c0a; red). ThrRS:tRNA (1qf6; orange) and mMbS-
erRS (2cim; blue) structures, resulting from the superpositions of
their entire catalytic domains. The backbone of the tRNA
Thr
accep-
tor arm is shown (tRNA
Asp
has been omitted for clarity). The open
conformation of motif 2 loop in apo-structure of mMbSerRS causes
clashes with tRNA, especially Ile342.
r.m.s.d
Serine
AT P
Ser-AMS
Residues around
serine loop
Beginning of
motif 3
Motif 2
loop
Helix 9
r.m.s.dr.m.s.d
Fig. 8. Conformational changes in the active site upon binding of
the substrates. The r.m.s.d of C
a
atoms after superposition of the
catalytic domains in complex with different substrates and the cata-
lytic domain of the apo-enzyme. The binding of the small substrates
in the active site leads to conformational changes in the motif 2
loop and SOL. The serine loop residues were excluded from calcu-

lation of the r.m.s.d. because they are not visible in the apo-struc-
ture or the structure of the enzyme in complex with ATP.
However, it can be seen that the residues that are enclosing this
loop are flexible and adopt a range of conformations.
mMbSerRS active site mutants S. Bilokapic et al.
2840 FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS
The conserved motif 2 loop arginine, Arg347, also
contributes to ATP binding. In the apo-structure, it
interacts with Glu338 from the motif 2 loop. Upon
ATP binding, Arg347 swings toward the bound sub-
strate and stabilizes its bend conformation by inter-
acting with the c-phosphate. Following the formation
of seryl-adenylate, Arg347 moves from its previous
position because it no longer interacts with the c-phos-
phate of the ATP molecule. Thus, the pyrophosphate
molecule can be released from the active site, terminat-
ing the amino acid activation. Unlike Arg336, Arg347
does not stabilize Ser-AMP in the active site. Interest-
ingly, successful completion of the serine activation
step releases Arg347 for interaction with the CCA-end
of bound tRNA
Ser
.
Our tRNA-binding model suggests that conforma-
tional changes in the SOL have to occur upon both
tRNA and serine binding. After superposition of the
ThrRS:tRNA
Thr
complex onto the catalytic core of
mMbSerRS, 3¢-terminal A is stacked, as expected, on

the class II-invariant Arg336. The same stacking inter-
actions have been observed in the ThrRS:tRNA
Thr
complex. However, in our model, A76 of mMbSerS
clashes with several residues from the SOL. Therefore,
it can be assumed that, upon tRNA binding, the SOL
has to adopt a new conformation, different from the
one with serine or Ser-AMS in the active site. The side
chain of Glu338 presumably forms interactions with
the incoming tRNA substrate, as in bacterial-type
SerRS, AspRS and phenyl-tRNA synthetase (PheRS).
The conformational change may aid release of the
by-product AMP after the second step of the aminoa-
cylation reaction. The new SOL conformation may
also lead to breaking of the interactions between
serine, Gln400 and Trp396, facilitating the release of
charged tRNA. Therefore, a conformational switch in
the SOL in complexes with serine or tRNA bound to
the active site, as well as the changes in the motif 2
loop residues with the respect to ATP and tRNA
binding, control the two steps of the aminoacylation
reaction.
Experimental procedures
Site-directed mutagenesis and purification of the
proteins
The seryl-tRNA synthetase expression vector (pET15b-
mMbSerRS) has been reported previously [15]. We carried
out site-directed mutagenesis using a Quick Change muta-
genesis kit (Stratagene; La Jolla, CA, USA) and primers
listed in Table 5. Mutations were confirmed by DNA

sequencing. Mutated proteins were expressed in E. coli and
purified as reported previously for the wild-type enzyme
[12].
tRNA cloning and preparation
Me. maripaludis tRNA
Ser
GCU
, Me. jannaschii tRNA
Sec
UCA
,
M. barkeri tRNA
Ser
CGA
and tRNA
Ser
GGA
were prepared by
in vitro transcription of their synthetic genes and purified
by electrophoresis on denaturing polyacrylamide gels [17].
The concentration of the tRNA transcript was estimated by
the absorption at 260 nm (1 absorption unit = 40 mgÆL
)1
)
and corrected by plateau charging with homologous
SerRSs.
Table 5. Sequence of PCR primers used for cloning of mutated mMbSerRS genes. Changes introduced with primer are shown in bold. Tri-
ple-mutant W396A ⁄ N435A ⁄ S437A was obtained using combination of primers for N435A ⁄ S437A and W396A.
Mutant Location in mMbSerRS Primer sequence
R267A HTH 5¢-CCTCCGCAGACA

GCGGATCCCGATTACTGG
5¢-CCAGTAATCGGGATCC
GCTGTCTGCGGAGG
E338A Motif 2 loop 5¢-CCCACAGGTATG
CGAGTGGTGGAATTCACGG
5¢-CCGTGAATTCCACCACTC
GCATACCTGTGGG
G340V ⁄ G341A Motif 2 loop 5¢-CCACAGGTATGAGAGTG
TTGCAATTCACGGAATCGAAAGG
5¢-CCTTTCGATTCCGTGAATT
GCAACACTCTCATACCTGTGG
R347A Motif 2 loop 5¢-CGGAATCGAA
GCGGTCGACGAGTTCCACAGG
5¢-CCTGTGGAACTCGTCGACC
GCTTCGATTCCG
W396A SOL 5¢-GAAAAGCAAGAGTTACCCCC
GCGTTTATGGCACAGGAAG
5¢-CTTCCTGTGCCATAAAC
GCGGGGGTAACTCTTGCTTTTC
N435A Between SOL and M3 5¢-GCTTGAGTTCCAG
GCTGTGAGCATCAATGGAGATAAGTATC
5¢-GATACTTATCTCCATTGATGCTCACA
GCCTGGAACTCAAGC
S437A Between SOL and M3 5¢-GGCTTGAGTTCCAGAATGTG
GCCATCAATGGAGATAAG
5¢-CTTATCTCCATTGATG
GCCACATTCTGGAACTCAAGCC
N435A ⁄ S437A Between SOL and M3 5¢-AATGGCTTGAGTTCCAG
GCTGTGGCCATCAATGGAGATAAGTATC
5¢-ATACTTATCTCCATTGATG

GCCACAGCCTGGAACTCAAGCCATTC
W396A ⁄ N435A ⁄ S437A SOL & between SOL and M3 Combination of primers for N435A ⁄ S437A and W396A
S. Bilokapic et al. mMbSerRS active site mutants
FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS 2841
Synthetic genes for M. barkeri tRNA
Ser
CGA
and
tRNA
Ser
GGA
with the isopropyl thio-b-d-galactoside-induc-
ible T7 promoter at their 5¢-end were constructed and
cloned between the SalI and BamHI sites of pET3a
transcription plasmid [35] and behind the terminator for
the T7 RNA polymerase, generating pET3aGCA and
pET3aCGA plasmid.
Production of in vivo-expressed wild-type tRNAs was
carried out by growing a culture of an expressing clone at
37 °C in Luria–Bertani medium supplemented with ampicil-
lin. When D = 0.7, expression of tRNA was induced by
adding isopropyl thio-b-d-galactoside to a final concentra-
tion of 1 mm followed by incubation at 37 °C. After 4 h of
incubation, 500 mL bacterial cultures overproducing
M. barkeri tRNA
Ser
were harvested by centrifugation and
resuspended in 16 mL of 0.3 m sodium acetate (pH 5.2).
The RNA fraction was extracted by phenol ⁄ chloroform
treatment followed by ethanol precipitation. A crude sam-

ple of total cellular tRNAs was deacylated and purified on
a DEAE–Sephacel column. Between 25% and 60% of total
tRNA could be charged with [
14
C]serine. The content of
endogenous E. coli tRNAs
Ser
in the preparation was deter-
mined by charging the total tRNA with overexpressed
E. coli SerRS. E. coli tRNAs
Ser
contributed 3.5% of the
total tRNA and was therefore negligible in comparison
with the overexpressed tRNA
Ser
. In addition, E. coli SerRS
did not detectably aminoacylate overexpressed M. barkeri
tRNA
Ser
or M. barkeri tRNA
Ser
transcripts.
ATP–PP
i
exchange assay
ATP–PP
i
exchange was measured at 37 °C in 100 mm He-
pes pH 7.6, 20 mm MgCl
2

,50mm KCl, 4 mm ATP, 4 mm
dithiothreitol, 1 mm
32
P-PP
i
(0.002–0.01 mCiÆmL
)1
). Serine
concentrations varied between 0.1 and 10 · K
m
. The
enzyme concentrations used were 200 nm for wild-type and
1 lm for serine-binding mutants. After stopping the reac-
tion, radioactive products were separated by TLC and
quantified as described previously [36]. Kinetic parameters
were determined by fitting the initial velocity and substrate
concentration data to the Michaelis–Menten equation using
nonlinear regression with the program graphpad prism 4
(GraphPad Software, Inc., San Diego, CA, USA). Final
individual kinetic parameters are the average of two or
more independent determinations.
Activation of noncognate amino acids (threonine, cyste-
ine, alanine, valine, glycine) with 1 lm wild-type and 2 lm
mutant enzymes was also examined. However, mutant
enzymes did not show any misactivation, even at high
amino acid concentrations (250 mm).
Aminoacylation assays
Aminoacylation of tRNA was performed at 37 °Cina
reaction mixture consisting of 50 mm Hepes ⁄ KOH
(pH 7.6), 4 mm 1,4-dithiothreitol, 10 mm MgCl

2
,5mm
ATP, 50 mm KCl. For all kinetic assays, the concentrations
of the substrate studied ([
14
C]serine or tRNA
Ser
) varied
between 0.1 and 10 · K
m
. The concentrations of substrates
not varied in specific assays were held at saturating levels:
35–40 lm E. coli tRNA with overexpressed tRNA
Ser
, which
corresponds to  20 lm M. barkeri tRNA
Ser
and 120 l m
14
C-labeled serine. Reactions were initiated with the addi-
tion of SerRS. Concentrations of wild-type and mutated
enzymes were determined experimentally in order to obtain
linear velocities (typically 13 nm for wild-type enzyme and
40–500 nm for the mutants). Kinetic parameters were deter-
mined from Hanes–Woolf plots. All values represent the
average of three or more independent determinations. In
reactions where M. barkeri tRNAs
Ser
transcripts were used,
the kinetic parameters were determined from two indepen-

dent experiments.
Active site titration
To determine the amount of active enzyme in the pre-
paration, active site titration was performed. Formation of
the enzyme tightly bound intermediate seryl-AMP was
followed in the presence of pyrophosphatase (PPiase) and
saturating amounts of
14
C-serine and ATP at 37 °C. The
reaction was performed in 50 mm Hepes ⁄ KOH, pH 7.6,
50 mm KCl, 10 mm MgCl
2
,4mm dithiothreitol and aliqu-
ots were spotted onto nitrocellulose filters at 1, 5, 10 and
20 min, washed with 25 mL of ice-cold buffer. The amount
of active sites varied between 20% and 40%.
CD spectroscopy
The overall structure of the wild-type and mutated SerRS
was examined by measuring the CD. Protein samples at a
concentration of 0.18 lm in 15 mm KP, pH 7.0, 50 mm
KCl, 2 mm dithiothreitol were analyzed on a Jasco J-815
spectropolarimeter at room temperature. A 2 mm path-
length cuvette was used, and spectra were accumulated over
five scans.
Native PAGE and gel-retardation assay
In order to check for complex formation between cognate
tRNA and wild-type or mutated mMbSerRS a constant
amount of purified protein was mixed with tRNA at differ-
ent molar ratios. tRNA was incubated with enzyme for
15 min at 37 °C, in a 15 lL volume containing 25 mm

Mes ⁄ KOH pH 5.8, 50 mm KCl, 5 mm MgCl
2
followed by
cooling on ice and was subjected to electrophoresis on a
native 6% acrylamide (w ⁄ v) gel of acrylamide:bis-acryl-
amide (19 : 1) containing 5% glycerol in electrophoresis
buffer (25 mm Mes⁄ KOH pH 5.8, 10 mm magnesium
acetate). Electrophoresis was at 4 °C for 2.5 h at 120 V and
mMbSerRS active site mutants S. Bilokapic et al.
2842 FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS
gels were stained with Coomassie Brilliant Blue to analyze
the components of the tRNA
Ser
:SerRS complex.
Crystalyzation of mMbSerRS mutants was performed as
described previously [12].
Acknowledgements
This work was supported by grants from Scientific
Co-operation between Eastern Europe and Switzerland
(SCOPES) program of the Swiss National Science
Foundation (SNSF), the Ministry of Science, Educa-
tion and Sports of the Republic of Croatia (project
119-0982913-1358), and Unity through Knowledge
Fund (project 09 ⁄ 07).
References
1 Ibba M & So
¨
ll D (2000) Aminoacyl-tRNA synthesis.
Annu Rev Biochem 69, 617–650.
2 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.
3 Arnez GJ & Moras D (1997) Structural and functional
considerations of the aminoacylation reaction. Trends
Biochem Sci 22, 211–216.
4 Ibba M, Francklyn C & Cusack S (2005) The Aminoacyl-
tRNA Synthetases. Landes Bioscience, Georgetown, TX.
5 Swairjo MA, Otero FI, Yang XL, Lovato MA, Skene
RJ, McRee DE, de Pouplana LR & Schimmel P (2004)
Alanyl-tRNA synthetase crystal structure and design
for acceptor-stem recognition. Mol Cell 13, 829–841.
6 Zhang C-M, Perona JJ, Ryu K, Francklyn C & Hou
Y-M (2006) Distinct kinetic mechanisms of the two
classes of aminoacyl-tRNA synthetases. J Mol Biol 361,
300–311.
7 Weygand-Durasevic I & Cusack S (2005) Structure,
function and evolution of seryl-tRNA synthetases. In
The Aminoacyl-tRNA Synthetases (Ibba M, Francklyn
C & Cusack S, eds), pp. 177–192. Landes Bioscience,
Georgetown, TX.
8 Cusack S, Berthet-Colominas C, Ha
¨
rtlein M, Nassar N
& Leberman R (1990) A second class of synthetase
structure revealed by X-ray analysis of Escherichia coli
seryl-tRNA synthetase at 2.5 A
˚
. Nature 347, 249–255.
9 Biou V, Yaremchuk A, Tukalo M & Cusack S (1994)

The 2.9 A
˚
crystal structure of T. thermophilus seryl-
tRNA synthetase complexed with tRNA
Ser
. Science 263,
1404–1410.
10 Cusack S, Yaremchuk A & Tukalo M (1996) The crys-
tal structure of the ternary complex of T. thermophilus
seryl-tRNA synthetase with tRNA
Ser
and a seryl-aden-
ylate analogue reveals a conformational switch in the
active site. EMBO J, 15 , 2834–2842.
11 Chimnaronk S, Jeppesen MG, Suzuki T, Nyborg J &
Watanabe K (2005) Dual-mode recognition of nonca-
nonical tRNAs
Ser
by seryl-tRNA synthetase in mamma-
lian mitochondria. EMBO J 24, 3369–3379.
12 Bilokapic S, Maier T, Ahel D, Gruic-Sovulj I, So
¨
ll D,
Weygand-Durasevic I & Ban N (2006) Structure of unu-
sual seryl-tRNA synthetase reveals a distinct zinc-
dependent mode of substrate recognition. EMBO J 25,
2498–2509.
13 Bult CJ, White O, Olsen GJ, Zhou L, Fleischmann RD,
Sutton GG, Blake JA, Fitzgerald LM, Clayton RA,
Gocayne JD et al. (1996) Complete genome sequence of

the methanogenic archaeon, Methanococcus jannaschii.
Science 273, 1058–1073.
14 Kim H-S, Vothknecht UC, Hedderich R, Celic I & So
¨
ll
D (1998) Sequence divergence of seryl-tRNA synrheta-
ses in Archaea. J Bacteriol 180, 6446–6449.
15 Korencic D, Polycarpo C, Weygand-Durasevic I & So
¨
ll
D (2004) Differential modes of transfer RNA
Ser
recog-
nition in Methanosarcina barkeri. J Biol Chem 279,
48780–48786.
16 Belrhali H, Yaremchuk A, Tukalo M, Larsen K,
Berthet-Colominas C, Leberman R, Beijer B, Sproat B,
Als-Nielsen J, Grubel G et al. (1994) Crystal structures
at 2.5 Angstrom resolution of seryl-tRNA synthetase
complexed with two analogs of seryl adenylate. Science
263, 1432–1436.
17 Bilokapic S, Korencic D, So
¨
ll D & Weygand-Durasevic
I (2004) The unusual methanogenic seryl-tRNA synthe-
tase recognizes tRNA
Ser
species from all three kingdoms
of life. Eur J Biochem 271, 694–702.
18 Ahel D, Slade D, Mocibob M, So

¨
ll D & Weygand-Du-
rasevic I (2005) Selective inhibition of divergent seryl-
tRNA synthetases by serine analogues. FEBS Lett 579,
4344–4348.
19 Gruic-Sovulj I, Jaric J, Dulic M, Cindric M & Wey-
gand-Durasevic I (2006) Shuffling of discrete tRNA
Ser
regions reveals differently utilized identity elements in
yeast and methanogenic archaea. J Mol Biol 361,
128–139.
20 Aberg A, Yaremchuk A, Tukalo M, Rasmussen B &
Cusack S (1997) Crystal structure analysis of the activa-
tion of histidine by Thremus thermophilus histidyl-tRNA
synthetase. Biochemistry 36, 3084–3094.
21 Schmitt E, Moulinier L, Fujiwara S, Imanaka
T, Thierry J-C & Moras D (1998) Crystal structure
of aspartyl-tRNA synthetase from Pyrococcus kodaka-
raensis KOD: archaeon specificity and catalytic mech-
anism of adenylate formation. EMBO J 17, 5227–
5237.
22 Ador L, Jaeger S, Geslain R, Martin F, Cavarelli J &
Eriani G (2004) Mutation and evolution of the magne-
sium-binding site of a class II aminoacyl-tRNA synthe-
tase. Biochemistry 43, 7028–7037.
S. Bilokapic et al. mMbSerRS active site mutants
FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS 2843
23 Lenhard B, Filipic S, Landeka I, Skrtic I, So
¨
ll D &

Weygand-Durasevic I (1997) Defining the active site of
yeast seryl-tRNA synthetase: mutations in motif 2 loop
residues affect tRNA-dependent amino acid recognition.
J Biol Chem 272, 1136–1141.
24 Gruic-Sovulj I, Landeka I, Soll D & Weygand-Duras-
evic I (2002) tRNA-dependent amino acid discrimina-
tion by yeast seryl-tRNA synthetase. Eur J Biochem
269, 5271–5279.
25 Jakubowski H (1997) Aminoacyl thioester chemistry of
class II aminoacyl-tRNA synthetases. Biochemistry 36,
11077–11085.
26 Jakubowski H (1998) Aminoacylation of coenzyme A
and pantetheine by aminoacyl-tRNA synthetases: possi-
ble link between noncoded and coded peptide synthesis.
Biochemistry 37, 5147–5153.
27 Eriani G & Gangloff J (1999) Yeast aspartyl-tRNA syn-
thetase residues interacting with tRNA(Asp) identity
bases connectively contribute to tRNA(Asp) binding in
the ground and transition-state complex and discriminate
against non-cognate tRNAs. J Mol Biol 291, 761–773.
28 Ador L, Camasses A, Erbs P, Cavarelli J, Moras D,
Gangloff J & Eriani G (1999) Active site mapping of
yeast aspartyl-tRNA synthetase by in vivo selection of
enzyme mutations lethal for cell growth. J Mol Biol
288, 231–242.
29 Steer BA & Schimmel P (1999) Different adaptations of
the same peptide motif for tRNA functional contacts
by closely homologous tRNA synthetases. Biochemistry
38, 4965–4971.
30 Guth EC & Francklyn CS (2007) Kinetic discrimination

of tRNA identity by the conserved motif 2 loop of a
class II aminoacyl-tRNA synthetase. Mol Cell 25,
531–542.
31 Maeder DL, Anderson I, Brettin TS, Bruce DC, Gilna
P, Han CS, Lapidus A, Metcalf WW, Saunders E,
Tapia R et al. (2006) The Methanosarcina barkeri gen-
ome: comparative analysis with Methanosarcina acetivo-
rans and Methanosarcina mazei reveals extensive
rearrangement within Methanosarcinal genomes. J Bac-
teriol 188, 7922–7931.
32 Praetorius-Ibba M, Stange-Thomann N, Kitabatake
M, Ali K, So
¨
ll I, Carter CW Jr, Ibba M & So
¨
ll D
(2000) Ancient adaptation of tryptophanyl-tRNA
synthetase for tryptophan binding. Biochemistry 39,
13136–13143.
33 Agou F, Quevillon S, Kerjan P & Mirande M (1998)
Switching the amino acid specificity of an aminoacyl-
tRNA synthetase. Biochemistry 37, 11309–11314.
34 Woese CR, Olsen GJ, Ibba M & So
¨
ll D (2000)
Aminoacyl-tRNA synthetases, the genetic code, and the
evolutionary process. Microbiol Mol Biol Rev 64, 202–
236.
35 Studier WF, Rosenberg AH, Dunn JJ & Dudendorf JW
(1990) Use of the T7 RNA polymerase to direct expres-

sion of cloned genes. Methods Enzymol 85, 60–89.
36 Uter NT, Gruic-Sovulj I & Perona JJ (2005) Amino
acid-dependent transfer RNA affinity in a class I ami-
noacyl-tRNA synthetase. J Biol Chem 280, 23966–
23977.
mMbSerRS active site mutants S. Bilokapic et al.
2844 FEBS Journal 275 (2008) 2831–2844 ª 2008 The Authors Journal compilation ª 2008 FEBS