Modeling of tRNA-assisted mechanism of Arg activation
based on a structure of Arg-tRNA synthetase, tRNA, and
an ATP analog (ANP)
Michiko Konno
1
, Tomomi Sumida
1,
*, Emiko Uchikawa
1
, Yukie Mori
1
, Tatsuo Yanagisawa
2,
*,
Shun-ichi Sekine
2
and Shigeuki Yokoyama
2
1 Department of Chemistry and Biochemistry, Graduate School of Humanities and Sciences, Ochanomizu University, Tokyo, Japan
2 Department of Biophysics and Biochemistry, Graduate School of Science, University of Tokyo, Japan
Introduction
Most aminoacyl-tRNA synthetases (aaRSs) catalyze the
formation of aminoacyl-AMP in the presence of
Mg
2+
[amino acid + ATP fi aminoacyl-AMP +
pyrophosphate (PP
i
)] and the reverse reaction (amino-
acyl-AMP + PP
i

CCU
)
(Received 18 March 2009, revisied 11 June
2009, accepted 26 June 2009)
doi:10.1111/j.1742-4658.2009.07178.x
The ATP–pyrophosphate exchange reaction catalyzed by Arg-tRNA, Gln-
tRNA and Glu-tRNA synthetases requires the assistance of the cognate
tRNA. tRNA also assists Arg-tRNA synthetase in catalyzing the pyro-
phosphorolysis of synthetic Arg-AMP at low pH. The mechanism by which
the 3¢-end A76, and in particular its hydroxyl group, of the cognate tRNA
is involved with the exchange reaction catalyzed by those enzymes has yet
to be established. We determined a crystal structure of a complex of Arg-
tRNA synthetase from Pyrococcus horikoshii, tRNA
Arg
CCU
and an ATP
analog with R
factor
= 0.213 ( R
free
= 0.253) at 2.0 A
˚
resolution. On the
basis of newly obtained structural information about the position of ATP
bound on the enzyme, we constructed a structural model for a mechanism
in which the formation of a hydrogen bond between the 2¢-OH group of
A76 of tRNA and the carboxyl group of Arg induces both formation of
Arg-AMP (Arg + ATP fi Arg-AMP + pyrophosphate) and pyrophos-
phorolysis of Arg-AMP (Arg-AMP + pyrophosphate fi Arg + ATP) at
low pH. Furthermore, we obtained a structural model of the molecular

reaction path of aminoacyl-AMP formation on the
aaRSs has not been attained yet.
In particular, the following detailed biochemical
findings on the formation of aminoacyl-AMP and clo-
sely related reactions have been reported for
Arg-tRNA synthetase (ArgRS; EC 6.1.1.19) [1–6]. No
success of consistent understanding has been achieved
for the molecular reaction paths of aminoacyl-AMP
formation and closely related reactions catalyzed by
ArgRS. New models of the reaction paths involved in
the formation of aminoacyl-AMP and closely related
reactions observed for ArgRS will lead to improved
understanding of the activated complex formed
between the amino acid and ATP in the reaction path
of aminoacyl-AMP formation on most aaRSs as well
as ArgRS.
For most aaRSs, the formation of aminoacyl-AMP
does not require tRNA. On the other hand, for ArgRS
from Escherichia coli, Bacillus stearothermophilus, Neu-
rospora crassa, and Saccharomyces cerevisiae [1–5],
Gln-tRNA synthetase (GlnRS) from E. coli W,
S. cerevisiae and porcine liver [7,8] and Glu-tRNA syn-
thetase (GluRS) from E. coli K12 [9,10], the ATP–PP
i
exchange reaction corresponding to the formation of
aminoacyl-AMP and its reverse reaction, amino
acid + ATP = aminoacyl-AMP + PP
i
, has never
been observed without tRNA. In the presence of cog-

mechanism and to clarify the orientation of the dihy-
drouridine (D) loop containing A20 of tRNA
Arg
inter-
acting with ArgRS, we determined crystal structures of
a binary complex of Pyrococcus horikoshii ArgRS and
tRNA
Arg
CCU
and a ternary complex also containing
the ATP analog adenosine-5¢-(b,c-imido)triphosphate
(ANP); we found one reasonable mechanism, based on
newly obtained structural information about the posi-
tion of ATP bound on ArgRS. In order to understand
the function of the N-terminal domain of ArgRS in
relation to the binding mechanism of tRNA
Arg
,we
constructed an ArgRS mutant lacking the N-terminal
domain (DN ArgRS). The experimental results showed
that the DN ArgRS protein retains sufficient catalytic
activity in the aminoacylation reaction for
tRNA
Arg
CCU
. Moreover, modeling of the relative posi-
tions of Arg, A76 of tRNA
Arg
and ATP on ArgRS
was undertaken to find the suitable position for the

(R
free
= 0.262) at 2.3 A
˚
, respectively. In the crystals
grown in the presence of l-Arg, l-Arg was not visible in
the electron density map. The overall structure of a ter-
nary complex of P. horikoshii ArgRS, tRNA
Arg
CCU
and
the ATP analog is shown in Fig. 1, and sequence align-
ments for ArgRSs from P. horikoshii , T. thermophilus
and S. cerevisiae on the basis on three-dimensional
structures are given in Fig. 2.
Structures of S. cerevisiae ArgRS-bound arginine
and tRNA
Arg
ICG
[12] (Protein Data Bank ID: 1F7V)
and ‘tRNA-free’ T. thermophilus ArgRS [13] (Protein
Data Bank ID: 1IQ0), the Rossmann fold and the
anticodon-binding domains of which were superim-
posed onto those of P. horikoshii ArgRS, are shown in
Fig. 3A,B. It has been reported that, in S. cerevisiae
ArgRS, the Asn fi Ala mutation of Asn153, corre-
sponding to Asn129 in P. horikoshii ArgRS (Fig. 2),
Modeling of tRNA-assisted mechanism on ArgRS M. Konno et al.
4764 FEBS Journal 276 (2009) 4763–4779 ª 2009 The Authors Journal compilation ª 2009 FEBS
gives a drastically decreased k

aminoacyl moiety have been observed, whereas no
ATP-bound protein of class Ia has been observed. In
P. horikoshii ArgRS, the ATP analog (ANP) molecule
was clearly found in the active site (Fig. 4). The obser-
vation of the ANP-bound protein is due to the high
hydrophobicity of ArgRS from the archaebacterium
P. horikoshii living at very high temperature and the
existence of the His417 residue. The adenine base of
ANP with small values of average B-factor is stacked
upon the aromatic ring of His417, which is specific to
P. horikoshii ArgRS. The adenine base is in close prox-
imity to the main chain of Val418 in the S16 strand,
the N1–Val418 N and N6–Val418 O distances being
3.19 A
˚
and 3.47 A
˚
, respectively, and the 2¢-OH of the
ribose is in close proximity to N of Gly384 and O
e1
of
Glu386 in the S14–H14 turn (Gly384–Ala385–Glu386–
Gln387 turn), the distances being 2.71 A
˚
and 2.77 A
˚
,
respectively. The distance between Ca of Glu386, the
third residue in the S14–H14 turn, and Ca of Val418
in the S16 strand is 12.8 A

Data Bank ID: 1IYW) [22], the distance between
Asp490 and Val521 is 11.9 A
˚
. These distances within
0.9 A
˚
of the distance of 12.8 A
˚
in ArgRS indicate that,
in these aaRSs, this space is the binding site of the
adenosine moiety of ATP, and in E. coli Cys-tRNA
synthetase bound to tRNA
Cys
(Protein Data Bank ID:
1U0B) [23], the distance between Asp229 and Val260
is 11.4 A
˚
. AMP weakly inhibits the binding of ATP in
a competitive manner in the aminoacylation reaction
[24].
The position of the Mg
2+
located between PbO
(2.45 A
˚
and 3.01 A
˚
) and PaO (2.97 A
˚
) of ANP is not

half of the electron density expected for an occupancy
of 1.0 for Mg
2+
. The presence of different orientations
for the PbNPc moiety of ANP attached and not
attached to Mg
2+
is manifested as low electron densi-
ties in the regions of Mg
2+
and the PbNPc moiety.
The salt bridge formed by Mg
2+
between PbO and
PaO may retard the conformational inversion at Pa of
ATP in the reversible process of the ATP–PP
i
exchange reaction, whereas the salt bridge formed by
Mg
2+
between PbO and PcO does not, by any means,
retard the conformational inversion at Pa of ATP.
The reported aminoacyl-AMP analogs have sulfa-
moyl (–NH–SO
2
–O–) or diaminosulfone (–NH–SO
2
–
NH–) in place of Pa [–O–PO(OH)–O–] of AMP.
Furthermore, the reported aminoacyl-AMP analogs

‘HIGH’, is close to PaO, PbO and PcO of ANP, with
Fig. 2. Sequence alignment of P. horikoshii ArgRS (PhRRS), T. thermophilus ArgRS (TtRRS) and S. cerevisiae ArgRS (ScRRS) on the basis
of three-dimensional structures. The residues exposed on the surface of a-helices (colored in red) and b-strands (colored in blue) are aligned
among the three ArgRSs. The Asn corresponding to Asn153 (ScRRS) is indicated by a green letter.
Modeling of tRNA-assisted mechanism on ArgRS M. Konno et al.
4766 FEBS Journal 276 (2009) 4763–4779 ª 2009 The Authors Journal compilation ª 2009 FEBS
distances of 2.67 A
˚
, 2.81 A
˚
, and 2.91 A
˚
, respectively.
The fact that in P. horikoshii ArgRS, where the
‘KMSK’ motif is replaced by Lys424-Phe425-Ser426-
Gly427, the first Lys424 of the current structure under-
goes no interaction with the PbNPc moiety of ANP
proves that, in P. horikoshii ArgRS, the ‘KFSG’ por-
tion does not contribute to the ATP–PP
i
exchange
reaction.
The 3¢-terminus of tRNA
In the 3¢-terminal G73–C74–C75–A76 sequence of
tRNA
Arg
CCU
, two transient forms were observed in
the ternary complex (Fig. 5A) as well in the binary
complex, depending on crystallization conditions; in

stage to the final stage. The base of C74 is found near
the surface of the connective polypeptide domain,
which is a transient position, i.e. the hydrophobic cleft
constructed by the side chains of Tyr300, Ala303,
Val321, Arg324 and Ser325 in the connective polypep-
tide domain. The relative orientation of G73 and C74
to the connective polypeptide domain is similar to that
N-terminal
domain
Anticodon-binding domain
'Stem contact fold'
domain
Catalytic
domain
Inserted domain 1
Connective polypeptide domain
Rossmann fold domain
A
B
N-terminal
domain
Anticodon-binding domain
'Stem contact fold'
domain
Catalytic
domain
Inserted domain 1
Connective polypeptide domain
Rossmann fold domain
Fig. 3. (A) Comparison between two overall structures of P. horiko-

obs
) F
calc
) cross-validated
r
A
-weighted omit map contoured at level
1.5r. The map was produced using the
complex model without ANP and all the
data from 40 A
˚
to 2.0 A
˚
resolution. A green
sphere shows the Mg
2+
.
M. Konno et al. Modeling of tRNA-assisted mechanism on ArgRS
FEBS Journal 276 (2009) 4763–4779 ª 2009 The Authors Journal compilation ª 2009 FEBS 4767
observed in tRNA
Arg
ICG
bound to S. cerevisiae
ArgRS in the tertiary complex. It is predicted that
the conformational change from the first stage to the
final stage takes place in the absence of Arg, and the
hydrophobic circumstance changes the hydration state
around the phosphodiester bridges in C72–G73–C74–
C75–A76. In T. thermophilus Val-tRNA synthetase
bound to tRNA

Arg
isoacceptors from S. cerevisiae have D or C.
Detailed experiments with E. coli and T. thermophilus
ArgRSs apparently suggested that the interaction with
the middle base of the anticodon (C35) and A20 of
tRNA
Arg
play an important role in tRNA
Arg
binding on
ArgRS [13,28,29]. Crystal structures of binary and
ternary complexes of ArgRS and tRNA
Arg
ICG
from
S. cerevisiae and Arg revealed that a base of D20 in the
D-loop, which is specific to S. cerevisiae tRNA
Arg
ICG
,is
positioned in close proximity to the side chains of
Asn106, Phe109, and Gln111, which are included in
the characteristic N-terminal domain of ArgRS [12].
A
C
B
Fig. 5. The structure of tRNA
Arg
CCU
on P. horikoshii ArgRS. (A) Two transient forms in the 3¢-terminal end of P. horikoshii tRNA

on the Asn106 fi Ala, Phe109 fi Ala
and Gln111 fi Ala mutant proteins of S. cerevisiae
ArgRS are the same as those on the wild-type ArgRS
[14]. This shows that the interaction between the N-ter-
minal domain of S. cerevisiae ArgRS and D20 of the
D-loop of tRNA
Arg
are not important for the binding
of tRNA
Arg
on ArgRS in the aminoacylation reaction.
The crystal structure of free ArgRS from T. thermophi-
lus has been determined, but that of the complex with
tRNA
Arg
has not been determined yet [13]. In T. ther-
mophilus ArgRS, the Tyr77 fi Ala and Asn79 fi Ala
mutants (Tyr77 and Asn79 correspond to Phe109 and
Gln111 of S. cerevisiae ArgRS, on the basis of struc-
tural comparison between S. cerevisiae ArgRS and
T. thermophilus ArgRS) showed a notable increase in
K
m
for tRNA
Arg
and a large decrease in V
max
in the am-
inoacylation reaction at pH 7.5. On the other hand, it is
very noticeable that the Asn79 fi Lys mutant, which

Gly84-Tyr85) between the S3 and S4 strands and the
hydrophobic side chains of Pro34 and Leu38. N1
and N6 of the base of A20 lie close to N
d2
and O
d1
of the side chain of Asn87 in the S4 strand, with
distances of 2.82 A
˚
and 2.97 A
˚
, respectively. The
plane of the base of A20 and the end plane of the
carbamoyl group of Asn87 are out of coplanar ori-
entation, and the dihedral angle between these two
planes was about 25°. In particular, O
d1
of the car-
bamoyl group of Asn87 is positioned far out of the
base plane. Large values of average B-factor of resi-
dues in the N-terminal domain (average B-factors of
residues 2–118 in the N-terminal domain and resi-
dues 119–629 in other domains are 49.9 A
˚
2
and
29.5 A
˚
2
, respectively) indicate that the D-loop does

)
2
SO
4
and 1,6-hexanediol are
used as precipitating agents [30]. This fact indicates
that, even though all of the substrates required for the
aminoacylation reaction are present at sufficient con-
centration in the crystallization solution, the aminoacy-
lation reaction does not occur during the long time
needed for crystal growth, which suggests that tRNA-
bound S. cerevisiae ArgRS observed in the ternary
complex is by no means in a conformation that is fit
to activate. The fact that k
cat
and K
m
for tRNA
Arg
in
the aminoacylation reaction do not change in the
Asn106 fi Ala, Gln111 fi Ala and Phe109 fi Ala
mutants of S. cerevisiae ArgRS [14] indicates that
those mutations have no influence on the orientation
of tRNA
Arg
in wild-type ArgRS.
When the Rossmann fold domain and the anticodon-
binding domain in P. horikoshii ArgRS were superim-
posed onto those of S. cerevisiae ArgRS, the C1s of A20

tRNA
Arg
CCU
, are within 1.3 A
˚
of the corresponding
M. Konno et al. Modeling of tRNA-assisted mechanism on ArgRS
FEBS Journal 276 (2009) 4763–4779 ª 2009 The Authors Journal compilation ª 2009 FEBS 4769
distances in S. cerevisiae tRNA
Arg
ICG
. These facts indi-
cate that the framework of tRNA
Arg
of the L-shape is
conserved in these two cases.
The anticodon loop of tRNA
In the complex of P. horikoshii ArgRS, the base of
C35 of tRNA
Arg
CCU
is located in the hydrophobic
pocket formed by the aromatic ring of Tyr587 at the
C-terminal end of H23 and the hydrophobic side
chains of Ile517 of H19 and Pro591, Val592 and
Leu593 of the loop between H23 and H24 (Fig. 5C).
N4H
2
and O2 of C35 are found within the distance of
the hydrogen bonds with the main chain CO of

ICG
on S. cerevisiae
ArgRS with Tyr565 replaced by Ala is identical to
that on the wild-type ArgRS [14] indicates that this
conserved Tyr makes little contribution to the recog-
nition of the base of C35. The report that a few
tRNA
Met
CAU
molecules are aminoacylated by Arg
with E. coli ArgRS [29] suggests that when the back-
bone of the anticodon stem is superimposed, the
anticodon bases of tRNA
Met
CAU
and tRNA
Arg
CCG
should be oriented in the same direction and bind to
almost the same region in the helix bundle structure
of E. coli ArgRS.
The reported structure of T. thermophilus ArgRS
also has a quite similar hydrophobic pocket to the
hydrophobic pocket accepting C35 of tRNA
Arg
CCU
in
the complex of P. horikoshii ArgRS and the hydro-
phobic pocket accepting C35 of tRNA
Arg

may successfully accept C35 and G36 of the mutated
tRNA
Met
CCG
in the formation of Arg-tRNA
Met
CCG
on
E. coli ArgRS [29].
On the other hand, in tRNA
Met
CAU
bound on
A. aeolicus Met-tRNA synthetase, the conformation of
the anticodon loop of C32–U33–C34–A35–U36–A37–
A38 of tRNA
Met
CAU
is largely different from that of
C32–U33–C34–C35–U36–A37–A38 of tRNA
Arg
CCU
bound on P. horikoshii ArgRS. It is worth noting that
the base of C32 is stacked on a C31ÆG39 base pair in
tRNA
Met
CAU
, but the base of A37 is stacked on a
C31ÆG39 base pair in the case of the observed complex
of P. horikoshii ArgRS and tRNA

CCU
to the N-terminal domain
contributes to the activation effect of tRNA on tRNA-
assisted Arg-AMP formation reaction or the amino-
acylation reaction, we constructed P. horikoshii ArgRS
(residues 92–629; DN ArgRS) lacking the core region
of the N-terminal domain from residue 1 to residue 91
in order to completely eliminate interactions between
the N-terminal domain and the D-loop of tRNA
Arg
,
and measured the kinetic parameters of the amino-
Modeling of tRNA-assisted mechanism on ArgRS M. Konno et al.
4770 FEBS Journal 276 (2009) 4763–4779 ª 2009 The Authors Journal compilation ª 2009 FEBS
acylation reaction for wild-type ArgRS and DN Ar-
gRS. For wild-type ArgRS and DN ArgRS, the K
m
values for tRNA
Arg
CCU
were 2.6 lm and 3.8 lm in
100 mm Hepes ⁄ NaOH buffer (pH 7.5), respectively,
and the measured ratio of the V-value of DN ArgRS
to that of wild-type ArgRS was [(8 ± 2) · 10
)2
]. This
indicates that the fixing of the D-loop of tRNA
Arg
with the N-terminal domain makes a minor contribu-
tion to the aminoacylation reaction, but is not essen-

loop, the H13 helix, the S14–H14 turn and the H14
helix in the Rossmann fold domain in P. horikoshii
ArgRS. Referring to the distances between the a-NH
2
group of Arg and the main chain C@O of Ser151 and
O of the side chain of Asn153, and the distances
between the guanidinium moiety and the side chains of
Glu148 and Asp351 in S. cerevisiae ArgRS, we pre-
dicted the possible site for Arg in P. horikoshii ArgRS.
In the predicted site, the distances between the a-NH
2
group of Arg and the main chain C@O of Ser127 and
O of the side chain of Asn129 are set at 3.15 A
˚
and
3.04 A
˚
, and the distances between the guanidinium
moiety and the side chain of Glu124 (S5) and Asp335
(H13) are set at 4.33 A
˚
and 3.66 A
˚
. Its carboxyl group
is located at the proper position relative to Pa of ANP
(ATP analog). In the ternary complex of S. cerevisiae
ArgRS, the base of A76 is stacked on the side chain of
Tyr347 on the helix corresponding to H13 in P. hori-
koshii ArgRS, whereas in P. horikoshii ArgRS, the
H13 helix deviates largely from that of S. cerevisiae

Arg
in the Arg-
NHOH formation reaction in the presence of tRNA
Arg
is shown in Fig. 7B.
In the case of modeling the deacylation reaction of
Arg-tRNA
Arg
, coordinations of NH
2
–Ca–C and Cb of
the Arg moiety of Arg-tRNA
Arg
were assumed to be
essentially identical to those of the Arg predicted
above, and the Arg moiety in the cyclic form was fitted
in the space that is provided by the rearrangement of
the side chain of Gln387 (Fig. 7C). We built a model
for the Glu-dependent ATP–PP
i
exchange reaction at
pH 6.0 in the absence of tRNA
Glu
on the basis of the
crystal structure of T. thermophilus GluRS (Protein
Data Bank IDs: 1N77, 1N78, and 2CV0) [32]. Coordi-
nations of NH
2
–Ca–C and Cb of the Glu moiety of
the intermediate of formed Glu-AMP were assumed to

observed at much higher concentrations of Glu,
whereas the tRNA-assisted ATP–PP
i
exchange reaction
was observed at lower concentrations of Glu. For
instance, the K
m
value for Glu measured in the tRNA-
assisted ATP–PP
i
exchange reaction decreases signifi-
cantly by 10
2
)10
3
-fold in comparison with that in the
tRNA-independent ATP–PP
i
exchange reaction (the
K
m
values for Glu are 0.4 m in the absence of tRNA
and 6.6 · 10
)4
m in the presence of tRNA at pH 7.7
for E. coli W, 0.2 m and 7 · 10
)3
m at pH 7.7 for
S. cerevisiae, 0.4 m and 4 · 10
)3

and Mg
2+
[6]. Further-
more, the pyrophosphorolysis of chemically synthesized
Arg-AMP and the ATP–PP
i
exchange reaction cata-
lyzed by ArgRS in the presence of tRNA have pH
optima of 6.2 and 6.5, respectively. The presence of PP
i
AB
CD
Fig. 7. Modeled intermediates on P. horikoshii ArgRS or T. thermophilus GluRS in the ATP–PP
i
exchange reaction, the Arg-NHOH formation
reaction, and the deacylation reaction. (A) Arg (cyan), ATP (orange) coordinated by Mg
2+
and A76 (green) of tRNA assisting the Arg-AMP for-
mation reaction on P. horikoshii ArgRS. The Mg
2+
is indicated by a green sphere. (B) HN
2
OH (cyan), enzymatically synthesized Arg-AMP
(orange) and A76 (green) of tRNA in P. horikoshii ArgRS in the Arg-NHOH formation reaction in the presence of tRNA. (C) Arg-A76 (green) of
Arg-tRNA with the Arg moiety with the cyclic configuration and AMP (orange) on P. horikoshii ArgRS in the deacylation reaction of Arg-tRNA.
The torsional angles of the side chain of Gln387 (cyan) were changed from the original structure. (D) A Glu with the cyclic configuration
(cyan) in the C–O
c)
–H–O2 form and ATP (orange) on T. thermophilus GluRS in the Glu-AMP formation reaction in the absence of tRNA.
Modeling of tRNA-assisted mechanism on ArgRS M. Konno et al.

interaction between the the 2¢-OH group of tRNA and
the a-carboxyl group of Arg ⁄ Gln will also take place
in the ATP–PP
i
exchange reaction on ArgRS and
GlnRS. The interaction between the 2¢-OH group
and the O of the a-carboxyl group of Arg is thought
to accelerate the pyrophosphorolysis reaction on Arg-
RS. In a ternary complex of S. cerevisiae ArgRS,
tRNA
Arg
ICG
, and Arg (Protein Data Bank ID: 1F7V)
[12], the O of the 2¢-OH group in the 3¢-end A76 of
tRNA
Arg
ICG
is 3.18 A
˚
, 3.71 A
˚
and 3.57 A
˚
distant
from the C, O1, and O2 (the O contacting Pa of
ATP is defined as O1 and the other as O2), respec-
tively, of the a-carboxyl group of Arg. The move-
ment of O2 by a rotation of 45° around the Ca–C
bond causes a decreased distance of 2.77 A
˚

i
. As O1 of the amino acid binds to
Pa, and PP
i
is released from Pa on the reverse
side, this reaction of Arg-AMP formation is an S
N
2
reaction.
The modeling of Arg, ATP and A76 of tRNA on
P. horikoshii ArgRS for the Arg-AMP formation reac-
tion at low pH indicates that when the straight side
chain of an Arg molecule is inserted into the hydro-
phobic pocket as in the Arg molecule bound to S. cere-
visiae ArgRS, its carboxyl group can locate in close
proximity to Pa of ATP (Fig. 7A). The a-carboxyl
group of the Arg molecule can assume such a confor-
mation that O2 forms a hydrogen bond with the
2¢-OH group of the ribose moiety of A76 of tRNA
Arg
by rotation around Ca–C. When Pa of ATP gains
access to O1 of C@O1 of the a-carboxyl group of Arg,
and two oxygen atoms of Pb and Pc are coordinated
by Mg
2+
on two parallel P@O bonds rather than
P–O
)
, as observed in the acetylacetonato [CH
3

'
Fig. 8. Reaction scheme of formation of Arg-AMP from Arg and
ATP in the presence of tRNA at low pH. The rotation of the a-car-
boxyl group around Ca–C induces the formation of a hydrogen bond
between the 2¢-OH group of the ribose of A76 and O2H of the
a-carboxyl group of Arg. Arrows indicate the sequential transfer of
the bond in the intermediate.
M. Konno et al. Modeling of tRNA-assisted mechanism on ArgRS
FEBS Journal 276 (2009) 4763–4779 ª 2009 The Authors Journal compilation ª 2009 FEBS 4773
configuration for the aminoacylation reaction without
a conformational rearrangement, but should be in a
much more preferable configuration to form a hydro-
gen bond with C@O2 of the Ca–(C@O2)–O1–Pa moi-
ety. Even though the pyrophosphorolysis reaction is
also an S
N
2 reaction, the hydrogen bond between the
2¢-OH group of tRNA and C@O2 is also required for
the pyrophosphorolysis reaction. The cleavage of
the bond between CO1 and Pa of Arg-AMP is acceler-
ated by this hydrogen bond. The formation of this
hydrogen bond leads to a double bond between C and
O1, and the intermolecular rearrangement reaction
proceeds in the reverse direction of the Arg-AMP for-
mation reaction shown in Fig. 8. Indeed, the amount
of neutral PP
i
, in which the lone pairs of double bonds
of Pb@O and Pc@O are coordinated by Mg
2+

released, the 2¢-OH group of the ribose moiety of
tRNA forms a hydrogen bond with O2 of C@O2 of
Arg-AMP (Fig. 7B), and the sp
2
hybrid orbital on the
C therefore changes to an sp
3
hybrid orbital. The C
with an sp
3
hybrid orbital has the ability to react with
NH
2
OH. Therefore, by an S
N
2 reaction, Arg-NHOH
is formed and AMP is released. The fact that Arg-
AMP reacts with NH
2
OH rather than with tRNA
shows that, in the ATP–PP
i
exchange reaction, tRNA
stays preferentially in the position suitable for the
formation of a hydrogen bond with O2 of C@O2.
If, in the ternary complex of S. cerevisiae ArgRS,
tRNA, and Arg, the carboxyl group of the Arg was
rotated by 45° around Ca–C, its O2 would come into
close proximity, with a distance of 2.77 A
˚

i
exchange reaction increases to up
to 10
2
)10
3
-fold that in the presence of tRNA at a pH
of 7.7 or 7.6 [7,8,33]. In the absence of tRNA, the
activity increases up to pH 8.0 as pH increases, and
decreases when the pH reaches over 8.0, whereas in the
presence of tRNA, the activity decreases when the pH
reaches over 6.2 [33]. The reported large difference in
the K
m
value is related to a difference in the form of
the bound Glu. In the presence of tRNA, the a-car-
boxyl group of the Glu interacts with the 2¢-OH group
of the ribose moiety of tRNA, and the side chain
assumes the straight form, with the alkyl group in the
all-trans form, allowing the ribose moiety to approach.
In the case of the ATP–PP
i
exchange reaction in the
absence of tRNA, the most likely candidate for the
form of the Glu is a rotational isomer in the cyclic
form, which contains a cis-alkyl group. In the cyclic
form, instead of the 2¢-OH group of tRNA, the c-car-
boxyl group of the side chain can form a hydrogen
bond with the a-carboxyl group in the C–O
c)

2
to an sp
3
hybrid
orbital requires the transfer of a proton to C@Oor
Modeling of tRNA-assisted mechanism on ArgRS M. Konno et al.
4774 FEBS Journal 276 (2009) 4763–4779 ª 2009 The Authors Journal compilation ª 2009 FEBS
the formation of a hydrogen bond. At pH 8, the find-
ing that, as no residue of the protein can donate a pro-
ton in the pyrophosphorolysis reaction, even though,
at low pH, the 2¢-OH group of tRNA is used, indicates
that there is, of course, no residue to donate a proton
in the first step at high pH. The guanidinium group of
Arg also forms the rotational isomer in a quite similar
manner to the side chain of Glu. The transfer of a pro-
ton from N
e
Hof–N
e+
H@C–(NH
2
)
2
to the C@Oof
the ester bond of Arg-tRNA converts the sp
2
hybrid
orbital of the C to an sp
3
hybrid orbital. A bond is

conformation for the formation of an intramolecular
hydrogen bond. The intramolecular hydrogen bond,
instead of tRNA, accelerates the pyrophosphorolysis
step of this deacylation reaction.
The reported results of the deacylation reaction indi-
cate that ArgRS has a binding affinity for Arg-tRNA
with the cyclic form of the Arg moiety that is compa-
rable to that for Arg-tRNA without the cyclic form.
Furthermore, as suggested above, ArgRS has a binding
affinity for Arg-AMP with the cyclic form of the Arg
moiety that is comparable to that for Arg-AMP with-
out the cyclic form.
The rate of transfer of the Arg moiety from the
chemically synthesized Arg-AMP to tRNA has an
optimum at pH 8.1 [6]. If the guanidium moiety of
Arg-AMP can donate a proton from N
e
H of the cyclic
form instead of the 2¢-OH group of tRNA to C@Oof
the Ca–(C@O)–O–Pa moiety of Arg-AMP, the 2¢-OH
group of tRNA becomes free. The acceptance of a
proton by the –(C@O)–O– moiety changes the C orbi-
tal from an sp
2
hybrid orbital to an sp
3
hybrid orbital,
and the bond is formed between this C and the 2¢-OH
group of tRNA. As Arg-AMP has the ionic guanidi-
nium group and the neutral amino group at pH 8, the

AMP
Arg-AMP
'
'
'
'
'
'
'
'
'
'
'
'
'
'
'
tRNA
Fig. 9. Reaction scheme of Arg transfer from Arg-tRNA to AMP in the deacylation reaction at pH 8.0. A proton of N
e
Hof–N
e+
H@C–(NH
2
)
2
of the Arg moiety with the cyclic configuration of Arg-tRNA is transferred to C@O of the ester bond of Arg-tRNA. The adenosine moiety of
AMP is omitted.
M. Konno et al. Modeling of tRNA-assisted mechanism on ArgRS
FEBS Journal 276 (2009) 4763–4779 ª 2009 The Authors Journal compilation ª 2009 FEBS 4775

m
values for Glu in the aminoacylation reaction
and in the ATP–PP
i
exchange reaction at pH 6.8 are
1.3 · 10
)4
m and 6.7 · 10
)4
m, respectively [33].
The newly determined structure of the ATP analog
(ANP) bound on P. horikoshii ArgRS has been used to
construct a model of the tRNA-assisted ATP–PP
i
exchange reaction and a model of the tRNA-assisted
pyrophosphorolysis reaction of Arg-AMP at low pH.
Modeling of the ArgRS-catalyzed deacylation of Arg-
tRNA at high pH has also been performed on the
basis of the common binding site of ATP and AMP
on P. horikoshii ArgRS. The newly determin ed struc-
ture of the tRNA
Arg
CCU
bound on P. horikoshii
ArgRS provides further information on the plausible
accepting structures for C35 and U36 of tRNA
Arg
CCU
.
In addition, the comparison of the reactivities of wild-

by PCR using plasmid pET28c ⁄ ArgRS as template. A
constructed plasmid (pET28c ⁄ DN ArgRS) was expressed in
E. coli BL21(DE3)codon+, and the crude protein was puri-
fied in the same way as the wild-type protein.
In P. horikoshii, codon usages for AGA and AGG co-
dons are 19 and 34, respectively, and they amount to 98%
among six codons for Arg. The D-loops of isoacceptor
tRNA
UCU
and tRNA
CCU
contain nine (AGCAGGAC
20a
A)
and 10 nucleotides (AGCCA
17a
GGAC
20a
A), respectively.
The P. horikoshii tRNA
Arg
CCU
gene (5¢-GGACCGGTAG
CCTAGCCA
17a
GGAC
20a
AGGG CGGCGGCCTCCTAAG
CCGCAGGTCCGGGGTTCAAATCCCCGCCGGTCCG
CCA-3¢) was cloned with the T7 promoter into the vector

quenched with 5 lL of 1% trichloroacetic acid and spot-
ted onto Whatman 3 MM disks. Radioactivity was quan-
tified in a scintillation counter. The kinetic constants were
derived from a Lineweaver plot.
Crystallization, data collection, and structure
determination
A solution of 10 mm Tris ⁄ HCl buffer (pH 7.5) containing
5mm MgCl
2
and 5–6 mgÆ mL
)1
tRNA
Arg
transcript was
heated at 80 °C for 5 min and then slowly cooled to room
temperature. Then, 10 mm Hepes ⁄ NaOH buffer (pH 7.5)
containing 5 m m MgCl
2
and 2 mgÆmL
)1
ArgRS was added,
so that the ArgRS ⁄ tRNA
Arg
molar ratio was 1 : 1.1. The
resulting mixture was heated at 80 °C for 5 min, and slowly
cooled to room temperature. In the case of the ternary com-
plex, ANP tetralithium salt hydrate (Sigma, St Louis, MO,
USA) was added to a final concentration of 5 mm. Crystals of
the complexes were grown at 20 °C by vapor diffusion in
hanging drops. A 2 lL drop of the above solution containing

P. horikoshii ArgRS from T. thermophilus ArgRS and
S. cerevisiae ArgRS, a large conformational change of the
anticodon loop, the D-loop, and the T-loop, and deviation
of the inclination of base planes of the acceptor stem and
the anticodon-binding stem of tRNA
Arg
CCU
from S. cerevi-
siae tRNA
Arg
ICG
. The repeats of the model building and
refinements using the o program [44] and the cns program
[45] gave a better electron density map, and an ANP mole-
cule was finally identified from the map. Crystallographic
statistics are summarized in Table 1. The final model for
the tertiary and binary complexes has good stereochemistry,
with 92.3% and 91.1% of residues, respectively, in the most
favored regions of the Ramachandran plot [46]. molscript
[47] and raster3d [48] were used for drawing Figs 1 and
3–7, and bobscript also [49] for Fig. 4.
Acknowledgements
This work was supported in part by a grant from the
National Project on Protein Structural and Functional
Analyses to M. Konno.
References
1 Mitra SK & Mehler AH (1966) The role of transfer
ribonucleic acid in the pyrophosphate exchange reaction
of arginine-transfer ribonucleic acid synthetase. J Biol
Chem 241, 5161–5162.

P
jj
F
obs
À
jj
F
calc
jj
=
P
F
obs
jj
, where F
obs
and F
calc
are the
observed and calculated structure factors, respectively (F
obs
> 0.01r). R
free
was calculated with 10% of randomly selected data. Ramachan-
dran plots analysis is performed using
PROCHECK [46].
Ternary complex Binary complex
Data sets ArgRS–tRNA
Arg
–ATP analog (ANP) ArgRS-tRNA

a
90.7 (79.3)
a
Refinement
R
factor
, R
free
(%) 21.3 (27.7)
a
, 25.3 (31.0)
a
20.1 (27.5)
a
, 26.2 (35.3)
a
Number of atoms
Protein, tRNA, ANP, water 5097
b
, 1626, 31, 362 5097
b
, 1626, none, 248
Averaged B-factor (A
˚
2
)
Overall 36.3 37.4
Protein, tRNA, ANP, waters 33.2, 46.5, 34.9, 35.3 35.2, 44.9, none, 33.4
rmsd from ideal geometry
Bonds (A

Reaction pathway and rate-determining step in the
aminoacylation of tRNA
Arg
catalyzed by the arginyl-
tRNA synthetase from yeast. Biochemistry 17, 3740–
3746.
12 Delagoutte B, Moras D & Cavarelli J (2000) tRNA
aminoacylation by arginyl-tRNA synthetase: induced
conformations during substrate binding. EMBO J 21,
5599–5610.
13 Shimada A, Nureki O, Goto M, Takahashi S & Yokoy-
ama S (2001) Structural and mutational studies of the
recognition of the arginine tRNA-specific major identity
element, A20, by arginyl-tRNA synthetase. Proc Natl
Acad Sci USA 98, 13537–13542.
14 Geslain R, Bey G, Cavarelli J & Eriani G (2003)
Limited set of amino acid residues in a class Ia amino-
acyl-tRNA synthetase is crucial for tRNA binding. Bio-
chemistry 42, 15092–15101.
15 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.
16 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.
17 Fukai S, Nureki O, Sekine S, Shimada A, Tao J, Vass-

24 Freist W, Sternbach H & Cramer F (1981) Arginyl-
tRNA synthetase from baker’s yeast. Eur J Biochem
119, 477–482.
25 Ramos A & Varani G (1997) Structure of the acceptor
stem of Escherichia coli tRNA
Ala
: role of the G3ÆU70
base pair in synthetase recognition. Nucleic Acids Res
25, 2083–2090.
26 Tukalo M, Yaremchuk A, Fukunaga R, Yokoyama S
& Cusack S (2005) The crystal structure of leucyl-tRNA
synthetase complexed with tRNA
Leu
in the post-trans-
fer-editing conformation. Nat Struct Mol Biol 12, 923–
930.
27 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.
28 McClain WH & Foss K (1988) Changing the acceptor
identity of a transfer RNA by altering nucleotides in a
‘variable pocket’. Science 241, 1804–1807.
29 Schulman LH & Pelka H (1989) The anticodon con-
tains a major element of the identity of arginine transfer
RNAs. Science 246, 1595–1597.
30 Delagoutte B, Keith G, Moras D & Cavarelli J
(2000) Crystallization and preliminary X-ray crystallo-
graphic analysis of yeast arginyl-tRNA synthetase–
yeast tRNA

mine. Structure 6, 439–449.
38 Geiger T & Clarke S (1987) Deamidation, isomeriza-
tion, and racemization at asparaginyl and aspartyl resi-
dues in peptides. J Biol Chem 262, 785–794.
39 Gangloff J, Schutz A & Dirheimer G (1976) Arginyl-
tRNA synthetase from baker’s yeast: purification and
some properties. Eur J Biochem 65, 177–182.
40 Li J, Yao Y, Liu M & Wang E (2003) Arginyl-tRNA
synthetase with signature sequence KMSK from Bacil-
lus stearothermophilus. Biochem J 376, 773–779.
41 Lin SX, Shi JP, Cheng XD & Wang YL (1988) Arginyl-
tRNA synthetase from Escherichia coli, purification by
affinity chromatography, properties, and steady-state
kinetics. Biochemistry 27, 6343–6348.
42 Otwinowski Z & Minor W (1997) Processing of X-ray
diffraction data collected in oscillation mode. Methods
Enzymol 276, 307–326.
43 CCP4 (1994) The CCP4 Suites: program for protein
crystallography. Acta Crystallogr D 50, 760–763.
44 Jones TA, Zou J-Y, Cowan SW & Kjeldgaard M (1991)
Improved methods for building protein models in elec-
tron density maps and the location of errors in these
models. Acta Crystallogr A 47, 110–119.
45 Bru
¨
nger AT, Adams PD, Clore GM, DeLano WL, Gros
P, Grosse-Kunstleve RW, Jiang J-S, Kuszewski J, Nilges
M, Pannu NS et al. (1998) Crystallography and NMR
system: a new software suite for macromolecular struc-
ture determination. Acta Crystallogr D 54, 905–921.