Substrate recognition and fidelity of maize seryl-tRNA synthetases

Article abstract of DOI:10.1016/j.abb.2012.11.014

Aminoacyl-tRNA synthetases (aaRSs) catalyze the attachment of amino acids to their cognate tRNAs to establish the genetic code. To obtain the high degree of accuracy that is observed in translation, these enzymes must exhibit strict substrate specificity

Full text of DOI:10.1016/j.abb.2012.11.014

Archives of Biochemistry and Biophysics 529 (2013) 122–130
Contents lists available at SciVerse ScienceDirect
Archives of Biochemistry and Biophysics
journal homepage: www.elsevier.com/locate/yabbi
Substrate recognition and fidelity of maize seryl-tRNA synthetases
⇑
Jasmina Rokov-Plavec , Sonja Lesjak, Ita Gruic-Sovulj, Marko Mocibob, Morana Dulic,
Ivana Weygand-Durasevic
Department of Chemistry, Faculty of Science, University of Zagreb, Horvatovac 102a, 10000 Zagreb, Croatia
a r t i c l e i n f o
a b s t r a c t
Article history:
Aminoacyl-tRNA synthetases (aaRSs) catalyze the attachment of amino acids to their cognate tRNAs to
establish the genetic code. To obtain the high degree of accuracy that is observed in translation, these
enzymes must exhibit strict substrate specificity for their cognate amino acids and tRNAs. We studied
the requirements for tRNASer recognition by maize cytosolic seryl-tRNA synthetase (SerRS). The enzyme
efficiently recognized bacterial and eukaryotic tRNAsSer indicating that it can accommodate various types
of tRNASer structures. Discriminator base G73 is crucial for recognition by cytosolic SerRS. Although cyto-
solic SerRS efficiently recognized bacterial tRNAsSer, it is localized exclusively in the cytosol. The fidelity
of maize cytosolic and dually targeted organellar SerRS with respect to amino acid recognition was com-
pared. Organellar SerRS exhibited higher discrimination against tested non-cognate substrates as com-
pared with cytosolic counterpart. Both enzymes showed pre-transfer editing activity implying their
high overall accuracy. The contribution of various reaction pathways in the pre-transfer editing reactions
by maize enzymes were different and dependent on the non-cognate substrate. The fidelity mechanisms
of maize organellar SerRS, high discriminatory power and proofreading, indicate that aaRSs in general
may play an important role in translational quality control in plant mitochondria and chloroplasts.
Ó 2012 Elsevier Inc. All rights reserved.
Received 12 October 2012
and in revised form 28 November 2012
Available online 7 December 2012
Keywords:
Aminoacyl-tRNA synthetase
tRNA
Editing
Plant
Mitochondria
Chloroplast
Introduction
tRNA proceeds with high accuracy based on binding and reaction
kinetics; in both cases the complexity of tRNA molecules allows
The fidelity of protein synthesis relies upon the interpretation of
the genetic code by aminoacyl-tRNA synthetases (aaRSs)1. Each en-
zyme from aaRS family catalyzes covalent attachment of the cognate
amino acid to the specific set of tRNA isoacceptors in a two-step
aminoacylation reaction [1]. The amino acid is first activated by
ATP to form an enzyme-bound aminoacyl-adenylate (aa-AMP) inter-
mediate, accompanied by the release of pyrophosphate. Subse-
quently, the amino acid moiety of the intermediate is transferred
to the 30-terminal ribose of the cognate tRNA, leading to the synthe-
sis of aminoacyl-tRNA and release of AMP.
establishment of many contacts between the tRNA and the protein
either in the ground or in the transition state [3,4]. However, at the
level of amino acid binding and selection step (prior to activation),
a number of aaRSs are unable to discriminate rigorously against
amino acids or their analogs of closely related structures. To avoid
potential mistakes and maintain the fidelity of protein synthesis,
about half of aaRSs evolved complex editing mechanisms [5,6].
Mistakes by aaRSs are generally cleared either by hydrolysis of
the non-cognate aminoacyl-adenylate intermediate in the syn-
thetic site (pre-transfer editing) or, if non-cognate aminoacyl moi-
ety is transferred to tRNA, by hydrolysis of the mischarged
aminoacyl-tRNA within a distinct editing domain (post-transfer
editing). In some cases, hydrolysis of the non-cognate aminoacyl-
adenylate intermediate may be stimulated by tRNA (tRNA-depen-
dent pre-transfer editing). Selective release of the non-cognate
aminoacyl-adenylate from the synthetic site may also contribute
to the pre-transfer editing mechanism. Released aa-AMP may
non-enzymatically hydrolyze in solution. AaRS editing deficiencies
cause mistranslation and can have significant physiological effects,
mainly due to an increased level of misfolded proteins [7–9].
SerRS enzymes from all three domains of life do not possess
editing domain and therefore were considered non-editing aaRS
enzymes [10]. Investigation of specificity of amino acid recognition
revealed weak misactivation of threonine by archeal SerRS from
To obtain the high degree of accuracy that is observed in trans-
lation (error rate of one misincorporated amino acid per 103–104
codons), these enzymes must exhibit strict substrate specificity
for their cognate amino acids and tRNAs [1,2]. Selection of cognate
⇑
Corresponding author. Fax: +385 1 4606 401.
E-mail addresses: rokov@chem.pmf.hr (J. Rokov-Plavec), slesjak@chem.pmf.hr
(S. Lesjak), gruic@chem.pmf.hr (I. Gruic-Sovulj), mocibob@chem.pmf.hr
(M. Mocibob), mdulic@chem.pmf.hr (M. Dulic), weygand@chem.pmf.hr
(I. Weygand-Durasevic).
1
Abbreviations used: aaRS, aminoacyl-tRNA synthetase; SerRS, seryl-tRNA synthe-
tase; ZmcSerRS, maize (Zea mays) cytosolic SerRS; ZmoSerRS, maize (Zea mays)
organellar SerRS; EcSerRS, Escherichia coli SerRS; aa-AMP, aminoacyl-adenylate; IPTG,
isopropyl-b-d-thiogalactopyranoside; TLC, thin layer chromatography; SerHX, serine
hydroxamate.
0003-9861/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.abb.2012.11.014

J. Rokov-Plavec et al. / Archives of Biochemistry and Biophysics 529 (2013) 122–130
123
Methanosarcina barkeri [11,12], as well as weak misactivation of
Supression of bacterial amber mutation
both threonine and cysteine by the yeast cytosolic enzyme [13].
In case of the yeast cytosolic SerRS, these misactivated near-cog-
nate amino acids were shown to be cleared by the pre-transfer
editing within the synthetic site of the enzyme [13]. This provided
the first example of proofreading activity of aaRS naturally lacking
a separate editing domain in all domains of life and it opened a
completely new perspective showing that aaRS may possess
hydrolytic correction even in the absence of the specialized
domain.
The fidelity of amino acid recognition by plant aaRS has not
been extensively studied. This issue is not only important from
the functional and evolutionary point of view, but it should also
be considered in biotechnology, namely in recombinant protein
production in plant cytosol or chloroplasts. Errors in translation
could potentially affect the quality of produced proteins, which is
especially relevant in the case of therapeutic recombinant proteins.
Several transgenic maize plants have been developed for various
biopharming applications [14,15]. Findings that yeast and archaeal
SerRSs showed moderate or low misactivation of near-cognate
amino acids, prompted us to explore the fidelity of maize cytosolic
and dually targeted organellar SerRSs.
Recognition of tRNA by aaRSs is facilitated by both positive and
negative identity elements contained within the tRNA structure
that ensure binding to the proper enzyme [3]. Recognition studies
performed with the components of serylation systems from differ-
ent species, revealed that some, but not all, determinants have
been conserved during evolution [16,17]. The main recognition ele-
ment required for productive tRNA:synthetase complex formation
is a long variable arm of tRNASer, present in all organisms except in
animal mitochondria. Length and orientation of the variable arm
were shown to be important for proper recognition of tRNASer by
SerRS enzymes [18–20]. Structural studies of bacterial tRNAs with
long variable arm (type II tRNAs) have shown that orientation of
the variable arm is influenced by the number of unpaired nucleo-
tides between the variable stem and the nucleotide at position
48 [21]. Furthermore, orientation of the variable arm of bacterial
tRNAsSer is maintained through its tertiary interactions with base
G20B in the loop of D-arm [22]. Another major identity region of
tRNASer is the acceptor arm comprising the discriminator base.
Although strictly conserved in all tRNAsSer, the role and signifi-
cance of the discriminator base G73 varies in different organisms,
even within the eukaryotic domain [16]. Thus far, the importance
of G73 for plant serylation has not been studied.
E. coli strain JR 104 (F0 trpA(UAG)211/glyV55
D
(tonB-trpAB)
argE(UAG) rpoB) was previously described [26]. Suppression of trpA
amber mutations in strain JR104 was tested in cells cotransformed
with pET15b plasmid carrying the gene for maize or bacterial SerRS
and pTech plasmid carrying the gene for tyrosine-specific amber
suppressor supF or its mutated variant supFA73G [27]. Supressor
tRNAsTyr were previously cloned in pTech plasmid under control of
the lpp promoter and the rrnC terminator [28]. Cotransformed cells
were plated on selective M9 minimal glucose plates supplemented
with IPTG (2 mM), arginine (40
lg/ml), chloramphenicol (20 lg/
ml) and ampicilline (100 g/ml), and grown at 30 °C for 20 h.
l
Green fluorescent protein targeting analysis in maize protoplasts
To determine subcellular localization of ZmcSerRS, ZmcSerRS-
GFP fusion construct was prepared. ZmcSerRS coding sequence
was amplified by PCR using pSKZmcSerRS [29] as a template and
cloned in frame with GFP(S65T) coding sequence, behind 35S
CaMV promoter in pTH-2 vector [30]. Protoplasts were isolated
from greening leaves of 10-day-old maize seedlings and trans-
formed by PEG-mediated transformation according to the protocol
described in [31]. The transiently transformed protoplasts were
analyzed with confocal laser scanning microscope Leica TCS SP2
(Leica, Solms, Germany). To stain mitochondria, protoplasts were
incubated with 0.5 lM MitoTracker Red CM-H2XRos (Molecular
Probes) for 30 min and washed before analysis by confocal micros-
copy. Both the GFP fluorophore and chlorophyll were excited with
argon laser at 488 nm. GFP fluorescence and far-red chlorophyll
fluorescence were collected between 500–535 nm and 650–
700 nm, respectively, in two separate channels. MitoTracker Red
CM-H2XRos fluorescence was detected between 600 and 610 nm
with excitation at 578 nm. The programs LCSLite and Adobe Photo-
shop were used for post-acquisition image processing.
ATP–PPi exchange
Proteins ZmcSerRS and ZmoSerRS were overexpressed and puri-
fied as previously described [23,32]. The samples of amino acids
and serine hydroxamate (SerHX) were checked by mass spectrom-
etry and contamination with serine was not detected. ATP–PPi ex-
change was measured at 37 °C in 100 mM HEPES pH 7.0, 20 mM
MgCl2, 25 mM KCl, 4 mM ATP and 1 mM
[
32P]PPi (0.002–
0.01 mCi/ml). Amino acid and analog concentrations varied be-
tween 0.1 and 10 ꢀ KM. Due to low solubility, the concentration
of cysteine in reactions with ZmoSerRS was varied between 0.1
and 5 ꢀ KM. Enzyme concentrations used were 100–200 nM in
Material and methods
reactions with serine, 500 nM in reactions with D,L-SerHX and 1–
Complementation of E. coli strain KL229
2 lM in reactions with threonine and cysteine. The reactions were
stopped with sodium acetate (pH 5.0) and SDS (final concentra-
tions 333 mM and 0.067%, respectively). The generated [32P]ATP
was separated from the remaining [32P]PPi by thin layer chroma-
tography (TLC) on polyethyleneimine plates, using 750 mM KH2-
PO4 pH 3.5 and 4 M urea as developer. The obtained signal was
visualized on Storm PhosphorImager and then quantified using
ImageQuant software. Kinetic parameters were determined by fit-
ting the data directly to the Michaelis–Menten equation using non-
linear regression. The parameters were obtained from three inde-
pendent measurements.
The cDNA encoding maize ZmcSerRS was amplified by PCR
using pET28bZmcSerRS [23] as a template and cloned into pET15b.
The gene for Escherichia coli SerRS was amplified by PCR from E. coli
genomic DNA and cloned into pET15b. E. coli strain KL229 carries a
mutated serS gene responsible for the temperature-sensitive phe-
notype [24]. KL229 cells were transformed with pET15bZmcSerRS,
pET15bEcSerRS or empty pET15b and tested for growth at permis-
sive (30 °C) and nonpermissive temperature (34 °C) on M9/glucose
minimal plates supplemented with ampicillin (100
lg/ml), thy-
mine (200 g/ml) and isopropyl-b- -thiogalactopyranoside (IPTG)
l
D
(2 mM) where indicated. The addition of IPTG ensures the dere-
pression of the T7 promoter, which is followed by the lac operator
in the pET15b vector. As already demonstrated [25], the T7 pro-
moter allows a low level of transcription even in bacterial strains
that do not contain T7 RNA polymerase.
Aminoacyl-adenylate synthesis assay
Aminoacyl-adenylate synthesis assay was done in 50 mM
HEPES pH 7.0, 25 mM KCl, 20 mM MgCl2, 0.5 mM ATP
(0.01–0.1 mCi/ml). Concentration of serine,
D,L-SerHX, threonine

124
J. Rokov-Plavec et al. / Archives of Biochemistry and Biophysics 529 (2013) 122–130
Table 1
and cysteine were 1 mM, 250 mM, 400 mM and 400 mM, respec-
tively. Enzymes were present in concentration of 2 M. The reac-
Steady-state kinetic parameters for ZmcSerRS with respect to tRNASer in aminoacy-
lation assay.
l
tion was started by addition of amino acid or analog and stopped
by quenching with 400 mM sodium acetate (pH 5.0) and 0.1%
SDS. Separation of [32P]adenylate from aminoacyl-[32P]adenylate
KM
(l
M)
kcat (sꢁ1
)
kcat/KM (l )
Mꢁ1sꢁ1
E. coli tRNASer
Maize tRNASer
Yeast tRNASer
0.8628
0.58 0.06
0.77 0.1
1.1
1.48 0.05
1.1 0.1
1.275
2.55
1.43
a
and
a
-[32P]ATP was achieved by TLC followed by quantification
a
and kinetic analysis as already described in [33]. The data were ob-
tained from three independent measurements. Control reactions
containing [32P]ATP and maize SerRSs showed that neither enzyme
possesses significant inherent ATPase activity.
E. coli tRNASer – total E. coli tRNA with quantified tRNASer content maize tRNASer
,
yeast tRNASer – total (mainly cytosolic) maize and yeast tRNA, respectively, with
quantified tRNASer content.
a
Data taken from Ref. [23].
Results
conversely, cytosolic tRNAs are poor substrates for organellar and
bacterial enzymes [34]. However, some dually targeted aaRSs can
recognize tRNAs of different origins [35]. Taking into account that
dual targeting of aaRSs is especially pronounced in plants [35] and
Maize cytosolic SerRS efficiently recognizes bacterial tRNASer, both
in vivo and in vitro
that maize cytosolic SerRS efficiently recognized E. coli tRNAsSer
,
We have previously shown that expression of maize cytosolic
SerRS (ZmcSerRS) cDNA efficiently complemented yeast cytosolic
SerRS knock-out strain [23]. Here we tested ZmcSerRS cDNA for
its ability to complement E. coli strain KL229 encoding thermola-
bile SerRS. Transformants of KL229 cells with pET15b plasmids,
carrying maize cytosolic or E. coli SerRS coding sequences, were
grown at a permissive temperature (30 °C) and then transferred
to a non-permissive temperature (34 °C) in the presence or ab-
sence of IPTG (Fig. 1). Addition of IPTG to the medium relieves
the repression of the lac operator located upstream of the cloned
ORFs and enables low levels of constitutive expression of the
cloned synthetases due to weak recognition of T7 promoter by bac-
terial RNA polymerase [25]. In the presence of IPTG, the transfor-
mants expressing maize cytosolic SerRS showed efficient growth
at non-permisive temperature (Fig. 1). The complementation with
ZmcSerRS was nearly as efficient as with wild type E. coli SerRS
(SerEC) (Fig. 1) and therefore we examined how well ZmcSerRS
recognizes bacterial tRNASer in vitro. Aminoacylation kinetic
parameters for ZmcSerRS with respect to E. coli tRNASer were al-
most the same as kinetic parameters previously determined for
yeast cytosolic tRNASer and comparable with parameters obtained
in the reaction using homologous maize cytosolic tRNASer as sub-
strate (Table 1, [23]).
which are very similar to maize organellar tRNAsSer [32,36], we
have tested cellular distribution of ZmcSerRS using transient
expression assay in maize protoplasts. Protoplasts were isolated
from maize seedling leaves and transformed with the construct
containing ZmcSerRS cDNA fused to the GFP reporter gene under
the control of the strong 35S CaMV promoter. After incubation in
the dark for 24 h, the expression of the fusion product was exam-
ined by confocal laser scanning microscopy (Fig. 2). As a control,
we used protoplasts transiently transformed with GFP construct
alone. To identify mitochondria, protoplasts were simultaneously
stained with the MitoTracker Red CM-H2XRos, a mitochondria spe-
cific fluorescent dye. Chloroplasts were identified by chlorophyll
fluorescence. In transformed protoplasts, transient expression of
the ZmcSerRS-GFP fusion protein resulted in GFP fluorescence sig-
nal throughout the cytosol and showed no association with mito-
chondria or chloroplasts, indicating that ZmcSerRS is localized in
the cytosol (Fig. 2A). The same fluorescence pattern was seen with
control protoplasts expressing GFP protein alone (Fig. 2B). In addi-
tion, we have performed in vitro import experiments of ZmcSerRS
protein into isolated mitochondria and chloroplasts (data not
shown). These experiments did not show import of ZmcSerRS into
the organelles, which confirmed exclusive cytosolic localization of
ZmcSerRS protein in maize.
Subcellular localization of ZmcSerRS as determined by targeting of GFP
fusion
Comparison of maize, yeast and E. coli tRNASer consensus sequences
This interesting broad tRNASer specificity of ZmcSerRS
prompted us to examine its cellular distribution, since cellular
localization of aaRS is often connected to its tRNA recognition pat-
tern. The general rule is that organellar, as well as E. coli tRNAs are
better substrates for organellar than for cytosolic aaRSs and
Our data (Table 1) shows that maize cytosolic SerRS serylates
E. coli and yeast cytosolic tRNAsSer in vitro with good efficiency,
implying that both bacterial and yeast cytosolic tRNAsSer possess
the identity determinants required for serylation by the maize en-
zyme. Moreover, efficient cross-species in vivo complementation
experiments (Fig. 1 and [23]) imply that ZmcSerRS recognizes all
serine isoacceptors needed to decode the complete repertoire of
serine codons in heterologous hosts, Saccharomyces cerevisiae and
E. coli. Therefore we have compared consensus sequences of E. coli
and yeast cytosolic tRNAsSer with the maize cytosolic tRNASer con-
sensus sequence. Inspection of the Genomic tRNA Database ([37],
http://gtrnadb.ucsc.edu) revealed that maize nuclear genome con-
tains 61 tRNASer genes, of which 23 show 95–100% identity to maize
organellar tRNASer genes (data not shown). Eukaryotic nuclear gen-
omes often contain mitochondrial and plastidial DNA fragments,
including organellar-like tRNA genes, that were integrated into
the nuclear genomes over the course of evolution [38]. It is pro-
posed that these nuclear tRNA genes of organellar ancestry are
not expressed [38] and therefore maize cytosolic tRNASer consensus
(Fig. 3) was compiled from sequences of nuclear tRNASer genes of
eukaryotic origin. Due to their eukaryotic nature, maize and yeast
cytosolic tRNAsSer have more conserved nucleotides in common
Fig. 1. Complementation of SerRS function in E. coli strain KL229 (encoding
thermolabile SerRS). KL229 was transformed with pET15bZmcSerRS, pET15bEcS-
erRS or empty pET15b. Three individual colonies obtained from each transforma-
tion were grown for 24 h at 30 °C and then restreaked to fresh plates (M9glu
containing 200
lg/ml thymine, 100 lg/ml ampicillin and 2 mM IPTG where
indicated) and grown at non-permissive temperature 34 °C.

J. Rokov-Plavec et al. / Archives of Biochemistry and Biophysics 529 (2013) 122–130
125
Fig. 2. Subcellular localization of ZmcSerRS. Transient expression of GFP constructs in maize protoplasts visualized by confocal microscopy. (A) ZmcSerRS–GFP. (B) GFP. GFP
fluorescence was detected in the green channel, chlorophyll autofluorescence in the far-red channel and Mitotracker Red CM-H2Xros fluorescence in the red channel.
GFP + chlorophyll and GFP + Mitotracker represent merging of GFP with chlorophyll autofluorescence, and GFP with Mitotracker Red CM-H2Xros fluorescence, respectively.
Images correspond to confocal single scans through protoplasts. Scale bars, 10
the web version of this article.)
lm. (For interpretation of the references to colour in this figure legend, the reader is referred to
than maize cytosolic and E. coli serine-specific isoacceptors (Fig. 3).
The main differences are due to different evolutionary origin of
maize cytosolic and bacterial tRNAsSer and are related to the length
and orientation of the variable arm (Fig. 3 and [36]). On the other
hand, nucleotides in consensus sequence of maize cytosolic
tRNAsSer that are identical to consensus nucleotides of both E. coli
and yeast cytosolic tRNAsSer are good candidates as minor or major
identity determinants for maize cytosolic SerRS (Fig. 3, indicated
with capital bold letters).
can only be suppressed with a limited set of amino acids, including
serine, but not tyrosine [39], and is therefore useful for testing of
tRNA requirements in serylation. Suppression was monitored in
strain JR104 co-expressing supF variants from pTech plasmid with
EcSerRS or ZmcSerRS from pET15b plasmid (Fig. 4). Low level of
supression, as indicated by weak growth, is observed when the
mutant supFA73G is expressed, most likely due to mischarging
by endogenous E. coli SerRS. When the bacterial enzyme is ex-
pressed from pET15b plasmid, supression is achieved with both
supF variants. As judged by the strength of growth, EcSerRS moder-
ately mischarges supF and more efficiently supFA73G, which is in
accordance with our previous results [28]. Coexpression of maize
cytosolic SerRS with supFA73G leads to efficient suppression, while
no suppression is observed when ZmcSerRS is coexpressed with
supF. This indicates a strong preference of the maize enzyme for
G at the position 73. Furthermore, the preference is stronger than
for E. coli SerRS which, when expressed from pET15b plasmid,
moderately recognized the original supF suppressor (Fig. 4).
Identification of G73 as a major identity element recognized by
ZmcSerRS
The recognition of tRNA is contextually dependent, and identity
requirements for a given system are influenced by interplay with
other isoaccepting systems with substantially similar tRNA sub-
strates. In the maize cytosol, as in other eukaryotic organisms,
there are only two type II tRNAs comprising long variable arm,
tRNASer and tRNALeu. We have analyzed maize cytosolic tRNALeu se-
quences extracted from Genomic tRNA Database (data not shown).
Both maize type II tRNAs have the same orientation of the variable
arm, since both have one unpaired nucleotide at the base of the
Near-cognate amino acid selectivity and pretransfer editing of maize
seryl-tRNA synthetases
variable arm. Since tRNAsLeu are structurally similar to tRNAsSer
they have to be accurately discriminated by ZmcSerRS. Therefore,
we assumed that among potential maize tRNASer identity elements,
the major determinants would be those that are not present in any
of the maize tRNALeu isoacceptors. The only difference between
consensus cytosolic tRNASer and all cytosolic tRNALeu isoacceptors
turned out to be the discriminator base position 73 (Fig. 3). There-
fore we set out to explore the importance of discriminator base for
cytosolic serylation in maize using previously developed bacterial
genetic suppression system based on E. coli tRNATyr [28].
,
Proper recognition of cognate and discrimination against non-
cognate amino acids is also a very important aspect of aaRS activ-
ity. In order to assess the specificity of amino acid recognition by
maize cytosolic SerRS several compounds were tested for misacti-
vation. We have also included in this investigation dually targeted
organellar maize SerRS (ZmoSerRS) because it was shown that the
requirements for translation quality control may differ among cel-
lular compartments [40]. Considering that ZmoSerRS serves its
function in both mitochondria and chloroplasts [31], it would be
of particular interest to explore its amino acid specificity.
In bacteria, unlike in eukaryotes and archaea, tRNATyr is a type II
tRNA with a long variable arm. The suppression system relies on
supF [27], an efficient tyrosine-specific amber suppressor tRNA,
comprising an adenine at the position of discriminator base and
its variant with mutation A73G. In E. coli strain JR104, trpA gene
carries an amber mutation at position 211. Mutation trpA(UAG)211
The specificity of active sites of both maize SerRSs was investi-
gated by steady-state ATP–PPi exchange assay that allows extrac-
tion of KM and kcat kinetic parameters for amino acid activation
(Table 2). Both enzymes exhibited similar catalytic efficiency to-
ward cognate serine. When near-cognate amino acids were tested,
weak activation of both threonine and cysteine by both enzymes

126
J. Rokov-Plavec et al. / Archives of Biochemistry and Biophysics 529 (2013) 122–130
aaRS generally needs to achieve at least 3300-fold specificity for
cognate over non-cognate amino acids to avoid elevating the error
rate of protein synthesis. Lower discrimination against non-cog-
nate substrates implies the need for editing by the aaRS or other
proteins. Specificity of ZmoSerRS for serine over both near-cognate
amino acids was above 3300-fold (Table 2). ZmcSerRS discrimi-
nated well against cysteine, while discrimination factor against
threonine was below 3300 (Table 2), indicating that ZmcSerRS
might need proofreading to enhance the accuracy in amino acid
recognition. To test this hypothesis, we investigated whether
ZmcSerRS is capable of hydrolytic editing. It is known that some
synthetases exhibit an editing function even if their discrimination
against non-cognate substrates is high [2], therefore we have also
tested the editing capability of ZmoSerRS.
Editing reactions are preferentially followed by the assay that
utilizes
a
-[32P]ATP (aminoacyl-adenylate synthesis assay, [33])
and thus allows tracking of both adenylate and aminoacyl-adenyl-
ate after separation by thin layer chromatography (TLC). Each time
non-cognate amino acid is activated and its respective aa-AMP is
hydrolyzed by aaRS, one molecule of ATP is converted into AMP.
In the absence of tRNA, successive rounds of activation and hydro-
lysis lead to accumulation of AMP which is a measure of tRNA-
independent hydrolytic pre-transfer editing. In the case of selective
release, aa-AMP dissociates from the enzyme and may accumulate
in the solution depending on rates of its dissociation and non-
enzymatic hydrolysis. A fraction of released aa-AMP that under-
goes non-enzymatic hydrolysis in solution also contributes to
AMP formation. Determination of rates for AMP formation, aa-
AMP and non-enzymatic aa-AMP hydrolysis enables better insight
into the contribution of various reaction pathways in the pre-
transfer editing reactions.
For both maize enzymes, the rate of adenylate formation in the
presence of near-cognate amino acids was at least an order of mag-
nitude higher than in the presence of cognate serine (Table 3,
Fig. 5). This confirmed that both enzymes possess active pre-trans-
fer editing toward threonine and cysteine. Observed catalytic con-
stants for adenylate formation by ZmcSerRS in the presence of
threonine and cysteine were 34-fold and 17-fold higher, respec-
tively, than in the presence of serine. Similarly, the obtained cata-
lytic rates of ZmoSerRS in the presence of threonine and cysteine
were 15-fold and 23-fold greater than that observed with serine.
Catalytic rate constants observed in the presence of near-cognate
amino acids are at least 10-fold higher than the rate constants of
spontaneous non-enzymatic hydrolysis of aminoacyl-adenylates
determined previously (1.7 ꢀ 10ꢁ4 sꢁ1–2 ꢀ 10ꢁ4 sꢁ1) [13,41]. This
implies that AMP predominantly originates from the intrinsic
ZmcSerRS and ZmoSerRS hydrolytic activity toward near-cognate
aa-AMP. The highest rate of adenylate formation was observed
with ZmcSerRS in the presence of threonine; threonine-dependent
stimulation of AMP formation was ꢂ2-fold higher as compared
with cysteine (Table 3, Fig. 5). Furthermore, ZmcSerRS was ꢂ4-fold
more efficient than ZmoSerRS in enzymatic hydrolysis of threonyl-
adenylate. Observed higher editing activity by ZmcSerRS is in
accordance with the predicted need for threonine editing (Table 2).
Interestingly, editing was also observed in cases where synthetic
reactions seem to be accurate enough (Table 2). Possible biological
rationale is discussed later. Addition of tRNASer did not stimulate
threonine-dependent AMP formation by ZmcSerRS (data not
shown).
Fig. 3. Maize cytosolic tRNASer consensus structure. Dots represent positions that
are not conserved in maize cytosolic tRNAsSer. Nucleotides conserved in all maize
cytosolic tRNASer isoacceptors are indicated by letters. Capital letters refer to
nucleotides conserved among maize and yeast cytosolic tRNAsSer. Bold small letters
indicate nucleotides conserved between maize cytosolic and E. coli tRNAsSer. Bold
capital letters indicate nucleotides conserved among maize cytosolic, yeast
cytosolic and E. coli isoacceptors. Underlined nucleotides indicate differences
among putative identity elements of maize cytosolic tRNASer and supF. Arrows show
positions of structural differences between maize cytosolic tRNAsSer, E. coli tRNAsSer
and supF. Plain arrow shows nucleotide at position 20B (none in maize cytosolic
tRNAsSer, G in E. coli tRNAsSer, A in supF). Dotted arrow shows number of nucleotides
at the base of variable arm (one in maize cytosolic tRNAsSer, none in E. coli tRNAsSer
,
two in supF). Dashed arrow shows number of base pairs in variable arm (4 bp in
cytosolic maize tRNAsSer, 5–7 bp in E. coli tRNAsSer, 3 bp in supF). Nucleotide G73,
that is absolutely conserved within maize, yeast and E. coli tRNAsSer and never
present in any of the maize cytosolic tRNALeu isoacceptors, is circled.
was observed. The kinetic constants showed an increased KM and a
decreased kcat as compared with serine in both cases. Since the ob-
served error rates in translation do not exceed one in 103–104 [2],
To evaluate the contribution of the selective release mechanism
in overall pre-transfer editing we monitored the formation of ami-
noacyl-adenylates. Aminoacyl-adenylate can be stably bound to the
enzyme, and therefore the formation of aa-AMP in the concentra-
tion that exceeds the concentration of the enzyme indicates its en-
hanced dissociation from the active site into the solution. In the
editing reaction with threonine and ZmcSerRS, steady state rate
Fig. 4. Suppression of amber trpA(UGA) mutation in E. coli strain JR104. Escherichia
coli strain JR104 was cotransformed with pTech plasmid carrying the amber
suppressor tRNATyr supF gene or its variant supFA73G and pET15b plasmid carrying
the serine-specific synthetase gene (EcSerRS or ZmcSerRS). Suppression of
trpA(UGA) was checked by streaking the transformants on M9 minimal glucose
plates supplemented with IPTG, arginine and required antibiotics. ZmcSerRS
displays strict dependence on discriminator base G73, unlike EcSerRS.

J. Rokov-Plavec et al. / Archives of Biochemistry and Biophysics 529 (2013) 122–130
127
Table 2
Steady-state kinetic parameters for ATP–PPi exchange for maize cytosolic and
organellar SerRSs.
KM (mM)
kcat (sꢁ1
)
kcat/KM
Discrimination
factora
(mMꢁ1 sꢁ1
)
ZmcSerRS
Ser
0.18 0.03
2.35 0.03
2.5 0.1
2.433 0.003
13.9
1.04
1
13
D,L-SerHX
Thr
Cys
23.4 0.2
0.14 0.01
0.078 0.005
0.0060
0.0027
2300
5100
29
3
ZmoSerRS
Ser
0.37 0.06
8.6 0.9
3.2 0.2
1.78 0.08
8.6
0.21
1
41
Fig. 6. Thin-layer chromatograms of reaction time courses of typical aminoacyl-
D,L-SerHX
adenylate synthesis reactions containing SerHX and ZmcSerRS or ZmoSerRS.
Thr
Cys
54 12
84 14
0.11 0.01
0.11 0.01
0.0020
0.0013
4300
6600
a
Discrimination factor represents ratio of kcat/KM for serine and kcat/KM for
SerHX behaves as a non-cognate substrate in activation and editing
reactions catalyzed by yeast cytosolic, bacterial and methanogenic
SerRSs [13,41,43]. In order to establish the influence of SerHX on
plant enzymes, we have tested maize enzymes for its activation.
Contrary to weak misactivation of near-cognate amino acids, both
maize enzymes very efficiently recognized and activated SerHX
(Table 2). The apparent affinity of both enzymes in terms of KM
was lower for SerHX as compared with serine (12.9-fold for ZmcS-
erRS and 23.3-fold for ZmoSerRS). On the contrary, kcat values were
of the same order of magnitude. In particular, kcat of ZmcSerRS for
serine and SerHX were almost identical. Therefore, under saturat-
ing conditions SerHX activation proceeds with the same rate as
the cognate reaction. The cytosolic SerRS was 5 times more effi-
cient in SerHX activation (as reflected in kcat/KM) than the organel-
lar enzyme. More pronounced activation of SerHX by ZmcSerRS
was also accompanied by 5-fold enhanced editing of SerHX com-
pared to ZmoSerRS (Table 3, Fig. 5, Fig. 6). In both cases AMP for-
mation (Table 3) was significantly faster than the rate of non-
enzymatic hydrolysis of SerHX-AMP (0.17 ꢀ 10ꢁ3 sꢁ1, [43]) sug-
gesting that both enzymes possess hydrolytic activity toward Ser-
HX-AMP. In contrast, measurable accumulation of SerHX-AMP
((1.8 0.1) ꢀ 10ꢁ3 sꢁ1), in the amount that exceeded the concen-
tration of enzyme used, was observed only in the presence of
ZmoSerRS (Fig. 5B and Fig. 6).
respective serine analog or near-cognate amino acid: SerHX, Thr or Cys.
Table 3
Observed steady-state rate constants for adenylate formation by maize cytosolic and
organellar SerRS.
kobs (10ꢁ3 sꢁ1
)
ZmcSerRS
ZmoSerRS
Ser
-SerHX
0.25 0.05
26.1 2.1
0.15 0.04
4.8 0.5
D,L
Thr
Cys
8.5 2.2
4.3 1.0
2.2 0.3
3.4 0.9
of Thr-AMP formation was (3.8 0.5) ꢀ 10ꢁ3 sꢁ1, with the end con-
centration exceeding that of the enzyme (Fig. 5A). The rate constant
for AMP formation (Table 3) is only 2-fold higher than the rate con-
stant for Thr-AMP formation by ZmcSerRS, indicating that en-
hanced dissociation from the enzyme active site contributes
significantly to overall elimination of near-cognate Thr-AMP. In
the presence of ZmoSerRS, Thr-AMP formation was rather slow
((0.41 0.08) ꢀ 10ꢁ3 sꢁ1) and its final concentration did not exceed
the concentration of enyzme used (data not shown). For both maize
enzymes, in the presence of cysteine, Cys-AMP was not accumu-
lated to the detectable level (data not shown), implying that selec-
tive release of Cys-AMP does not contribute to cysteine editing.
Discussion
Serine hydroxamate is efficiently activated, but its adenylate is
differentially edited by two maize SerRSs
Broad tRNASer specificity of maize cytosolic SerRS
In the present and previous study [23], we demonstrated that
the maize cytosolic SerRS can efficiently aminoacylate bacterial
and eukaryotic cytosolic tRNAsSer in vitro and in vivo. Thus, despite
its eukaryotic features revealed by a strong similarity to other
Serine hydroxamate (SerHX) is an artificial serine analog with
hydroxamate group instead of carboxylate that was shown to be
a competitive inhibitor of many SerRS enzymes [42]. In addition,
Fig. 5. Representative plots of adenylate synthesis in pre-transfer editing assay by (A) ZmcSerRS or (B) ZmoSerRS in the presence of SerHX (d), threonine (j), cysteine (N) and
serine (.). Dashed lines indicate formation of aminoacyl-adenylates whose final concentrations exceeded the concentrations of enzyme used (2
AMP (h) in the presence of ZmcSerRS. (B) Formation of SerHX-AMP (s) in the presence of ZmoSerRS.
lM). (A) Formation of Thr-

128
J. Rokov-Plavec et al. / Archives of Biochemistry and Biophysics 529 (2013) 122–130
eukaryotic cytosolic SerRSs [32], maize cytosolic SerRS was able to
efficiently recognize bacterial tRNAsSer, that are clearly distinguish-
able from maize cytosolic tRNAsSer which show typical eukaryotic
characteristics (Fig. 3 and [36]). This finding is in clear contrast
with tRNA specificity of maize organellar SerRS that showed clear
preference for substrates of bacterial origin, while no or poor
aminoacylation was observed for plant and yeast cytosolic eukary-
otic tRNAsSer, respectively [32,44]. The main differences between
eukaryotic cytosolic and bacterial tRNAsSer are related to the length
and orientation of the variable arm (Fig. 3 and [36]). Although the
crystal structure of eukaryotic SerRS:tRNA complex has not yet
been reported, it is presumed that bacterial and eukaryotic
cytosolic tRNAsSer have a different spatial orientation of the vari-
able arm due to differences in the occupation of position 20B in
the D-loop and in the number of unpaired nucleotides at the base
of the variable arm (Fig. 3 and [36]). If this is the case, then maize
cytosolic SerRS has considerable flexibility in regard to substrate
recognition by ZmcSerRS. Positions G9, G13, U20A and A21 were
shown to be important for tertiary core packing of bacterial tRNASer
[22]. Others, except for A38, are conserved or semiconserved in all
canonical tRNAs [53]. We have focused our attention to G73 since
it is not present in any of the maize cytosolic tRNAsLeu, which are
structurally very similar to maize cytosolic tRNAsSer
.
Our in vivo experiments show that ZmcSerRS efficiently sup-
presses the amber mutation in trpA gene only when it is coexpres-
sed with mutant form of suppressor tRNATyr supFA73G, while no
suppresion is observed with the original supF, containing A73
(Fig. 4). Although supF contains 21 out of 24 nucleotides that are
conserved within maize, yeast and E. coli serine isoacceptors
(Fig. 3), it is not a substrate for maize cytosolic SerRS. Differences
are at positions 9, 20A and 73. In addition, variable arm of supF
comprises three base pairs and two unpaired nucleotides at its
base (Fig. 3). Interestingly, aminoacylation in vivo is established
when the discriminator base of supF is changed from A to G indi-
cating that a shorter variable arm in a different spatial orientation
is not an obstacle for in vivo aminoacylation of supF by maize cyto-
solic SerRS as long as the discriminator base remains G73. So, the
exchange of the discriminator base A73 for G is alone sufficient
to convert tyrosine-specific supF into a serine-acceptor in vivo, sug-
gesting that the discriminator base G73 is essential for recognition
by maize cytosolic SerRS. This sensitivity to the discriminator base
is reminiscent of the recognition mechanism described for the hu-
man serylation system, where replacement of the leucine-specific
discriminator base A73 in human tRNALeu by serine-specific G73
alone was sufficient for an identity switch to serine acceptance
in vitro [51]. Similarly, methanogenic archaeal and Trypansoma cru-
zi SerRSs adopted G73 as one of the major elements of serine iden-
tity [20,52]. In contrast, in bacteria and yeast, G73 is not important
for SerRS recognition but serves only as an antideterminant for leu-
cyl-tRNA synthetase [18,54]. In order to unambiguously demon-
strate the importance of G73 and other putative minor or major
elements of tRNASer identity, experiments may be performed using
tRNASer in vitro transcripts and their variants. All maize cytosolic
tRNALeu species invariably contain adenine at the discriminator po-
sition (data not shown). Strict conservation of A73 in all maize
tRNALeu species taken together with our result that supF (which
contains A73) is not aminoacylated in vivo by ZmcSerRS suggest
that the discriminator base A73 of maize tRNAsLeu may play a role
of the negative identity element for maize cytosolic SerRS.
tRNASer
We have shown previously that another eukaryotic SerRS, the
yeast cytosolic SerRS aminoacylates in vitro both types of tRNASer
.
,
but its complementation in vivo of a thermosensitive variant of
E. coli SerRS was very weak [25]. Using in vitro experiments it
has been shown that atypical methanogenic-type archaeal SerRS
recognizes tRNAsSer from all three domains of life [45]. This was
explained by the fact that archaeal tRNAsSer show mixed bacterial
and eukaryotic features. However, this SerRS was unable to com-
plement the function of bacterial and yeast SerRSs in vivo [28].
Canonical archaeal SerRS from Pyrococcus horikoshii efficiently
aminoacylated bacterial tRNAsSer, but its function in vivo was
not tested [46]. In contrast, the bacterial SerRSs strictly discrimi-
nate the bacterial tRNASer species from the eukaryotic and archa-
eal ones [32,46,47]. Similarly, maize organellar SerRS is specific
only for bacterial-type tRNASer [32]. Thus far, maize cytosolic
SerRS is the only SerRS with broad tRNASer specificity that has
been efficiently demonstrated both in vivo and in vitro. Therefore,
it could be possible that this broad tRNASer specificity has func-
tional implications with regard to recognition of tRNASer of bacte-
rial origin. Our inspection of 61 maize nuclear tRNASer genes
revealed that more than 1/3 of them are of organellar origin dis-
playing typical bacterial-like structural features. It is assumed
that organellar-like tDNAs integrated into a nuclear genome are
not functional [38]. Due to the lack of experimental evidence
[48], we can not exclude the possibility that some of the trans-
ferred organellar tRNASer genes are expressed and that their cor-
responding bacterial-like tRNAsSer are recognized and
aminoacylated by ZmcSerRS in the maize cytosol, possibly giving
a certain advantage to maize. If that is the case, the observed
broad tRNASer specificity of maize cytosolic SerRS may also have
functional consequences. This possibility certainly has to be fur-
ther investigated. In addition, phylogenetic analysis has shown
that Arabidopsis cytosolic and organellar SerRS genes originated
from a duplication event of an inherited gene from Actinobacteria
[49]. It is possible that during evolution cytosolic SerRS retained
efficient recognition of bacterial tRNASer species, a functional
characteristic that was inherent to the gene product of the bacte-
rial ancestral gene.
Fidelity of maize cytosolic and organellar SerRS
Research on the role of aaRSs in maintaining translational accu-
racy in plants is very limited. Proofreading activities were shown
for bean cytosolic and chloroplastic phenylalanyl-tRNA syntheta-
ses, yellow lupin seed valyl-tRNA synthetase and rice methionyl-
tRNA synthetase [55–57]. Amino acid discrimination of plant argi-
nyl-tRNA synthetases depends on interaction of structural features
of the tRNA with the enzymes [58,59]. Here we investigated the
fidelity of maize cytosolic and organellar seryl-tRNA synthetases.
We assesed the specificity of the SerRS active sites towards threo-
nine and cysteine, two proteinogenic near-cognate amino acids
that are structurally similar to serine and could therefore be poten-
tially misactivated by SerRS. Both maize SerRS enzymes discrimi-
nated against cysteine better than threonine. In general, the
discrimination threshold [based on (kcat/KM, cognate)/(kcat/KM,
noncognate)] taken as a requirement for editing is approximately
1 error in 3300. In the presence of threonine, maize cytosolic SerRS
did not meet this threshold (Table 2). However, ZmcSerRS edited
Thr-AMP, implying that impaired amino acid discrimination could
be compensated by its corrective activity. In contrast, our data
revealed that discrimination of ZmcSerRS against cysteine and
ZmoSerRS against both threonine and cysteine was satisfactory
Importance of discriminator base G73 for maize cytosolic serylation
Comparison of tRNASer substrates efficiently aminoacylated by
maize cytosolic SerRS indicated nucleotides that could be putative
minor or major elements of serine identity (Fig. 3, indicated with
capital bold letters). Of those, discriminator base G73 and base
pairs G1:C72 and C11:G24 were previously identified as tRNASer
identity elements of varying importance in serine system of
different organisms [20], [50–52], and may also be important for

J. Rokov-Plavec et al. / Archives of Biochemistry and Biophysics 529 (2013) 122–130
129
Table 4
implying no need for editing. Nevertheless, we observed that both
enzymes display editing toward both threonine and cysteine inde-
pendently of the specificity obtained in the synthetic reaction (Ta-
ble 3, Fig. 5). It is, however, important to underline that the most
efficient editing was observed for the pair with the smallest spec-
ificity in activation: ZmcSerRS and threonine.
In contrast to weak misactivation of proteinogenic near-cognate
amino acids, both maize enzymes very efficiently recognized and
activated artificial serine analog SerHX (Table 2). Therefore, pertur-
bations of the side chain on serine molecule influenced activation
more efficiently than substitution of carboxylate by larger
hydroxamate group. Both maize enzymes hydrolyzed SerHX-AMP
more efficiently as compared with near-cognate aa-AMPs, but
the difference was more pronounced for ZmcSerRS (Table 3,
Fig. 5). This is analogous to the previous data showing that SerHX
is the preferential editing substrate for eukaryotic yeast cytosolic
but not for the bacterial E. coli SerRS [13].
Predicted selectivity of ZmcSerRS and ZmoSerRS in maize. Note that in the case of Ser/
Thr concentration ratio which is observed in pea embryo, selectivity of both SerRS
enzymes is predicted to be much lower than needed fidelity threshold (1 in 3300).
Specificity
Ser/Thr
Concentration
Ser/Thr
Selectivity
Ser/Thr
ZmcSerRS
Maize cytosol
Pea embryo
2300
2300
3.5:1a
1:2.7b
8050
850
ZmoSerRS
Maize chloroplasts
Mitochondria
Pea embryo
4300
4300
4300
3.5:1a
1.35:1c
1:2.7b
15,050
5800
1590
a
b
c
Data taken from Ref. [63].
Data taken from Ref. [68].
Since there are no data on plant mitochondria, data on animal mitochondria are
used [78].
Taking into account that both maize enzymes do not possess
editing domains, the observed hydrolytic activity towards near-
cognate aa-AMPs and SerHX-AMP seems to reside within the syn-
thetic site, as was shown for other SerRSs [13,43]. In this work we
focused our attention to the pre-transfer editing activity of maize
enzymes in the absence of tRNA. Influence of tRNA on editing,
determined in the reaction with threonine and ZmcSerRS, was neg-
ligible. This is in a good agreement with previous findings showing
that SerRSs from eukaryotes, bacteria and methanogenic archaea
display only tRNA-independent editing of non-cognate substrates
[13,43]. In addition, it was shown that tRNA does not stimulate
pre-transfer editing in other synthetases lacking specialized edit-
ing domain [60,61]. Nevertheless, as synthetases interact with
tRNA in vivo, it would be of interest to evaluate activities of maize
enzymes in all editing reactions in the presence of tRNA, as well as
the extent of tRNASer misacylation with threonine and cysteine,
thus completing the picture of editing and fidelity of maize
enzymes.
significantly to overall pre-transfer editing (Fig. 5A), while in the
case of ZmoSerRS importance of this proofreading mechanism
was negligible. The difference was also seen with editing of Ser-
HX-AMP, where SerHX-AMP was partially eliminated through en-
hanced dissociation only by ZmoSerRS (Fig. 5B and Fig. 6).
Selective release of the non-cognate aa-AMP usually represents
a minor pre-transfer editing pathway for many aaRSs [13,61,69–
72]. This is not unexpected since a high portion of the cognate
aa-AMP would also be destroyed if this route is effective, as dis-
cussed by Jakubowski and Fersht [69]. Interestingly, we did not ob-
serve that involvement of this pathway depends on the nature of
the enzyme active site alone. It is more an interplay between the
non-cognate substrate and the environment provided by the syn-
thetic active site. All aa-AMPs possess the same adenylate part of
the structure that establishes multiple interactions with amino
acid residues within the synthetic site [73]. This suggests that
the ability of the amino acid portion to modulate aa-AMP affinity
is a feature of the enzyme:amino acid cross talk.
Fidelity of aaRS relies on amino acid selectivity that takes into
account both aaRS specificity and the relative cellular concentra-
tions of the corresponding cognate and near-cognate substrates
[40]. Indeed, it has been shown for several species that cellular via-
bility is compromised in editing-deficient aaRS mutants when in-
creased levels of non-cognate amino acids are present that could
potentially increase tRNA mischarging [8,9,62]. The intracellular
concentration of serine in maize and other plants is usually higher
than the concentration of threonine (Table 4, [63,64]) or cysteine
[65], therefore enhancing the selectivity of maize SerRSs. However,
the usual amino acid concentrations can rapidly change, caused by
stress or various programmed developmental conditions [66,67].
For example, in pea embryo 15 days after pollination the ratio of
serine vs. threonine is 1:2.7 [68]. It is plausible that a similar rever-
sal of serine and threonine concentrations occurs in maize cells
which could dramatically change SerRS selectivity and potentially
jeopardize translational fidelity (Table 4). We propose that this
problem could be at least partially resolved by the pre-transfer
editing activities of maize SerRSs that can act as a buffer against
sudden changes in cellular amino acid pools.
Implications for fidelity in plant organelles
Observation that some mitochondrial aaRSs lack the editing do-
main found in cytosolic counterparts led to an initial proposition
that they are more error-prone and that translation in mitochon-
dria may be somewhat less accurate [74,75]. However, it has been
reported that yeast mitochondrial phenylalanyl-tRNA synthetase
and human mitochondrial leucyl-tRNA synthetase are highly selec-
tive against non-cognate amino acids [40,76], while yeast mito-
chondrial ThrRS displays pre-transfer proofreading of serine,
thereby assuring fidelity [61]. In addition, in vivo study of the yeast
mitochondrial phenylalanyl-tRNA synthetase showed that its re-
duced accuracy caused severe mitochondrial dysfunction [40].
Here we show that maize organellar SerRS exhibits higher discrim-
ination against all tested non-cognate substrates compared to
cytosolic SerRS (Table 2) and also possesses pre-transfer editing
activity (Table 3, Fig. 5B). Therefore it appears that maize organel-
lar SerRS is more accurate synthetase than cytosolic enzyme. To
our knowledge this is the first report on the fidelity of a synthetase
serving its role in both mitochondria and plastids. Considering that
ZmoSerRS is dually targeted to both mitochondria and chloroplasts
[31], its fidelity must meet the requirements of both organelles for
translational quality control. Two observed fidelity mechanisms of
maize organellar SerRS, high discriminatory power and proofread-
ing, indicate that plant organelles require a high level of transla-
tional quality control and that aaRSs in general may play an
important role in plant organellar quality control.
Non-cognate substrates can be edited through different mechanisms
Pre-transfer editing mechanisms can involve corrective enzy-
matic hydrolysis and selective release of the non-cognate aa-AMP
that may be followed by its non-enzymatic hydrolysis. Our data
show that depending on the SerRS enzyme and the substrate, par-
titioning between the two mechanisms can be different. Both
maize enzymes edited Cys-AMP preferentially through enzymatic
hydrolysis. In the presence of cytosolic enzyme, enhanced
dissociation of Thr-AMP from the enzyme active site contributed

130
J. Rokov-Plavec et al. / Archives of Biochemistry and Biophysics 529 (2013) 122–130
[23] M. Mocibob, I. Weygand-Durasevic, Arch. Biochem. Biophys. 470 (2008) 129–
Conclusions
138.
[24] B. Low, F. Gates, T. Goldstein, D. Söll, J. Bacteriol. 108 (1971) 742–750.
[25] I. Weygand-Durasevic´, N. Ban, D. Jahn, D. Söll, Eur. J. Biochem. 214 (1993) 869–
877.
[26] I. Weygand-Durasevic, M. Nalaskowska, D. Söll, J. Bacteriol. 176 (1994) 232–
239.
[27] J.M. Sherman, M.J. Rogers, D. Söll, Nucleic Acids Res. 20 (1992) 2847–2852.
[28] S. Lesjak, I. Weygand-Durasevic, FEMS Microbiol. Lett. 294 (2009) 111–118.
[29] J. Rokov, I. Weygand-Durasevic, Period. Biol. 101 (1999) 137–142.
[30] W. Chiu, Y. Niwa, W. Zeng, T. Hirano, H. Kobayashi, J. Sheen, Curr. Biol. 6 (1996)
325–330.
The observed broad specificity towards tRNAsSer of eukaryotic
and bacterial origin indicates that maize cytosolic SerRS can
accommodate various types of long variable arms as well as vari-
ous types of tertiary structures, displaying an exceptional flexibil-
ity in tRNASer substrate recognition and interaction. The major
identity element for ZmcSerRS is the discriminator base G73,
which probably also plays a role in discrimination against tRNALeu
.
[31] J. Rokov-Plavec, M. Dulic, A.-M. Duchêne, I. Weygand-Durasevic, Plant Cell Rep.
27 (2008) 1157–1168.
[32] J. Rokov-Plavec, S. Lesjak, I. Landeka, I. Mijakovic, I. Weygand-Durasevic, Arch.
Biochem. Biophys. 397 (2002) 40–50.
[33] I. Gruic-Sovulj, N. Uter, T. Bullock, J.J. Perona, J. Biol. Chem. 280 (2005) 23978–
23986.
[34] A. Steinmetz, J.H. Weil, Meth. Enzymol. 118 (1986) 212–231.
[35] A.-M. Duchêne, C. Pujol, L. Maréchal-Drouard, Curr. Genet. 55 (2009) 1–18.
[36] J. Rokov-Plavec, S. Bilokapic, I. Gruic-Sovulj, M. Mocibob, F. Glavan, M. Brgles, I.
Weygand-Durasevic, Period. Biol. 106 (2004) 147–154.
[37] P.P. Chan, T.M. Lowe, Nucleic Acids Res. 37 (2009) D93–97.
[38] M. Michaud, V. Cognat, A.-M. Duchêne, L. Maréchal-Drouard, Plant J. 66 (2011)
80–93.
[39] E.J. Murgola, Annu. Rev. Genet. 19 (1985) 57–80.
[40] N.M. Reynolds, J. Ling, H. Roy, R. Banerjee, S.E. Repasky, P. Hamel, M. Ibba, Proc
Natl. Acad. Sci. USA 107 (2010) 4063–4068.
Both maize cytosolic and organellar SerRS enzymes display high
fidelity in amino acid recognition. This is established by highly
selective synthetic pathways in corroboration with pre-transfer
editing possibly acting as a buffer against sudden changes in cellu-
lar amino acid pools. However, it is possible that some misacyla-
tion of tRNASer does actually occur if the aminoacyl transfer step
is comparable to or faster than the aa-AMP hydrolysis rate
[71,72,77]. Contrary to previous thoughts, it has been recently ar-
gued that mistranslation does not necessarily have lethal conse-
quences for the cell and that under some conditions increased
levels of mistranslation can be advantageous [2]. Using in vitro
[71,72] and in vivo [28] methods developed in our laboratory we
will try to get a better insight into the connection between pre-
transfer editing and possible tRNASer misacylation as well as into
the requirements for translational fidelity in the plant cytosol,
mitochondria and plastids.
[41] I. Gruic-Sovulj, M. Dulic, J. Jaric, N. Cvetesic, K. Majsec, I. Weygand-Durasevic,
Croat. Chem. Acta 83 (2010) 163–169.
[42] D. Ahel, D. Slade, M. Mocibob, D. Söll, I. Weygand-Durasevic, FEBS Lett. 15
(2005) 4344–4348.
[43] I. Gruic´-Sovulj, M. Dulic´, I. Weygand-Ðuraševic´, Croat. Chem. Acta 84 (2011)
179–184.
[44] J. Rokov, D. Söll, I. Weygand-Durasevic, Plant Mol. Biol. 38 (1998) 497–502.
[45] S. Bilokapic, D. Korencic, D. Söll, I. Weygand-Durasevic, Eur. J. Biochem. 271
(2004) 694–702.
[46] Y. Itoh, S. Sekine, C. Kuroishi, T. Terada, M. Shirouzu, S. Kuramitsu, RNA Biol. 5
(2008) 169–177.
Acknowledgments
´
The authors thank Nevena Cvetešic for support and helpfull
comments. This work was supported by the grant from the Minis-
try of Science, Education and Sports of the Republic of Croatia (pro-
ject 119-0982913-1358).
[47] A. Soma, H. Himeno, Nucleic Acids Res. 26 (1998) 4374–4381.
[48] X. Tian, J. Zheng, S. Hu, J. Yu, J. Mol. Evol. 64 (2007) 299–307.
[49] M.M. Brandão, M.C. Silva-Filho, Biol. Evol. 28 (2011) 79–85.
[50] S. Cusack, A. Yaremchuk, M. Tukalo, EMBO J. 15 (1996) 2834–2842.
[51] K. Breitschopf, H.J. Gross, EMBO J. 13 (1994) 3166–3169.
[52] R. Geslain, E. Aeby, T. Guitart, T.E. Jones, M. Castro de Moura, F. Charrière, A.
Schneider, L. Ribas de Pouplana, J. Biol. Chem. 281 (2006) 38217–38225.
[53] R. Giegé, F. Jühling, J. Pütz, P. Stadler, C. Sauter, C. Florentz, Wiley Interdiscip.
Rev. RNA 3 (2011) 37–61.
[54] H. Himeno, S. Yoshida, A. Soma, K. Nishikawa, J. Mol. Biol. 268 (1997) 704–711.
[55] R. Rauhut, H.J. Gabius, F. Cramer, Biochemistry 24 (1985) 4052–4057.
[56] H. Jakubowski, Biochemistry 19 (1980) 5071–5078.
[57] H. Jakubowski, A. Guranowski, J. Biol. Chem. 278 (2003) 6765–6770.
[58] G.L. Igloi, E. Schiefermayr, FEBS J. 276 (2009) 1307–1318.
[59] C.A. Aldinger, A.K. Leisinger, G.L. Igloi, FEBS J. 279 (2012) 3622–3638.
[60] P.J. Beuning, K. Musier-Forsyth, J. Biol. Chem. 276 (2001) 30779–30785.
[61] J. Ling, K.M. Peterson, I. Simonovic, D. Söll, M. Simonovic, J. Biol. Chem. 287
(2012) 28518–28525.
[62] L.A. Nangle, C.M. Motta, P. Schimmel, Chem. Biol. 13 (2006) 1091–1100.
[63] H. Weiner, S. Blechschmidtschneider, H. Mohme, W. Eschrich, H.W. Heldt,
Plant Physiol. Biochem. 29 (1991) 19–23.
[64] H. Winter, G. Lohaus, H.W. Heldt, Plant Physiol. 99 (1992) 996–1004.
[65] S. Boggio, J. Palatnik, H. Heldt, E. Valle, Plant Sci. 159 (2000) 125–133.
[66] R.C. Sicher, J.Y. Barnaby, Physiol. Plant 144 (2012) 238–253.
[67] Z. Kovács, L. Simon-Sarkadi, I. Vashegyi, G. Kocsy, Different accumulation of
free amino acids during short- and long-term osmotic stress in wheat, Sci.
World J. (2012) 10. Article ID 216521.
References
[1] M. Ibba, S. Dieter, Annu. Rev. Biochem. 69 (2000) 617–650.
[2] N.M. Reynolds, B. Lazazzera, M. Ibba, Nat. Rev. Microbiol. 8 (2010) 849–856.
[3] R. Giegé, M. Sissler, C. Florentz, Nucleic Acids Res. 26 (1998) 5017–5035.
[4] M. Ibba, S. Sever, M. Praetorius-Ibba, D. Söll, Nucleic Acids Res. 27 (1999)
3631–3637.
[5] H. Jakubowski, E. Goldman, Microbiol. Rev. 56 (1992) 412–429.
[6] J. Ling, N. Reynolds, M. Ibba, Annu. Rev. Microbiol. 63 (2009) 61–78.
[7] J.W. Lee, K. Beebe, L.A. Nangle, J. Jang, C.M. Longo-Guess, S.A. Cook, M.T.
Davisson, J.P. Sundberg, P. Schimmel, S.L. Ackerman, Nature 443 (2006) 50–55.
[8] J.M. Bacher, V. de Crécy-Lagard, P.R. Schimmel, Proc Natl. Acad. Sci. USA 102
(2005) 1697–1701.
[9] V. Karkhanis, A.P. Mascarenhas, S. Martinis, J. Bacteriol. 189 (2007) 8765–8768.
[10] S. Bilokapic, N. Ban, I. Weygand-Durasevic, Croat. Chem. Acta 82 (2009) 493–
501.
[11] S. Bilokapic, T. Maier, D. Ahel, I. Gruic-Sovulj, D. Söll, I. Weygand-Durasevic, N.
Ban, EMBO J. 25 (2006) 2498–2509.
[12] S. Bilokapic, J. Rokov Plavec, N. Ban, I. Weygand-Durasevic, FEBS J. 275 (2008)
2831–2844.
[13] I. Gruic-Sovulj, J. Rokov-Plavec, I. Weygand-Durasevic, FEBS Lett. 581 (2007)
5110–5114.
[14] O.O. Obembe, J.O. Popoola, S. Leelavathi, S.V. Reddy, Biotechnol. Adv. 29 (2011)
210–222.
[15] P. Ahmad, M. Ashraf, M. Younis, X. Hu, A. Kumar, N.A. Akram, F. Al-Qurainy,
Biotechnol. Adv. 30 (2012) 524–540.
[16] B. Lenhard, O. Orellana, M. Ibba, I. Weygand-Durasevic, Nucleic Acids Res. 27
(1999) 721–729.
[17] I. Weygand-Durasevic, S. Cusack, Seryl-tRNA synthetases, in: M. Ibba, C.
Francklyn, S. Cusack (Eds.), The Aminoacyl-tRNA Synthetases, Landes
Bioscience, Georgetown TX, 2005, pp. 177–192.
[18] H. Asahara, H. Himeno, K. Tamura, N. Nameki, T. Hasegawa, M. Shimizu, J. Mol.
Biol. 236 (1994) 738–748.
[19] X.Q. Wu, H.J. Gross, Nucleic Acids Res. 21 (1993) 5589–5594.
[20] D. Korencic, C. Polycarpo, I. Weygand-Durasevic, D. Söll, J. Biol. Chem. 279
(2004) 48780–48786.
[68] K. Weigelt, H. Küster, R. Radchuk, M. Müller, H. Weichert, A. Fait, A.R. Fernie, I.
Saalbach, H. Weber, Plant J. 55 (2008) 909–926.
[69] H. Jakubowski, A.R. Fersht, Nucleic Acids Res. 9 (1981) 3105–3117.
[70] K.E. Splan, M.E. Ignatov, K. Musier-Forsyth, J. Biol. Chem. 283 (2008) 7128–
7134.
[71] M. Dulic, N. Cvetesic, J.J. Perona, I. Gruic-Sovulj, J. Biol. Chem. 285 (2010)
23799–23809.
[72] N. Cvetesic, J.J. Perona, I. Gruic-Sovulj, J. Biol. Chem. 287 (2012) 25381–25394.
[73] J.G. Arnez, D. Moras, Trends Biochem. Sci. 22 (1997) 211–216.
[74] H. Roy, J. Ling, J. Alfonzo, M. Ibba, J. Biol. Chem. 280 (2005) 38186–38192.
[75] V. Karkhanis, M.T. Boniecki, K. Poruri, S. Martinis, J. Biol. Chem. 281 (2006)
33217–33225.
[76] S.W. Lue, S.O. Kelley, Biochemistry 44 (2005) 3010–3016.
[77] A. Minajigi, C.S. Francklyn, J. Biol. Chem. 285 (2010) 23810–23817.
[78] C. Ross-Inta, C.-Y. Tsai, C. Giulivi, Biosci. Rep. 28 (2008) 239–249.
[21] A. Yaremchuk, I. Kriklivyi, M. Tukalo, S. Cusack, EMBO J. 21 (2002) 3829–3840.
[22] V. Biou, A. Yaremchuk, M. Tukalo, S. Cusack, Science 263 (1994) 1404–1410.