Journal of Experimental Botany, Vol. 55, No. 397, pp. 623±630, March 2004
DOI: 10.1093/jxb/erh079 Advance Access publication 13 February, 2004
RESEARCH PAPER
A glycolate dehydrogenase in the mitochondria of
Arabidopsis thaliana
Ra®jul Bari1, Rashad Kebeish1, Rainer Kalamajka1, Thomas Rademacher2 and Christoph PeterhaÈnsel1,*
1
Institute for Biology I, Aachen University, Worringer Weg 1, D-52056 Aachen, Germany
2
Fraunhofer Institute of Molecular Biology and Applied Ecology, D-52074 Aachen, Germany
Received 12 September 2003; Accepted 28 November 2003
Abstract
The ®xation of molecular O2 by the oxygenase activity of Rubisco leads to the formation of phosphoglycolate in the chloroplast that is further metabolized in
the process of photorespiration. The initial step of
this pathway is the oxidation of glycolate to glyoxylate. Whereas in higher plants this reaction takes
place in peroxisomes and is dependent on oxygen as
a co-factor, most algae oxidize glycolate in the mitochondria using organic co-factors. The identi®cation
and characterization of a novel glycolate dehydrogenase in Arabidopsis thaliana is reported here. The
enzyme is dependent on organic co-factors and
resembles algal glycolate dehydrogenases in its
enzymatic properties. Mutants of E. coli incapable of
glycolate oxidation can be complemented by overexpression of the Arabidopsis open reading frame. The
corresponding RNA accumulates preferentially in illuminated leaves, but was also found in other tissues
investigated. A fusion of the N-terminal part of the
Arabidopsis glycolate dehydrogenase to red ¯uorescent protein accumulates in mitochondria when overexpressed in the homologous system. Based on
these results it is proposed that the basic photorespiratory system of algae is conserved in higher
plants.
Key words: Arabidopsis thaliana, dehydrogenase, glycolate,
mitochondrion, photorespiration.
Introduction
All plants and algae share the same bifunctional enzyme,
ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco,
EC 4.1.1.39) for photosynthetic CO2 ®xation. The ®rst stable
product of the carboxylase reaction that enters the Calvin
cycle is 3-phosphoglycerate (3-PGA). In the oxygenase
reaction, one molecule of 3-PGA and one molecule of
2-phosphoglycolate are formed. The latter is recycled by
higher plants in the process of photorespiration (Leegood
et al., 1995). After dephosphorylation, glycolate is
exported from the chloroplast and oxidized to glyoxylate
by the peroxisomal glycolate oxidase (GO; EC 1.1.3.15).
This enzyme uses inorganic oxygen as an electron acceptor
and H2O2 is formed that has to be detoxi®ed by a catalase
(Tolbert, 1997). In a complex reaction cascade that takes
place in peroxisomes, mitochondria, and chloroplasts,
glyoxylate is recycled to glycerate that can be reintegrated
into the plastidal carbon metabolism. A major inef®ciency
of the photorespiratory pathway is the loss of one-quarter
of the already ®xed CO2 inside the mitochondria.
Moreover, reduced nitrogen is lost during photorespiration
and has itself to be re®xed (Ogren, 1984). On the other
hand, important functions have been assigned to photorespiration in the protection of photosystems from excessive light energy (Kozaki and Takeba, 1996) as well as in
nitrogen metabolism (Leegood et al., 1995).
Instead of the peroxisomal glycolate oxidase, algal
species lacking leaf-type peroxisomes mostly contain a
mitochondrial glycolate dehydrogenase activity (GDH, EC
1.1.99.14) This enzyme is dependent on organic co-factors
and shows deviant enzymatic properties compared with the
oxygen-dependent form, including a preference for Dlactate instead of L-lactate as an alternative substrate and a
high sensitivity for cyanide ions (Nelson and Tolbert,
1970). A similar activity has also been described for algal
as well as higher plant chloroplasts and has been proposed
to be involved as an electron donor in cyclic photophosphorylation (Goyal and Tolbert, 1996). The capacity of
the reactions dependent on organic co-factors seems to be
* To whom correspondence should be addressed. Fax: +49 241 8022637. E-mail: [email protected]
Journal of Experimental Botany, Vol. 55, No. 397, ã Society for Experimental Biology 2004; all rights reserved
624 Bari et al.
limited compared with the peroxisomal reaction and many
algae excrete glycolate into the medium under photorespiratory growth conditions (Colman et al., 1974).
Escherichia coli is capable of growing on glycolate as
the sole carbon source. Here, glycolate is oxidized to
glyoxylate by an enzyme traditionally named glycolate
oxidase that is, however, dependent on organic co-factors.
The enzyme is encoded by the glc operon and mutation
analysis has shown that the subunits glcD, glcE, and glcF
are necessary and suf®cient to form the active enzyme
(Lord, 1972; Pellicer et al., 1996). Two molecules of
glyoxylate are converted to one molecule of tartronic
semialdehyde by the release of CO2 catalysed by
glyoxylate carboligase (GCL) and this compound is
reduced to glycerate by the action of tartronic semialdehyde reductase (TSR) (Chang et al., 1993; Pellicer et al.,
1996). These reactions constitute what is known as the
glycerate pathway.
In this study, the identi®cation and characterization of a
novel glycolate dehydrogenase in Arabidopsis are described. The protein shows enzymatic properties resembling those of algal glycolate dehydrogenases and is also
targeted to mitochondria.
Materials and methods
Plant growth
Arabidopsis thaliana ecotype Columbia was grown in soil (ED 73,
Werkverband, Sinntal-Jossa, Germany) containing 30% sand at
24 °C with a photon ¯ux density of approximately 100 mmol m±2 s±1
light intensity. Long-day conditions were 16 h in the light and 8 h in
the dark, whereas short-day conditions were 8 h in the light and 16 h
in the dark.
Cloning of AtGDH and E. coli glycolate oxidase
RNA was prepared from the young leaves of Arabidopsis plants
following the TRIZOL protocol (Chomczynski, 1993). The integrity
of the preparation and the presence of genomic DNA were tested by
gel electrophoresis. Approximately 1 mg of RNA were mixed with
10 pmol oligo-dT primer, heated for 5 min to 68 °C and cooled down
on ice before adding 200 U of MMLV-RT (Promega, Mannheim,
Germany) and 1 mM dNTPs in reaction buffer supplied by the
manufacturer. AtGDH was ampli®ed by PCR using the cDNA as a
template and gene speci®c primers (5¢-ACGGATCCAATGTTTAGGTCCGAAGAAGAA-3¢ and 5¢-ACCTCGAGGAAACATACATGAGGAGGAATT-3¢). The primers contain extensions with an
additional BamHI site and SalI site, respectively. PCR fragments
were ligated into the plasmid pET22b(+) (Novagen, Darmstadt,
Germany) between the BamHI and XhoI sites, resulting in a
C-terminal translational fusion containing six histidines.
For complementation analyses (see below), AtGDH was also
cloned into pTrc99a (Amersham Biosciences, Freiburg, Germany).
pET-AtGDH was restricted with Bpu1102I, protruding ends were
blunted using Klenow polymerase and the construct was restricted
with BamHI. The fragment was cloned into the vector restricted with
SalI, blunted with Klenow polymerase, and cut with BamHI.
The subunits glcD, glcE, and glcF of the E. coli glycolate oxidase
operon constituting the active enzyme were ampli®ed from genomic
DNA using the oligonucleotides 5¢-TCGACCATGGGCATCTTGTACGAAGAGC-3¢ and 5¢-TCGACTCGAGTTATTCCTTTTC-
AAGGGC-3¢. The primers contain extensions with an additional
NcoI site and XhoI site, respectively. Otherwise, the cloning strategy
was identical as for the AtGDH constructs.
For the subcellular localization of AtGDH, the ®rst 241
nucleotides of the coding sequence were cloned in translational
fusion to a modi®ed dsRED (Jach et al., 2001) and the gene construct
was inserted into the binary plant expression vector pTRAK, a
derivative of pPAM (GenBank: AY027531). The expression cassette
was ¯anked by the scaffold attachment region of the tobacco RB7
gene (GenBank: U67919). The nptII cassette of pPCV002 (Koncz
and Schell, 1986) was used for the selection of transgenic plants.
Expression of the construct is under the constitutive control of the
transcriptionally enhanced CaMV 35S promoter.
Recombinant protein expression and enzymatic assays
The pET-AtGDH construct was transformed into the bacterial strain
ER2566 (Novagen, Darmstadt, Germany). For protein expression, a
2.0 l culture of LB-medium was grown up to an OD600 of 0.4. The
expression of the proteins was induced by adding 1 mM IPTG and
culture growth was continued for 2 h at 37 °C. The cells were washed
once in 10 mM potassium phosphate (pH 8.0) and were resuspended
in the same buffer to give a 15±20% (v/v) cell suspension. After the
cells were lysed by sonication on ice (40% duty cycle, 231 min,
Sonicator Bandelin Sonopuls GM 70, Berlin, Germany), the extract
was centrifuged for 30 min at 30 000 g. The presence of AtGDH in
the supernatant was tested by Western Blot using an Anti-His-HRP
conjugate (Qiagen, Hilden, Germany) following the protocol
provided by the supplier.
Glycolate dehydrogenase activity was assayed according to Lord
(1972). A bacterial cell extract containing 100 mg of protein was
added to 100 mmol potassium phosphate (pH 8.0), 0.2 mmol DCIP,
0.1 ml 1% (w/v) PMS, and 10 mmol potassium glycolate in a ®nal
volume of 2.4 ml. At ®xed time intervals, individual assays were
terminated by the addition of 0.1 ml of 12 M HCl. After standing for
10 min, 0.5 ml of 0.1 M phenylhydrazine-HCl was added. The
mixture was allowed to stand for a further 10 min, and then the
extinction due to the formation of glyoxylate phenylhydrazone was
measured at 324 nm.
Glyoxylate reductase activity was assayed according to Gia® and
Rumsby (1998). A cell extract containing 20 mg of protein was added
to a 1 cm cuvette containing 0.6 mM NADPH and 6 mM glyoxylate
in 50 mM potassium phosphate buffer pH 7.6, in a ®nal volume of 1
ml. The activity was determined spectophotometrically at 340 nm.
Lactate dehydrogenase activity was assayed according to Gutheil
(1998). A cell extract containing 50 mg of protein was added to a
1 cm cuvette containing 4.8 mM NAD+ and 1.0 mM D-lactate or
L-lactate, respectively, in 0.1 mM TRIS-HCl, pH 8.5, in a ®nal
volume of 1 ml and the activity was determined from the progress
curve at 340 nm. As an inhibitor, 1 mM KCN was added to the
reaction mixture before the addition of the protein extract.
Complementation analysis
Complementation analyses were done with mutants of E. coli
de®cient in any of the three subunits forming the active endogenous
glycolate oxidase (Pellicer et al., 1996). All of the transformed
bacteria were grown for 2 d in minimal medium (Miller, 1972) using
glycolate as a sole carbon source supplemented with the appropriate
antibiotics (25 mg ml±1 chloramphenicol or 100 mg ml±1 ampicillin)
and 1 mM IPTG. The cells were diluted in fresh medium and allowed
to grow until the OD reached 0.7. The cells were spun down at 2000 g
and resuspended in 100 ml of minimal medium and were serially
diluted from 10±1 to 10±5. 10 ml of the cells from each dilution were
spotted on plates containing minimal medium, antibiotics, and 1 mM
IPTG for the induction of recombinant protein expression. The
plates were then incubated at 37 °C for 3 d.
Arabidopsis glycolate dehydrogenase 625
Analysis of transcript abundance
RNA was isolated and ®rst strand cDNA synthesis was performed as
described above for the cloning of AtGDH. Quantitative PCR
reactions were performed on an ABI PRISMâ 7700 Sequence
Detection System (Applied Biosystems, Weiterstadt, Germany)
following the manufacturer's instructions. Ampli®cations were
performed in the presence of SYBR Green as the ¯uorescent dye
using 2% of a reverse transcription reaction as a template. Reaction
kits were derived from Eurogentec (Cologne, Germany) and
oligonucleotides were purchased from Metabion (Planegg,
Germany). For the detection of AtGDH transcripts, the primers
were 5¢-CCTTGCAGAACTCATATCAAGATC-3¢ and 5¢-CATGAGGAGGAAGAATTAACTTTCC-3¢ with a ®nal primer concentration of 300 nM in the reaction mixture. For the detection of Actin2
transcripts, primers were 5¢-GGTAACATTGTGCTCAGTGGTGG3¢ and 5¢-GGTGCAACGACCTTAATCTTCAT-3¢ with a ®nal
primer concentration of 900 nM in the reaction mixture. The
MgCl2 concentration was always 2 mM and the dNTP concentration
200 mM. Ampli®cation conditions were 10 min of initial denaturation at 95 °C, followed by 40 cycles of denaturation at 95 °C for 15 s
and combined annealing and extension at 60 °C for 1 min.
Subcellular localization
Plant expression vectors containing the ®rst 241 nucleotides of the
AtGDH coding sequence in translational fusion to dsRED (see
above) were transferred into A. tumefaciens GV3101 by electroporation. The recombinant A. tumefaciens were selected in YEB
medium containing appropriate antibiotics (rifampicin 100 mg ml±1,
kanamycin 25 mg ml±1, and carbenicillin 50 mg ml±1). The construct
was transferred to Arabidopsis by ¯oral dip transformation (Clough
and Bent, 1998) and regeneration of kanamycin-resistant transgenic
plants was carried out according to standard protocols. For dsRED
localization and mitochondrial staining, leaves were collected from
4-week-old plants and cut into small pieces. Protoplasts were
isolated by incubating the leaf pieces in protoplast isolation buffer
(0.5 M D-mannitol, 5 mM MES pH 5.8, 10 mM CaCl2, 3% (v/v)
Rohalase 7069 (RoÈhm, Darmstadt, Germany), 2% (v/v) Rohament
PL (RoÈhm, Darmstadt, Germany) and 0.12% (w/v) Mazeroenzyme
R-10 (Serva, Heidelberg, Germany)) at 30 °C for 2 h. The
supernatant containing the protoplasts was carefully collected,
followed by staining with MitoTracker Green (Molecular Probes,
Leiden, Netherlands) as described by the supplier. Images of the
protoplasts were taken with a Leica TCS-SP spectral confocal
microscope (Leica, Heidelberg, Germany). Images were acquired
with a 1.2 N.A. 633 oil immersion PLAN-APO objective. The entire
sample was excited with the 488 nm and 568 nm laser line. The
confocal sections were collected using a 515±535 nm emission
setting for MitoTracker Green, 570±610 nm emission setting for
dsRED, and 660±720 nm emission setting for chlorophyll ¯uorescence.
Results
Arabidopsis encodes a glycolate dehydrogenase
Sequence analysis of the Arabidopsis genome revealed an
open reading frame (At5g06580) with homology to the
subunit D of Escherichia coli glycolate oxidase (EcGLCD)
and to yeast D-lactate dehydrogenase (ScDLDH). As
shown in Fig. 1, all three enzymes share signi®cant
homology over the complete sequence length apart from
the ®rst 100 amino acids. The putative Arabidopsis protein
shares a 40% similarity and 30% identity to EcGLCD and a
50% similarity and 41% identity to ScDLDH. According to
RPS-BLAST (Altschul et al., 1997), the Arabidopsis open
reading frame contains homologous regions to a FADbinding domain in between amino acids 150 and 300 and a
C-terminal FAD-oxidase domain with a ferredoxin-like
fold from amino acid 320. Both domains are frequently
found in oxygen-dependent oxygenases as well as
dehydrogenases dependent on organic co-factors (Dym
and Eisenberg, 2001).
The respective cDNA was cloned from RNA isolated
from young green leaves. The construct was overexpressed
in E. coli for enzymatic assays and complementation
analysis. The gene was provisionally named AtGDH for
Arabidopsis thaliana glycolate dehydrogenase.
Enzymatic assays were performed with crude extracts
from a bacterial strain overexpressing the gene as an
N-terminal fusion to six histidine residues. Crude extracts
were used because, for unknown reasons, it was not
possible to purify the enzyme from the extract successfully
without losing most of the activity. Figure 2 and Table 1
show the results of the assays. The extract containing
overexpressed AtGDH shows an appreciable formation of
glyoxylate from glycolate as speci®cally determined by the
formation of glyoxylate phenylhydrazone in a coupled
reaction. These levels are reduced to background if an
unrelated protein (with similar accumulation levels after
the induction of expression) is used. The same reduction is
also found when the substrate (glycolate) or the organic
electron acceptors (DCIP/PMS) are omitted from the
reaction. Thus, AtGDH catalyses the formation of
glyoxylate from glycolate dependent on organic cofactors.
Further characterization of the enzymatic properties
revealed that the enzyme shows an apparent KM for
glycolate of approximately 3.5310±2 mM whereas the KM
for glyoxylate is appproximately 4310±1 mM. The pH
optimum for glycolate oxidation is in between pH 8.0 and
pH 8.5 (data not shown). The values obtained are within
the range of the enzymatic properties described for algal
glycolate dehydrogenases (Nelson and Tolbert, 1970).
Bacterial and algal glycolate dehydrogenases show a
high preference for D-lactate over L-lactate and are
sensitive to cyanide ions, whereas the plant oxygendependent glycolate oxidases show a preferential af®nity
for L-lactate and are cyanide-insensitive (Nelson and
Tolbert, 1970). As listed in Table 1, crude bacterial
extracts overexpressing AtGDH show D-lactate dehydrogenase activity that is reduced to a background activity of
approximately one-third by the addition of cyanide ions.
No L-lactate dehydrogenase activity can be detected in this
assay compared with controls with the empty vector
construct. Taken together, AtGDH shows all the properties
of a glycolate dehydrogenase and can be clearly discriminated from the so-far-described higher plant glycolateoxidizing enzymes.
626 Bari et al.
Fig. 1. AtGDH shows homology to glycolate dehydrogenases and lactate dehydrogenases. The ®gure shows an alignment of the novel identi®ed
sequence and related sequences. Black shaded boxes indicate identical amino acids and grey shaded boxes indicate similar amino acids. AtGDH,
Arabidopsis thaliana glycolate dehydrogenase; ScDLDH, Saccharomyces cerevisiae D-lactate dehydrogenase; EcGLCD, Subunit D of Escherichia
coli glycolate oxidase.
AtGDH can complement E. coli glycolate oxidase
mutants
We wanted to determine whether AtGDH is capable of
complementing glycolate oxidase mutants of E. coli. As
speci®ed in the introduction, the E. coli enzyme is identical
to algal glycolate dehydrogenases in its speci®city for
organic co-factors and its enzymatic properties and only
traditionally named an oxidase. The mutants JA155,
JA156, and JA157 carry transposon insertions in the
glcD, glcE, and glcF subunits of the glc operon and are
Arabidopsis glycolate dehydrogenase 627
Table 1. Substrate speci®city and potassium cyanide sensitivity
of AtGDH
Construct
a
AtGDH
AtGDH
AtGDH
AtGDH
Empty vector
a
Fig. 2. AtGDH shows glycolate dehydrogenase activity in vitro.
Protein extracts from E. coli cultures overexpressing AtGDH or an
unrelated protein were tested for glycolate dehydrogenase activity by
direct measurement of the glyoxylate formation. Each data point is
based on data from three different experiments. Vertical bars show
standard deviations. AtGDH, overexpression of the putative
Arabidopsis thaliana glycolate dehydrogenase; ±glycolate, assay as
above, but in the absence of exogenously added glycolate, ±DCIP/
PMS, assay as above, but in the absence of exogenously added
organic electron acceptors; unrelated, overexpression of the unrelated
maize transcription factor DOF1 as a negative control.
incapable of growing on glycolate as the sole carbon
source (Pellicer et al., 1996). As shown in Fig. 3,
overexpression of AtGDH in any of these mutants restored
the growth of bacteria on the medium containing glycolate
as the sole carbon source. Comparing the dilution series of
bacteria transformed with AtGDH or the intact E. coli
glycolate oxidase (EcGO), similar growth rates are
observed, whereas no colonies are detectable when
bacteria were transformed with the empty expression
vector under identical conditions. However, replicate
platings on rich medium were capable of growing independently of the transformed construct (data not shown).
The results indicate that AtGDH can complement for all
three subunits of the active EcGO enzyme in vivo and thus
encodes a functional equivalent.
AtGDH shows preferential transcript accumulation in
illuminated leaf tissues
In order to determine the gene expression pro®le of the
AtGDH gene, the accumulation of the corresponding
transcripts in different plant tissues and the diurnal rhythm
of transcript accumulation was measured by Real-Time
RT-PCR. The amount of AtGDH signal in the different
RNA preparations was standardized for the abundance of
Substrate
KCN
Activity (%)
D(±)-lactate
D(±)-lactate
L(+)-lactate
L(+)-lactate
D(±)-lactate
±
+
±
+
±
100%
37%
29%
29%
37%
50 mg of crude extract overexpressing AtGDH.
the Actin2 transcript (Igarashi et al., 2003). As shown in
Fig. 4A, the transcript accumulates in leaves as well as
stems, ¯owers, and roots when samples were taken in the
middle of the light period. The accumulation is approximately 2-fold in illuminated leaves compared with leaves
from plants that were kept for 2 d in the dark or to the other
tissues. When the transcript abundance is compared at
different time points after the onset of light (Fig. 4B), a
constitutive low level of expression and an increase during
the light period is observed. This increase only becomes
apparent after 5 h of illumination, but not after 1 h. Taken
together, AtGDH is expressed in all of the tissues
investigated, but transcripts preferentially accumulate in
illuminated leaves.
The N-terminal domain of AtGDH targets proteins to
mitochondria
Sequence analysis of AtGDH using the TargetP algorithm
(Emanuelsson et al., 2000) identi®ed high scores both for a
putative chloroplast transit sequence of 56 amino acids and
a putative mitochondrial target sequence of 30 amino
acids. In order to test whether this sequence stretch targets
proteins to plastids or mitochondria in vivo, Arabidopsis
plants were transformed with a plant expression construct
carrying the ®rst 77 amino acids of AtGDH in translational
fusion to red ¯uorescent protein (dsRED) (Jach et al.,
2001). Microscopic analysis of transgenic lines revealed an
accumulation of the red ¯uorescence in particles of
approximately 1 mm in size resembling mitochondria
(data not shown). For further analysis, protoplasts were
prepared from leaves of the transgenic plants and
mitochondria were speci®cally stained with the
MitoTracker Green dye.
The confocal laser images in Fig. 5 show the chlorophyll
¯uorescence of chloroplasts in blue, the stained mitochondria in green, and the dsRED-mediated ¯uorescence of the
fusion protein in red. Care was taken that the green
¯uorescence of the MitoTracker Green dye was not visible
in the red channel detecting the ¯uorescence of the dsRED
protein and vice versa (see Materials and methods). The
mitochondrial stain shown in Fig. 5A almost perfectly colocalizes with the dsRED ¯uorescence shown in Fig. 5B
indicating that the N-terminal domain of AtGDH targets
proteins to mitochondria in vivo.
628 Bari et al.
Fig. 3. Complementation of E. coli glycolate oxidase mutants with AtGDH. Bacterial strains de®cient in glycolate oxidase subunits were
transformed with the listed constructs and a dilution series was spotted on agar plates containing glycolate as the sole carbon source (as speci®ed
in the Materials and methods). The ®gure shows photographs of the plates after 3 d of growth at 37 °C. GO, overexpression of E. coli glycolate
oxidase subunits glcD±F; AtGDH, overexpression of the Arabidopsis thaliana glycolate dehydrogenase; pTrc, transformation with empty vector.
Discussion
Fig. 4. Tissue speci®city and diurnal rhythm of AtGDH transcript
accumulation. The amount of AtGDH mRNA was measured by RealTime PCR and calculated in arbitrary units by comparison with a
standard dilution series. Each value is the relative accumulation of the
respective RNA compared with the Actin2 levels measured in the
preparation. Each data point is based on at least three independent
RNA preparations and for each preparation the quanti®cation was
repeated at least three times. Vertical bars show standard deviations.
(A) Tissue samples were collected from the youngest fully expanded
leaf or the indicated organ 5 h after the onset of illumination. (B) Leaf
samples were collected from the youngest fully expanded leaf 1 h
before the onset of light, 1 h, 5 h, and 7 h after the onset of light, and
6 h after the beginning of the dark period.
Enzymes catalysing the oxidation of glycolate to glyoxylate in plants have been grouped into two main classes
(Frederick et al., 1973). Higher plants use an enzyme that
is located in the peroxisomes where the H2O2 produced in
the course of the oxygen-dependent oxidation reaction can
be directly detoxi®ed by catalases. By contrast, the ®rst
electron acceptor for most green algal enzymes seems to be
a so-far-unidenti®ed organic compound that can be
replaced by DCIP in vitro (Nelson and Tolbert, 1970).
Several studies propose a subcellular location of the algal
enzyme in mitochondria and coupling of glycolate oxidation to the respiratory electron transport chain (Husic and
Tolbert, 1987; Stabenau, 1992; Stabenau et al., 1984).
Beside the peroxisomal and the mitochondrial glycolate
oxidoreductases, glycolate dehydrogenase activity has also
been described for algal as well as higher plant
chloroplasts (Goyal, 2002; Goyal and Tolbert, 1996).
The identi®cation and characterization of AtGDH in this
study proposes the existence of mitochondrial glycolate
oxidation in higher plants that resembles the algal
photorespiratory pathway. First, AtGDH shares enzymatic
properties with its algal counterparts like the speci®city for
substrates and co-factors as well as the sensitivity to
inhibitors and the pH-optimum. This clearly discriminates
the enzyme from the so-far-described peroxisomal glycolate oxidases of higher plants and no sequence homology is
found to this class of enzymes (data not shown). Second,
the corresponding cDNA is capable of complementing
bacterial mutants that are de®cient in glycolate oxidation
dependent on organic co-factors. These analyses revealed
that AtGDH is capable of replacing any of the three
subunits of E. coli glycolate dehydrogenase. It was not
possible to test whether the Arabidopsis enzyme also
restores growth in mutants where all three subunits have
been deleted simultaneously. Therefore, it can not formally
be excluded that AtGDH can only replace any individual
subunit of EcGO, but not the complete enzyme, and forms
active complexes with the remaining subunits. However,
this seems improbable because the enzymatic activity can
be measured in crude extracts of overexpressing wild-type
Arabidopsis glycolate dehydrogenase 629
Fig. 5. Subcellular localization of AtGDH. The picture shows epi¯uorescence photographs of protoplasts from Arabidopsis plants that were
transformed with a construct expressing the N-terminal part of AtGDH in translational fusion to red ¯uorescent protein (dsRED). Excitation wave
lengths were 488 nm and 568 nm and emission ®lters were 515±535 nm (MitoTracker Green, shown in green), 570±610 nm (dsRED, shown in
red), and 660±720 nm (chlorophyll ¯uorescence, shown in blue). Magni®cation is 6303. (A) Overlay of chlorophyll ¯uorescence and MitoTracker
Green-mediated ¯uorescence. (B) Overlay of chlorophyll ¯uorescence and dsRED-mediated ¯uorescence.
E. coli strains that show only a very low background
glycolate dehydrogenase activity in the absence of AtGDH.
A third evidence for the identity of AtGDH with a
mitochondrial glycolate dehydrogenase is the ability of the
N-terminal domain of AtGDH to direct red ¯uorescent
protein to mitochondria. The recombinant protein clearly
co-localized with mitochondria that were identi®ed using
the MitoTracker dye. Although computer predictions also
identi®ed a high score for a chloroplast transit signal,
¯uorescence of the overexpressed enzyme was only
occasionally found in leaf chloroplasts probably due to
mistargeting (data not shown). Cytochemical analyses
provided evidence that the mitochondrial glycolate
dehydrogenases of Chlamydomonas is located in between
the inner and the outer membrane (Beezley et al., 1976)
and antibodies to AtGDH are currently being generated in
order to determine the targeting of AtGDH within the
mitochondrion.
Which possible functions can be assigned to a
mitochondrial glycolate dehydrogenase in higher plants
based on the available information? Most obvious, the
identi®cation of AtGDH could indicate the conservation of
the algal photorespiratory pathway in higher plants.
Dependent on the green algal species, the glyoxylate
produced by the mitochondrial glycolate dehydrogenase is
metabolized to glycine in the mitochondrion and afterwards by different pathways to glycerate, serine, or related
compounds (Igamberdiev and Lea, 2002; Winkler and
Stabenau, 1992). However, no glyoxylate aminotransferase activity catalysing the conversion of glyoxylate to
glycine has been described in higher plant mitochondria so
far. Furthermore, mutations in peroxisomal photorespiration usually lead to lethal phenotypes that can only be
rescued by non-photorespiratory growth conditions
(Somerville, 1984). These observations make the presence
of a second alternative photorespiratory pathway improbable in the ®rst instance. However, the available data are
also consistent with a scenario where enzymes are shared
by both pathways or where the capacity of the mitochon-
drial pathway is simply insuf®cient to complement mutations in peroxisomal photorespiration.
It is not clear which enzymes might be shared by both
pathways and whether the relative capacity of the putative
mitochondrial pathway is suf®cient to complement mutations in peroxisomal photorespiration. The activity of the
mitochondrial glycolate dehydrogenase in algae is inadequate under photorespiratory conditions and algae can,
alternatively, excrete excess glycolate (Colman et al.,
1974).
In an alternative scenario, AtGDH might act as a lactate
dehydrogenase in vivo that is involved in anaerobic energy
production. However, in plants, anoxic conditions have
only been described for roots and seedlings and lead to the
formation of L-lactate (Drew, 1997) whereas AtGDH
shows a preferential transcript accumulation in leaves and
can only use D-lactate as a substrate.
For the alga Euglena gracilis, the involvement of a
mitochondrial glycolate dehydrogenase in a glycolate/
glyoxylate shuttle has been proposed (Yokota et al., 1985;
Yokota and Kitaoka, 1979). In this case, excess NADPH
from the light reactions of photosynthesis is transferred to
the mitochondrion by the malate-oxaloacetate shuttle and
used for the reduction of glyoxylate to glycolate.
Glyoxylate is, in turn, recycled by glycolate dehydrogenase and the electrons are transferred into the respiratory
electron transport chain and used for ATP synthesis. A
similar reaction has been suggested for plant chloroplasts
where electrons could be shuttled from NADPH into the
photosynthetic transport chain (Goyal and Tolbert, 1996).
Nevertheless, whereas higher plant chloroplasts and
Chlamydomonas mitochondria show glyoxylate reductase
activity (Givan and Kleczkowski, 1992; Yokota and
Kitaoka, 1979), such an enzyme has not been described
for higher plant mitochondria until now.
The hypothesis of an ancient algal photorespiratory
system in higher plants is favoured by the available data.
Further studies of plants either lacking functional GDH or
overexpressing the enzyme will help in understanding the
630 Bari et al.
role of a mitochondrial glycolate dehydrogenase in
photorespiration or alternate biochemical pathways.
Acknowledgements
The authors thank Nikolaus Schlaich and Nicolas Sauerbrunn for
helping with the cultivation of Arabidopsis and Dagmar Weier,
Rainer HaÈusler, and Veronica Maurino for discussions during this
investigation. Maria-Teresa Pellicer kindly provided us with E. coli
mutants de®cient in glycolate oxidase. Neill Emans put the confocal
¯uorescence microscope at our disposal. This work was supported
by a grant from the Deutsche Forschungsgemeinschaft to CP.
References
Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z,
Miller W, Lipman DJ. 1997. Gapped BLAST and PSI-BLAST:
a new generation of protein database search programs. Nucleic
Acids Research 25, 3389±3402.
Beezley BB, Gruber PJ, Frederick SE. 1976. Cytochemical
localization of glycolate dehydrogenase in mitochondria of
Chlamydomonas reinhardi. Plant Physiology 58, 315±319.
Chang YY, Wang AY, Cronan Jr JE. 1993. Molecular cloning,
DNA sequencing, and biochemical analyses of Escherichia coli
glyoxylate carboligase. An enzyme of the acetohydroxy acid
synthase-pyruvate oxidase family. Journal of Biological
Chemistry 268, 3911±3919.
Chomczynski P. 1993. A reagent for the single-step simultaneous
isolation of RNA, DNA and proteins from cell and tissue samples.
Biotechniques 15, 532±534.
Clough SJ, Bent AF. 1998. Floral dip: a simpli®ed method for
Agrobacterium-mediated transformation of Arabidopsis thaliana.
The Plant Journal 16, 735±743.
Colman B, Miller AG, Grodzinski B. 1974. A study on the control
of glycolate excretion in Chlorella pyrenoidosa. Plant Physiology
53, 395±397.
Drew MC. 1997. Oxygen de®ciency and root metabolism: injury
and acclimation under hypoxia and anoxia. Annual Review of
Plant Physiology and Plant Molecular Biology 48, 223±250.
Dym O, Eisenberg D. 2001. Sequence-structure analysis of FADcontaining proteins. Protein Science 10, 1712±1728.
Emanuelsson O, Nielsen H, Brunak S, von Heijne G. 2000.
Predicting subcellular localization of proteins based on their
N-terminal amino acid sequence. Journal of Molecular Biology
300, 1005±1016.
Frederick SE, Gruber PJ, Tolbert NE. 1973. The occurrence of
glycolate dehydrogenase and glycolate oxidase in green plants: an
evolutionary survey. Plant Physiology 52, 318±323.
Gia® CF, Rumsby G. 1998. Kinetic analysis and tissue distribution
of human D-glycerate dehydrogenase/glyoxylate reductase and its
relevance to the diagnosis of primary hyperoxaluria type 2.
Annual Review of Clinical Biochemistry 35, 104±109.
Givan CV, Kleczkowski LA. 1992. The enzymic reduction of
glyoxylate and hydroxypyruvate in leaves of higher plants. Plant
Physiology 100, 552±556.
Goyal A. 2002. Glycolate metabolism in algal chloroplasts:
inhibition by salicylhydroxamic acid (SHAM). Physiologia
Plantarum 116, 264±270.
Goyal A, Tolbert NE. 1996. Association of glycolate oxidation
with photosynthetic electron transport in plant and algal
chloroplasts. Proceedings of the National Academy of Sciences,
USA 93, 3319±3324.
Gutheil WG. 1998. A sensitive equilibrium-based assay for D-
lactate using D-lactate dehydrogenase: application to penicillinbinding protein/DD-carboxypeptidase activity assays. Analytical
Biochemistry 259, 62±67.
Husic DW, Tolbert NE. 1987. Inhibition of glycolate and D-lactate
metabolism in a Chlamydomonas reinhardtii mutant de®cient in
mitochondrial respiration. Proceedings of the National Academy
of Sciences, USA 84, 1555±1559.
Igamberdiev AU, Lea PJ. 2002. The role of peroxisomes in the
integration of metabolism and evolutionary diversity of
photosynthetic organisms. Phytochemistry 60, 651±674.
Igarashi D, Miwa T, Seki M, Kobayashi M, Kato T, Tabata S,
Shinozaki K, Ohsumi C. 2003. Identi®cation of photorespiratory
glutamate:glyoxylate aminotransferase (GGAT) gene in
Arabidopsis. The Plant Journal 33, 975±987.
Jach G, Binot E, Frings S, Luxa K, Schell J. 2001. Use of red
¯uorescent protein from Discosoma sp. (dsRED) as a reporter for
plant gene expression. The Plant Journal 28, 483±491.
Koncz C, Schell J. 1986. The promoter of TL-DNA gene 5 controls
the tissue-speci®c expression of chimaeric genes carried by a
novel type of Agrobacterium binary vector. Molecular and
General Genetics 204, 383±396.
Kozaki A, Takeba G. 1996. Photorespiration protects C3 plants
from photooxidation. Nature 384, 557±560.
Leegood RC, Lea PJ, Adcock MD, HaÈusler RE. 1995. The
regulation and control of photorespiration. Journal of
Experimental Botany 46, 1397±1414.
Lord JM. 1972. Glycolate oxidoreductase in Escherichia coli.
Biochimica et Biophysica Acta 267, 227±237.
Miller JH. 1972. Experiments in molecular genetics. New York:
Cold Spring Harbor Laboratory Press.
Nelson EB, Tolbert NE. 1970. Glycolate dehydrogenase in green
algae. Archives of Biochemistry and Biophysics 141, 102±110.
Ogren WL. 1984. Photorespiration: pathways, regulation and
modi®cation. Annual Review of Plant Physiology and Plant
Molecular Biology 35, 415±442.
Pellicer MT, Badia J, Aguilar J, Baldoma L. 1996. glc locus of
Escherichia coli: characterization of genes encoding the subunits
of glycolate oxidase and the glc regulator protein. Journal of
Bacteriology 178, 2051±2059.
Somerville CR. 1984. The analysis of photosynthetic carbon
dioxide ®xation and photorespiration by mutant selection. Oxford
Surveys of Plant Molecular and Cell Biology 1, 103±131.
Stabenau H. 1992. Evolution of peroxisomes in green algae. In:
Stabenau H, ed. Phylogenetic changes in peroxisomes of algae.
Phylogeny of plant peroxisomes. Oldenburg: University of
Oldenburg, 63±79.
Stabenau H, Winkler U, Saftel W. 1984. Mitochondrial
metabolism of glycolate in the alga Eremosphaera viridis.
Zeitschrift fuÈr P¯anzenernaehrung und Bodenkunde 114, 413±
420.
Tolbert NE. 1997. The C2 oxidative photosynthetic carbon cycle.
Annual Review of Plant Physiology and Plant Molecular Biology
48, 1±25.
Winkler U, Stabenau H. 1992. Metabolism of glycolate in
mitochondria and peroxisomes of green algae. In: Stabenau H,
ed. Phylogenetic changes in peroxisomes of algae. Phylogeny of
plant peroxisomes. Oldenburg: University of Oldenburg, 80±91.
Yokota A, Haga S, Kitaoka S. 1985. Puri®cation and some
properties of glyoxylate reductase (NADP+) and its functional
location in mitochondria in Euglena gracilis Z. Biochemical
Journal 227, 211±216.
Yokota A, Kitaoka S. 1979. Occurrence and operation of the
glycollate±glyoxylate shuttle in mitochondria of Euglena gracilis
Z. Biochemical Journal 184, 189±192.