Plant Physiology Preview. Published on October 9, 2014, as DOI:10.1104/pp.114.248526
1 2 3 4 5 6 7 8 9 10 11
Running head: Cytochrome c oxidase mutant in Arabidopsis
12
Phone: +33-1-30-83-30-70
13
E-mail: [email protected]
14 15 16 17 18
Research area: Biochemistry and Metabolism
Corresponding author: Hakim Mireau UMR 1318 INRA/AgroParisTech – Institut Jean-Pierre Bourgin Route de Saint-Cyr 78026 Versailles Cedex France
1 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
Copyright 2014 by the American Society of Plant Biologists
19
Disruption of the CYTOCHROME C OXIDASE DEFICIENT 1 gene leads to
20
cytochrome c oxidase depletion and reorchestrated respiratory metabolism
21
in Arabidopsis thaliana
22 23
Jennifer Dahan1,2§, Guillaume Tcherkez3,4,5, David Macherel6,7,8, Abdelilah Benamar6,7,8,
24
Katia Belcram1,2, Martine Quadrado1,2, Nadège Arnal1,2, Hakim Mireau1,2*
25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
1
41
One-sentence summary: Depletion of the mitochondrial respiratory complex IV through
42
knockout of the COD1 PPR gene led to activation of the alternative respiratory pathway and
43
important switches in metabolic fluxes.
INRA, UMR1318, Institut Jean-Pierre Bourgin, RD10, F-78000 Versailles, France AgroParisTech, Institut Jean-Pierre Bourgin, RD10, F-78000 Versailles, France 3 Institut de Biologie des Plantes, CNRS UMR 8618, Bâtiment 630, Université Paris-Sud, 91405 Orsay Cedex, France 4 Institut Universitaire de France, 103 Boulevard Saint-Michel, 75005 Paris, France 5 Plateforme Métabolisme-Métabolome, IFR87, Université Paris-Sud, 91405 Orsay cedex, France 6 Université d’Angers, LUNAM Université, UMR 1345 Institut de Recherche en Horticulture et Semences, SFR 4207 QUASAV, Angers 49045, France 7 INRA, UMR 1345 Institut de Recherche en Horticulture et Semences, SFR 4207 QUASAV, Angers 49045, France 8 Agrocampus-Ouest, UMR 1345 Institut de Recherche en Horticulture et Semences, SFR 4207 QUASAV, Angers 49045, France 2
44
2 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
45
Footnotes: This work was supported by INRA and the French Agence Nationale de la
46
Recherche (ANR- 09-BLAN-0244 grant).
47 48 49 50 51 52 53 54
* Corresponding author, e-mail: [email protected] §
Current address: Department of Plant, Soil, and Entomological Sciences, University of Idaho, Moscow, Idaho 83844-2339
3 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
55
ABSTRACT
56
Cytochrome c oxidase is the last respiratory complex of the electron transfer chain in
57
mitochondria and is responsible for transferring electrons to oxygen, the final acceptor, in the
58
classical respiratory pathway. The essentiality of this step makes that depletion in complex IV
59
leads to lethality, thereby impeding studies on complex IV assembly and respiration plasticity
60
in plants. Here, we characterized Arabidopsis embryo-lethal mutant lines impaired in the
61
expression of the COD1 gene, which encodes a mitochondria-localized PPR protein.
62
Although unable to germinate under usual conditions, cod1 homozygous embryos could be
63
rescued from immature seeds and developed in vitro into slow-growing bush-like plantlets,
64
devoid of root system. cod1 mutants were defective in C-to-U editing events in cox2 and nad4
65
transcripts, encoding subunits of respiratory complex IV and I, respectively, and consequently
66
lacked cytochrome c oxidase activity. We further show that respiratory oxygen consumption
67
by cod1 plantlets is exclusively associated with alternative oxidase activity, and that
68
alternative NADH dehydrogenases are also upregulated in these plants. The metabolomics
69
pattern of cod1 mutants was also deeply altered, suggesting that alternative metabolic
70
pathways compensated for the probable resulting restriction in NADH oxidation. Being the
71
first complex IV-deficient mutants described in higher plants, cod1 lines should be
72
instrumental to future studies on respiration homeostasis.
73 74
4 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
75
INTRODUCTION
76
Mitochondria are vital cellular organelles responsible for energy production through aerobic
77
respiration. They are also the place of numerous crossroads between metabolic pathways and
78
signal transduction. Mitochondria are thus crucial actors in the coordination of cellular
79
activities to maintain metabolic homeostasis. Since the endosymbiotic event at the origin of
80
these organelles, mitochondria have become semi-autonomous, possessing their own genetic
81
material and protein synthesis machinery. However, to carry out all their activities, they rely
82
heavily on the import of nuclear-encoded proteins. In fact, plant mitochondrial proteome is
83
believed to represent ~2500 proteins, among which only a handful (32 in Arabidopsis) is
84
encoded by the mitochondrial genome (Millar et al., 2005; Lee et al., 2013).
85
ATP is produced in mitochondria through oxidative phosphorylation reactions, that rely on
86
the oxidation of NAD(P)H and succinate and involve electron flow through respiratory
87
complexes I, II, III and IV. These redox reactions are coupled with the establishment of a
88
proton gradient across the inner membrane whose dissipation by complex V allows ATP
89
synthesis. Alternative electron transfers through NADH dehydrogenases (NDH) and
90
alternative oxidase (AOX) that are not coupled to proton translocation and thus ATP synthesis
91
also occur. They are believed to play the role of an energy dissipating pathway, allowing cells
92
to cope with oxidative stress and reactive oxygen species (ROS) (Rasmusson et al., 2008;
93
Vanlerberghe, 2013).
94
Due to its essentiality in nature, functional perturbations of the mitochondrial electron
95
transport chain (mETC) cause profound alterations at early stages of embryo development,
96
except for those affecting respiratory complex I activity. Although lethal in animals, plants
97
with disrupted complex I are usually viable and functionally rescued by the expression of
98
alternative NDH. Plant mutants affected in the expression of nuclear- as well as mitochondria-
99
encoded complex I subunits display a diversity of phenotypes, such as germination
100
deficiencies, growth retardation, developmental defects and altered responses to hormones or
101
stresses (Lee et al., 2002; Perales et al., 2005; Longevialle et al., 2007; Zsigmond et al., 2008;
102
Meyer et al., 2009; Murayama et al., 2012; Toda et al., 2012; Yuan and Liu, 2012; Haïli et al.,
103
2013; Zhu et al., 2014). Plant complex I mutants have been largely used to study complex I
104
architecture, respiration plasticity and metabolism adaptation (Dutilleul et al., 2003; Noctor et
105
al., 2004; Meyer et al., 2011; Braun et al., 2014). In contrast to the large amount of data
106
available on complex I function and mutants, very little has been described on other
107
components of mETC, likely because of lethality, which prevents physiological
108
characterization. Indeed, complex II depletion results in gametophyte-lethality in Arabidopsis 5 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
109
thaliana (León et al., 2007) and mutants affected in complex III and IV are thought to be
110
embryo-lethal (Meyer et al., 2005). Previous attempts to maintain and grow complex IV
111
mutants were performed in maize (Zea mays), using the heteroplasmic nonchromosomal
112
stripe 6 (NCS6) line (Lauer et al., 1990). This line possesses mitochondria harboring a partial
113
deletion in the mitochondrial cox2 gene, encoding one of the three complex IV core subunits.
114
Very few embryogenic calli homoplasmic for the cox2 deletion could be produced by tissue
115
culture (Gu et al., 1994), and none of these homoplasmic mutant calli was able to regenerate
116
complex IV-deficient plantlets that could have been used to understand the assembly of
117
complex IV as well as the plasticity of the respiratory chain.
118
The expression of mitochondrial mRNA, such as those encoding subunits of mETC
119
complexes, implies an astonishing number of post-transcriptional RNA processing steps,
120
comprising cis- and trans-splicing, 5’ and 3’ trimming, stabilization and editing (Binder and
121
Brennicke, 2003). Although very little is known about the molecular actors involved in these
122
steps, the PentatricoPeptide Repeat (PPR) proteins have been shown to represent fundamental
123
players in these aspects. PPR proteins are RNA-binding proteins characterized by tandem
124
repeats of a poorly conserved 35-aminoacid motif, and are for the vast majority targeted to
125
plastids and/or mitochondria (Small and Peeters, 2000; Lurin et al., 2004). They have been
126
shown to take part in virtually all post-transcriptional RNA processing steps necessary for
127
proper expression of organellar genomes, including C-to-U editing (Barkan and Small, 2013).
128
PPR proteins achieve their functions by specific binding to cis-elements within their target
129
RNA, acting presumably as platforms for additional factors bearing enzymatic activities
130
(Nakamura et al., 2012; Ke et al., 2013; Yin et al., 2013). Accordingly, disruption of PPR
131
genes leads to alterations in the normal maturation process of the targeted organellar
132
transcripts, impairing their correct expression and associated activities. Considering the
133
inability to transform mitochondria to generate mitochondrial mutants, PPR gene mutants
134
represent invaluable tools to study mitochondrial activities, regulation and metabolism, by
135
allowing indirect knockout of mitochondrial genes.
136
In the present study, we analyzed the cod1 mutants, impaired in the expression of a
137
mitochondria-targeted E/E+ PPR and which are characterized by a late embryo development
138
arrest leading to seed abortion. Using an embryo rescue approach, we succeeded in growing
139
homozygous cod1 plantlets in vitro, which displayed a complete loss of mitochondrial
140
respiratory complex IV. We further showed that this absence is due to the lack of C-to-U
141
editing at two sites within the cox2 transcript. We took advantage of this first complex IV
6 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
142
mutant identified in plants to conduct respiratory and metabolomics analyses, so as to
143
characterize the metabolic response to such a severe respiratory alteration.
144 145
RESULTS
146
Phenotypic characterization of cod1 mutants
147
The cytochrome c oxidase deficient 1 (cod1) mutant was identified in the course of a reverse
148
genetic screen analyzing a collection of Arabidopis thaliana mutants having T-DNA
149
insertions in nuclear genes encoding PPR proteins putatively localized to mitochondria. Two
150
heterozygous insertion lines (SALK_000882 (cod1-1) and SALK_615308 (cod1-2)) affected
151
in the At2g35030 gene were phenotyped as potential embryo-lethal mutants, as no
152
homozygous mutant individuals could be identified in the progeny of self-fertilized
153
heterozygous plants. Both insertions were mapped within the COD1 gene coding sequence,
154
precisely 990 (cod1-2) and 1707 nucleotides (cod1-1) downstream of the start codon (Fig.
155
1A). Segregation analysis further revealed that about a quarter of seeds recovered from self-
156
pollinated heterozygous cod1 plants were not able to germinate in vitro on MS medium,
157
suggesting a germination deficiency associated with a single recessive mutation (Fig. 1B).
158
Consistent with this result, immature siliques from heterozygous plants contained about a
159
quarter of yellow to white seeds (Fig. 1C). Confocal analysis of the embryos contained within
160
mutant mature seeds revealed that their size was greatly reduced compared to the wild type.
161
Nevertheless, they reached the cotyledon stage and no obvious morphological defects in the
162
embryo cellular architecture could be observed at this developmental stage (Fig. 1D). They
163
did not eventually complete full maturation, and ~25% of mature dried seeds produced by
164
COD1/cod1 heterozygous plant were shrunken and could not germinate on classical MS
165
medium (Supplemental Fig. S1). This embryo developmental arrest phenotype clearly
166
indicated an essential role of the COD1 protein in seed development and survival.
167
To generate enough mutant biological material to work on and study the function of the
168
COD1 protein, we tentatively rescued immature mutant embryos from developing siliques of
169
heterozygous plants. Dried mutant seeds appeared to be improper for such rescue, as they did
170
not germinate in any of the tested conditions. We established optimal conditions to germinate
171
and grow cod1 mutants in vitro from immature seeds, by plating them successively on a high-
172
and low-sugar medium, supplemented with cofactors and nutrients. After several weeks in
173
growth chamber, homozygous mutant seeds were able to germinate. However, their
174
subsequent growth was retarded and very slow, as compared with wild type seeds grown
175
under the same conditions. When genotyped, these slow-growing seedlings were all 7 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
176
homozygous for their respective T-DNA insertions in COD1. The development of these
177
seedlings was quite anarchical, with proliferation of short leaves with limited stems, leading
178
to a small bush-like structure (Fig. 1E). Strikingly, root development was severely
179
compromised in most of the plantlets. Several weeks to months were needed to obtain a fully
180
developed plantlet and a sufficient amount of biological material for further analysis.
181
Occasionally, some mutant plants were able to produce flowers, which did not produce viable
182
pollen (Supplemental Fig. S2). The lack of coordinated development of organs (leaves, stem
183
and roots) was not due to possible detrimental effects of our in vitro culture conditions, since
184
immature wild-type seeds developed normally under the same conditions (Fig. 1E).
185 186
COD1 is a mitochondrial protein from the E/E+ subclass of PPR family
187
The At2g35030 gene encodes a PPR protein of 71 kDa, containing 13 PPR repeats. It belongs
188
to the PLS subfamily, due to the tandem repeat of P, L and S-type PPR motifs (Supplemental
189
Fig. S3A). Domain structure prediction further indicated that COD1 contains the C-terminal
190
extension E/E+, which has been correlated to C-to-U RNA editing (Schmitz-Linneweber and
191
Small, 2008). To determine the subcellular localization of COD1, its amino acid sequence
192
was analyzed with TargetP (Emanuelsson et al., 2007) and Predotar (Small et al., 2004)
193
programs. Both predicted a localization signal peptide at the N-terminus putatively driving the
194
protein to mitochondria. To experimentally verify this subcellular localization, the full-length
195
COD1 protein was fused to the GFP reporter at its C-terminus and the fusion protein was
196
transiently expressed in epidermal cells of Nicotiana benthamiana. The COD1-GFP fusion
197
protein construct was co-expressed with a mitochondria-targeted mCherry marker used as a
198
control (Nelson et al., 2007). Analysis by confocal laser scanning microscopy showed an
199
almost perfect overlap of the GFP and mCherry fluorescent signals, strongly supporting the
200
mitochondrial localization of COD1 (Supplemental Fig. S3B). Additionally, the GFP
201
fluorescence did not co-localize with the plastid auto-fluorescence, indicating that COD1 is
202
not likely transported into plastids in vivo.
203 204
Complex IV activity and assembly are compromised in cod1 mutants
205
The Arabidopsis thaliana mitochondrial genome encodes 32 proteins amongst which 19 are
206
subunits of mitochondrial respiratory complexes, except for complex II which is entirely
207
encoded by nuclear genes (Unseld et al., 1997). Due to the mitochondrial localization of the
208
COD1 proteins, the abundance of mitochondrial respiratory complexes was investigated by
209
blue native polyacrylamide gel electrophoresis (BN-PAGE). Since cod1 mutants developed 8 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
210
only as miniature plants, crude membrane extracts enriched in mitochondria instead of
211
purified mitochondria were used to perform this analysis. Based on in-gel activity staining,
212
complex I and V showed no apparent accumulation defects in cod1-1 and cod1-2 plantlets.
213
However, complex IV activity could not be detected at all in both mutants, indicating that
214
disruption of the COD1 gene leads to a deficiency in complex IV activity (Fig. 2A). Complex
215
III presented an intermediate situation with a slight but significant decrease in the mutants as
216
compared to the wild type (Fig. 2B). To investigate whether the lack of cytochrome c oxidase
217
activity could result from a loss of complex IV assembly, membranes derived from BN-
218
PAGE were hybridized with antibodies directed against COX2, one of the 3 mitochondria-
219
encoded complex IV subunits. Antibodies against NAD9, a mitochondria-encoded complex I
220
subunit, was also used as control. Complex I could be detected with the anti-NAD9 antibodies
221
in extracts prepared from cod1 mutants. By contrast, no signal corresponding to native
222
complex IV could be detected using anti-COX2 antibodies in extracts from both cod1-1 and
223
cod1-2 plants (Fig. 2B). This result confirmed that cod1 mutants corresponded to complex IV
224
defective plants, strongly suggesting that the expression of at least one mitochondria-encoded
225
COX subunits was altered in cod1 plants.
226 227
COD1 is required for editing of two critical sites in cox2 mRNA and one in nad4
228
transcript
229
The presence of an E/E+ domain at the end of the COD1 protein suggested a possible role of
230
COD1 in C-to-U RNA editing in mitochondria. Since disruption of the COD1 gene led to a
231
lack of detectable complex IV (see above) with no other significant deficiencies in other
232
respiratory complexes, we first investigated the editing status of the mitochondrial RNA
233
encoding subunits of cytochrome c oxidase that are known to be edited, namely cox2 and
234
cox3. While no change in the editing status of cox3 mRNA was observed in cod1 mutants, the
235
editing of two distant sites in cox2 transcripts was completely abolished: cox2-253 and cox2-
236
698 (Fig. 3).
237
Since a number of PPR proteins involved in RNA editing have been shown to target
238
numerous sites, the editing status of the other mitochondrial RNAs was also investigated.
239
Amongst the 451 sites monitored, only the nad4-1129 site appeared to be unedited in the cod1
240
mutants as compared with the wild type (Fig. 3). The three editing losses (2 sites in cox2 and
241
1 in nad4) are likely not due to side effects of altered RNA processing or stability, since no
242
differences in the pattern of the cox2 and nad4 transcripts were detected in RNA gel blot
243
analysis, apart from an over-accumulation of mature and premature forms (Supplemental Fig. 9 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
244
S4A). The up-regulation of cox genes was confirmed by quantitative RT-PCR analysis, which
245
also revealed that the up-regulation concerned almost all mitochondria-encoded transcripts in
246
cod1 mutants (Supplemental Fig. S4B, S5).
247
The absence of C-to-U editing at the three detected sites resulted in amino acid changes
248
within the corresponding protein sequences. In the nad4 transcript, the lack of C-to-U editing
249
at site nad4-1129 converts Phe377 into Leu (Supplemental Fig. S6). At the cox2-253 site, it
250
converts Trp85 to Arg, and at the cox2-698 site it changes Met233 to Thr (Supplemental Fig.
251
S6). Sequence alignments of plant, yeast and animal COX2 and NAD4 sequences indicated
252
that these three amino acid positions correspond to highly conserved amino acids in their
253
respective proteins (Supplemental Fig. S6).
254
Since the loss of editing at the cox2-253 and cox2-698 sites did not lead to premature stop
255
codon formation, the presence of COX2 in crude membrane extracts from wild type and
256
mutants was investigated by immunoblot analysis. Interestingly, the COX2 subunit did not
257
accumulate to detectable levels in mutant plant extracts. As a control, the COX1 subunit was
258
also detected in the same samples and appeared to substantially over-accumulate in both
259
mutants (Fig. 4). This result was likely due to the observed over-accumulation of the
260
corresponding transcript (Supplemental Fig. S4B). The amino acid substitutions associated
261
with the cod1 mutations are probably responsible for the severe destabilization of COX2,
262
hence leading to the observed complex IV default (see above). By contrast, since no
263
accumulation defects were observed for complex I, it is likely that the Phe377Leu conversion
264
had no detrimental effects on NAD(P)H-quinone dehydrogenase assembly and activity.
265 266
Respiration is altered and the alternative respiratory pathway is strongly induced in
267
cod1 mutants
268
Cytochrome c oxidase is the last complex of the electron transfer chain, transferring electrons
269
from oxidized cytochrome c to the final acceptor O2. The severe developmental phenotype of
270
cod1 mutants and the lack of detectable complex IV activity in cod1-1 and cod1-2 extracts
271
suggested strong alterations of respiratory activities. We therefore investigated the uptake of
272
O2 in freshly harvested Col-0 and cod1 mutant seedlings using a Clark’s electrode (Fig. 5A
273
and Supplemental Fig. S7). Intriguingly, cod1-1 and cod1-2 extracts showed O2 consumption
274
rate in the dark that are not significantly different from the wild type. To verify whether
275
complex IV was responsible for O2 consumption in cod1 mutants, we next measured the O2
276
uptake in the dark in the presence of KCN, a potent complex IV-specific inhibitor. As
277
expected, the O2 consumption rate was strongly reduced in the presence of cyanide in wild 10 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
278
type samples (about 66% decrease). On the contrary, the inhibitor had no significant effect on
279
O2 uptake by cod1-1 and cod1-2 explants. These results correlate with the absence of active
280
complex IV as observed on BN-PAGE analysis (Fig. 2A). We next investigated the origins of
281
the O2 consumption in cod1 mutants by looking at the alternative, cyanide-insensitive
282
respiratory pathway, mediated by AOX. The inhibitor of AOX, n-propylgallate, was added to
283
the reaction medium and the O2 consumption rate was measured (Fig. 5A). In both mutant
284
explants, the addition of n-propylgallate caused a considerable reduction of O2 consumption,
285
while it reduced wild type consumption rates of only about 20%. When n-propylgallate was
286
used in conjunction with KCN, the respiration rates decreased to low, basal levels in all
287
genotypes. Furthermore, these basal O2 consumption rates were not significantly different
288
from the level reached in the cod1 mutants when n-propylgallate was used alone (Student t-
289
test, P
1 2 3 4 5 6 7 8 9 10 11
Running head: Cytochrome c oxidase mutant in Arabidopsis
12
Phone: +33-1-30-83-30-70
13
E-mail: [email protected]
14 15 16 17 18
Research area: Biochemistry and Metabolism
Corresponding author: Hakim Mireau UMR 1318 INRA/AgroParisTech – Institut Jean-Pierre Bourgin Route de Saint-Cyr 78026 Versailles Cedex France
1 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
Copyright 2014 by the American Society of Plant Biologists
19
Disruption of the CYTOCHROME C OXIDASE DEFICIENT 1 gene leads to
20
cytochrome c oxidase depletion and reorchestrated respiratory metabolism
21
in Arabidopsis thaliana
22 23
Jennifer Dahan1,2§, Guillaume Tcherkez3,4,5, David Macherel6,7,8, Abdelilah Benamar6,7,8,
24
Katia Belcram1,2, Martine Quadrado1,2, Nadège Arnal1,2, Hakim Mireau1,2*
25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
1
41
One-sentence summary: Depletion of the mitochondrial respiratory complex IV through
42
knockout of the COD1 PPR gene led to activation of the alternative respiratory pathway and
43
important switches in metabolic fluxes.
INRA, UMR1318, Institut Jean-Pierre Bourgin, RD10, F-78000 Versailles, France AgroParisTech, Institut Jean-Pierre Bourgin, RD10, F-78000 Versailles, France 3 Institut de Biologie des Plantes, CNRS UMR 8618, Bâtiment 630, Université Paris-Sud, 91405 Orsay Cedex, France 4 Institut Universitaire de France, 103 Boulevard Saint-Michel, 75005 Paris, France 5 Plateforme Métabolisme-Métabolome, IFR87, Université Paris-Sud, 91405 Orsay cedex, France 6 Université d’Angers, LUNAM Université, UMR 1345 Institut de Recherche en Horticulture et Semences, SFR 4207 QUASAV, Angers 49045, France 7 INRA, UMR 1345 Institut de Recherche en Horticulture et Semences, SFR 4207 QUASAV, Angers 49045, France 8 Agrocampus-Ouest, UMR 1345 Institut de Recherche en Horticulture et Semences, SFR 4207 QUASAV, Angers 49045, France 2
44
2 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
45
Footnotes: This work was supported by INRA and the French Agence Nationale de la
46
Recherche (ANR- 09-BLAN-0244 grant).
47 48 49 50 51 52 53 54
* Corresponding author, e-mail: [email protected] §
Current address: Department of Plant, Soil, and Entomological Sciences, University of Idaho, Moscow, Idaho 83844-2339
3 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
55
ABSTRACT
56
Cytochrome c oxidase is the last respiratory complex of the electron transfer chain in
57
mitochondria and is responsible for transferring electrons to oxygen, the final acceptor, in the
58
classical respiratory pathway. The essentiality of this step makes that depletion in complex IV
59
leads to lethality, thereby impeding studies on complex IV assembly and respiration plasticity
60
in plants. Here, we characterized Arabidopsis embryo-lethal mutant lines impaired in the
61
expression of the COD1 gene, which encodes a mitochondria-localized PPR protein.
62
Although unable to germinate under usual conditions, cod1 homozygous embryos could be
63
rescued from immature seeds and developed in vitro into slow-growing bush-like plantlets,
64
devoid of root system. cod1 mutants were defective in C-to-U editing events in cox2 and nad4
65
transcripts, encoding subunits of respiratory complex IV and I, respectively, and consequently
66
lacked cytochrome c oxidase activity. We further show that respiratory oxygen consumption
67
by cod1 plantlets is exclusively associated with alternative oxidase activity, and that
68
alternative NADH dehydrogenases are also upregulated in these plants. The metabolomics
69
pattern of cod1 mutants was also deeply altered, suggesting that alternative metabolic
70
pathways compensated for the probable resulting restriction in NADH oxidation. Being the
71
first complex IV-deficient mutants described in higher plants, cod1 lines should be
72
instrumental to future studies on respiration homeostasis.
73 74
4 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
75
INTRODUCTION
76
Mitochondria are vital cellular organelles responsible for energy production through aerobic
77
respiration. They are also the place of numerous crossroads between metabolic pathways and
78
signal transduction. Mitochondria are thus crucial actors in the coordination of cellular
79
activities to maintain metabolic homeostasis. Since the endosymbiotic event at the origin of
80
these organelles, mitochondria have become semi-autonomous, possessing their own genetic
81
material and protein synthesis machinery. However, to carry out all their activities, they rely
82
heavily on the import of nuclear-encoded proteins. In fact, plant mitochondrial proteome is
83
believed to represent ~2500 proteins, among which only a handful (32 in Arabidopsis) is
84
encoded by the mitochondrial genome (Millar et al., 2005; Lee et al., 2013).
85
ATP is produced in mitochondria through oxidative phosphorylation reactions, that rely on
86
the oxidation of NAD(P)H and succinate and involve electron flow through respiratory
87
complexes I, II, III and IV. These redox reactions are coupled with the establishment of a
88
proton gradient across the inner membrane whose dissipation by complex V allows ATP
89
synthesis. Alternative electron transfers through NADH dehydrogenases (NDH) and
90
alternative oxidase (AOX) that are not coupled to proton translocation and thus ATP synthesis
91
also occur. They are believed to play the role of an energy dissipating pathway, allowing cells
92
to cope with oxidative stress and reactive oxygen species (ROS) (Rasmusson et al., 2008;
93
Vanlerberghe, 2013).
94
Due to its essentiality in nature, functional perturbations of the mitochondrial electron
95
transport chain (mETC) cause profound alterations at early stages of embryo development,
96
except for those affecting respiratory complex I activity. Although lethal in animals, plants
97
with disrupted complex I are usually viable and functionally rescued by the expression of
98
alternative NDH. Plant mutants affected in the expression of nuclear- as well as mitochondria-
99
encoded complex I subunits display a diversity of phenotypes, such as germination
100
deficiencies, growth retardation, developmental defects and altered responses to hormones or
101
stresses (Lee et al., 2002; Perales et al., 2005; Longevialle et al., 2007; Zsigmond et al., 2008;
102
Meyer et al., 2009; Murayama et al., 2012; Toda et al., 2012; Yuan and Liu, 2012; Haïli et al.,
103
2013; Zhu et al., 2014). Plant complex I mutants have been largely used to study complex I
104
architecture, respiration plasticity and metabolism adaptation (Dutilleul et al., 2003; Noctor et
105
al., 2004; Meyer et al., 2011; Braun et al., 2014). In contrast to the large amount of data
106
available on complex I function and mutants, very little has been described on other
107
components of mETC, likely because of lethality, which prevents physiological
108
characterization. Indeed, complex II depletion results in gametophyte-lethality in Arabidopsis 5 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
109
thaliana (León et al., 2007) and mutants affected in complex III and IV are thought to be
110
embryo-lethal (Meyer et al., 2005). Previous attempts to maintain and grow complex IV
111
mutants were performed in maize (Zea mays), using the heteroplasmic nonchromosomal
112
stripe 6 (NCS6) line (Lauer et al., 1990). This line possesses mitochondria harboring a partial
113
deletion in the mitochondrial cox2 gene, encoding one of the three complex IV core subunits.
114
Very few embryogenic calli homoplasmic for the cox2 deletion could be produced by tissue
115
culture (Gu et al., 1994), and none of these homoplasmic mutant calli was able to regenerate
116
complex IV-deficient plantlets that could have been used to understand the assembly of
117
complex IV as well as the plasticity of the respiratory chain.
118
The expression of mitochondrial mRNA, such as those encoding subunits of mETC
119
complexes, implies an astonishing number of post-transcriptional RNA processing steps,
120
comprising cis- and trans-splicing, 5’ and 3’ trimming, stabilization and editing (Binder and
121
Brennicke, 2003). Although very little is known about the molecular actors involved in these
122
steps, the PentatricoPeptide Repeat (PPR) proteins have been shown to represent fundamental
123
players in these aspects. PPR proteins are RNA-binding proteins characterized by tandem
124
repeats of a poorly conserved 35-aminoacid motif, and are for the vast majority targeted to
125
plastids and/or mitochondria (Small and Peeters, 2000; Lurin et al., 2004). They have been
126
shown to take part in virtually all post-transcriptional RNA processing steps necessary for
127
proper expression of organellar genomes, including C-to-U editing (Barkan and Small, 2013).
128
PPR proteins achieve their functions by specific binding to cis-elements within their target
129
RNA, acting presumably as platforms for additional factors bearing enzymatic activities
130
(Nakamura et al., 2012; Ke et al., 2013; Yin et al., 2013). Accordingly, disruption of PPR
131
genes leads to alterations in the normal maturation process of the targeted organellar
132
transcripts, impairing their correct expression and associated activities. Considering the
133
inability to transform mitochondria to generate mitochondrial mutants, PPR gene mutants
134
represent invaluable tools to study mitochondrial activities, regulation and metabolism, by
135
allowing indirect knockout of mitochondrial genes.
136
In the present study, we analyzed the cod1 mutants, impaired in the expression of a
137
mitochondria-targeted E/E+ PPR and which are characterized by a late embryo development
138
arrest leading to seed abortion. Using an embryo rescue approach, we succeeded in growing
139
homozygous cod1 plantlets in vitro, which displayed a complete loss of mitochondrial
140
respiratory complex IV. We further showed that this absence is due to the lack of C-to-U
141
editing at two sites within the cox2 transcript. We took advantage of this first complex IV
6 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
142
mutant identified in plants to conduct respiratory and metabolomics analyses, so as to
143
characterize the metabolic response to such a severe respiratory alteration.
144 145
RESULTS
146
Phenotypic characterization of cod1 mutants
147
The cytochrome c oxidase deficient 1 (cod1) mutant was identified in the course of a reverse
148
genetic screen analyzing a collection of Arabidopis thaliana mutants having T-DNA
149
insertions in nuclear genes encoding PPR proteins putatively localized to mitochondria. Two
150
heterozygous insertion lines (SALK_000882 (cod1-1) and SALK_615308 (cod1-2)) affected
151
in the At2g35030 gene were phenotyped as potential embryo-lethal mutants, as no
152
homozygous mutant individuals could be identified in the progeny of self-fertilized
153
heterozygous plants. Both insertions were mapped within the COD1 gene coding sequence,
154
precisely 990 (cod1-2) and 1707 nucleotides (cod1-1) downstream of the start codon (Fig.
155
1A). Segregation analysis further revealed that about a quarter of seeds recovered from self-
156
pollinated heterozygous cod1 plants were not able to germinate in vitro on MS medium,
157
suggesting a germination deficiency associated with a single recessive mutation (Fig. 1B).
158
Consistent with this result, immature siliques from heterozygous plants contained about a
159
quarter of yellow to white seeds (Fig. 1C). Confocal analysis of the embryos contained within
160
mutant mature seeds revealed that their size was greatly reduced compared to the wild type.
161
Nevertheless, they reached the cotyledon stage and no obvious morphological defects in the
162
embryo cellular architecture could be observed at this developmental stage (Fig. 1D). They
163
did not eventually complete full maturation, and ~25% of mature dried seeds produced by
164
COD1/cod1 heterozygous plant were shrunken and could not germinate on classical MS
165
medium (Supplemental Fig. S1). This embryo developmental arrest phenotype clearly
166
indicated an essential role of the COD1 protein in seed development and survival.
167
To generate enough mutant biological material to work on and study the function of the
168
COD1 protein, we tentatively rescued immature mutant embryos from developing siliques of
169
heterozygous plants. Dried mutant seeds appeared to be improper for such rescue, as they did
170
not germinate in any of the tested conditions. We established optimal conditions to germinate
171
and grow cod1 mutants in vitro from immature seeds, by plating them successively on a high-
172
and low-sugar medium, supplemented with cofactors and nutrients. After several weeks in
173
growth chamber, homozygous mutant seeds were able to germinate. However, their
174
subsequent growth was retarded and very slow, as compared with wild type seeds grown
175
under the same conditions. When genotyped, these slow-growing seedlings were all 7 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
176
homozygous for their respective T-DNA insertions in COD1. The development of these
177
seedlings was quite anarchical, with proliferation of short leaves with limited stems, leading
178
to a small bush-like structure (Fig. 1E). Strikingly, root development was severely
179
compromised in most of the plantlets. Several weeks to months were needed to obtain a fully
180
developed plantlet and a sufficient amount of biological material for further analysis.
181
Occasionally, some mutant plants were able to produce flowers, which did not produce viable
182
pollen (Supplemental Fig. S2). The lack of coordinated development of organs (leaves, stem
183
and roots) was not due to possible detrimental effects of our in vitro culture conditions, since
184
immature wild-type seeds developed normally under the same conditions (Fig. 1E).
185 186
COD1 is a mitochondrial protein from the E/E+ subclass of PPR family
187
The At2g35030 gene encodes a PPR protein of 71 kDa, containing 13 PPR repeats. It belongs
188
to the PLS subfamily, due to the tandem repeat of P, L and S-type PPR motifs (Supplemental
189
Fig. S3A). Domain structure prediction further indicated that COD1 contains the C-terminal
190
extension E/E+, which has been correlated to C-to-U RNA editing (Schmitz-Linneweber and
191
Small, 2008). To determine the subcellular localization of COD1, its amino acid sequence
192
was analyzed with TargetP (Emanuelsson et al., 2007) and Predotar (Small et al., 2004)
193
programs. Both predicted a localization signal peptide at the N-terminus putatively driving the
194
protein to mitochondria. To experimentally verify this subcellular localization, the full-length
195
COD1 protein was fused to the GFP reporter at its C-terminus and the fusion protein was
196
transiently expressed in epidermal cells of Nicotiana benthamiana. The COD1-GFP fusion
197
protein construct was co-expressed with a mitochondria-targeted mCherry marker used as a
198
control (Nelson et al., 2007). Analysis by confocal laser scanning microscopy showed an
199
almost perfect overlap of the GFP and mCherry fluorescent signals, strongly supporting the
200
mitochondrial localization of COD1 (Supplemental Fig. S3B). Additionally, the GFP
201
fluorescence did not co-localize with the plastid auto-fluorescence, indicating that COD1 is
202
not likely transported into plastids in vivo.
203 204
Complex IV activity and assembly are compromised in cod1 mutants
205
The Arabidopsis thaliana mitochondrial genome encodes 32 proteins amongst which 19 are
206
subunits of mitochondrial respiratory complexes, except for complex II which is entirely
207
encoded by nuclear genes (Unseld et al., 1997). Due to the mitochondrial localization of the
208
COD1 proteins, the abundance of mitochondrial respiratory complexes was investigated by
209
blue native polyacrylamide gel electrophoresis (BN-PAGE). Since cod1 mutants developed 8 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
210
only as miniature plants, crude membrane extracts enriched in mitochondria instead of
211
purified mitochondria were used to perform this analysis. Based on in-gel activity staining,
212
complex I and V showed no apparent accumulation defects in cod1-1 and cod1-2 plantlets.
213
However, complex IV activity could not be detected at all in both mutants, indicating that
214
disruption of the COD1 gene leads to a deficiency in complex IV activity (Fig. 2A). Complex
215
III presented an intermediate situation with a slight but significant decrease in the mutants as
216
compared to the wild type (Fig. 2B). To investigate whether the lack of cytochrome c oxidase
217
activity could result from a loss of complex IV assembly, membranes derived from BN-
218
PAGE were hybridized with antibodies directed against COX2, one of the 3 mitochondria-
219
encoded complex IV subunits. Antibodies against NAD9, a mitochondria-encoded complex I
220
subunit, was also used as control. Complex I could be detected with the anti-NAD9 antibodies
221
in extracts prepared from cod1 mutants. By contrast, no signal corresponding to native
222
complex IV could be detected using anti-COX2 antibodies in extracts from both cod1-1 and
223
cod1-2 plants (Fig. 2B). This result confirmed that cod1 mutants corresponded to complex IV
224
defective plants, strongly suggesting that the expression of at least one mitochondria-encoded
225
COX subunits was altered in cod1 plants.
226 227
COD1 is required for editing of two critical sites in cox2 mRNA and one in nad4
228
transcript
229
The presence of an E/E+ domain at the end of the COD1 protein suggested a possible role of
230
COD1 in C-to-U RNA editing in mitochondria. Since disruption of the COD1 gene led to a
231
lack of detectable complex IV (see above) with no other significant deficiencies in other
232
respiratory complexes, we first investigated the editing status of the mitochondrial RNA
233
encoding subunits of cytochrome c oxidase that are known to be edited, namely cox2 and
234
cox3. While no change in the editing status of cox3 mRNA was observed in cod1 mutants, the
235
editing of two distant sites in cox2 transcripts was completely abolished: cox2-253 and cox2-
236
698 (Fig. 3).
237
Since a number of PPR proteins involved in RNA editing have been shown to target
238
numerous sites, the editing status of the other mitochondrial RNAs was also investigated.
239
Amongst the 451 sites monitored, only the nad4-1129 site appeared to be unedited in the cod1
240
mutants as compared with the wild type (Fig. 3). The three editing losses (2 sites in cox2 and
241
1 in nad4) are likely not due to side effects of altered RNA processing or stability, since no
242
differences in the pattern of the cox2 and nad4 transcripts were detected in RNA gel blot
243
analysis, apart from an over-accumulation of mature and premature forms (Supplemental Fig. 9 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
244
S4A). The up-regulation of cox genes was confirmed by quantitative RT-PCR analysis, which
245
also revealed that the up-regulation concerned almost all mitochondria-encoded transcripts in
246
cod1 mutants (Supplemental Fig. S4B, S5).
247
The absence of C-to-U editing at the three detected sites resulted in amino acid changes
248
within the corresponding protein sequences. In the nad4 transcript, the lack of C-to-U editing
249
at site nad4-1129 converts Phe377 into Leu (Supplemental Fig. S6). At the cox2-253 site, it
250
converts Trp85 to Arg, and at the cox2-698 site it changes Met233 to Thr (Supplemental Fig.
251
S6). Sequence alignments of plant, yeast and animal COX2 and NAD4 sequences indicated
252
that these three amino acid positions correspond to highly conserved amino acids in their
253
respective proteins (Supplemental Fig. S6).
254
Since the loss of editing at the cox2-253 and cox2-698 sites did not lead to premature stop
255
codon formation, the presence of COX2 in crude membrane extracts from wild type and
256
mutants was investigated by immunoblot analysis. Interestingly, the COX2 subunit did not
257
accumulate to detectable levels in mutant plant extracts. As a control, the COX1 subunit was
258
also detected in the same samples and appeared to substantially over-accumulate in both
259
mutants (Fig. 4). This result was likely due to the observed over-accumulation of the
260
corresponding transcript (Supplemental Fig. S4B). The amino acid substitutions associated
261
with the cod1 mutations are probably responsible for the severe destabilization of COX2,
262
hence leading to the observed complex IV default (see above). By contrast, since no
263
accumulation defects were observed for complex I, it is likely that the Phe377Leu conversion
264
had no detrimental effects on NAD(P)H-quinone dehydrogenase assembly and activity.
265 266
Respiration is altered and the alternative respiratory pathway is strongly induced in
267
cod1 mutants
268
Cytochrome c oxidase is the last complex of the electron transfer chain, transferring electrons
269
from oxidized cytochrome c to the final acceptor O2. The severe developmental phenotype of
270
cod1 mutants and the lack of detectable complex IV activity in cod1-1 and cod1-2 extracts
271
suggested strong alterations of respiratory activities. We therefore investigated the uptake of
272
O2 in freshly harvested Col-0 and cod1 mutant seedlings using a Clark’s electrode (Fig. 5A
273
and Supplemental Fig. S7). Intriguingly, cod1-1 and cod1-2 extracts showed O2 consumption
274
rate in the dark that are not significantly different from the wild type. To verify whether
275
complex IV was responsible for O2 consumption in cod1 mutants, we next measured the O2
276
uptake in the dark in the presence of KCN, a potent complex IV-specific inhibitor. As
277
expected, the O2 consumption rate was strongly reduced in the presence of cyanide in wild 10 Downloaded from www.plantphysiol.org on October 15, 2014 - Published by www.plant.org Copyright © 2014 American Society of Plant Biologists. All rights reserved.
278
type samples (about 66% decrease). On the contrary, the inhibitor had no significant effect on
279
O2 uptake by cod1-1 and cod1-2 explants. These results correlate with the absence of active
280
complex IV as observed on BN-PAGE analysis (Fig. 2A). We next investigated the origins of
281
the O2 consumption in cod1 mutants by looking at the alternative, cyanide-insensitive
282
respiratory pathway, mediated by AOX. The inhibitor of AOX, n-propylgallate, was added to
283
the reaction medium and the O2 consumption rate was measured (Fig. 5A). In both mutant
284
explants, the addition of n-propylgallate caused a considerable reduction of O2 consumption,
285
while it reduced wild type consumption rates of only about 20%. When n-propylgallate was
286
used in conjunction with KCN, the respiration rates decreased to low, basal levels in all
287
genotypes. Furthermore, these basal O2 consumption rates were not significantly different
288
from the level reached in the cod1 mutants when n-propylgallate was used alone (Student t-
289
test, P
