Subcomplexes of ancestral respiratory complex I subunits rapidly turn over in vivo as productive ...

23 Pages·2012·5.46 MB·English

JBC Papers in Press. Published on December 27, 2012 as Manuscript M112.432070
The latest version is at http://www.jbc.org/cgi/doi/10.1074/jbc.M112.432070

Subcomplexes of ancestral respiratory complex I subunits rapidly turn over in vivo as
productive assembly intermediates in Arabidopsis
Lei Lia,b, Clark J. Nelsona,b, Chris Carriea, Ryan M.R. Gawrylukc, Cory Solheim a,b,
Michael W. Grayc, James Whelana, A. Harvey Millara,b,*

aARC Centre of Excellence in Plant Energy Biology, M316, The University of Western
Australia, 35 Stirling Highway, Crawley WA 6009, Western Australia, Australia.
bCentre for Comparative Analysis of Biomolecular Networks (CABiN), M316, The
University of Western Australia, 35 Stirling Highway, Crawley WA 6009, Western Australia,
Australia.
cCentre for Comparative Genomics and Evolutionary Bioinformatics, Department of
Biochemistry and Molecular Biology, Dalhousie University, Halifax, Nova Scotia B3H 4R2,
Canada.
D
o
w
n
Running Title: Assembly of mitochondrial complex I in Arabidopsis lo
a
d
e
d
To whom correspondence should be addressed: fro
m
A. Harvey Millar e-mail: [email protected] Tel: [phone] Fax: [phone] http
://w
Key words: plant mitochondria, complex I, blue native electrophoresis, 15N incorporation. ww
.jbc
.o
brg/
y
Background: Plant complex I contains γ-CA anhydrase-like and 20.9-kDa subunits, had a g
u
e
subunits whose role is unclear. significantly higher turnover rate than intact CI st o
Results: 15N labeling and import show CI or CI+CIII2. In vitro import of precursors for n A
p
assembly via a rapidly turning over γ-CA these CI subunits demonstrated rapid generation ril 1
subcomplex. of subcomplexes and revealed that their specific 1, 2
0
Conclusions: γ-CAs form an ancient pathway abundance varied when different ancestral 1
9
for CI assembly. subunits were imported. Time-course studies of
Significance: The assembly pathway of precursor import showed the further assembly
complex I in plants is different to animals and of these subcomplexes into CI and CI+CIII ,
2
more closely represents the ancestral enzyme. indicating that the subcomplexes are productive
intermediates of assembly. The strong transient
SUMMARY incorporation of new subunits into the 200-kDa
subcomplex in a γ-CA mutant is consistent with
Subcomplexes of mitochondrial respiratory
this subcomplex being a key initiator of CI
complex I (CI; EC 1.6.5.3) are shown to turn
assembly in plants. This evidence alongside the
over in vivo and we propose a role in an
pattern of coincident occurrence of genes
ancestral assembly pathway. By progressively
encoding these particular proteins broadly in
labeling Arabidopsis cell cultures with 15N and
eukaryotes, outside opisthokonts, provides a
isolating mitochondria, we have identified CI
framework for the evolutionary conservation of
subcomplexes through differences in 15N
these accessory subunits and evidence of their
incorporation into their protein subunits. The
function in ancestral CI assembly.
200-kDa subcomplex, containing the ancestral
γ-carbonic anhydrase (γ-CA), γ-carbonic
1

Copyright 2012 by The American Society for Biochemistry and Molecular Biology, Inc.
INTRODUCTION considerably different among major taxonomic
groups of eukaryotes. In CI mutants of
NADH-ubiquinone oxidoreductase (Complex I
Neurospora crassa, analysis of CI assembly
(CI), EC 1.6.5.3) is the largest (~1000 kDa)
using radiolabeling pulse chase has revealed
protein complex of the oxidative
that the matrix and membrane arms assemble
phosphorylation (OXPHOS) assembly and is
independently via separate pathways (13-17).
also the major entry point of electrons from the
The membrane subunit 20.9, which is absent
oxidation of NADH into the respiratory chain.
from mammalian CI, was identified some years
CI contains 14 core subunits that represent the
ago as a factor essential for assembly of the
minimal active form and that are conserved
membrane arm in fungi (18). Based on a
from prokaryotic to eukaryotic organisms. In
combination of radiolabeling pulse-chase
the mitochondria of eukaryotes, CI contains an
experiments, in vitro mitochondrial import and
additional 31-38 accessory subunits, a number
monitoring of tagged CI subunits in assembly-
of which differ among organisms (1-4). Single-
disturbed systems, several CI assembly models
particle electron microscope imaging analysis of
have been proposed for human cells (14,19,20).
isolated CI from Arabidopsis and Polytomella
In this process, mammalian assembly factors
identified a specific matrix-exposed domain
not conserved in other species have been
formed by γ-carbonic anhydrase (γ-CA)
identified (13,14,19). Subcomplexes containing
subunits that are attached to the membrane arm D
CA2 have been reported using antibodies (11) o
w
of mitochondrial CI by the integral membrane n
in detergent disassembly studies of CI in plants lo
a
protein CA2 (5). In Arabidopsis, this protein d
(21), in Arabidopsis CI subunit knockout ed
domain contains at least three different γ-CA mutants (22), and in wild-type Arabidopsis fro
m
panrodt eCinAs3, C(AAt15 g(6A6t511g01)9, 5h8a0v)i,n gC Aco2n s(eArvt1egd4 a7c2t6iv0e) mofi ttowcoh-odnidmreianls ipornoatel obmluees nina tisvaetu (rBatNio)n g melasp (p2i3n)g. http://w
site regions, and two less well conserved γ-CA- w
Knockout of CA2 and CA3 do not yield a clear w
like proteins, CAL1 (At5g66510) and CAL2 phenotype in Arabidopsis (5,11), which might .jbc
(At3g48680), which contain alterations to the be a consequence of functional redundancy in .org
primary structure of typical γ-CAs (6). Links activities between these proteins. However, by/
g
between these subunits and photosynthetic u
suspension cell culture lines derived from the es
function have been postulated (7), a role in t o
mutant ca2 showed reduced growth rates and n
A
pbhyo tomgeonrpe hogkennoecsikso uint plasntutsd iheas s be(e8n), shoawnnd lγo-CwAer CsuI baubnuintsd anccoeu ld(1 1b),e wihmicpho ritnadnitc atfeosr thCaIt pril 11
overexpression leads to male sterility in plants , 2
assembly or structural stability in plants. 0
1
(9), but a CA activity from these proteins has 9
not been experimentally confirmed. While at However, reports of studies to date have not
first these subunits were thought to be specific been able to distinguish between productive and
to CI from photosynthetic organisms (10,11), non-productive intermediates, an essential
evidence has emerged of γ-CA proteins in distinction in building a rigorous CI assembly
mitochondria and in isolated CI from a range of model. Specifically we needed to determine if
other eukaryotic lineages including slime subcomplexes were present in vivo or if they
moulds and amoebae that lack photosynthesis were formed in vitro due to detergent
(12). Broader phylogenetic studies have also solubilisation or electrophoresis; if they were
found homology between other CI accessory incorporated into the small/early assembly
subunits once thought to be species-or lineage- intermediates; and if they were productive
specific (4). intermediates that go on to form CI or are
simply non-productive subcomplexes that
Given the divergence of CI composition
accumulated in mitochondria. We have
amongst eukaryotes, both the assembly pathway combined a progressive 15N labeling strategy in
and the biological importance of accessory
vivo (24) with BN polyacrylamide gel
subunits in assembly intermediates could be
electrophoresis (BN-PAGE) and in vitro
2
[35S]Met import and assembly assays to address was centrifuged at 800 g for 10 min. The
these issues. This approach has allowed analysis supernatant was decanted and the protoplast
of the incorporation rate of subunits into pellet resuspended in digestion solution without
subcomplexes, mature CI, and supercomplexes enzymes by swirling and repeated rounds of
of plant mitochondria to investigate their role in centrifugation. Washed protoplasts were then
assembly and to provide key evidence for the resuspended in extraction buffer (0.4 M sucrose,
existence of a productive CI subcomplex in vivo 50 mM Tris, 3 mM EDTA and 0.1% [w/v], pH
in plants. The subunits of the identified 7.5) and homogenized 5-9 times with a Dounce
assembly subcomplex, lost in animal lineages, 0.12 mm homogenizer. Extraction buffer
are widely conserved in other eukaryotic containing broken protoplasts was then
lineages, providing evidence for their role as centrifuged at 2500 g for 5 min, the supernatant
ancestral accessory subunits. recovered, and then this supernatant subjected
to centrifugation at 20,000 g for 20 min. The
EXPERIMENTAL PROCEDURES
resultant pellet was resuspended in 0.5-1 ml
mannitol wash buffer (0.3 M mannitol, 10 mM
Arabidopsis cell culture growth and 15N
TES and 0.1% [w/v] BSA pH 7.5) using a paint
labeling
brush and the suspension pipetted onto two 35
Arabidopsis cell suspension was cultured in ml Percoll step gradients comprising 40%, 25%
D
growth medium (1×Murashige and Skoog and 18% [w/v] Percoll in mannitol wash buffer. ow
n
medium without vitamin, 3% sucrose [w/v], 0.5 The gradients were centrifuged at 40,000 g for lo
a
d
mg/L naphthalene acetic acid, 0.05 mg/L 60 min and decelerated with slow braking. ed
kinetin, pH 5.8) at 22° C under continuous light Mitochondria, which form a yellow band at the from
conditions with a light intensity of 90 μmol m-2 40% and 25% interface, were collected and http
s-1 and with orbital shaking at 120 rpm. Cultures diluted in sucrose wash buffer (0.3 M sucrose, ://w
w
were maintained in 250 ml Erlenmeyer flasks 10 mM TES, pH 7.5). Diluted mitochondria w
by the inoculation of 20-25 ml of 7-d old cells were centrifuged at 24,000 g for 15 min and .jbc
.o
into 100 ml of new medium. The same growth decelerated with slow braking. Pellet was brg/
medium without nitrogen (i.e. no ammonium collected and diluted again in wash buffer. y g
u
nitrate or potassium nitrate) was used to wash Washed mitochondria were diluted in 0.5-1 ml es
t o
the 7-d old cell culture three times before it was of sucrose wash buffer. Isolated mitochondria n A
t1r5aNn snfietrrareted- 1t5oN 1 5(N1. 6g9ro gw/Lth) manedd ipuomta.s Asimumm nointiruamte-- wweerree thqeuna nutsifeide df owr iimthp othrte a sBsaraydsf, oorrd m aasdsea yin taon d1 pril 11
, 2
15N (1.92 g/L) (98% 15N Sigma) were added to mg protein aliquots and run on BN gels or 01
9
growth medium without nitrogen to make the stored at -80 o C for further experiments.
15N growth medium. Cells labeled for 24, 120

and 168 h with 15N were collected onto filter
paper using vacuum filtration. Three biological
Arabidopsis plant growth and mitochondrial
replicates for each time point were performed.
preparation

Wild type Arabidopsis (Col-0) or ca2 mutant
seeds (～30 μg) (11) were washed in 70% (v/v)
Protoplast preparation and mitochondrial
ethanol for 2 min and then the seeds placed in a
isolation from cell culture
sterilization solution (5% [v/v] bleach, 0.1%
Fresh cells (20-25 g) for each single time point [v/v] Tween-20) for 15 min with periodical
replicate were dispersed in digestion solution shaking. Seeds were then washed in sterile
(0.4% [w/v] Onozuka cellulase RS, 0.05% [w/v] water 5 times before being dispensed into a 250
Kyowa pectolyase Y-23, 0.4 M mannitol, MES ml plastic vessel containing 80 ml of growth
0.7 g/L, pH 5.7) and gently shaking in the dark medium (1/2 strength Murashige and Skoog
for 3 hours at room temperature. The solution medium without vitamins, 1⁄2 strength Gamborg
3
B5 vitamin solution, 5 mM MES, 2.5% [w/v] and subjected to in-gel digestion with trypsin
sucrose, pH 5.8). Arabidopsis seedlings were using previously described methods (28), but
grown under a 16/8-h light/dark period with using only a quarter of the trypsin concentration
light intensity of 100–125 μmol m-2 s-1 at 22 °C in order to decrease the effect of trypsin
for 2 weeks. Ten vessels of wild type or ca2 autolysis peaks on MS data interpretation.
mutant seedlings were collected for Peptides were analyzed with an UltraFlex III
mitochondrial preparations. Seedlings (50-100 MALDI-TOF/TOF mass spectrometer (Bruker
g) were homogenized with mortar and pestle Daltonics). In-gel digested peptides were
and mitochondria isolated as previously reconstituted with 5 μL of a solution containing
described (25). The protein content of freshly 5% acetonitrile (ACN v/v), 0.1% trifluoroacetic
isolated mitochondria was quantified with a acid (TFA v/v) in ddH O. Two-μL of each
2
Bradford assay protocol before being used for sample was spotted onto a MTP 384 MALDI
import experiments. target plate and mixed with 2 μL of spotting
matrix (90% ACN, 10% saturated HCCA [α-

Cyano-4-hydroxycinnamic acid] in TA90 [90%
ACN,0.01% TFA]). Dried spots were overlaid
Two dimensional blue-native(BN)/SDS
with 10 μL cold washing buffer (10 mM
PAGE
NH H PO , 0.1% TFA) and allowed to stand for
4 2 4 D
Mitochondrial aliquots were washed with 2 ml 10 sec before removal by pipette. Spots were ow
n
sucrose wash buffer (0.3 M sucrose, 10 mM analyzed at 50-85% laser intensity with up to lo
a
d
TES, pH 7.5) and then centrifuged at 14,300 g 1200 shots for MS analysis per spot. Ions ed
for 10 min. Pellets were dissolved in 100 μL of between 700-4000 m/z were selected for from
digitonin solubilization buffer (30 mM HEPES MS/MS experiments using 3% additional laser http
pH 7.4, 150 mM potassium acetate, 10% [v/v] power. Masses corresponding to trypsin ://w
w
glycerol, digitonin 50 mg/ml). Samples were autolysis products were excluded from analysis. w
centrifuged at 18,000g for 30 min after A maximum of 15 molecular ions were .jbc
.o
incubation for 20 min on ice to remove analysed by MS/MS from each gel spot. brg/
insoluble material and were subsequently Tandem mass spectrometry data were analyzed y g
u
supplemented with 5 μL of Serva Blue G250 using Biotools (Bruker Daltonics) and an in- es
t o
(5% [w/v] Coomassie Blue in 750 mM house Arabidopsis database comprising n A
apmroitneoinc asparmoipcl esa cwide)r.e dCiroeocmtlyas sloiea debdl uoen-ttroe aBteNd AInTfoHrm1.apteipo n (relReaesseo u9rc) e from(T ATIhRe) Araanbdid optshies pril 11
, 2
gels. Two dimensional BN/SDS PAGE was Arabidopsis mitochondrial and plastid protein 01
9
carried out as described previously (26). sets (33621 sequences; 13487170 residues),
Gradient gels (4.5%-16% [w/v] acrylamide) using the Mascot search engine version 2.3.02
were used for the first BN dimension. Second and utilizing error tolerances of ±1.2 Da for MS
dimension Tricine-SDS PAGE gels, comprised and ±0.6 Da for MS/MS; “Max Missed
of a 4% stacking gel and a 12% separation gel, Cleavages” set to 1; variable modifications of
were used for separation of protein complex oxidation (Met) and carbamidomethyl (Cys).
subunits. Electrophoresis parameters and Only protein matches with more than 2 peptides
Coomassie staining of BN and BN/SDS-PAGE and with ion scores greater than 40 were used
were carried out as previously described (27). for analysis (p<0.05). For all spots, the 0 and 24
h time points were used for both MS and
MS/MS analyses for protein IDs while only the
MS data were analysed for 120 and 168 h.
In gel digestion and MALDI TOF/TOF
Search results were exported to CSV format for
analysis
each spot.
Protein spots in CI+CIII , CI and CI
2
subcomplex lanes from BN gels were excised
4
Quantification of protein turnover as used imports of the epsilon subunit of ATP
NA/(NA+H) and protein degradation as K synthase (At1g51650), which does not readily
D
and protein half-life. get incorporated into native complexes in plant
mitochondria, as a control for radiolabelled
In order to calculate heavy label incorporation
patterns in the isolated mitochondria. Equivalent
and the relative abundance of NA (natural
quantities of mitochondria isolated from cell
abundance) and H (heavy labeled) peptide
culture, wild type and ca2 knockout
populations, MALDI-TOF/TOF MS data were
Arabidopsis were used in import reactions as
exported to mzXML files (CompassXport 3.0
described previously (30).
(Bruker)), and were subsequently converted to
text files using the ProteomeCommons.org IO
Framework 6.21

(http://proteomecommons.org/current/531/).
Text files were then parsed using a script

written in Mathematica 7.0 (Wolfram Research)
for further analysis. Isotopic envelopes were RESULTS
fitted using a two-population model consisting
of natural abundance and 15N-labeled fractions Higher turnover rates of specific subunits in
with the open source program Isodist the 200- and 650-kDa CI subcomplexes. D
o
w
(http://williamson.scripps.edu/isodist/) (29) as n
lo
Arabidopsis cells were grown for 24, 120 and a
we have previously described (24). d
168 h in 98% [15N]-enriched MS medium (24), ed fro
Protein degradation rates (K ) were calculated mitochondrial OXPHOS complexes prepared m
D h
as previously described (24).The total protein and separated by BN-PAGE, and the major ttp
abundance in the cell culture at different protein spots identified by mass spectrometry. ://w
w
timepoints was quantified using Image J of These data were consistent with the reported w.jb
stained protein gels and by Amido black protein complex proteome of plant c.o
rg
analysis as described previously (24). An mitochondria ( 23). This confirmed several b/
y
average KD among subunits within CI+III2, CI, reported CI subunit-containing protein gue
650 and 200 kDa subcomplexes was separately complexes including; CI+CIII2 and the mature st o
n
determined at 24, 120 and 168 h. Half-lives of CI holoenzyme (10,21), (Figure 1). Peptide A
p
CI+III2, CI, 650 and 200 kDa subcomplexes at mass spectrometry analysis revealed 22 ril 1
1
24, 120 and 168 h were determined from KD as different CI-specific subunits in the CI (ND1, , 20
half-life=ln2/K . The final K values and half- ND7, ND9, 75 kDa, 51 kDa, 39 kDa, 20.9 kDa, 19
D D
lives for CI+III , CI, 650 and 200 kDa 18 kDa, ASHI, B14, B13, B14.7, B16.6, B17.2,
2
subcomplexes were determined by averaging CA2, CA3, CAL1, NDU10, NDU12, PDSW,
the values obtained at 24, 120 and 168 h. PSST, TYKY) and 23 different subunits in the
CI+CIII (ND1, ND7, ND9, 75 kDa, 51 kDa, 39
2
kDa, 20.9 kDa, 18 kDa, ASHI, B14, B13, B14.7,
B16.6, B17.2, CA2, CA3, CAL1, NDU10,
Import in cell culture and plant
NDU11, NDU12, PDSW, PSST, TYKY). All of
mitochondria
these subunits are confirmed subunits in the
plant CI holoenzyme (3,10,23,31). Interestingly,
The full cDNAs of A. thaliana CA1
the CI subunit CA2 was also found in spot 37
(At1g19580), CA2 (At1g472660), CA3
which was a visible stained region in our BN-
(At5g66510), 20.9-(At4g16450), CAL1
gels (Figure 1). CI subunits have previously
(At5g63510), CAL2 (At3g48680) and ATPE
been observed at 650-kDa (22), at 400-450 kDa
(At1g51650) were cloned into pDest14
(3,22,23) and at 200-kDa (3,22,23) in
(Invitrogen) using gateway-cloning techniques
Arabidopsis mitochondrial BN-gel analyses.
(Invitrogen) for in vitro transcription and
Analysis of gel regions at these native
translation as described previously (30). We
5
molecular masses, and at masses of the complex averaged K value for the 200 kDa subcomplex
D
I subunits observed in the CI holoenzyme and (0.42±0.04 d-1) was about 10 times faster than
CI+CIII supercomplex, was then carried out. for CI+III (0.04±0.01d-1) and CI (0.04±0.02 d-1)
2 2
This revealed that CI subunits CAL1 and 20.9- and about 1.6 times faster than for the 650 kDa
kDa could be found along with CA2 at 650 kDa subcomplex (0.26±0.07d-1). The averaged
and 200 kDa (Figure 1, Dataset S1). A wider protein half-life for the 200 kDa subcomplex
range of protein bands were analysed, but only (1.7±0.3 d) was thus about 8% of that of CI+III
2
those including significant identifications of CI (22.0±5.7d) and CI (23.9±11.9 d) and 60% of
subunits are annotated in Figure 1. that of the 650 kDa subcomplex (3.0±0.7d). The
lower NA/(NA+H) values of subunits in
High-scoring peptides by the Mascot algorithm
subcomplexes are thus due to their significant in
(p<0.05) where further analyzed to determine
vivo turnover during the course of the seven
what proportion of each peptide in mass spectra
days of the experiment, rather than just a
was derived from the natural abundance (NA)
process of dilution through growth.
of N in the original cell culture, and what
proportion derived from the incorporation of Matrix-arm subunits have higher turnover
15N (H) into new protein synthesised after the rates than membrane-associated and
growth medium switch. These values (as mitochondrion-encoded CI subunits
outlined in Materials and Methods and in (24)) D
o
To investigate whether subunits have different w
were used to calculate the NA/(NA+H) ratio for n
lo
turnover rates compared to each other within CI a
specific peptides from each gel spot. d
e
d
Calculations of NA/(NA+H) for all the peptides and CI+CIII2, the NA/(NA+H) ratios for a fro
variety of CI subunits after 120 h and 168 h 15N m
of a given protein were averaged to determine a h
final NA/(NA+H) ratio ± SE. This value was laarbme lsiunbgu wnietrse i nccolmudpianrge d7 5(F kigDuar,e 5 31A k-DDa),. 3M9a ktrDixa-, ttp://w
used as a proxy for each protein’s turnover rate w
w
(Figure 2; Dataset S2). Subunits, including CA2, 18 kDa and B17.2 exhibited relatively low .jb
c
CAL1 and 20.9-kDa, had a significantly lower NA/(NA+H) ratios and thus relatively higher .org
NA/(NA+H) ratio when the peptides were turnover rates than other subunits in both CI and by/
derived from the 200-kDa and 650-kDa CI+CIII2. The specific matrix-exposed arm gue
s
subcomplexes than when the same peptides for components including CA2, CAL1 and 20.9 t o
n
kDa and mitochondrion-encoded subunits A
the same proteins were derived from CI and p
CI+CIII . These results indicate a significantly including ND1, ND7 and ND9 had relatively ril 1
2 1
higher turnover rate for the 200- and 650-kDa high NA/(NA+H) ratios, and thus lower , 2
0
subcomplexes (p<0.05) than for CI and turnover rates. The plant 20.9-kDa subunit, 19
which has been found in both matrix and
CI+CIII . These differences were clearly
2
apparent after just 24 h of 15N labeling and were membrane modules of CI (3,22), does not have
a direct homolog in mammals, but is related to
very evident after 120 h of labeling (Figure 2).
fungal NUXM subunits and the 13-kDa and 21-
Notably, the differences between CI and
kDa CI subunits in other eukaryotes, and was
CI+CIII were not statistically significant for
2
the most stable protein subunit studied in both
any of the three subunits examined (Figure 2;
CI and CI+CIII .
Dataset S2). 2
To determine if these observations of varying
The turnover rates indicated by NA/(NA+H)
15N incorporation rates for different modules of
values do not define the specific contribution of
synthesis of 15N protein and the degradation of the complex were statistically significant, CI
pre-existing 14N proteins to turnover. However, subunits were divided into four groups as
detailed in Figure 3. One-way ANOVA analysis
data on the growth rate of the cell culture
and Tukey Post Hoc multiple comparisons
allowed us to subsequently calculate the
among the four groups revealed that the matrix
degradation rate (K ) of CI and subcomplexes
D group had a significantly higher 15N
and convert this to protein half-life. The
incorporation rate than every other group
6
(P<0.01). The membrane/matrix interaction accumulate in Figure 1; however, they were
group also had a significantly higher 15N similar in size to the multiple subcomplexes
incorporation rate than the specific γ-CA- containing γ-CAs found by CI disassembly (3).
domain group (P=0.04 in CI, P<0.01 in I+III ). The γ-CAL and 20.9-kDa proteins differed with
2
(Figure 3 E,F; Dataset S3). Notably, there was respect to the subcomplexes into which they
no difference between the incorporation rate of progressed, compared to the γ-CA proteins.
15N into the CI holoenzyme and the CI+CIII Overall, these results demonstrated that the CA-
2
supercomplex, indicating the two appear to be domain subunits were not simply added as
in equilibrium. accessory components to fully or near-fully
assembled CI, but to different degrees they were
γ-CAs, γ-CALs and 20.9-kDa subunit
present in the full range of subcomplexes that
precursor proteins are imported and
are known to include components of the
assembled into a variety of subcomplexes in
membrane arm of CI (3,22). Import of the
mitochondria
epsilon subunit of ATP synthase which fails to
assemble into a protein complex was used as a
To determine whether γ-CA-domain
radiolabeled control import for comparison of
subcomplexes are produced rapidly after import,
labeling on gel images.
we performed in vitro import assays with
[35S]Met-labeled CA1, CA2, CA3, CAL1, CA2 is imported and assembled into a D
o
CAL2 and 20.9-kDa subunit precursors into w
transient 200-kDa subcomplex in ca2 mutant n
lo
isolated Arabidopsis cell culture mitochondria. a
Arabidopsis mitochondria d
e
d
Analysis of autoradiographs of SDS-PAGE- fro
separated samples showed that all subunits were To further understand the nature of the γ-CA- m
h
imported into mitochondria during 60-min domain subcomplexes in imports, time-series ttp
incubation period, and that imported products imports were compared between mitochondria ://w
w
w
were resistant to proteinase K (PK) digestion isolated from wild-type and ca2 knockout plants .jb
c
(Figure 4A). The CAL1 subunit was clearly that have been previously reported by others (11) .o
rg
processed on import, leading to a ~2 kDa and have been propagated in our laboratory. b/
y
decrease in its apparent molecular mass, and Analysis in the mutant had the potential gu
e
only the processed forms were PK-protected. advantage that [35S]Met-labeled precursors did st o
n
The other subunits assayed were not clearly not have to compete with unlabeled ones, and A
p
processed leading to changes in MW, but initial assembly processes and assembly factors ril 1
1
membrane-bound proteins were PK-protected might be more abundant in ca2 knockouts , 2
0
after 60 min of the import reaction (Figure 4A). where the abundance of fully-assembled CI is 19
Separation of PK-protected proteins by BN- reduced (11). A time series of 5, 10, 20, 30, and
PAGE revealed that CA3, CAL1-2 and 20.9 60 min of CA2 precursor import was conducted
kDa could be clearly assembled into 200- to and the samples separated by SDS-PAGE and
300-kDa subcomplexes and the 650-kDa BN-PAGE. More CA2 precursor was imported
subcomplex. CA1-3 were assembled to differing and PK-protected in mitochondria isolated from
degrees into CI and CI+CIII (Figure 4B). This Arabidopsis ca2 knockouts than in the
2
efficient assembly of labeled subunits might equivalent wild-type mitochondria (Figure 5A).
explain why the smaller subcomplexes are less On BN gels it was clear that CA2 precursors
defined in CA1 and CA2. Comparison of the were rapidly assembled into a discrete 200-kDa
radiolabeled import gels with Coomassie- subcomplex in mitochondria from the ca2
stained gels showed that the 200- to 300-kDa knockout (Figure 5B). In comparison,
complexes (Figure 4B *) seen in the import mitochondria from wild type showed assembly
experiments were not exactly the same size, of CA2 into the same 200- to 300-kDa range of
when compared to the abundant 200-kDa complexes observed during imports into cell
subcomplex comprising CA2, CAL-1 and 20.9 culture mitochondria (Figure 4B v Figure 5B).
kDa subunits that had been observed to While the [35S]Met incorporation into CI
7
steadily increased in the ca2 knockout during and Trypanosoma brucei (36,37) (Table 1).
the import time course, the abundance of the Evidently, the subunits of the subcomplex
200-kDa band peaked at 20 min and declined described here in Arabidopsis are widely
thereafter, indicating that it was further conserved and coincident in CI across
assembled into intact CI and was thus a eukaryotes.
productive intermediate of assembly (Figure

5B). CI+CIII could not be visualized until 20
2

min and then it exhibited the same steadily
increasing pattern of [35S]Met incorporation DISCUSSION

seen for CI. A similar but weaker pattern was
CI matrix-arm and mitochondrion-encoded
observed in wild-type mitochondria for
subunits have different turnover rates
progressive labeling from subcomplexes to CI
The assembly of the largely nucleus-encoded
and CI+CIII (Figure 5B).
2
matrix arms and the set of mitochondrion-

encoded proteins that are mainly in the
Conservation and losses of 20.9, γ-CA and γ-
membrane arm of CI occur via independent
CAL proteins among eukaryotic lineages.
processes in other organisms (13-17,19,20,38);
While γ-CAs were initially considered plant-
thus, it is conceivable that these subcomplexes
specific subunits of CI (6,10,11), more recent
might turn over at different rates in plants. Our D
work has demonstrated the widespread o
w
data show that the matrix arm (including n
occurrence of γ-CA homologs throughout lo
a
members of both the N-module [51 kDa and 75 d
eukaryotes, with the notable exception of ed
opisthokonts (animals + fungi) (12). Extending kDa] and the Q-module [39 kDa] as defined for fro
m
tihni sw ahniaclhy sidse, ewp e gheanvoem feo uanndd /tohra t tirna nosrcgraipntiosmmes tthuern omvaemr mraatlei anth acno mepitlhexer (3th9e) ) mhiatvoec hao nhdirgiohner- http://w
encoded subunits or the γ-CA-domain w
sequencing has been performed, multiple γ-CA w
genes are often found (Table 1 and Dataset S4). components of the membrane arm. Elevated .jbc
Where the predicted protein sequences are protein turnover can arise from high synthesis .org
deemed to be complete at the N-terminus, many or high degradation rates or a combination of by/
g
both processes (24). As matrix-arm subunits u
of these γ-CAs have a predicted mitochondrial es
form the major ROS production sites in t o
localization (Table 1). A bioinformatic survey n
A
t2o0 .9a-skseDsas tphreo tepihny lsoegqeuneentcice dsihsotrwibsu taio nsi moifl atrhlye rdeasmpiargaetodr ym ocrhea irne,a dtihlye saen dc ormeqpuoinree nrtesp lamcaeym ebnet pril 11
more frequently. ROS-mediated protein , 2
wide occurrence of the corresponding gene 0
1
oxidation can sensitize proteins to proteolytic 9
across the various eukaryotic supergroups
attack (40,41). It is possible the ROS released
(Table 1). However, in contrast to what is seen
from CI could lead to increased oxidation of
with animals, the 20.9-kDa subunit is very
subunits in the vicinity, or that ROS damage in
widely present in fungi (where it has a known
the matrix is generally greater in magnitude
role in CI assembly (18)). Outside of
than in the membrane (41). The matrix-arm
opisthokonts, γ-CAs and the 20.9-kDa subunit
subunits are also likely to be more accessible to
exhibit a strikingly similar pattern of co-
matrix proteases than the mainly membrane
occurrence (Table 1). Aside from Arabidopsis,
subunits. It is also possible that the natural
direct evidence for the presence of both γ-CAs
breakdown of intact CI simply begins with the
and the 20.9-kDa subunit in mitochondrial CI is
degradation of matrix-arm subunits, irrespective
available for the green alga, Chlamydomonas
of damage or accessibility. Whichever the
reinhardtii (32), as well as A. castellanii (33).
reason, the higher turnover rate of matrix-arm
Various of these components have also been
subunits than of both mitochondrion-encoded
identified in proteomic screens of purified
and the specific γ-CA-domain subunits
mitochondria from additional non-
demonstrates that different functional modules
photosynthetic protists, including Tetrahymena
thermophila (34), Dictyostelium discoideum (35)
8
of CI have distinct protein half-lives in plant stage of CI assembly (22), and γ-CA-containing
cells. subcomplexes have been detected in BN-PAGE
gels of wild-type mitochondria (23). Using a
15N labeling approach, we have shown the
Supercomplex formation does not higher 15N incorporation rate of subunits
significantly stabilise Arabidopsis complex I
contained in subcomplexes than in CI and
It has previously been proposed that formation
CI+CIII revealing that these structures were
2,
of supercomplex I+III might stabilize complex
2 not created in vitro by solubilisation or
I in Arabidopsis (42). Lowering protein
electrophoresis artefacts, but rather are in vivo
turnover by supercomplex formation could
intermediates of disassembly or assembly of CI.
provide a functional significance to these super-

assemblies in the electron transport chain,
The CI subcomplexes could conceivably be
beyond their role in electron transport catalysis.
disassembly intermediates that have a high
This proposal is supported by evidence from degradation rate. However, the higher 15N
mouse that knockout of complex III (43) or
incorporation in subcomplexes as compared to
complex IV (44) subunits decrease the content
CI and CI+CIII means this explanation only
2
of complex I, suggesting supercomplex
holds if dissociation of the complex is biased
formation stabilizes Complex I. It has also been
toward newly assembled CI or CI+CIII . It is
proposed in human mitochondria that 2 D
not clear what could cause such a bias. Also, o
w
supercomplex formation is required for n
our calculations based on Coomassie staining lo
a
assembly of Complex I as well as for its d
abundance together with the proportion of ed
stabilization (45). In addition, structural peptide 15N reveals that the amount of 15N- fro
m
sctoambiplilzeaxteiso ni s porfo polsaebdil ea s am emmabjorra nfeu nctpiroont eoinf ltahbee lteodta ls uabmuonuitnst woift h1i5nN sulabbceolm ipnl ebxoetsh eCxIc eaendds http://w
supercomplex formation in Paracoccus w
CI+CIII2 combined. Hence the balance of the w
denitrificans (46). As we pointed out in the data, from 15N incorporation alone, is in favour .jbc
results, however, we did not show any impact of of the 200-kDa subcomplexes found in wild .org
supercomplex formation on protein turnover type being assembly rather than disassembly by/
g
rate in Arabidopsis cells (Figure 3). Hence these u
intermediates. es
direct measurements are not able to confirm a t o
n
A
rsotaleb ilfiosar tisounp eorfc ocmopmlepxle xf oIr.m aRtaiothne ri,n ite nahpapnecaerds Tanhde 2im0.p9o-krtD eax spuebruimnietns tssh uoswinegd tγh-aCtA asll, oγ-f CthAeLses pril 11
that plant complex I is in equilibrium between , 2
subunits can be imported into cell culture 0
1
the Complex I and the supercomplex I+III 9
2 mitochondria and that they form a series of
states. Higher order complexes and
subcomplexes ranging in size from 100-300
respirasomes could not be detected in our
kDa, although the 200-kDa complex found in
analysis of Arabidopsis membranes, so we are
wild type did not predominate (Figure 1 vs
not able to determine if these might have Figure 4&5). Imported [35S]Met-labeled
enhanced stability, but this technique could be
precursors must compete with the
used in systems where larger supercomplexes
corresponding subunits present in vivo and
are evident and provide in vivo evidence of
could lessen the subcomplex signal or even
relative stability.
change the original assembly process, especially

when redundant subunits compete with subunits

having high sequence similarity. Mitochondria
A specific subcomplex of unknown function
from CI subunit mutants have proved to be a
proteins has a productive role in CI assembly
useful system for detection of assembly
A range of knock-out mutants for CI subunits
intermediates in mammals (14,19). We found
leading to the accumulation of certain
that more CA2 precursor could be imported into
intermediates has previously been exploited to
ca2 mutant Arabidopsis mitochondria and that a
proposed that CA2 is incorporated at the early
strong 200-kDa subcomplex band could be
9
detected which weakened after 20 min as the CI, present at the earliest stages of eukaryotic
radiolabel was transferred into CI and CI+CIII . cell evolution; notably, however, this co-
2
Thus γ-CAs and γ-CALs appear to assemble occurrence pattern is not preserved in
into a productive membrane 200-kDa opisthokonts (fungi + animals) (Table 1). Of
subcomplex which serves as the initiator of CI particular note is the finding that several protists
assembly. Our data provide key experimental included within the assemblage Holozoa
evidence to justify the model recently proposed (animals and their specific unicellular relatives;
in plants (22) that subunits (including at least (47)) are seen to encode either γ-CAs or the
ND1 and ND7 and matrix subunits 75 kDa and 20.9-kDa subunit, although (like fungi) not both
39 kDa) are added to a plant-specific 200-kDa (Table 1). Intriguingly, genes for both γ-CAs
subcomplex to form a 650-kDa subcomplex that and the 20.9-kDa subunit can also be retrieved
is further assembled into CI. from the recently determined genome sequence
of the apusomonad Thecamonas trehans,
Our data do not prove that CA, CAL and 20.9 arguably the closest neighbour to opisthokonts
kDa are the only subunits of this 200 kDa (48). These observations indicate that
subcomplex. Previous surveys of BN-PAGE progressive loss of components of this putative
have found evidence for At2g28430 and ND9 in assembly subcomplex began in opisthokonts at
this region on BN gels (23). While we were not the base of this lineage, after its divergence
D
able to confirm this in our data, the possibility from its closest related lineages, Apusozoa and o
w
n
of other CI subunits in this 200 kDa complex Amoebozoa. In Holozoa, loss of all the lo
a
d
cannot be excluded and is likely given its mass. subunits known to be in the subcomplex appears ed
However the low abundance of the complex has to have been complete before the emergence of fro
m
not allowed us to date to make any progress in multicellular animals (Metazoa). Outside of http
finding other rapid turnover subunits in this eukaryotes, γ-CAs are prominent among α- ://w
w
mass range. Proteobacteria and the γ-CAs of these w
organisms are the most similar in sequence to .jbc
.o
the eukaryotic γ-CAs, although in this case they rg
Conservation of the subunits of 200-kDa do not appear to be associated with CI (49). by/
g
subcomplex suggests an ancestral feature of Considering the α-proteobacterial ancestry of ues
CI important for CI assembly t o
mitochondria (50), it is possible (perhaps n
A
The patterns of conservation of genes for γ-CAs pevroobluatbiloen)a rtyh ast outhrcise boaf ctmeriitaolc hgornodurpia lw γa-sC Athse, pril 11
and the 20.9 subunit and the evidence for their , 2
which were then recruited to serve in CI 0
conserved presence in CI (Table 1) argue that 19
assembly early in eukaryotic cell evolution.
both protein types are ancestral components of

REFERENCES

1. Walker, J. E., Skehel, J. M., and Buchanan, S. K. (1995) Structural analysis of NADH:
ubiquinone oxidoreductase from bovine heart mitochondria. Meth Enzymol 260, 14-34
2. Videira, A., and Duarte, M. (2001) On complex I and other NADH:ubiquinone reductases of
Neurospora crassa mitochondria. J Bioenerg Biomembr 33, 197-203
3. Klodmann, J., Sunderhaus, S., Nimtz, M., Jansch, L., and Braun, H. P. (2010) Internal
architecture of mitochondrial complex I from Arabidopsis thaliana. Plant Cell 22, 797-810
4. Cardol, P. (2011) Mitochondrial NADH:ubiquinone oxidoreductase (complex I) in eukaryotes:
a highly conserved subunit composition highlighted by mining of protein databases. Biochim
Biophys Acta 1807, 1390-1397
5. Sunderhaus, S., Dudkina, N. V., Jansch, L., Klodmann, J., Heinemeyer, J., Perales, M.,
Zabaleta, E., Boekema, E. J., and Braun, H. P. (2006) Carbonic anhydrase subunits form a
10