We also found that Cox6a and Cox7a are incorporated into a novel intermediate complex of approximately 250 kDa, and that transition of subunits from this complex to the mature holoenzyme
Trang 1mitochondrial complex IV, and their preferential
integration into supercomplex forms in patient
mitochondria
Michael Lazarou1, Stacey M Smith1,2, David R Thorburn2, Michael T Ryan1and
Matthew McKenzie1
1 Department of Biochemistry, La Trobe University, Melbourne, Australia
2 Murdoch Children’s Research Institute and Genetic Health Services Victoria, Royal Children’s Hospital, and Department of Pediatrics, University of Melbourne, Australia
Introduction
The mitochondrial respiratory chain consists of four
multi-subunit complexes (I–IV) that, through
electron-transfer reactions, generate a proton gradient for the
F1Fo-ATPase (complex V) to synthesize ATP
Cyto-chrome c oxidase (complex IV) catalyzes the final step
of the electron transfer chain, in which electrons are
transferred from reduced cytochrome c to molecular oxygen Mammalian complex IV comprises 13 differ-ent subunits, and the crystal structure of the bovine complex was solved over 10 years ago [1] The three largest and most hydrophobic subunits (CO1–3) form the catalytic core, and all are encoded by
mitochon-Keywords
complex IV; cytochrome c oxidase;
membrane proteins; mitochondria; oxidative
phosphorylation
Correspondence
M T Ryan, Department of Biochemistry,
La Trobe University, 3086 Melbourne,
Australia
Fax: +61 3 94792467
Tel: +61 3 94792156
E-mail: m.ryan@latrobe.edu.au
(Received 4 June 2009, revised 18 August
2009, accepted 16 September 2009)
doi:10.1111/j.1742-4658.2009.07384.x
Complex IV is the terminal enzyme of the mitochondrial respiratory chain
In humans, biogenesis of complex IV involves the coordinated assembly of
13 subunits encoded by both mitochondrial and nuclear genomes The early stages of complex IV assembly involving mitochondrial DNA-encoded subunits CO1 and CO2 have been well studied However, the latter stages, during which many of the nuclear DNA-encoded subunits are incorporated, are less well understood Using in vitro import and assembly assays, we found that subunits Cox6a, Cox6b and Cox7a assembled into pre-existing complex IV, while Cox4-1 and Cox6c subunits assembled into subcomplexes that may represent rate-limiting intermediates We also found that Cox6a and Cox7a are incorporated into a novel intermediate complex of approximately 250 kDa, and that transition of subunits from this complex to the mature holoenzyme had stalled in the mitochondria of patients with isolated complex IV deficiency A number of complex IV subunits were also found to integrate into supercomplexes containing combinations of complex I, dimeric complex III and complex IV Subunit assembly into these supercomplexes was also observed in mitochondria of patients in whom monomeric complex IV was selectively reduced We con-clude that newly imported nuclear DNA-encoded subunits can integrate into the complex IV holoenzyme and supercomplex forms by associating with pre-existing subunits and intermediate assembly complexes
Abbreviations
BN-PAGE, blue-native polyacrylamide gel electrophoresis; CAP, chloramphenicol; DDM, n-dodecyl-b- D -maltoside; Dw m, membrane potential, LSI, late-stage intermediate.
Trang 2drial DNA (mtDNA) The remaining 10 subunits are
encoded by the nuclear genome, and, like other
pro-teins, must be synthesized on cytosolic ribosomes
before being imported and subsequently assembled at
the mitochondrial inner membrane [2] The subunits of
complex IV encoded by nuclear DNA (nDNA) do not
harbor enzymatic activity but function in the structural
integrity, regulation and dimerization of the enzyme
For example, subunits Cox6a and Cox5 play roles in
the regulation of enzymatic activity [3,4], while Cox6b,
along with Cox6a, provides contacts sites for
dimeriza-tion [1,5]
Mammalian complex IV forms a complex of
approx-imately 200 kDa on blue native (BN) PAGE, but is
also found in supercomplexes together with complex I
and dimeric complex III [6] More recently, it has been
suggested that this supercomplex can associate with
complex II, cytochrome c and complex V in a
func-tional respirasome [7] It has been suggested that
supercomplexes enhance respiration due to coordinated
channeling of the electron carriers ubiquinol and
cyto-chrome c [6,8,9]
Isolated complex IV deficiency is one of the most
common respiratory chain defects in humans, and is
associated with various clinical phenotypes such as
mitochondrial encephalomyopathy, lactic acidosis and
stroke-like episodes (MELAS), Leigh disease and lactic
acidosis [10,11] Enzymatic deficiencies of complex IV
that cause disease are often due to defects in the
assembly and⁄ or stability of the enzyme To assemble
complex IV, subunits translated from both genomes
must come together in a coordinated and regulated
manner Studies carried out in the model organism
Saccharomyces cerevisiae have provided insights into
the biogenesis of complex IV, and led to the
identifica-tion of over 20 assembly factors [12] These
nDNA-encoded assembly factors are not associated with
assembled complex IV but instead act at various
func-tional levels, ranging from subunit insertion and
co-factor attachment to regulation of transcription⁄
translation Although yeast COX is a close model of
its human counterpart, significant differences exist For
example, almost half of the yeast assembly factors do
not appear to have human orthologs, and two subunits
(Cox7b and Cox8) are found in the human assembly
but not in yeast In addition, as yeast lacks complex I,
the supercomplex assemblies differ substantially from
those seen in mammalian mitochondria By tracking
the subunit composition of subcomplexes using
trans-lational inhibitors and metabolic labeling of
mtDNA-encoded proteins, a model for the assembly of human
complex IV was proposed [13] This model was later
refined through studies analyzing the composition of
sub-assemblies in the mitochondria of patients deficient
in complex IV [14,15] The model proposes that com-plex IV is assembled via pre-formed intermediates in a stepwise process Of the 10 nDNA-encoded subunits, only Cox4 and Cox5a are integrated at the early stages
of assembly that involve formation of the catalytic core The remaining nDNA-encoded subunits are thought to be added during the last steps of assembly; however, this part of the assembly pathway is the least studied and requires clarification Furthermore, with the model focusing on the de novo synthesis of com-plex IV, it is unclear how newly imported nDNA-encoded subunits are incorporated in the presence of pre-existing holo-complex IV
In this study, we address the assembly of newly imported complex IV subunits in mitochondria from control cells and from cells of patients with defects in complex IV biogenesis We found that, in the presence
of pre-existing complex IV, newly imported nDNA-encoded subunits can integrate into the holoenzyme as well as into its supercomplex forms Furthermore, sub-units Cox6a and Cox7a integrate into a novel late-stage intermediate complex of approximately 250 kDa Assembly into, and progression from, this intermediate was defective in the mitochondria of patients deficient
in complex IV, suggesting that it represents an impor-tant step in assembly of the holoenzyme
Results
In vitro import and assembly of nDNA-encoded complex IV subunits
The assembly of a number of nDNA-encoded com-plex IV subunits was investigated by importing them into isolated mitochondria and monitoring their assem-bly using BN-PAGE As isolated mitochondria are used and protein synthesis does not take place under the conditions used (data not shown), the integration
of a select newly imported subunit into a complex occurs through its association with pre-existing units within the organelle [16,17] Representative sub-units with cleavable (Cox4-1, Cox6a, Cox7a) and non-cleavable (Cox6b and Cox6c) presequences were selected for investigation With the exception of Cox4-1, these nDNA-encoded complex IV subunits are postulated to integrate late in the assembly pathway [13,15] Based on the crystal structure of complex IV [1], the selected subunits are positioned peripherally within the complex (Fig 1A), and, apart from Cox6b, all contain a single transmembrane-spanning domain
35S-labeled complex IV subunit precursor proteins were generated in vitro using rabbit reticulocyte lysate,
Trang 3and incubated with mitochondria isolated from
cul-tured human fibroblasts for 10 or 60 min in the
pres-ence or abspres-ence of a membrane potential (Dwm) After
import, external proteinase K was added to half of
each sample to degrade non-imported protein Samples
were then subjected to SDS–PAGE, and radiolabeled
subunits were detected using phosphorimage analysis
(Fig 1B) 35S-labeled complex IV subunits bound to
mitochondria, with the signal increasing over time
(Fig 1B, lanes 2 and 3) For the
presequence-contain-ing subunits Cox4-1, Cox6a and Cox7a, an additional
faster-migrating species accumulated, representing the
mature form of the protein (Fig 1B, lanes 2 and 3)
An additional band seen after proteinase K treatment
of Cox7a samples most likely represents a
protease-resistant domain of the precursor, as it is also present
in the absence of import (lane 7) Successful import of
all subunits was determined by their protection from
externally added proteinase K (Fig 1B, lanes 5 and 6),
their dependence on the membrane potential (Dwm) for
this protection (Fig 1B, lanes 4 and 7), and, in the
case of Cox4-1, Cox6a and Cox7a, processing of their
presequences
We next tested the assembly of newly imported
radiolabeled complex IV subunits using BN-PAGE
Radiolabeled subunits were imported into isolated mitochondria and all samples were treated with pro-teinase K before solubilization in n-dodecyl-b-d-malto-side (DDM) and BN-PAGE analysis (Fig 2A) Use of this detergent results in partial dissociation of respira-tory chain supercomplexes, liberating monomeric (holo-) complex IV and a complex III2⁄ complex IV supercomplex [18] The migration of complex IV, dimeric complex III, complex I and their super-complexes (CIII2⁄ CIV and CI ⁄ CIII2) was determined
by western blot analysis (Fig 2A, right panels) Imported Cox6a and Cox7a were incorporated into both monomeric and supercomplex (CIII2⁄ CIV) forms
of complex IV, with the additional presence of an approximately 250 kDa complex (marked with an asterisk) that resolved after 10 min of import (Fig 2A, lanes 3, 4, 9 and 10) Radiolabeled Cox6b also appeared to be incorporated into the holoenzyme, albeit weakly (Fig 2A, lanes 5 and 6), while
35S-labeled Cox4-1 was not incorporated into any distinct complexes, although some high-molecular-weight smearing was evident (lanes 1 and 2) Two dis-tinct complexes ranging between approximately 100 and 150 kDa were seen with newly imported Cox6c (Fig 2A, lanes 7 and 8), although neither of these
A
B
Fig 1 Import of nDNA-encoded complex IV
subunits (A) Structural position of subunits
Cox4 (magenta), Cox6b (red), Cox6c (green),
Cox7a (blue) and Cox6a (orange) within the
crystal structure of bovine complex IV [1].
(B) SDS–PAGE analysis of imported
radiola-beled complex IV subunits Precursor (p)
and mature (m) forms of the subunits are
identified Samples were imported into
mito-chondria in the presence or absence of a
membrane potential (Dwm), and treated with
or without externally added proteinase K
(Prot K) A sample of lysate (representing
20% of added protein ⁄ import) is also shown
(lane 1) Radiolabeled proteins were
detected by phosphorimage analysis.
Trang 4complexes co-migrated with holo-complex IV Based
on the BN-PAGE analysis of nDNA-encoded
com-plex IV subunits, it can be concluded that some newly
imported subunits can assemble with pre-existing
subunits into holo-complex IV Other subunits do not
integrate into complex IV, perhaps because there is
impaired progression of intermediates along the
assem-bly pathway due to the use of isolated mitochondria
Our results are consistent with those of a previous
study that analyzed the assembly of yeast complex IV
subunits Cox4p, Cox5ap and Cox10p (equivalent to
human Cox5b, Cox4-1 and Cox6a, respectively) [19]
In order to characterize the assembly of
nDNA-encoded complex IV subunits into the various
super-complex forms of super-complex IV, mitochondria were sol-ubilized in digitonin after subunit import In this case, the respiratory chain components are also found in their supercomplex forms [18] The relative positions
of supercomplexes comprising complex I, complex III and complex IV (CI⁄ CIII2⁄ CIV), complex III and complex IV (CIII2⁄ CIV), as well as dimeric com-plex III (CIII2) and monomeric complex IV (CIV), were shown by western blot analysis (Fig 2B, right panels) Radiolabeled Cox6c was incorporated into a complex of approximately 200 kDa that resolved at a slightly lower position than the holo-complex IV, and also into a larger complex at approximately 1 MDa Likewise, 35S-labeled Cox4-1 was found in large
A
CIV
*
CI 669
440
kDa
Cox4-1Cox6aCox6b Cox6c Cox7a
BN-PAGE 0.65% DDM
Time (min)10 60 10 60 10 60 10 60 10 60
α-comple
x I
α-comple
x III
α-comple
x IV
134
67
B
669
440
kDa
Cox4-1Cox6aCox6b Cox6c Cox7a
CIV
BN-PAGE 1% Digitonin
Time (min)10 60 10 60 10 60 10 60 10 60
CIII
α-comple
x I
α-comple
x III
α-comple
x IV
134
67
*
Fig 2 BN-PAGE analysis of imported radio-labeled complex IV subunits 35 S-labeled complex IV subunits were individually incu-bated with isolated fibroblast mitochondria for increasing times as indicated Samples were treated with proteinase K, and solubi-lized in either (A) DDM-containing buffer or (B) digitonin-containing buffer Radiolabeled proteins were detected by phosphorimage analysis Right panels: complex IV (CIV), complex I (CI), complex III (CIII 2 ), and their supercomplex forms (CI ⁄ CIII 2 ), (CI ⁄ CIII 2 ⁄ CIV) and (CIII 2 CIV) were identified by western blot analysis using antibodies to the complex I subunit NDUFA9 (a-com-plex I), the core I subunit of com(a-com-plex III (a-complex III) and the COI subunit of complex IV (a-complex IV) The asterisk indicates a complex of approximately
250 kDa.
Trang 5complexes in the range of approximately 700–1000
kDa (lanes 1 and 2), but these did not co-migrate with
any of the complex IV-containing supercomplexes The
identity of these complexes is unknown The complex
of approximately 250 kDa observed for newly
imported subunits Cox6a and Cox7a after DDM
solu-bilization was poorly resolved using digitonin (Fig 2B,
marked with an asterisk) However, Cox6a and Cox7a
as well as Cox6b assembled into holo-complex IV and
into complexes co-migrating with the supercomplex
forms CIII2⁄ CIV and CI ⁄ CIII2⁄ CIV Of note, the
intensity of assembled35S-labeled Cox6b was stronger
in mitochondria solubilized in digitonin than those
sol-ubilized in DDM (compare lane 6 in Fig 2A and 2B)
This is consistent with a previous report indicating that
Cox6b can be preferentially stripped from complex IV
in the presence of DDM [20] due to the peripheral
nat-ure of this subunit From these results, we conclude
that, in isolated mitochondria, some newly imported
subunits have the capacity to integrate into the
holoen-zyme as well as into supercomplexes containing
com-plex IV Furthermore, given that yeast mitochondria
lack complex I, this analysis also characterized
nDNA-encoded subunit integration into more complicated
supercomplexes that additionally contain pre-existing
complex I
Assembly profile of imported Cox6a in human
and yeast mitochondria
At early stages of import, the nDNA-encoded
com-plex IV subunits Cox6a and Cox7a were found also to
be incorporated into a complex of approximately
250 kDa, slightly higher than monomeric complex IV
Both subunits are thought to integrate into the
assem-bly pathway at a late stage [13,21] By studying the
import and assembly of Cox6a (see below), we have
established that the complex of approximately
250 kDa represents a novel intermediate, and have
termed it the late-stage intermediate (LSI) complex
(Fig 3) 35S-labeled Cox6a was imported into isolated
mitochondria for various times before solubilization in
DDM and BN-PAGE analysis (Fig 3A) At early time
points, Cox6a was predominantly incorporated into
the LSI complex, while a minor amount accumulated
into holo-complex IV and its supercomplex, CIII2⁄
CIV Over the time course of the experiment, Cox6a
accumulated into holo-complex IV, while the signal for
the LSI complex remained relatively constant Western
blot analysis confirmed the relative positions of
com-plex IV and CIII2⁄ CIV (Fig 3A, right-hand panel);
however, the LSI complex was not seen Assembly
intermediates generally cannot be resolved with
anti-bodies due to their low steady-state levels, but import using small amounts of radiolabeled protein can detect their presence [16,22–24] Next, an in vitro import and chase experiment was performed 35S-labeled Cox6a was imported for 5 min to accumulate the subunit at the LSI complex before mitochondria were re-isolated and then further incubated in the absence of additional radiolabeled precursor (Fig 3B, lanes 1–4) Over the chase period, the intensity of 35S-labeled Cox6a in the LSI complex decreased, with a concomitant increase in the intensity of labeling of holo-complex IV
35S-labeled Cox6a was also imported into mitochon-dria isolated from cells that had been pre-incubated with chloramphenicol (CAP) Under these conditions, complex IV assembly intermediates containing mtDNA-encoded subunits are unlikely to be present, and a pool of unassembled nDNA-encoded subunits may accumulate Any 35S-labeled Cox6a assembly into holo-complex IV is therefore likely to occur via inte-gration into the pre-existing complex as opposed to new assemblies As can be seen in Fig 3C, 35S-labeled Cox6a was incorporated into the LSI complex and holo-complex IV in mitochondria from both control (lanes 1–3) and CAP-treated cells (lanes 4–6) Given that the import of 35S-labeled Cox6a was unaffected (Fig 3D), the decreased signal of assembled Cox6a observed in the CAP-treated samples is probably a result of decreased levels of fully assembled com-plex IV as shown by western blot analysis (bottom panels in Fig 3C) As in organello labeling is inefficient under the conditions used (data not shown) and Cox6a assembly occurs even in mitochondria isolated from CAP-pretreated cells, these results support the possibil-ity that late-assembling subunits such as Cox6a have the capacity to assemble into complex IV by cycling with pre-existing subunits
In order to eliminate the possibility that the precur-sor form is found in the LSI complex, 35S-labeled Cox6a was imported into mitochondria for various times (without proteinase K treatment) and subjected
to BN-PAGE followed by SDS–PAGE in the second dimension (Fig 3E) At early time points, the precur-sor form of 35S-labeled Cox6a (as judged by its co-migration with the lysate control sample) was found
in a high-molecular-weight smear, presumably bound
to molecular chaperones and⁄ or the translocase of the outer mitochondrial membrane (TOM) machinery [25] The mature form of 35S-labeled Cox6a was initially found in the more slowly migrating LSI complex, and over time assembled into mature complex IV The position of mature complex IV was confirmed by immunostaining (bottom panel) Based on these results, we conclude that the LSI complex represents a
Trang 6novel intermediate for the biogenesis of Cox6a, and
hence Cox7a, into human complex IV Coomassie
staining of mitochondria did not reveal the presence of
the LSI complex even though complex IV could be
detected (data not shown), consistent with the LSI
complex being a short-lived intermediate
As an ortholog of Cox6a is found in yeast [termed
Cox10p (COX13)], we determined whether the LSI
complex is evolutionarily conserved In vitro import of
Cox10p has previously been used to analyze
com-plex IV assembly in yeast, but digitonin-solubilized
mitochondria were employed and hence only
super-complexes were observed [19] Furthermore, a time
course of the import was not performed Radiolabeled Cox10p was imported into isolated yeast mitochondria, and all samples were treated with proteinase K before solubilization in DDM and BN-PAGE analysis (Fig 4A) The time-course analysis revealed that the assembly pathway of Cox10p resembles that of its human counterpart The intensity of the LSI complex remained consistent over time, and the holo-com-plex IV of approximately 200 kDa accumulated Import and chase analysis of 35S-labeled Cox10p (Fig 4B, lanes 1–4), revealed that, like its human Cox6a counterpart, Cox10p initially accumulated in the LSI complex and was chased over time into holo-complex IV Evolutionary conservation of the LSI complex containing Cox6a⁄ Cox10p suggests that it is
an important step in the biogenesis of these subunits and subsequently the holoenzyme
Assembly analysis of35S-labeled Cox6a in patient mitochondria
The assembly of subunit Cox6a was investigated fur-ther using fibroblast mitochondria from two patients
LSIC CIV
669 440 kDa
α-complex IV
669
440
kDa
LSIC CIII2/CIV
CIII2/CIV
CIII2/CIV
CIV
α-complex I V
35 S-Cox6a (Human)
Time (min)10 30 601 2 3 Chase (min) 0 15 30 601 2 3 4
134
67
134 67
35 S-Cox6a (Human)
669
440
kDa
134
67
LSIC CIV Control + CAP
1 2 3 4 5 6
Time (min) 10 30 60 10 30 60
1 2 3 4 5 6
α -complex IV
α -complex II CIV
CII
Control
+ CAP
1 2 3 4 5 6 Time (min) 10 30 60 10 30 60
p m
p m
- Prot K + Prot K
CIV
BN-PAGE
Lysate
control
5 min
10 min
60 min
30 min
α -complex IV
E
Fig 3 Import and assembly of Cox6a into pre-existing complex IV (A) 35S-labeled Cox6a was incubated for various times with mito-chondria isolated from human fibroblasts Samples were treated with proteinase K and subjected to DDM solubilization, BN-PAGE and phosphorimaging Right lane: the migration of CIV and CIII 2 ⁄ CIV supercomplexes was identified by western blot analysis using antibodies to the complex IV subunit COI (a-complex IV) (B)
35
S-labeled Cox6a was incubated for 5 min with mitochondria after removal of free 35 S-labeled Cox6a and chase of assembly for vari-ous times as indicated Mitochondria were treated as in (A) (C)
35 S-labeled Cox6a was incubated for 10–60 min with mitochondria isolated from control fibroblasts that had been pre-treated with or without chloramphenicol (CAP) for 12 h Samples were treated with proteinase K before being solubilized in DDM-containing buffer and subjected to BN-PAGE, western transfer and phosphorimage analy-sis (top panel), followed by immunodecoration using antibodies to COI (a-complex IV) and 70 kDa subunit (a-complex II) (bottom panel) (D) 35S-labeled Cox6a was imported into mitochondria as described in (C), with and without proteinase K treatment, before SDS–PAGE and phosphorimage analysis (E) 35 S-labeled Cox6a was imported into control mitochondria for increasing times as indi-cated Samples were solubilized in DDM-containing buffer, and sub-jected to BN-PAGE in the first dimension followed by SDS–PAGE
in the second dimension Gels were subjected to phosphorimaging The position of complex IV was confirmed based on immunoblot analysis using antibodies against COI (bottom panel) The lysate control shows the position of the 35 S-labeled Cox6a precursor spe-cies after one-dimensional SDS–PAGE p,35S-labeled Cox6a precur-sor form; m, 35 S-labeled Cox6a mature form; CIV, complex IV; LSIC, late-stage intermediate complex; CIII 2 ⁄ CIV, complex III 2 ⁄ com-plex IV supercomcom-plex.
Trang 7with Leigh syndrome involving isolated complex IV
deficiency [18] Patient 1 was homozygous for a
patho-genic frameshift mutation in the gene encoding the
complex IV assembly factor Surf1 Patient 2 had
iso-lated complex IV deficiency and an as yet unknown
nuclear gene mutation, although mutations in
com-plex IV subunits and known assembly factors have
been excluded (data not shown) This analysis served
to test whether defects in formation of the LSI
com-plex may be involved in altered comcom-plex IV assembly
and human disease, and also to assess the possible involvement of Surf1 in the formation and assembly of the LSI complex BN-PAGE and western blot analysis
of mitochondria solubilized in DDM indicated that the fibroblasts of patients 1 and 2 had very low levels of mature complex IV (Fig 5A, right panels) In addi-tion, the CIII2⁄ CIV supercomplex was not detected in patient mitochondria, although the levels of complex I, complex III and the CI⁄ CIII2supercomplex were simi-lar to control The amount of assembled 35S-labeled Cox6a in the mitochondria of patient 1 (Fig 5A, lanes 5–7) was lower relative to control (lanes 1–3), with reduced levels of both the LSI complex and holo-com-plex IV This is most likely due to the substantially decreased levels of complex IV in these mitochondria
In the mitochondria of patient 2, 35S-labeled Cox6a was initially incorporated into the LSI complex (Fig 5A, lane 13) where it accumulated over time (lanes 14 and 15) The radioactivity below the LSI complex and at the position of the CIII2⁄ CIV complex may represent low levels of assembled complex IV (lane 15) Thus, it appears that progression from the LSI complex to holo-complex IV is defective in the mitochondria of patient 2 compared to the assembly profile for the control (compare lanes 9–11 and 13–15)
As expected, disruption of the membrane potential (Dwm) abolished assembly in all samples due to inhibi-tion of import SDS–PAGE was also performed in order to eliminate the possibility that the assembly defects for Cox6a in patient mitochondria were a result
of impaired import As shown in Fig 5B, 35S-labeled Cox6a was efficiently imported into both patient and control mitochondria in a Dwm-dependent manner
Time (min)10 30 60 10 30 60
Patient 1 Control
60 60
CIII
2 /CIV CIV LSIC
669
440
kDa
10 30 60 10 30 60
Patient 2 Control
60 60
+ + + + + + –
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
α -complex IV α -complex III α -complex I
CIII2 CI/CIII2
Time (min)10 30 60 60 10 30 60 60
SDS-PAGE
Control
Patient 1
p m
p m –Prot K +Prot K
Lysate
Patient 2 p
m
1 2 3 4 5 6 7 8 9
BN-PAGE 0.65% DDM
Control Patient 1 Patient 2 Contro
l
Patient 1 Pati ent 2 Contro l Patient 1 Patient 2
134
67
Fig 5 Import and assembly of Cox6a in control and patient mitochondria Mitochondria from control or patient cells were incubated with
35
S-labeled Cox6a for increasing times in the presence or absence of a membrane potential (Dw m ) Half of each sample was treated with proteinase K before being split in two and (A) solubilized in DDM-containing buffer and subjected to BN-PAGE (protease-treated samples only), or (B) SDS–PAGE and phosphorimaging The right panels in (A) show western blot analysis of complex IV, complex III and complex I
in mitochondrial preparations CI, complex I; CIII 2 , complex III dimer; CIV, complex IV; LSIC, late-stage intermediate complex; CIII 2 ⁄ CIV, complex III 2 ⁄ complex IV supercomplex; CI ⁄ CIII 2 , complex I ⁄ complex III supercomplex.
669
440
kDa
Time (min) 10 30 60
LSIC CIV
134
67
35S-Cox10p (Yeast)
669
440 kDa
Chase (min) 0 15 30 60
35S-Cox10p (Yeast)
LSIC CIV
134 67
Fig 4 Import and assembly of the yeast Cox6a ortholog Cox10p
in yeast mitochondria (A) 35 S-labeled Cox6a was incubated for
vari-ous times with mitochondria isolated from yeast cells Samples
were treated with proteinase K and subjected to DDM
solubiliza-tion, BN-PAGE and phosphorimaging (B) 35 S-labeled Cox6a was
incubated for 5 min with mitochondria after removal of free
35 S-labeled Cox6a and chase of assembly for various times
Mito-chondria were treated as in (A) CIV, complex IV; LSIC, late-stage
intermediate complex.
Trang 8These results suggest that incorporation of Cox6a into
the LSI complex and its progression to complex IV is
not dependent on Surf1, and the slowed rate of
assem-bly into complex IV is most likely a result of reduced
levels of holo-complex IV In the mitochondria of
patient 2, the progression of Cox6a into complex IV is
effectively stalled, suggesting that the underlying defect
may be related to a late stage in complex IV biogenesis
Assembly of nDNA-encoded complex IV subunits
into supercomplexes in control and patient
mitochondria
When we analyzed respiratory complexes in patient
mitochondria after digitonin solubilization and
BN-PAGE, we found that the residual complex IV was predominantly found in supercomplexes Using anti-bodies against the subunits of complexes I, III and IV (Fig 6A, right panels), it appeared that only the super-complex form (CI⁄ CIII2⁄ CIV) was present in patient mitochondria An additional faster-migrating species was also detected in patient mitochondria, and is likely
to represent the CI⁄ CIII2 supercomplex In control mitochondria, complex IV was also found in its super-complex forms, but its most predominant form was as
a monomer As complex IV was only detected in the
CI⁄ CIII2⁄ CIV supercomplex in patient mitochondria, this suggests that this form is particularly stable and⁄ or crucial for respiratory function However, we cannot exclude the possibility that complex IV found
Cox4-1 Cox6b Cox6c Cox7a Cox6a
Time (min) 10 60 10 60 10 60 10 60 10 60
669 440
kDa
CI/CIII2/CIV
CIV CIII2
BN-PAGE 1% Digitonin
α-complex I
α-complex III
α-complex IV
CI/CIII2
669
440
kDa
A
B
BN-PAGE 1% Digitonin α -complex IV α -complex III α -complex I
Control Patient 1 Patient 2
CIII2/CIV
CIV CIII2
CI/CIII2/CIV
CI/CIII2 CI/CIII2/CIV
Control Patient 1 Patient 2 Control Patient 1 Patient 2 Time (min)10 30 60 10 30 60
Patient 1 Control
60 60
+
Patient 2 Control
60 60
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
134
67
134
67
*
LSIC
Fig 6 nDNA-encoded subunits assemble into supercomplexes of control and patient mitochondria (A) Mitochondria from control or patient cells were incubated with 35 S-labeled Cox6a for increasing times in the presence or absence of a membrane potential (Dwm) Each sample was treated with proteinase K before being solubilized in digitonin-containing buffer and subjected to BN-PAGE and phosphorimaging (B)
35 S-labeled complex IV subunits were incubated with mitochondria from patient 1 for 10 and 60 min and treated as described in (A) The right panels in (A) and (B) show the migration of complex IV (CIV), dimeric complex III (CIII2), CIII2⁄ CIV supercomplex, complex I (CI) ⁄ CIII 2 ⁄ CIV supercomplex and CI ⁄ CIII 2 supercomplex by western blot analysis The asterisk indicates the complex of approximately 100 kDa.
Trang 9in supercomplexes of patient mitochondria is not all
fully assembled
We determined whether the integration of newly
imported Cox6a into supercomplexes was defective in
patient mitochondria Radiolabeled Cox6a was
imported into mitochondria isolated from patient 1,
patient 2 and control fibroblasts, solubilized in
digito-nin and analyzed using BN-PAGE (Fig 6A) As
observed above (Fig 5), the assembly of newly
imported Cox6a into monomeric complex IV was
reduced in mitochondria from cells of both patient 1
(Fig 6A, lanes 5–7) and patient 2 (lanes 13–15)
Although the LSI complex was not clearly resolved
after solubilization using digitonin, mitochondria from
cells of patient 2 showed smearing above complex IV
(lanes 13–15), and this may represent the impaired
pro-gression of35S-labeled Cox6a from the LSI complex to
holo-complex IV Incorporation of 35S-labeled Cox6a
into the CIII2⁄ CIV supercomplex was reduced in
mito-chondria from both patient 1 and patient 2; however,
its incorporation into the CI⁄ CIII2⁄ CIV supercomplex
was not impaired, with an overall signal comparable to
that in controls (Fig 6A)
The relatively efficient assembly of newly imported
Cox6a into the CI⁄ CIII2⁄ CIV supercomplex in patient
mitochondria led us to investigate the assembly of
additional nDNA-encoded subunits Mitochondria
iso-lated from fibroblasts from patient 1 were chosen for
this study because of the better growth rate of these
cells in culture As shown in Fig 6B, newly imported
Cox4-1, Cox6b, Cox6c, Cox7a and Cox6a all
effi-ciently assembled into the CI⁄ CIII2⁄ CIV supercomplex
in the mitochondria of patient 1 The subunit assembly
profile also differed to that observed in control
mito-chondria, in that Cox4-1 and Cox6c did not assemble
into any complex IV-containing supercomplexes (see
Fig 2B, lanes 1–2 and lanes 5–6) However, in
addi-tion to its assembly into supercomplexes in the
mito-chondria of patient 1, Cox4-1 also assembled into a
complex of approximately 100 kDa (Fig 6B, lanes 1
and 2) that was not observed in control mitochondria
Cox4-1 is believed to form an early assembly
interme-diate with CO1 [15], and the species of approximately
100 kDa may represent such an intermediate; however,
further characterization is required Additional
low-molecular-weight complexes were also seen for Cox6c
import in the mitochondria of patient 1 that were not
seen in control mitochondria (see Fig 2B, lanes 5 and
6) These smaller complexes may represent
rate-limit-ing intermediates due to assembly defects in the
mito-chondria of patient 1 It is possible that, in the absence
of monomeric complex IV, a portion of these
inter-mediates can combine with complex I and complex IV
to form the supercomplexes that are seen under these conditions
Discussion
The current model of complex IV assembly follows a sequential pathway that begins with the mitochondrial translation of CO1 followed by integration of addi-tional subunits and co-factors through a set of defined intermediates (for reviews on complex IV assembly, see [12,21,26,27]) A number of assembly factors have been identified that assist in the process and act at the levels
of regulation [28], co-factor biosynthesis and insertion [29,30] and chaperoning of assembly intermediates [31] Much of our current knowledge regarding complex IV biogenesis has been provided by studies using the model organism S cerevisiae Most studies have focused on the early stages of assembly that involve formation of the catalytic core consisting of the mtDNA-encoded subunits CO1, CO2 and CO3 Thus details of the latter stages in which the majority of nDNA-encoded subunits that surround the core are assembled remain largely unknown In particular, it is not clear how a newly imported nDNA-encoded subunit assembles in the presence of pre-existing complex IV Furthermore, there are a number of dif-ferences between mammalian and yeast mitochondria that affect complex IV biogenesis, and these may be relevant to disease These include the presence of struc-tural subunits and assembly factors in yeast that do not appear to have homologs in mammals, and differences in supercomplex forms The consequence of these differences is that further analysis of the biogene-sis of the mammalian enzyme is required
Assembly of nDNA-encoded complex IV subunits
Of the five nDNA-encoded subunits investigated in this study, Cox6a, Cox7a and Cox6b were found to assemble into both monomeric and supercomplex forms of complex IV (Fig 1) According to the current model of complex IV assembly, subunit Cox6b assem-bles into the S3 subcomplex together with a host of other nDNA-encoded subunits as well as CO3 [13,15] However, it has been suggested more recently that Cox6b is incorporated at the very last step of com-plex IV assembly [32], possibly after addition of the late-assembling subunits Cox6a and Cox7a Based on the results presented here, it appears that subunits that are incorporated late in the assembly pathway have a greater propensity to assemble into pre-existing com-plex IV in isolated mitochondria Previously, we have shown that individual, newly imported complex I
Trang 10subunits can dynamically exchange with their
pre-exist-ing counterparts to assemble into the complex [16]
Similar findings have also been reported for other
complexes [24,33–36] It is therefore possible that these
newly imported nDNA-encoded complex IV subunits
are assembling into the pre-existing complex via a
similar mechanism
The remaining two subunits, Cox4-1 and Cox6c,
were not found to assemble into holo-complex IV;
however, they were found to integrate into additional
complexes Subunit Cox6c resolved into two distinct
complexes in the range of approximately 100–150 kDa
when the detergent DDM was used The assembly
pro-file of Cox6c differed in digitonin-solubilized samples,
where it was found to assemble into a complex of
approximately 160 kDa and also a large species of
approximately 1 MDa Import of subunit Cox4-1
revealed that it also assembled into large complexes
ranging from approximately 700–1000 kDa, although
it was not found in any distinct complexes when DDM
was used This subunit is considered to be one of the
first nDNA-encoded subunits to integrate into
sub-assemblies of complex IV [13,14] In a previous study,
import analysis of the yeast ortholog of human Cox4-1
revealed that it assembles into a number of complexes
ranging from 250 to 450 kDa in size [19] These
com-plexes were found to contain assembly factors such as
Cox14p, Coa1p and Shy1p as well as complex IV
su-bunits Of these assembly factors, only Shy1p has a
human homolog, termed Surf1 [37,38], and this could
account for the different complexes observed in yeast
and mammalian mitochondria Nevertheless, the
com-plexes identified here may represent rate-limiting
inter-mediates that require additional factors not found in
yeast mitochondria for Cox4-1 and Cox6c subunit
maturation and⁄ or assembly
A novel late-stage intermediate complex for
Cox6a and Cox7a assembly into complex IV
In mitochondria solubilized with DDM, newly
imported subunits Cox7a and Cox6a were found to
integrate into the LSI complex of approximately
250 kDa prior to their incorporation into
holo-com-plex IV and its supercomholo-com-plex forms Similar results
were observed in yeast mitochondria when the
assem-bly of the Cox6a ortholog, Cox10p was analyzed, thus
indicating that the assembly profile for this subunit is
evolutionarily conserved In comparison to
com-plex IV, the LSI comcom-plex is not visible on
Coomassie-stained 2D gels (data not shown), supporting the
conclusion that it is a low-level intermediate complex
that is only detected with radiolabeled proteins
Previous import and assembly analysis of Cox10p using digitonin solubilization did not reveal the pres-ence of a complex of approximately 250 kDa [19], con-sistent with our findings that this complex is not well resolved in digitonin Analysis of mitochondria from a patient harboring an isolated complex IV deficiency of unknown etiology (patient 2) revealed that Cox6a assembly into the LSI complex had stalled Although a complex of similar size has previously been described
in yeast mitochondria [19,39], this complex is likely to differ as yeast contains assembly factors that are not present in humans Furthermore, the studies in yeast used digitonin for solubilization, and the LSI complex
is not clearly resolved in this detergent Therefore, we conclude that the LSI complex represents a novel com-plex for maturation of at least Cox6a and Cox7a into complex IV, and that this assembly process is per-turbed in human disease A possible explanation for the larger size of the LSI complex relative to the holo-enzyme is that a specific accessory factor is associated with this late-stage intermediate that is displaced after integration of the subunits into the final complex Alternatively, Cox6a and Cox7a may integrate into the LSI complex and then be transferred into the maturing complex IV
Integration of nDNA-encoded subunits into supercomplexes in patient mitochondria Using digitonin solubilization of mitochondria to visu-alize supercomplexes, it was found that subunit Cox6a assembled into the CI⁄ CIII2⁄ CIV supercomplex in the mitochondria of patients 1 and 2 Immunoblot analysis
of patient mitochondria revealed that all detectable complex IV was present in a supercomplex with com-plexes I and III This is in contrast to control mito-chondria in which the majority of complex IV is not associated with supercomplexes and instead resolves in its monomeric form [18,40,41] As that complex IV may be important for the assembly⁄ stability of com-plex I [42,43] and functions within respirasomes [7], it follows that limiting amounts of complex IV (in partial
or fully assembled forms) could be sequestered within supercomplexes, as observed in patient mitochondria All subunits tested here (Cox4-1, Cox6a, Cox6b, Cox6c and Cox7a) were found to efficiently assemble into the CI⁄ CIII2⁄ CIV supercomplex Of particular interest were the newly imported subunits Cox4-1 and Cox6c, as they did not assemble into complex IV or its supercomplex forms in control mitochondria, but had the ability to integrate into the CI⁄ CIII2⁄ CIV super-complex in patient mitochondria As the subunits Cox4-1, Cox6b, Cox6c, Cox7a, and Cox6a all