Báo cáo khoa học: Assembly of nuclear DNA-encoded subunits into mitochondrial complex IV, and their preferential integration into supercomplex forms in patient mitochondria doc

Assembly of nuclear DNA-encoded subunits into
mitochondrial complex IV, and their preferential
integration into supercomplex forms in patient
mitochondria
Michael Lazarou
1
, Stacey M. Smith
1,2
, David R. Thorburn
2
, Michael T. Ryan
1
and
Matthew McKenzie
1
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
F
1
F
o
-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:
(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.
FEBS Journal 276 (2009) 6701–6713 ª 2009 The Authors Journal compilation ª 2009 FEBS 6701
drial 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 sub-
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.
35
S-labeled complex IV subunit precursor proteins
were generated in vitro using rabbit reticulocyte lysate,

Mitochondrial complex IV assembly M. Lazarou et al.
6702 FEBS Journal 276 (2009) 6701–6713 ª 2009 The Authors Journal compilation ª 2009 FEBS
and incubated with mitochondria isolated from cul-
tured human fibroblasts for 10 or 60 min in the pres-
ence or absence of a membrane potential (Dw
m
). 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).
35
S-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 (Dw
m
) 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 III
2
⁄ complex IV
supercomplex [18]. The migration of complex IV,
dimeric complex III, complex I and their super-
complexes (CIII
2
⁄ CIV and CI ⁄ CIII
2
) was determined
by western blot analysis (Fig. 2A, right panels).
Imported Cox6a and Cox7a were incorporated into
both monomeric and supercomplex (CIII
2
⁄ 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

35
S-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 (Dw
m
), 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.
M. Lazarou et al. Mitochondrial complex IV assembly
FEBS Journal 276 (2009) 6701–6713 ª 2009 The Authors Journal compilation ª 2009 FEBS 6703
complexes 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 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 ⁄ CIII
2
⁄ CIV), complex III and
complex IV (CIII
2
⁄ CIV), as well as dimeric com-
plex III (CIII
2
) 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,
35
S-labeled Cox4-1 was found in large
A
CIV
*
CI
669
440
kDa
Cox4-1 Cox6b Cox6c Cox7aCox6a
BN-PAGE
0.65% DDM
Time (min)10
60
10
60 10
60
10
60
10
60
12345678910
α-complex I
α-complex III
α-complex IV
134
67

B
669
440
kDa
Cox4-1 Cox6b Cox6c Cox7aCox6a
CIV
BN-PAGE
1% Digitonin
Time (min)10
60
10
60 10
60
10
60
10
60
12345678910
CIII
α-complex I
α-complex III
α
-complex 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 complex III
(a-complex III) and the COI subunit of
complex IV (a-complex IV). The asterisk
indicates a complex of approximately
250 kDa.
Mitochondrial complex IV assembly M. Lazarou et al.
6704 FEBS Journal 276 (2009) 6701–6713 ª 2009 The Authors Journal compilation ª 2009 FEBS
complexes 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 CIII
2
⁄ CIV and CI ⁄ CIII
2
⁄ CIV. Of note, the
intensity of assembled
35
S-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).
35
S-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, CIII
2
⁄
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 CIII
2
⁄ 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.
35
S-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
35
S-labeled Cox6a in the
LSI complex decreased, with a concomitant increase in
the intensity of labeling of holo-complex IV.
35
S-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
35
S-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,
35

S-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
35
S-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,
35
S-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
35
S-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
35
S-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
M. Lazarou et al. Mitochondrial complex IV assembly
FEBS Journal 276 (2009) 6701–6713 ª 2009 The Authors Journal compilation ª 2009 FEBS 6705
novel 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
35
S-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 of
35
S-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
CIII
2
/CIV
CIII
2
/CIV
CIII
2
/CIV
CIV
α
-complex I
V

35
S-Cox6a (Human)
Time (min)10 30 60
12
3
Chase (min) 0 15 30 60
1
2
3
4
AB
134
67
134
67

35
S-Cox6a (Human)
669
440
kDa
134
67
LSIC
CIV
Control + CAP
12 345 6
Time (min)
10 30 60 10 30 60
123
4
56
α-complex IV
α-complex II
CIV
CII
Control
+ CAP
1
23456
Time (min)
10 30 60 10 30 60
p
m
p

m
- Prot. K
+ Prot. K
CIV
BN-PAGE
SDS-PAGE
Lysate
control
5 min
10 min
60 min
30 min
α-complex IV
p
LSIC
CD
E
Fig. 3. Import and assembly of Cox6a into pre-existing complex IV.
(A)
35
S-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)
35
S-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,
35

S-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 supercomplex.
Mitochondrial complex IV assembly M. Lazarou et al.
6706 FEBS Journal 276 (2009) 6701–6713 ª 2009 The Authors Journal compilation ª 2009 FEBS
with 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 complex 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 CIII
2

⁄ CIV supercomplex was not detected in
patient mitochondria, although the levels of complex I,
complex III and the CI ⁄ CIII
2
supercomplex were simi-
lar to control. The amount of assembled
35
S-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,
35
S-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 CIII
2
⁄ 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
(Dw
m

) 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,
35
S-labeled
Cox6a was efficiently imported into both patient and
control mitochondria in a Dw
m
-dependent manner.
Time (min)
10 30 60 10 30 60
Patient 1
Control
6060
++
+– – –
Δ
m
+
++
CIII
2
/CIV
CIV
LSIC
669
440
kDa

A B
10 30 60 10 30 60
Patient 2
Control
6060
++
++
++–
123 56784
91011 1314151612
α-complex IV α-complex III α-complex I
CIII
2
CI/CIII
2
Time (min)10 30 60 10 30 60 6060
+++– –
Δ
+++
SDS-PAGE
Control
Patient 1
p
m
p
m
–Prot. K
+Prot. K
Lysate
Patient 2

p
m
123 56784
9
BN-PAGE 0.65% DDM
Control
Patient 1
Patient 2
Control
Patient 1
Pati
ent 2
Control
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
AB
Time (min) 10 30 60
LSIC
CIV
123
134
67
35
S-Cox10p (Yeast)
669
440
kDa
Chase (min)
0153060
35
S-Cox10p (Yeast)
LSI
C
CIV
12
3
4
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.
M. Lazarou et al. Mitochondrial complex IV assembly
FEBS Journal 276 (2009) 6701–6713 ª 2009 The Authors Journal compilation ª 2009 FEBS 6707
These 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⁄ CIII
2
⁄ CIV) was present in patient
mitochondria. An additional faster-migrating species
was also detected in patient mitochondria, and is likely
to represent the CI ⁄ CIII
2
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 ⁄ CIII
2
⁄ 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
Cox4-1Cox6b Cox6c Cox7a Cox6a
Time (min) 10
60
10
60 10
60
10
60
10

60
12345678910
669
440
kDa
CI/CIII
2
/CIV
CIV
CIII
2
BN-PAGE 1% Digitonin
α
-complex I
α
-complex III
α
-complex IV
CI/CIII
2
669
440
kDa
A
B
BN-PAGE 1% Digitonin
α-complex IV α-complex III α-complex I
Control
Patient 1
Patient 2

CIII
2
/CIV
CIV
CIII
2
CI/CIII
2
/CIV
CI/CIII
2
CI/CIII
2
/CIV
Control
Patient 1
Patient 2
Control
Patient 1
Patient 2
Time (min)
10 30 60 10 30 60
Patient 1
Control
6060
+
m
+
+
Δ

+
++
10 30 60 10 30 60
Patient 2
Control
6060
++
+
+
++
123 56784 9 10 11 13 14 15 1612
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 (Dw
m
). 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 (CIII
2
), CIII
2

⁄ 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.
Mitochondrial complex IV assembly M. Lazarou et al.
6708 FEBS Journal 276 (2009) 6701–6713 ª 2009 The Authors Journal compilation ª 2009 FEBS
in 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 of
35
S-labeled Cox6a from the LSI complex to
holo-complex IV. Incorporation of
35
S-labeled Cox6a

into the CIII
2
⁄ CIV supercomplex was reduced in mito-
chondria from both patient 1 and patient 2; however,
its incorporation into the CI ⁄ CIII
2
⁄ 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 ⁄ CIII
2
⁄ 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 ⁄ CIII
2
⁄ 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
M. Lazarou et al. Mitochondrial complex IV assembly
FEBS Journal 276 (2009) 6701–6713 ª 2009 The Authors Journal compilation ª 2009 FEBS 6709
subunits 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 supercomplex 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 complex 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 ⁄ CIII
2
⁄ 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 ⁄ CIII
2
⁄ 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 ⁄ CIII
2
⁄ CIV super-
complex in patient mitochondria. As the subunits
Cox4-1, Cox6b, Cox6c, Cox7a, and Cox6a all
Mitochondrial complex IV assembly M. Lazarou et al.
6710 FEBS Journal 276 (2009) 6701–6713 ª 2009 The Authors Journal compilation ª 2009 FEBS
assembled into the supercomplex, this suggests that
none of these subunits are rate-limiting for assembly.
Although it is not clear why these subunits assemble
into the supercomplex in patient but not control mito-
chondria, it has been shown previously that the
steady-state levels of various complex IV subunits are

reduced in the mitochondria of patients with a muta-
tion in the SURF1 gene [14]. A small pool of remain-
ing subunits was found predominantly in stalled
assembly intermediates. Therefore, it is possible that
newly imported subunits are channeled more efficiently
into assembly intermediates in patient mitochondria as
they do not have to compete with a large pool of pre-
existing subunits. In addition, these assembly inter-
mediates may already be in supercomplex forms, as
recently proposed by Mick et al. [19] for the biogenesis
of yeast complex IV.
In summary, the work presented here shows that
newly imported complex IV subunits can integrate into
complex IV and its supercomplexes by associating with
pre-existing subunits and integrating into intermediate
complexes. A novel late-stage intermediate complex
was identified that is evolutionarily conserved. The
exact composition of this complex awaits further char-
acterization.
Experimental procedures
Cloning procedures
The cDNAs encoding human Cox4-1, Cox6b, Cox6c,
Cox6a and Cox7a (accession numbers AAB94819,
AAP35591, AAC73061, AAH70186 and NP_001856,
respectively) were amplified by PCR from a mixed-tissue
cDNA library (BD Biosciences ⁄ Clontech, San Jose, CA,
USA) and cloned into pGEM-4Z (Promega, Madison, WI).
The cDNA encoding Cox13p (accession number
NP_011324) was amplified by PCR from a yeast genomic
DNA library and cloned into pGEM-4Z.

Cell culture and mitochondrial isolation
Primary skin fibroblasts grown from patient skin biopsy
material (obtained by consent) were cultured in Dulbecco’s
modified Eagle’s medium (Invitrogen, Carlsbad, CA, USA)
containing 10% v ⁄ v fetal bovine serum at 37 °C under an
atmosphere of 5% CO
2
⁄ 95% air supplemented with
50 lgÆmL
)1
uridine. For chloramphenicol (CAP) pre-treat-
ment, cells were incubated with or without 50 lgÆmL
)1
CAP in Dulbecco’s modified Eagle’s medium containing
10% v ⁄ v fetal bovine serum for 12 h before harvesting. For
mitochondrial isolation, harvested cells were homogenized
in 20 mm Hepes (pH 7.6), 220 mm mannitol, 70 mm
sucrose, 1 mm EDTA, 0.5 mm phenylmethylsulfonyl fluo-
ride, and centrifuged at 800 g at 4 °C for 10 min to obtain
a postnuclear supernatant. Mitochondria were pelleted by
centrifugation at 10 000 g at 4 °C for 20 min.
In vitro import and assembly assays
Generation of radiolabeled nDNA-encoded precursor pro-
teins was performed by in vitro transcription, followed by
translation using rabbit reticulocyte lysates (Promega) in
the presence of
35
S-methionine ⁄ cysteine protein labeling
mix (Perkin-Elmer, Waltham, MA, USA) as previously
described [44]. Translation products were incubated with

freshly isolated mitochondria in import buffer (20 mm
Hepes ⁄ KOH pH 7.4, 250 mm sucrose, 80 mm KOAc, 5 mm
MgOAc, 10 mm sodium succinate and 5 mm methionine) at
37 °C for various times as indicated in the figure legends.
For chase experiments, mitochondria were re-isolated using
centrifugation (see above), washed and resuspended in
import buffer lacking
35
S-labeled translation products. Dis-
sipation of membrane potential (Dw
m
) was performed
using10 lm valinomycin (Sigma, St Louis, MO, USA).
Samples subjected to protease treatment were incubated on
ice for 10 min with 100 lgÆmL
)1
proteinase K (Sigma),
before protease inactivation with 1 mm phenylmethane-
sulfonyl fluoride. For Tris-Tricine SDS–PAGE [45], mito-
chondrial pellets (25 lg) were first precipitated with
trichloroacetic acid as described previously [46]. For
BN-PAGE, mitochondrial pellets (50 lg protein) were
resuspended in either 50 lL1%w⁄ v digitonin (Merck,
Whitehouse Station, NJ, USA) or 0.65% w ⁄ v n-dodecyl-b-
d-maltoside (Sigma) in 50 mm NaCl, 10% v ⁄ v glycerol,
20 mm Bis ⁄ Tris pH 7.0, and subjected to 4–13% gradient
BN-PAGE, with or without SDS–PAGE in the second
dimension [18]. Thyroglobulin (669 kDa), ferritin (440 kDa)
and bovine serum albumin (134 and 67 kDa) were used as
markers for BN-PAGE.

The YPH499 haploid yeast strain was cultured in 1%
w ⁄ v yeast extract, 2% w ⁄ v bacto-peptone and 3% v ⁄ v glyc-
erol (pH 4.9), and mitochondria were isolated as described
previously [47]. Import of Cox10p was performed as
described by Ryan et al. [44].
Miscellaneous
Antibodies against the 70 kDa subunit, the core I subunit
and CO1 (Invitrogen)were used to detect complexes II, III
and IV, respectively, and NDUFA9 polyclonal antibodies
were used to detect complex I [16]. Radiolabeled proteins
were detected by phosphorimage analysis (GE Healthcare,
Piscataway, NJ, USA). Western blotting was performed
using a semi-dry transfer method [48]. Horseradish peroxi-
dase-coupled secondary antibodies and ECL chemilumines-
cent substrate (GE Healthcare) were used to detect
M. Lazarou et al. Mitochondrial complex IV assembly
FEBS Journal 276 (2009) 6701–6713 ª 2009 The Authors Journal compilation ª 2009 FEBS 6711
immunoreactive proteins on blots. Images were captured
using a ChemiGenius chemiluminescence detection system
(SynGene, Cambridge, UK).
Acknowledgements
This work was supported by gra from the Australian
National Health and Medical Research Council
(NHMRC) and the Australian Research Council. D.T.
is supported by an NHMRC Principal Research
Fellowship, and M.M. by an NHMRC Career Devel-
opment Award fellowship.
References
1 Tsukihara T, Aoyama H, Yamashita E, Tomizaki T,
Yamaguchi H, Shinzawa-Itoh K, Nakashima R, Yaono

R & Yoshikawa S (1996) The whole structure of the
13-subunit oxidized cytochrome c oxidase at 2.8 A
˚
.
Science 272, 1136–1144.
2 Rehling P, Brandner K & Pfanner N (2004) Mitochon-
drial import and the twin-pore translocase. Nat Rev
Mol Cell Biol 5, 519–530.
3 Beauvoit B, Bunoust O, Guerin B & Rigoulet M (1999)
ATP-regulation of cytochrome oxidase in yeast mito-
chondria: role of subunit VIa, Eur. J Biochem 263 ,
118–127.
4 Allen LA, Zhao XJ, Caughey W & Poyton RO (1995)
Isoforms of yeast cytochrome c oxidase subunit V affect
the binuclear reaction center and alter the kinetics of
interaction with the isoforms of yeast cytochrome c.
J Biol Chem 270, 110–118.
5 Yoshikawa S, Shinzawa-Itoh K & Tsukihara T (1998)
Crystal structure of bovine heart cytochrome c oxi-
dase at 2.8 A
˚
resolution. J Bioenerg Biomembr 30,
7–14.
6 Schagger H & Pfeiffer K (2000) Supercomplexes in the
respiratory chains of yeast and mammalian mitochon-
dria. EMBO J 19 , 1777–1783.
7 Acin-Perez R, Fernandez-Silva P, Peleato ML, Perez-
Martos A & Enriquez JA (2008) Respiratory active
mitochondrial supercomplexes. Mol Cell 32, 529–539.
8 Schafer E, Seelert H, Reifschneider NH, Krause F,

Dencher NA & Vonck J (2006) Architecture of active
mammalian respiratory chain supercomplexes. J Biol
Chem 281, 15370–15375.
9 Schafer E, Dencher NA, Vonck J & Parcej DN (2007)
Three-dimensional structure of the respiratory chain
supercomplex I
1
III
2
IV
1
from bovine heart mitochon-
dria. Biochemistry 46, 12579–12585.
10 Thorburn DR, Sugiana C, Salemi R, Kirby DM, Wor-
gan L, Ohtake A & Ryan MT (2004) Biochemical and
molecular diagnosis of mitochondrial respiratory chain
disorders. Biochim Biophys Acta 1659, 121–128.
11 Robinson BH (2000) Human cytochrome oxidase defi-
ciency. Pediatr Res 48, 581–585.
12 Fontanesi F, Soto IC, Horn D & Barrientos A (2006)
Assembly of mitochondrial cytochrome c-oxidase, a
complicated and highly regulated cellular process. Am J
Physiol Cell Physiol 291, C1129–C1147.
13 Nijtmans LG, Taanman JW, Muijsers AO, Speijer D &
Van den Bogert C (1998) Assembly of cytochrome-c
oxidase in cultured human cells. Eur J Biochem 254,
389–394.
14 Williams SL, Valnot I, Rustin P & Taanman JW (2004)
Cytochrome c oxidase subassemblies in fibroblast
cultures from patients carrying mutations in COX10,

SCO1, or SURF1. J Biol Chem 279, 7462–7469.
15 Stiburek L, Vesela K, Hansikova H, Pecina P, Tesarova
M, Cerna L, Houstek J & Zeman J (2005) Tissue-
specific cytochrome c oxidase assembly defects due to
mutations in SCO2 and SURF1. Biochem J 392,
625–632.
16 Lazarou M, McKenzie M, Ohtake A, Thorburn DR &
Ryan MT (2007) Analysis of the assembly profiles for
mitochondrial- and nuclear-DNA-encoded subunits into
complex I. Mol Cell Biol 27, 4228–4237.
17 Lazarou M, Thorburn DR, Ryan MT & McKenzie M
(2009) Assembly of mitochondrial complex I and
defects in disease. Biochim Biophys Acta 1793, 78–88.
18 McKenzie M, Lazarou M, Thorburn DR & Ryan MT
(2007) Analysis of mitochondrial subunit assembly into
respiratory chain complexes using Blue Native poly-
acrylamide gel electrophoresis. Anal Biochem 364,
128–137.
19 Mick DU, Wagner K, van der Laan M, Frazier AE,
Perschil I, Pawlas M, Meyer HE, Warscheid B & Reh-
ling P (2007) Shy1 couples Cox1 translational regulation
to cytochrome c oxidase assembly. EMBO J 26, 4347–
4358.
20 Weishaupt A & Kadenbach B (1992) Selective removal
of subunit VIb increases the activity of cytochrome c
oxidase. Biochemistry 31, 11477–11481.
21 Barrientos A, Gouget K, Horn D, Soto IC & Fontanesi
F (2009) Suppression mechanisms of COX assembly
defects in yeast and human: insights into the COX
assembly process. Biochim Biophys Acta 1793, 97–107.

22 Dunning CJ, McKenzie M, Sugiana C, Lazarou M,
Silke J, Connelly A, Fletcher JM, Kirby DM, Thorburn
DR & Ryan MT (2007) Human CIA30 is involved in
the early assembly of mitochondrial complex I and
mutations in its gene cause disease. EMBO J 26, 3227–
3237.
23 Johnston AJ, Hoogenraad J, Dougan DA, Truscott
KN, Yano M, Mori M, Hoogenraad NJ & Ryan MT
(2002) Insertion and assembly of human tom7 into the
preprotein translocase complex of the outer mitochon-
drial membrane. J Biol Chem 277, 42197–42204.
Mitochondrial complex IV assembly M. Lazarou et al.
6712 FEBS Journal 276 (2009) 6701–6713 ª 2009 The Authors Journal compilation ª 2009 FEBS
24 Model K, Meisinger C, Prinz T, Wiedemann N, Trus-
cott KN, Pfanner N & Ryan MT (2001) Multistep
assembly of the protein import channel of the mito-
chondrial outer membrane. Nat Struct Biol 8, 361–370.
25 Ryan MT, Muller H & Pfanner N (1999) Functional
staging of ADP ⁄ ATP carrier translocation across the
outer mitochondrial membrane. J Biol Chem 274,
20619–20627.
26 Stiburek L, Hansikova H, Tesarova M, Cerna L &
Zeman J (2006) Biogenesis of eukaryotic cytochrome c
oxidase. Physiol Res 55(Suppl. 2), S27–S41.
27 Fernandez-Vizarra E, Tiranti V & Zeviani M (2009)
Assembly of the oxidative phosphorylation system in
humans: what we have learned by studying its defects.
Biochim Biophys Acta 1793, 200–211.
28 Barrientos A, Zambrano A & Tzagoloff A (2004)
Mss51p and Cox14p jointly regulate mitochondrial

Cox1p expression in Saccharomyces cerevisiae . EMBO J
23, 3472–3482.
29 Leary SC, Kaufman BA, Pellecchia G, Guercin GH,
Mattman A, Jaksch M & Shoubridge EA (2004)
Human SCO1 and SCO2 have independent, cooperative
functions in copper delivery to cytochrome c oxidase.
Hum Mol Genet 13, 1839–1848.
30 Leary SC, Cobine PA, Kaufman BA, Guercin GH,
Mattman A, Palaty J, Lockitch G, Winge DR, Rustin
P, Horvath R et al. (2007) The human cytochrome c
oxidase assembly factors SCO1 and SCO2 have regula-
tory roles in the maintenance of cellular copper homeo-
stasis. Cell Metab 5, 9–20.
31 Hell K, Tzagoloff A, Neupert W & Stuart RA (2000)
Identification of Cox20p, a novel protein involved in
the maturation and assembly of cytochrome oxidase
subunit 2. J Biol Chem 275, 4571–4578.
32 Massa V, Fernandez-Vizarra E, Alshahwan S, Bakhsh
E, Goffrini P, Ferrero I, Mereghetti P, D’Adamo P,
Gasparini P & Zeviani M (2008) Severe infantile
encephalomyopathy caused by a mutation in COX6B1,
a nucleus-encoded subunit of cytochrome c oxidase.
Am J Hum Genet 82, 1281–1289.
33 Minai L, Fish A, Darash-Yahana M, Verchovsky L &
Nechushtai R (2001) The assembly of the PsaD subunit
into the membranal photosystem I complex occurs via
an exchange mechanism. Biochemistry 40, 12754–12760.
34 Leake MC, Chandler JH, Wadhams GH, Bai F, Berry
RM & Armitage JP (2006) Stoichiometry and turnover
in single, functioning membrane protein complexes.

Nature 443, 355–358.
35 Stricker J, Maddox P, Salmon ED & Erickson HP
(2002) Rapid assembly dynamics of the Escherichia coli
FtsZ-ring demonstrated by fluorescence recovery after
photobleaching. Proc Natl Acad Sci USA 99, 3171–
3175.
36 Daley DO (2008) The assembly of membrane proteins
into complexes. Curr Opin Struct Biol 18, 420–424.
37 Tiranti V, Hoertnagel K, Carrozzo R, Galimberti C,
Munaro M, Granatiero M, Zelante L, Gasparini P,
Marzella R, Rocchi M et al. (1998) Mutations of
SURF-1 in Leigh disease associated with cytochrome c
oxidase deficiency. Am J Hum Genet 63, 1609–1621.
38 Zhu Z, Yao J, Johns T, Fu K, De Bie I, Macmillan C,
Cuthbert AP, Newbold RF, Wang J & Chevrette M
et al. (1998) SURF1, encoding a factor involved in the
biogenesis of cytochrome c
oxidase, is mutated in Leigh
syndrome. Nat Genet 20, 337–343.
39 Nijtmans LG, Artal Sanz M, Bucko M, Farhoud MH,
Feenstra M, Hakkaart GA, Zeviani M & Grivell LA
(2001) Shy1p occurs in a high molecular weight com-
plex and is required for efficient assembly of cyto-
chrome c oxidase in yeast. FEBS Lett 498, 46–51.
40 McKenzie M, Lazarou M, Thorburn DR & Ryan MT
(2006) Mitochondrial respiratory chain supercomplexes
are destabilized in Barth syndrome patients. J Mol Biol
361, 462–469.
41 Schagger H & Pfeiffer K (2001) The ratio of oxidative
phosphorylation complexes I–V in bovine heart mito-

chondria and the composition of respiratory chain
supercomplexes. J Biol Chem 276, 37861–37867.
42 Li Y, D’Aurelio M, Deng JH, Park JS, Manfredi G,
Hu P, Lu J & Bai Y (2007) An assembled complex IV
maintains the stability and activity of complex I in
mammalian mitochondria. J Biol Chem 282, 17557–
17562.
43 Diaz F, Fukui H, Garcia S & Moraes CT (2006) Cyto-
chrome c oxidase is required for the assembly ⁄ stability
of respiratory complex I in mouse fibroblasts. Mol Cell
Biol 26, 4872–4881.
44 Ryan MT, Voos W & Pfanner N (2001) Assaying pro-
tein import into mitochondria. Methods Cell Biol 65,
189–215.
45 Schagger H & von Jagow G (1987) Tricine–sodium
dodecyl sulfate–polyacrylamide gel electrophoresis for
the separation of proteins in the range from 1 to
100 kDa. Anal Biochem 166, 368–379.
46 Alconada A, Gartner F, Honlinger A, Kubrich M &
Pfanner N (1995) Mitochondrial receptor complex from
Neurospora crassa and Saccharomyces cerevisiae. Meth-
ods Enzymol 260, 263–286.
47 Daum G, Gasser SM & Schatz G (1982) Import of
proteins into mitochondria. Energy-dependent, two-step
processing of the intermembrane space enzyme cyto-
chrome b2 by isolated yeast mitochondria. J Biol Chem
257, 13075–13080.
48 Harlow E & Lane D (1999) Using Antibodies: A Labo-
ratory Manual. Cold Spring Harbor Laboratory Press,
Cold Spring Harbor, NY.

M. Lazarou et al. Mitochondrial complex IV assembly
FEBS Journal 276 (2009) 6701–6713 ª 2009 The Authors Journal compilation ª 2009 FEBS 6713