Báo cáo khoa học: Transcription of mammalian cytochrome c oxidase subunit IV-2 is controlled by a novel conserved oxygen responsive element pptx

Transcription of mammalian cytochrome c oxidase
subunit IV-2 is controlled by a novel conserved oxygen
responsive element
Maik Hu
¨
ttemann, Icksoo Lee, Jenney Liu and Lawrence I. Grossman
Molecular Medicine and Genetics, Wayne State University School of Medicine, Detroit, MI, USA
Oxygen sensing and the adaptation to varying oxygen
concentrations are fundamental for the survival of
species from bacteria to humans. Oxygen regulation
can occur at both the intake stage and the usage
stage. In higher organisms, which depend largely on
aerobic energy metabolism, usage takes place largely
at the last step of the mitochondrial respiratory
chain, the transfer of electrons from cytochrome c to
oxygen, catalyzed by cytochrome c oxidase (CcO; EC
1.9.3.1).
In vertebrates, oxygen supply is also regulated via a
unique mechanism in the lungs for hypoxic response.
Other tissues (and, indeed, the bronchial circulatory
system of the lung) react to a hypoxic trigger by vaso-
dilation, thereby increasing blood flow to under-oxy-
genated regions. In the pulmonary circulation of the
lungs, however, the converse effect, hypoxic vasocon-
striction, is critical for shunting blood to more highly
ventilated regions to help optimize the ventilation of
deoxygenated blood.
Keywords
electron transport chain; hypoxia; isoform;
lung; mitochondria
Correspondence

L. Grossman or M. Hu
¨
ttemann, Molecular
Medicine and Genetics, Wayne State
University School of Medicine, 540 E.
Canfield Ave., Detroit, MI 48201, USA
Fax: + 1 313 5775218
Tel: + 1 313 5775326 or +1 3135779150
E-mail: ,

Database
CcO4-2 promoter sequences, including
exon I, have been submitted to the
GenBank data library under the accession
numbers: AY219183 (human), AY219183
(cow), AY219183 (rat) and AY219183
(mouse)
(Received 18 July 2007, revised 30 August
2007, accepted 5 September 2007)
doi:10.1111/j.1742-4658.2007.06093.x
Subunit 4 of cytochrome c oxidase (CcO) is a nuclear-encoded regulatory
subunit of the terminal complex of the mitochondrial electron transport
chain. We have recently discovered an isoform of CcO 4 (CcO4-2) which is
specific to lung and trachea, and is induced after birth. The role of CcO as
the major cellular oxygen consumer, and the lung-specific expression of
CcO4-2, led us to investigate CcO4-2 gene regulation. We cloned the
CcO4-2 promoter regions of cow, rat and mouse and compared them with
the human promoter. Promoter activity is localized within a 118-bp proxi-
mal region of the human promoter and is stimulated by hypoxia, reaching
a maximum (threefold) under 4% oxygen compared with normoxia. CcO4-

2 oxygen responsiveness was assigned by mutagenesis to a novel promoter
element (5¢-GGACGTTCCCACG-3¢) that lies within a 24-bp region that is
79% conserved in all four species. This element is able to bind protein, and
competition experiments revealed that, within the element, the four core
bases 5¢-TCNCA-3¢ are obligatory for transcription factor binding. CcO
isolated from lung showed a 2.5-fold increased maximal turnover compared
with liver CcO. We propose that CcO4-2 expression in highly oxygenated
lung and trachea protects these tissues from oxidative damage by accelerat-
ing the last step in the electron transport chain, leading to a decrease in
available electrons for free radical formation.
Abbreviations
CcO, cytochrome c oxidase; CcO4-2, CcO subunit 4-2 gene; Egr1, early growth response factor 1; EMSA, electrophoretic mobility shift
assay; HIF-1a, hypoxia-inducible factor 1a; ORE, oxygen responsive element; OREF, ORE binding factor; RACE, rapid amplification of cDNA
ends; ROS, reactive oxygen species.
FEBS Journal 274 (2007) 5737–5748 ª 2007 The Authors Journal compilation ª 2007 FEBS 5737
At the gene level, a major breakthrough for the
understanding of oxygen sensing was the discovery
of hypoxia-inducible factor 1a (HIF-1a) [1], which,
during hypoxia, activates at least 60 genes involved
in energy metabolism, erythropoiesis and vasculariza-
tion. It is clear, however, that HIF-1 is not adequate
to explain ischemic cellular pathophysiology. For
example, ferrets develop profound pulmonary hyper-
tension when chronically exposed to oxygen concen-
trations of 10%, whereas, in ex vivo ferret lungs,
HIF-1a expression is not observed until oxygen con-
centration goes below 7%, rises somewhat at 4%
but still does not increase dramatically until 0–1.3%
[2]. In addition, the deposition of fibrin in the lung
vasculature, which is a consequence of the induction

of procoagulant tissue factor, is independent of
HIF-1 [3], and has been shown to result from
hypoxia-mediated induction or activation of the tran-
scription factor early growth response factor 1 (Egr1)
[4]. Our finding, reported here, that a new isoform
of CcO subunit 4 (CcO4-2), which is expressed in
lung and trachea [5], is regulated by oxygen concen-
tration adds a new component to the mechanism of
pulmonary oxygen sensing. Furthermore, the recent
demonstration that HIF-1 stimulates CcO4-2 expres-
sion under hypoxia [6], whilst down-regulating mito-
chondrial metabolism [7–9], underscores the role of
mitochondria in the hypoxic response. The mecha-
nism for oxygen sensing in the lung is still unre-
solved. Considerable data have suggested that the
sensor resides in the mitochondrial electron transport
chain [10–13], although contrary evidence has
recently been presented [14]. Whilst this work was in
progress, Horvat et al. [15] suggested that CcO4-2 is
present in several mouse brain cell types under
hypoxic conditions.
We demonstrate in this report that CcO4-2 is oxy-
gen-regulated at the transcriptional level. It is induced
under hypoxia via a conserved, novel, oxygen respon-
sive element (ORE) in the proximal promoter that
can be separated from the well-characterized HIF-1
system.
Results
CcO4-2 is a respiratory system isoform
Strong transcription signals of CcO4-2 in smooth mus-

cle of the lung [5] led us to investigate transcript levels
in other nonlung smooth muscle tissues by quantitative
PCR. Compared with lung (100%), small intestines
showed only trace amounts (< 0.5%), aorta 4.8% and
trachea 100%, which extends CcO4-2 expression in the
respiratory system but not to the other smooth muscle
tissues examined.
Phylogenetic footprinting utilizing ‘one-way PCR’
reveals a 24-bp conserved region in the CcO4-2
promoter
We cloned and sequenced the promoter region of cow,
rat and mouse utilizing one-way PCR (see Experimen-
tal procedures; Fig. 1A). The alignment of the proxi-
mal promoter sequences, including human, revealed
low identities, except for a 24-bp region conserved in
all four species (Fig. 1B, black bar). Within the 400-bp
Fig. 1. (A) Schematic representation of ‘one-way PCR’. This method is an extension of RACE PCR, and allows the rapid characterization of
unknown upstream or downstream genomic DNA sequences (see Experimental procedures). A short sequence length must be known, e.g.
an exonic sequence derived from a cDNA. Three subsequent primers directed to the region of interest are generated from the known
sequence; here, the CcO4-2 promoter sequences of cow, mouse and rat were amplified. The outermost primer P-1 is used to linearly
amplify into the unknown region using genomic DNA as template. The length-heterogeneous single-stranded fragments are polyadenylated.
In one reaction, a poly(T) primer, which contains an appended sequence (outer ⁄ inner) for later PCRs, is annealed to the poly(A) tails of the
fragments, counter strand synthesis is performed as a single PCR cycle, followed by an outer PCR with the next inner primer P-2 and Q
outer
.
To increase specificity, a nested PCR is performed with primers P-3 and Q
inner
and 1 lL of a 1 : 50 dilution of the outer PCR as template.
PCR products, separated by agarose gel electrophoresis, usually appear as a smear. Fragments of the desired size are gel extracted, cloned
and sequenced. (B) Alignment of the CcO4-2 promoter sequences. Proximal promoter sequences, including exon I from human, cow, rat

and mouse, were aligned with the program
MEGALIGN using the CLUSTAL algorithm. Identical bases are indicated with an asterisk. Transcription
factor binding sites are underlined and specified for the human promoter. Probes used for EMSA experiments are boxed. Mutations used
for control probes for EMSA and mutations introduced by site-directed mutagenesis for reporter gene analysis are indicated above the
underlined elements in italics. Exon I sequences are italicized. The start ATG is located in exon II (not shown). The starting points of reporter
gene constructs 4–7 in Fig. 1C are indicated with arrows and fragment sizes. A 24-bp region, conserved in all four species (black bar), is
composed of a novel oxygen responsive element (ORE) and the adjacent Sp1
A
site. A HIF-1a element that has recently been suggested to
regulate CcO4-2 [6] is underlined. (C) Human CcO4-2 gene activity is driven by the ) 140-bp proximal promoter region. Firefly luciferase
reporter gene activity was normalized to cotransfected Renilla luciferase reporter gene activity, and relative reporter activities were standard-
ized against the wild-type 579-bp construct reporter gene activity (set to 100%) after transfection and incubation under 20% oxygen for
40 h. Control, reporter gene vector without inserted DNA (basal activity).
Cytochrome c oxidase subunit IV-2 hypoxic response M. Hu
¨
ttemann et al.
5738 FEBS Journal 274 (2007) 5737–5748 ª 2007 The Authors Journal compilation ª 2007 FEBS
promoter regions, including exon I, human and cow
share 52% and rat and mouse share 72% sequence
identity, whereas overall identities become insignificant
between the human ⁄ cow and rat ⁄ mouse groups (26–
28%).
Human CcO4-2 gene activity is driven by the
118-bp proximal promoter region
Seven deletion fragments from the human CcO4-2 pro-
moter were cloned in front of the luciferase gene, and
AC
B
M. Hu
¨

ttemann et al. Cytochrome c oxidase subunit IV-2 hypoxic response
FEBS Journal 274 (2007) 5737–5748 ª 2007 The Authors Journal compilation ª 2007 FEBS 5739
reporter activity was tested in H460 human lung cells
(Fig. 1C). Constructs starting from 2650 to 118 bp
revealed similar reporter activities, followed by an 80%
drop in activity obtained with the 76-bp construct. We
concluded that the main regulatory elements are within
the )118-bp proximal region (see below).
A novel ORE mediates hypoxia-induced human
CcO4-2 gene activity
In order to investigate oxygen as a potential regulator
for CcO4-2 gene activity, we analyzed the 579- and
203-bp reporter constructs under both standard 20%
oxygen cell culture conditions and at 2% oxygen. Both
reporter constructs showed more than twofold induc-
tion under hypoxia (Fig. 2A, constructs 1 and 2).
Analysis of the human promoter with the program
TFsearch [16] for potential transcription factor binding
sites revealed three Sp1-like elements (Sp1
A,B,C
), one of
which is part of the 24-bp conserved region (Fig. 1B).
The three elements were altered by site-directed muta-
genesis of the 579-bp promoter construct, and activity
was evaluated at 2% and 20% oxygen. These experi-
ments revealed that each is necessary for maximal pro-
moter activity, but none is abolished by hypoxic
induction after mutation (Fig. 2A, constructs 4, 5 and
6). An additional construct was generated that con-
tained mutations in the conserved 24-bp region (ORE in

Fig. 1B) upstream of the Sp1
A
site. Mutation of this
element significantly reduced reporter activity and,
importantly, also eliminated CcO4-2 hypoxic induction,
assigning oxygen responsiveness to the new element
(Fig. 2A, construct 3). Double mutations (Fig. 2A, con-
structs 7 and 8) are approximately additive and, when
ORE is one of them, show a loss of hypoxic stimulation.
CcO4-2 hypoxic response is threefold induced
at 4% oxygen
We analyzed the CcO4-2 hypoxic response in H460
cells between zero (set to 100%) and 20% oxygen.
350
20% Oxygen
4% Oxygen
Reporter gene activity (%)
300
250
200
150
100
50
0
A
B
C
Fig. 2. (A) Human CcO4-2 promoter activity is stimulated by hypoxia
and mediated by a novel oxygen responsive element (ORE). Wild-
type and site-directed mutagenesis constructs of the human CcO4-

2 promoter were transfected into human H460 cells that were
grown under 20% (hatched bars) or 2% (filled bars) oxygen, and
reporter gene activity was determined as in Fig. 1. A 2.2-fold induc-
tion in promoter activity was observed for the 579-bp wild-type
reporter (construct 1). Hypoxic stimulation is abolished when ORE is
mutated (black rectangle, construct 3). All indicated elements contrib-
ute to maximum promoter activity (open circles, Sp1-like). (B) CcO4-2
reporter gene activity is about threefold increased at 4% oxygen
compared with normoxia, whereas the HIF-1a construct shows
highest activity under anoxia. The reporter gene activity of the 579-
bp human CcO4-2 (see Fig. 1C) and the HIF-1a response element
(HRE) reporter constructs was examined in H460 cells under various
oxygen concentrations, and standardized against identically treated
cells incubated at 20% oxygen in parallel. (C) The 17-bp ORE was
cloned in a promoterless reporter gene vector. A one- (left) and four-
(middle) copy-containing construct with the correctly oriented
sequence and the control plasmid (right) were transfected into
H460 cells and incubated under 4% and 20% oxygen concentra-
tions, revealing that the ORE by itself is able to both stimulate tran-
scription and mediate the hypoxic response. Reporter activity for
the one-copy-containing construct was 8.9% (± 0.5%) relative to
the 579-bp construct (construct 1 in Fig. 2A) at 20% oxygen.
Cytochrome c oxidase subunit IV-2 hypoxic response M. Hu
¨
ttemann et al.
5740 FEBS Journal 274 (2007) 5737–5748 ª 2007 The Authors Journal compilation ª 2007 FEBS
Compared with normoxia, reporter gene activity was
induced threefold at 4% oxygen, declining to about
1.7-fold induced at 0% (Fig. 2B, filled diamonds). To
compare the CcO4-2 response with that of HIF-1, a

reporter containing three HIF-1a regulatory elements
was also introduced under the same conditions
(Fig. 2B, open squares).
In both cases, the reporter assay was performed after
40 h of incubation at each indicated oxygen concentra-
tion. To determine the time delay necessary for 1.1 mL
of fresh oxygen-saturated medium overlying the cells in
a 24-well plate to equilibrate with the gas atmosphere,
we positioned an oxygen electrode into a standard well
and recorded media oxygen concentration over time
(data not shown). Constant oxygen equilibrium concen-
trations were obtained after about 3 h for 20% and 4%
oxygen atmospheres, but not until more than 8 h for
0% oxygen. Thus, the incubation period of 40 h prior
to reporter gene analysis allowed > 24 h exposure to
the experimental oxygen concentration.
We next cloned ORE in a promoterless vector to
test whether its presence is sufficient to mediate a hyp-
oxic response in the absence of other cis elements in
the human promoter. A construct containing one copy
of ORE produced a 2.8-fold induction of reporter gene
activity at 4% oxygen compared with normoxia
(Fig. 2C, left columns). However, the presence of three
additional copies of the element in the same orienta-
tion did not further increase the hypoxic response
(Fig. 2C, second column pair). A construct containing
mutations in ORE produced background reporter gene
activity at both 4% and 20% oxygen, similar to the
empty vector (Fig. 2C, right two column pairs). Thus,
the element alone can act as a promoter and can medi-

ate the hypoxic response.
Electrophoretic mobility shift assay (EMSA) with
ORE probe
The analysis of protein–DNA interaction for ORE was
performed with a probe containing part of the adjacent
Sp1
A
site and nuclear extract derived from H460 cells
grown under 2% oxygen. Two complexes were formed
(Fig. 3A, lane 2) that could be competed with unla-
beled probe (lane 3). The upper band was shown to be
a nonspecific artifact that could be competed with any
oligonucleotide (Fig. 3D). Competition with the probe,
but with ORE mutated again, competed the nonspe-
cific upper band (Fig. 3A, lane 4, triangle) but not
ORE with its binding factor (OREF) in the lower band
(arrow). Antibodies against Sp3, Sp4, HIF-1a and
CREB protein, to examine whether the cognate protein
was part of the shifted band complex, gave negative
results in all cases (Fig. 3A, lanes 8–12). Antibodies
against transcription factors c-Rel and Ikaros, which
were suggested by Genomatix software as candidates
for binding to this sequence, also did not produce a
supershift or interference with binding (not shown).
Furthermore, neither the use of nuclear extract derived
from H460 cells grown under 20% instead of 2% oxy-
gen (not shown), nor the addition of potential regula-
tory nucleotides ATP, ADP (not shown), NAD
+
and

NADH (Fig. 3A, lanes 13–15), affected the relative
intensity of the specific band.
A partial characterization of the 5¢-upstream bases
required for the interaction of OREF with ORE
(Fig. 3B, lane 2) expands the core sequence identified
by site-directed mutagenesis (Fig. 2A). We conclude
from the above experiments and phylogenetic foot-
printing (Fig. 1B) that the sequence 5¢-GGA(C ⁄ T)
GTTCCCACGT-3¢ represents the minimum OREF
recognition sequence. To further narrow the core bind-
ing site, we performed experiments with competitor
oligonucleotides for each position of the 15-bp region:
for each nucleotide, a mixture of three competitors
was generated containing the three bases not present
in the human sequence. By applying a large excess
(100-fold) of competitor mixture, only those reactions
in which the particular base is absolutely required for
protein binding will produce a shifted band (Fig. 3C,
upper panel). Applying this method, we identified four
essential bases 5¢-TCCCA-3¢ in the middle of the ORE
sequence (Fig. 3C, boxed). A reduction in stringency
by decreasing the amount of competitor DNA to a
10-fold excess revealed the participation of other bases
with different signal intensities, and two bases
(italicized) that do not seem to be required for factor
binding (Fig. 3C, lower panel). Thus, the final consen-
sus sequence from the above data is 5¢-GGA(C ⁄ T)
NTTCNCACG(C ⁄ T).
Lung CcO is a high-activity isozyme
We isolated CcO from cow lung for the first time to

our knowledge and also from cow liver, in both cases
following our previous protocol [17]. Both isoforms of
CcO subunit IV are expressed in lung. However, we
obtained only a single band in the size range of sub-
unit IV (Fig. 4A), most probably because of the very
similar sizes of both isoforms. We then performed
activity measurements. Functional differences between
the lung and liver isozymes became obvious after
kinetic analysis. CcO activity (turnover number) was
measured by the polarographic method (Fig. 4B).
Lung CcO is more than 2.5 times as active at maximal
turnover than liver CcO.
M. Hu
¨
ttemann et al. Cytochrome c oxidase subunit IV-2 hypoxic response
FEBS Journal 274 (2007) 5737–5748 ª 2007 The Authors Journal compilation ª 2007 FEBS 5741
Discussion
Our discovery of a new mammalian CcO isoform for
regulatory subunit 4 [5], whose expression is specific to
lung and trachea amongst the tissues examined, which
is induced after birth, and whose transcription is
dependent on oxygen concentration, points to a role
for CcO in respiratory physiology. Strong expression
within lung has been localized to smooth muscle, in
both arteriole and bronchiole walls. Each of these
locations may involve a different functional role and
thus, consequently, a different regulatory circuitry. In
bronchiole walls, our previous observation of CcO
inhibition by the anti-asthma drug theophylline,
A

B
C
D
Cytochrome c oxidase subunit IV-2 hypoxic response M. Hu
¨
ttemann et al.
5742 FEBS Journal 274 (2007) 5737–5748 ª 2007 The Authors Journal compilation ª 2007 FEBS
followed by decreased ATP levels [17], suggests a role
in airway constriction.
In arteriole walls, muscle can regulate blood flow in
response to hypoxic signals to ameliorate potential
ischemic damage. In most cases, hypoxia triggers vaso-
dilation in order to increase blood flow to under-oxy-
genated tissue regions. Uniquely, in the pulmonary
arterioles of the lung, however, hypoxia triggers vaso-
constriction in order to shunt blood to better ventilated
(apical) regions to improve the ventilation ⁄ perfusion
ratio. As alveolar sacs do not contain a muscle compo-
nent, hypoxic signals must be transmitted upstream to
the terminal bronchioles. The nature of the initial hyp-
oxic signal is unresolved, although considerable data tie
it to the mitochondrial electron transport chain and,
in particular, to the free radicals generated there
[12,18,19]. Two general factors are known to increase
free radical production in the mitochondria, the oxygen
concentration and the redox state of the electron trans-
port chain, because both oxygen and electrons are
required substrates for radical formation. Lung cells
face the highest (atmospheric) oxygen concentrations
compared with other tissues. Therefore, special protec-

tive adaptations would be expected that prevent excess
free radical formation. One way to prevent electron
build-up in the mitochondria would be to increase CcO
activity, which has been shown to be the rate-limiting
step in the electron transport chain under physiological
conditions [20,21]. We have shown here that isolated
CcO from lung is 2.5 times more active than liver CcO,
and recent experiments with cells overexpressing
CcO4-2 have shown that CcO4-2 is superior to CcO4-1
at dissipating H
2
O
2
build-up [6]. Taken together, the
expression of CcO4-2 in the highly oxygenated tissues
of lung and trachea, the increased activity of the iso-
lated enzyme and the finding that CcO4-2-expressing
cells produce fewer free radicals suggest a role in pro-
tecting these tissues from radical damage. Interestingly,
CcO4-2 evolved by gene duplication about 320 million
years ago [5], a time when atmospheric oxygen concen-
trations dramatically increased from an estimated hyp-
oxic level of 13% in the Devonian to 35% hyperoxia
Fig. 3. (A) EMSA with the [
32
P]-labeled oxygen responsive element (ORE) probe yields two specific bands. The19-bp [
32
P]-labeled ORE probe
(bottom) contains four G nucleotides of the adjacent Sp1 site. The nuclear extract was prepared from H460 lung cancer cells grown under
2% (+) oxygen for 2 days. In each binding reaction, 60 lg of nuclear extract was applied. The two strongly shifted bands (lane 2) can be

removed with a 50-fold excess of unlabeled competitor DNA (lane 3). Competitor DNA that contains mutations in ORE (indicated on the top
of the probe sequence; lane 4) or competition experiments with an Sp1 fragment derived from the middle Sp1 site of the human promoter
(Fig. 1B) in wild-type (lane 5) or mutated (lane 6) form largely eliminates the upper band (triangle), and thus reveals specific interaction of
ORE with a corresponding transcription factor (arrow). Supershift experiments with Sp antibodies show interference only with an Sp1 anti-
body (lane 8). Normal goat serum (N) and Sp3, Sp4, HIF-1a and CREB binding protein antibodies show no effect (lanes 7, 9–12). Similarly,
the addition of 2 m
M NAD
+
(lane 14) or NADH (lane 15) shows no effect on the lower band. Free probe at the bottom of each lane is not
shown. (B) EMSA experiments performed with probes differing in sequence length revealed that, in addition to the core sequence as shown
above the 5¢-upstream bases, GGA is necessary for transcription factor binding, because the lower band representing OREF–ORE interaction
is abolished on its removal (lane 2). Probe sequences are indicated. Stars show bases conserved in the human, cow, rat and mouse promot-
ers. (C) The definition of bases indispensable for transcription factor binding to ORE was performed via competition experiments using 15
oligonucleotides containing mutations at each position of the binding site as defined in (B) (lanes 2–16). Each competitor consists of a mix-
ture of three oligonucleotides containing the bases not present in the ORE sequence (H, C ⁄ A ⁄ T; B, T ⁄ C ⁄ G; V, A ⁄ C ⁄ G; D, G ⁄ A ⁄ T). For exam-
ple, the competitor in lane 2 contains three double-stranded oligonucleotides with C, A and T in the first position, excluding G present in the
wild-type ORE. EMSA was performed under the conditions given in (A) using a 10- or 100-fold excess of unlabeled competitor as indicated.
Only the lower band specific for OREF binding is shown (see (A), arrow). The absence of competitor DNA produces a similar signal, as
observed with unspecific competitor under 10- or 100-fold excess (compare lanes 1 and 17; see Table 1 for oligonucleotide sequences).
Using a 100-fold excess of competitor DNA reveals that four bases (boxed) are indispensable for OREF binding (lanes 8, 9, 11 and 12),
because the corresponding competitor DNAs cannot compete with complex formation. Reducing the stringency by applying a 10-fold excess
of competitor reveals signals with varying intensities for most positions (lanes 2–5, 7–9, 11–16), except for two bases (italicized) that do not
seem to contribute to OREF binding (lanes 6 and 10). (D) EMSA experiments with ORE and Sp1 probes reveal that the upper band is non-
specific for Sp1. The higher molecular weight complex could be competed with an unlabeled Sp1
B
probe or the addition of Sp1 antibody
[see (A), triangle; lanes 5 and 8, respectively]. Thus, in order to test whether the upper band is specific and contains Sp1, side-by-side com-
petition experiments were performed using the ORE probe and an Sp1 consensus probe identical in length and GC ⁄ AT content (Table 1)
with nuclear extract from H460 cells grown at 20% oxygen. In comparison with nuclear extract grown under hypoxia, normoxic nuclear
extract leads to an increase in the upper band, but does not affect the intensity of the lower (specific) band. The Sp1 probe shows three

bands, with the lowest band migrating at the position of the upper band obtained with the ORE probe (triangle; compare lanes 9 and 2,
respectively). Using both probes, this band could be competed with wild-type and mutated Sp1 oligonucleotides (lanes 3, 4, 10, 11). In addi-
tion, a mixture of unspecific oligonucleotides that do not contain Sp1-like sequences (see Table 1) efficiently abolishes the band already at
low excess (lanes 6, 7, 13 and 14), indicating that the upper band obtained with ORE and the lowest band obtained with the Sp1 probe
result from unspecific (sequence-independent) protein binding. This band is weakened on addition of Sp1 antibody (lanes 5 and 12); how-
ever, using the Sp1 probe only, the top band (arrow) produces a specific supershift (star; compare lanes 9 and 12).
M. Hu
¨
ttemann et al. Cytochrome c oxidase subunit IV-2 hypoxic response
FEBS Journal 274 (2007) 5737–5748 ª 2007 The Authors Journal compilation ª 2007 FEBS 5743
during the Permian and Carboniferous, compared with
the present 21%. Compared with other CcO subunit
isoforms, mammalian CcO subunit 4 led to the genera-
tion of the earliest isoform pair during a period with
dramatic changes in oxygen concentration, the sub-
strate of CcO, which further suggests an adaptation to
higher oxygen concentrations.
A recent report has suggested that CcO4-2 regula-
tion is mediated by HIF-1a [6], and the authors identi-
fied two such elements, one in the promoter (Fig. 2,
double underline) and one in intron 1. The involve-
ment of HIF-1a or HIF-2a in CcO4-2 regulation needs
further evaluation because both elements are not
conserved in mammals, and it remains to be shown
whether, under physiological conditions, lung cells face
such low oxygen concentrations under which HIF reg-
ulation is operating (Fig. 2B). We found maximal
CcO4-2 reporter gene activity at 4% oxygen, condi-
tions under which regulation by HIF does not occur
(Fig. 2B). Oxygen concentration in lung is clearly

higher than that in other tissues, and 4% oxygen, as
applied during our cell culture experiments, might rep-
resent the physiological equivalent range present at the
cellular level in lung tissue.
The properties of CcO4-2 in withstanding oxidative
stress (faster ability to utilize O
2
and produce ATP,
whilst producing less H
2
O
2
and caspase activation)
suggest that its expression may be found in other cell
types where survival is critical. The question then
arises as to why CcO4-2 has not become the dominant,
tissue-unspecific isoform. The answer may be that it is
not very conservative of energy because of its increased
basal activity (Fig. 4B). Mitochondrial reactive oxygen
species (ROS) have been shown to trigger hypoxia-
stimulated responses, including transcription and cal-
cium increases in pulmonary arterial myocytes [18,22].
Inhibitor studies to localize the ROS-producing seg-
ment of the electron transport chain place it proximal
to the ubisemiquinone site of complex III [11]. How-
ever, inhibitors acting distal to ubisemiquinone, such
as the CcO inhibitors cyanide and azide, can augment
ROS generation by increasing the ubisemiquinone pool
[10]. As discussed above, regulation of electron flux at
CcO by subunit 4-2, in analogy with ATP regulation

of CcO activity by subunit 4-1 [23], would be a way of
modulating the redox state of the ubiquinone pool.
Our results stimulate the determination of how oxy-
gen concentration regulates CcO4-2 expression. The
discovery of a novel 24-bp region in the proximal pro-
moter, conserved between human, cow, rat and mouse,
containing an element (ORE) shown by mutagenesis to
be required for oxygen regulation, leads to the ques-
tion of what factor binds to this element and how it
mediates this response. The use of nuclear extract
obtained from H460 cells grown under normoxia or
hypoxia did not show differences in ORE–OREF bind-
ing, indicating that the amount of OREF does not
change as a function of the oxygen concentration. Pos-
sibly, the interaction of OREF with other factors is
modified as a function of the oxygen concentration. If
OREF is not the oxygen sensor, but a downstream
factor of the actual oxygen sensor, OREF could be
A
B
Fig. 4. (A) SDS-PAGE of isolated CcO from cow lung in comparison
with liver CcO. CcO samples from liver and lung were isolated
side-by-side and applied to SDS-PAGE. Lane M, molecular size mar-
ker; lanes 1 and 2, 37% and 45% ammonium sulfate-precipitated
lung CcO; lane 3, 45% ammonium sulfate-precipitated cow liver
CcO. (B) Respiration kinetics of solubilized cow lung CcO in com-
parison with liver CcO. CcO activity was measured with the polaro-
graphic method at 25 °C by increasing the amount of substrate
cytochrome c. CcO activity (TN, turnover number) is defined as the
amount of O

2
consumed (lmol) per second per amount of CcO
(lmol). The data shown were obtained with the 45% ammonium
sulfate-precipitated fractions. CcO activity was analyzed in a closed
200 lL chamber containing a micro-Clark-type oxygen electrode
(Oxygraph system, Hansatech). Representative data from a total of
four independent experiments are shown.
Cytochrome c oxidase subunit IV-2 hypoxic response M. Hu
¨
ttemann et al.
5744 FEBS Journal 274 (2007) 5737–5748 ª 2007 The Authors Journal compilation ª 2007 FEBS
targeted for phosphorylation, altering its interaction
with other factors involved in complex formation. As
several Sp1-like sites are present in the CcO4-2 pro-
moter, indirect regulation is possible, involving modifi-
cations that modulate the potential OREF–Sp1
interaction, such as Sp1 phosphorylation [24].
In lower organisms there are two examples of oxygen-
mediated CcO isoform gene expression. In yeast, there
are isoforms of subunit 5, called 5a and 5b, which are
expressed under normoxia and hypoxia, respectively,
and have been proposed to be analogous to mamma-
lian CcO subunit 4 [25]. However, mammalian CcO4-
1 ⁄ 2 and yeast CcO5a ⁄ b do not share any homology at
the protein level. In the slime mold Dictyostelium dis-
coideum, there is an isoform pair of CcO subunit 7.
Here, CcO7e, which is expressed under normoxic con-
ditions, is replaced by CcO7s under hypoxia, and this
switching is mediated by an oxygen-dependent tran-
scriptional element located in the short intergenic

region between the two adjacent genes [26]. Analysis of
the yeast and Dictyostelium CcO hypoxia-regulated
promoters revealed that the mammalian ORE sequence
is absent in both organisms, which agrees with our
functional model that the expression of CcO4-2 repre-
sents a unique protective adaptation found in the
highly oxygenated respiratory system in higher organ-
isms, rather than being an adaptation to very low oxy-
gen levels, under which yeast CcO5b and Dictyostelium
CcO7s are expressed.
Experimental procedures
Cell lines and reagents
Human lung adenocarcinoma-derived cell line H460 was
grown in RMPI 1640 medium (Gibco BRL, Carlsbad, CA,
USA) containing 0.1% glucose supplemented with 10%
fetal bovine serum and 1% penicillin–streptomycin in a 5%
CO
2
atmosphere. Parallel experiments involving varying O
2
concentrations were performed in a hypoxic chamber under
the control of ProOx 110 oxygen and ProCO
2
carbon
dioxide controllers (BioSperix, Redfield, NY, USA).
Media were supplemented with 50 mgÆmL
)1
uridine and
110 mgÆmL
)1

pyruvate in experiments performed at 0%
oxygen [27]. HeLa cells were grown in DMEM (Gibco
BRL) containing 0.1% glucose supplemented with 10%
fetal bovine serum and 1% penicillin–streptomycin.
RNA isolation and quantitative PCR
Lung, heart, small intestine, aorta and trachea RNA isola-
tions and quantitative PCR with specific primers for CcO4-
2 and CcO4-1 were performed as described previously [5].
Cloning of the promoter regions with a novel
method: ‘one-way PCR’ (Fig. 1A)
In order to amplify the unknown CcO4-2 5¢-genomic pro-
moter region from cow, mouse and rat, we developed a
novel method, which is an extension of 5¢-rapid amplifica-
tion of cDNA ends (5¢-RACE) used to generate 5¢-cDNAs
[28]. This approach utilizes a dT
17
-oligonucleotide (Q
T
pri-
mer) which contains an appended sequence that allows the
use of specific primers (Q
inner
and Q
outer
) in subsequent
PCRs. Genomic DNA, isolated from muscle tissue of all
species using the Wizard Genomic DNA Isolation Kit (Pro-
mega, Madison, WI, USA), was used as template instead of
RNA in the RACE protocol. Three primers directed to the
unknown 5¢-region (P-1

cow
, P-2
cow
, P-3
cow
, P-1
rat
, P-2
rat
,
P-3
rat
, P-1
mouse
, P-2
mouse
and P-3
mouse
) were derived from
known cDNA exon I (cow and rat) or intron I (mouse)
sequences [5]. In a first linear PCR amplification, the outer-
most primer P-1 was used without a counter primer in a
50 lL PCR for each species employing the Expand Long
Template PCR System (Roche, Indianapolis, IN, USA) in
combination with the kit’s buffer 3 and a 0.5 mm final con-
centration of each dNTP. Initial denaturation at 93 °C for
2 min was followed by 5 s at 93 °C, 30 s at 65 °C and
2 min at 68 °C, with 30 cycles in total. Buffer components
and dNTPs were removed from the mix, a poly(A) tail was
appended to the single-stranded DNA (equivalent to the

cDNA first strands in the 5¢-RACE method) and 5 lL (of
25 lL) of the previous reaction was used to anneal the Q
T
primer and for the extension reaction, as described previ-
ously [29]. Primers P-2 and Q
outer
(15 pm each) were added
to the mixture, and the outer PCR was performed with an
initial denaturation at 93 °C for 1 min, followed by 30 s at
93 °C, 30 s at 58 °C and 2 min at 72 °C, with 30 cycles in
total. A 50 lL nested PCR was then carried out with prim-
ers P-3 and Q
inner
, and 1 lL of a 1 : 30 dilution of the pre-
vious reaction as template, using similar conditions as in
the outer PCR (summarized in Fig. 1A). The amplifications
yielded a smeary size distribution on agarose gel electro-
phoresis, ranging from 500 bp to 3 kb, because of the
absence of a counter primer in the initial linear amplifica-
tion reaction. DNA between 1 and 2 kb was cut out of the
gel and purified using the Nucleotrap Gel Extraction Kit
(Clontech, Mountain View, CA, USA). DNA fragments
were cloned and sequenced as described previously [5].
Reporter gene constructs
A 2.8-kb genomic fragment of the human CcO4-2 promoter,
including exon I, was amplified from human genomic DNA
with primers P
promoter-forward
and P
promoter-reverse

in a 50 lL
touchdown PCR () 1 °C ⁄ cycle), with denaturation for 35 s
at 94 °C, annealing for 30 s at 63–58 °C, extension for
4 min at 70 ° C, and 32 cycles in total, using the Expand
M. Hu
¨
ttemann et al. Cytochrome c oxidase subunit IV-2 hypoxic response
FEBS Journal 274 (2007) 5737–5748 ª 2007 The Authors Journal compilation ª 2007 FEBS 5745
High Fidelity PCR System (Roche). Seven different-sized
promoter fragments (2646, 579, 399, 293, 203, 118, 76 bp,
all including 56 bp of exon I) were generated in separate
nested PCRs using 1 lL of a 1 : 40 dilution of the previous
PCR as template and primers P
prom
+1,2,3,4,5,6,7 in combi-
nation with P
promreverse
under similar touchdown conditions.
All P
prom
primers contained a XhoI5¢-adapter sequence AG-
TCTATTCTCGAG (Table 1). Fragments were gel-purified
(see above), digested with XhoI (Promega) and gel-purified
once more. The longest fragment was only partially digested
as it contained an internal XhoI site. The pGL3basic lucifer-
ase reporter vector (Promega) was digested with XhoI,
dephosphorylated with shrimp alkaline phosphatase (Roche)
and used for ligation of the seven promoter fragments. Cor-
rect orientation of individual clones was tested by PCR, and
positive clones were verified by sequencing.

The HIF-1a construct was a kind gift from Dr Navdeep
Chandel (Northwestern University, Evanston, IL, USA). It
contains three copies of the 5¢-RCGTG-3¢ motif in front
of the luciferase gene in the pGL2 reporter vector (Promega).
A promoterless reporter gene vector was generated by
removing the SV40 promoter from the pGL2 Promoter vec-
tor (Promega) via a HindIII ⁄ XhoI double digestion. The
vector fragment lacking the SV40 promoter was purified as
above, DNA ends were filled using Pfu polymerase (Strata-
gene, La Jolla, CA, USA) and the vector was treated with
shrimp alkaline phosphatase (Roche). The ORE-containing
sequence 5¢-GGACGTTCCCACGCTGG-3¢ and the mutated
sequence 5¢-GGTCGTAACCACGCTGG-3¢ were cloned into
the vector in various configurations and confirmed by
sequencing.
Site-directed mutagenesis
The 579-bp promoter construct was used for the generation
of all further constructs. Primers P
ORE mut
,P
Sp1 distal mut
,
P
Sp1 middle mut
and P
Sp1 proximal mut
(Table 1) were used for
site-directed mutagenesis with the GeneEditor site-directed
mutagenesis kit (Promega), according to the supplier’s pro-
tocol.

Transfection and luciferase assay
H460 cells were plated onto 24-well plates at 4 · 10
4
cells ⁄
well and grown overnight. Cells were transfected using
TransFast (Promega) with 1 lg of the promoter firefly lucif-
erase construct and 0.04 lg of the pRL-SV40 control vector
(Promega), which contains the Renilla luciferase cDNA
downstream of the SV40 promoter. Cells were harvested
40 h after transfection and both luciferase activities were
analyzed with the Dual-Luciferase Reporter Assay System
(Promega), according to the supplier’s protocol, with an
Optocomp 1 luminometer (MGM Instruments, Sparks, NV,
USA). At least four replicates were performed for each.
Preparation of nuclear extract and EMSA
Nuclear protein extracts were prepared from H460 cells as
described previously [30], HeLa nuclear extracts were pur-
chased from Promega and protein concentrations were
determined using the Bradford assay (Bio-Rad, Hercules,
CA, USA). The oligonucleotide primers P
ORE
,P
ORE mut
,
P
Sp1
,P
Sp1 mut
,P
Sp1 19bp

,P
Sp1 19bp mut
and Punspecific com-
petitor, and the 15 primer mixes containing mutations in
each position of the core ORE sequence (see Fig. 3C),
together with their reversed and complemented primers,
were heated to 85 °C and slowly cooled to room tempera-
ture in annealing buffer (10 mm MgCl
2
,50mm NaCl,
20 mm Tris ⁄ Cl, pH 7.5). [c-
32
P]-labelling of double-
stranded oligonucleotides, their purification and subsequent
nuclear extract binding reactions were carried out as
described previously [31]. The DNA-bound complexes were
Table 1. Sequences of oligonucleotides used in RACE, ‘one-way
PCR’, site-directed mutagenesis and EMSA.
Primer ID Sequence (5’- to 3’)
P-1
cow
TCTTGCGGCTTGGAGAGAGCCAG
P-2
cow
CCAGAACGCGACCCAGGTC
P-3
cow
CAGGTCTGCAGAGCAAGCAACAG
P-1
rat

TAGTTGCAAGCTGAAGACCG
P-2
rat
GCTGAAGACCGCGGAGGTAC
P-3
rat
GAGGTACCCAGAACTGCCCTG
P-1
mouse
GATAGTCAGTGGGGGAAACCTCAG
P-2
mouse
CAGCAAAAGAGGGCTGTGTGGTG
P-3
mouse
TGGCCGCCACGAACATCCCATC
P
promoter forward
GTTGCCCAGGTTGGAGTGCAG
P
promoter reverse
CTCGCGGGCTCGGCAGTGGGAG
P
prom+1
AGTCTATTCTCGAGCACCTGGGACTACAGG
P
prom+2
AGTCTATTCTCGAGCCCAAAGCGCTGAGATTACAG
P
prom+3

AGTCTATTCTCGAGATGCTTCTGGAGTAGGAGGCA
P
prom+4
AGTCTATTCTCGAGGTGTGGAGGAGGCAGGGAGAC
P
prom+5
AGTCTATTCTCGAGGAGGCGCTCTGCAGTGCCTC
P
prom+6
AGTCTATTCTCGAGAAGCAGGACGTTCCCACGCTG
P
prom+7
AGTCTATTCTCGAGGGGGCGGGCGCCCGCACTCAG
P
promreverse
AGTCTATTCTCGAGCGCGACCTGGGTCTGCCCAG
P
ORE mut
GGCCGCCCCAGCGTGGTTACGACCTGCTTCGGCAGG
GCGTGG
P
Sp1 distal mut
GCCTTTCTCGGGGCCGCTTCAGCGTGGGAACG
P
Sp1 middle mut
GGCGCCCGCCCCCGGCCATACCACAGCCTTTCTCGG
P
Sp1 proximal mut
GCGGGCGCCCGAACCCTGCCCGCCCCACA
P

intron IIIfoward
ATATTCTAGGATCCTGGCTCATTCACTGCTGTCAC
P
intron IIIreverse
ATATTCTAGGATCCCGGCTTCCCCCTCCCTGCAG
P
HIF-1a mut
GCAAATTCTTACTGAGCTTTTACTATATGCACAGC
P
ORE
GGACGTTCCCACGCTGGGG
P
ORE mut
GGTCGTAACCACGCTGGGG
P
Sp1
GGCTGTGGGGCGGGCCGG
P
Sp1 mut
GGCTGTGGGTATGGCCGG
P
Sp1 19bp
TTCGATCGGGGCGGGGCGA
P
Sp1 19bp mut
TTCGATCGGTTCGGGGCGA
P
unspecific competitor
CTAGCAANNATNNTTGCTAG
Cytochrome c oxidase subunit IV-2 hypoxic response M. Hu

¨
ttemann et al.
5746 FEBS Journal 274 (2007) 5737–5748 ª 2007 The Authors Journal compilation ª 2007 FEBS
resolved by electrophoresis through 8% polyacrylamide gels
(37.5 : 1, acrylamide ⁄ bisacrylamide) using TGE buffer
(25 mm Tris base, 190 mm glycine, 1 mm EDTA). Gels
were vacuum-dried, exposed to storage phosphor screens
overnight and analysed with a Storm 840 Phosphorimager
(Molecular Dynamics, Sunnyvale, CA, USA). Competition
analysis and supershift experiments were performed as
described previously [31]. Antibodies Sp1 (1C6), Sp3 (D-
20), Sp4 (V-20), HIF-1a (H-206), CBP (451), Ikaros (E-20),
c-Rel (B-6), and normal rabbit IgG as control, were
obtained from Santa Cruz Biotech (Santa Cruz, CA, USA).
Quantification of EMSA bands
After the gels had been scanned as described above, the
intensities of individual bands were analysed with image-
quant software (version 5, Molecular Dynamics) and cor-
rected for background signal.
Isolation of CcO and enzymatic activity
measurements
CcO was isolated from cow lung and liver under standard
conditions as described previously [17]. Three micromolar
CcO was dialyzed in the presence of 0.1 mm ATP and
120 lm cardiolipin to remove cholate and to replace poten-
tially damaged cardiolipin, in 10 mm K-Hepes (pH 7.4),
40 mm KCl, 1% Tween 20, 2 mm EGTA and 10 mm KF.
CcO activity was analyzed in a closed 200 lL chamber con-
taining a micro-Clark-type oxygen electrode (Oxygraph sys-
tem, Hansatech, Kings Lynn, UK). Measurements were

carried out using 250 nm CcO at 25 °C after the addition
of ascorbic acid (20 mm) and increasing amounts of Cyt c
from 0 to 30 lm. Oxygen consumption was recorded on a
computer and analyzed with Oxygraph software. The turn-
over number is defined as the amount of oxygen consumed
(lmol) per second per amount of CcO (lmol).
Acknowledgements
We gratefully thank Dr Bruce Berkowitz and Robin
Roberts for providing rat lung tissue, Dr Michelle Ku-
rpakus-Weather for the use of her hypoxic chamber,
Dr Dennis Goebel for rat tissues, Dr Fazlul Sarkar for
H460 cells, Dr Jeffrey Potts and Esta Grossman for
discussions on lung physiology, Dr Navdeep Chandel
for the HIF-1a construct and Dr Barry Rosen for
manuscript suggestions. The research described in this
article was supported in part by NIH grant GM48517
(LG), Philip Morris USA Inc. and Philip Morris Inter-
national (MH), a fellowship (HU885⁄ 2-1) from the
Deutsche Forschungsgemeinschaft (MH) and the Cen-
ter for Molecular Medicine and Genetics of Wayne
State University.
References
1 Semenza GL (2000) HIF-1 and human disease: one
highly involved factor. Genes Dev 14, 1983–1991.
2 Yu AY, Frid MG, Shimoda LA, Wiener CM, Stenmark
K & Semenza GL (1998) Temporal, spatial, and oxy-
gen-regulated expression of hypoxia-inducible factor-1
in the lung. Am J Physiol 275, L818–L826.
3 Yan SF, Lu J, Zou YS, Soh-Won J, Cohen DM,
Buttrick PM, Cooper DR, Steinberg SF, Mackman N,

Pinsky DJ & Stern DM (1999) Hypoxia-associated
induction of early growth response-1 gene expression.
J Biol Chem 274, 15 030–15 040.
4 Yan SF, Lu J, Zou YS, Kisiel W, Mackman N, Leitges
M, Steinberg S, Pinsky D & Stern D (2000) Protein
kinase C-beta and oxygen deprivation. A novel Egr-1-
dependent pathway for fibrin deposition in hypoxemic
vasculature. J Biol Chem 275, 11 921–11 928.
5Hu
¨
ttemann M, Kadenbach B & Grossman LI (2001)
Mammalian subunit IV isoforms of cytochrome c oxi-
dase. Gene 267, 111–123.
6 Fukuda R, Zhang H, Kim JW, Shimoda L, Dang CV &
Semenza GL (2007) HIF-1 regulates cytochrome oxidase
subunits to optimize efficiency of respiration in hypoxic
cells. Cell 129, 111–122.
7 Zhang H, Gao P, Fukuda R, Kumar G, Krishnamach-
ary B, Zeller KI, Dang CV & Semenza GL (2007)
HIF-1 inhibits mitochondrial biogenesis and cellular
respiration in VHL-deficient renal cell carcinoma by
repression of C-MYC activity. Cancer Cell 11, 407–420.
8 Semenza GL (2007) HIF-1 mediates the Warburg effect
in clear cell renal carcinoma. J Bioenerg Biomembr
doi:10.1007/s10863-007-9081-2.
9 Kim JW, Tchernyshyov I, Semenza GL & Dang CV
(2006) HIF-1-mediated expression of pyruvate dehydro-
genase kinase: a metabolic switch required for cellular
adaptation to hypoxia. Cell Metab 3, 177–185.
10 Chandel NS & Schumacker PT (2000) Cellular oxygen

sensing by mitochondria: old questions, new insight.
J Appl Physiol 88, 1880–1889.
11 Waypa GB, Chandel NS & Schumacker PT (2001)
Model for hypoxic pulmonary vasoconstriction involv-
ing mitochondrial oxygen sensing. Circ Res 88, 1259–
1266.
12 Paddenberg R, Ishaq B, Goldenberg A, Faulhammer P,
Rose F, Weissmann N, Braun-Dullaeus RC & Kummer
W (2003) Essential role of complex II of the respiratory
chain in hypoxia-induced ROS generation in the pulmo-
nary vasculature. Am J Physiol Lung Cell Mol Physiol
284, L710–L719.
13 Schroedl C, McClintock DS, Budinger GR & Chandel
NS (2002) Hypoxic but not anoxic stabilization of
HIF-1alpha requires mitochondrial reactive oxygen
species. Am J Physiol Lung Cell Mol Physiol 283,
L922–L931.
M. Hu
¨
ttemann et al. Cytochrome c oxidase subunit IV-2 hypoxic response
FEBS Journal 274 (2007) 5737–5748 ª 2007 The Authors Journal compilation ª 2007 FEBS 5747
14 Ortega-Saenz P, Pardal R, Garcia-Fernandez M &
Lopez-Barneo J (2003) Rotenone selectively occludes
sensitivity to hypoxia in rat carotid body glomus cells.
J Physiol 548, 789–800.
15 Horvat S, Beyer C & Arnold S (2006) Effect of hypoxia
on the transcription pattern of subunit isoforms and the
kinetics of cytochrome c oxidase in cortical astrocytes
and cerebellar neurons. J Neurochem 99, 937–951.
16 Heinemeyer T, Wingender E, Reuter I, Hermjakob H,

Kel AE, Kel OV, Ignatieva EV, Ananko EA, Pod-
kolodnaya OA, Kolpakov FA et al. (1998) Databases
on transcriptional regulation: TRANSFAC, TRRD and
COMPEL. Nucleic Acids Res 26, 362–367.
17 Lee I, Salomon AR, Ficarro S, Mathes I, Lottspeich F,
Grossman LI & Hu
¨
ttemann M (2005) cAMP-dependent
tyrosine phosphorylation of subunit I inhibits cyto-
chrome c oxidase activity. J Biol Chem 280, 6094–6100.
18 Waypa GB, Marks JD, Mack MM, Boriboun C, Mun-
gai PT & Schumacker PT (2002) Mitochondrial reactive
oxygen species trigger calcium increases during hypoxia
in pulmonary arterial myocytes. Circ Res 91, 719–726.
19 Michelakis ED, Hampl V, Nsair A, Wu X, Harry G,
Haromy A, Gurtu R & Archer SL (2002) Diversity in
mitochondrial function explains differences in vascular
oxygen sensing. Circ Res 90, 1307–1315.
20 Villani G, Greco M, Papa S & Attardi G (1998) Low
reserve of cytochrome c oxidase capacity in vivo in the
respiratory chain of a variety of human cell types. J Biol
Chem 273, 31 829–31 836.
21 Villani G & Attardi G (1997) In vivo control of respira-
tion by cytochrome c oxidase in wild-type and mito-
chondrial DNA mutation-carrying human cells. Proc
Natl Acad Sci USA 94, 1166–1171.
22 Chandel NS, Maltepe E, Goldwasser E, Mathieu CE,
Simon MC & Schumacker PT (1998) Mitochondrial
reactive oxygen species trigger hypoxia-induced tran-
scription. Proc Natl Acad Sci USA 95, 11 715–11 720.

23 Arnold S & Kadenbach B (1997) Cell respiration is con-
trolled by ATP, an allosteric inhibitor of cytochrome-c
oxidase. Eur J Biochem 249, 350–354.
24 Jackson SP, MacDonald JJ, Lees-Miller S & Tjian R
(1990) GC box binding induces phosphorylation of Sp1
by a DNA-dependent protein kinase. Cell 63, 155–165.
25 Burke PA & Poyton RO (1998) Structure ⁄ function of
oxygen-regulated isoforms in cytochrome c oxidase.
J Exp Biol 201, 1163–1175.
26 Bisson R, Vettore S, Aratri E & Sandona D (1997) Sub-
unit change in cytochrome c oxidase: identification of
the oxygen switch in Dictyostelium. Embo J 16, 739–
749.
27 King MP & Attardi G (1989) Human cells lacking
mtDNA: repopulation with exogenous mitochondria by
complementation. Science
246, 500–503.
28 Frohman MA (1995) Rapid amplification of cDNA
ends. In PCR Primer, a Laboratory Manual (Dieffen-
bach CW & Dveksler GS, eds), pp. 381–409. Cold
Spring Harbor Laboratory Press, Cold Spring Harbor,
NY.
29 Hu
¨
ttemann M (2002) Partial heat denaturation step
during reverse transcription and PCR screening yields
full-length 5¢-cDNAs. Biotechniques 32, 730, 732, 734,
736.
30 Dignam JD, Lebovitz RM & Roeder RG (1983) Accu-
rate transcription initiation by RNA polymerase II in a

soluble extract from isolated mammalian nuclei. Nucleic
Acids Res 11, 1475–1489.
31 Yu M, Jaradat SA & Grossman LI (2002) Genomic
organization and promoter regulation of human cyto-
chrome c oxidase subunit VII heart ⁄ muscle isoform
(COX7AH). Biochim Biophys Acta 1574 , 345–353.
Cytochrome c oxidase subunit IV-2 hypoxic response M. Hu
¨
ttemann et al.
5748 FEBS Journal 274 (2007) 5737–5748 ª 2007 The Authors Journal compilation ª 2007 FEBS