Electrogenic events upon photolysis of CO from fully reduced cytochrome c oxidase
Marko Rintanen, Ilya Belevich⁎, Michael I. Verkhovsky1
Helsinki Bioenergetics Group, Institute of Biotechnology, University of Helsinki, PB 65 (Viikinkaari 1), 00014, Helsinki, Finland
a b s t r a c t a r t i c l ei n f o
Received 5 August 2011
Received in revised form 7 November 2011
Accepted 9 November 2011
Available online 18 November 2011
CO photolysis from fully reduced Paracoccus denitrificans aa3-type cytochrome c oxidase in the absence of O2
was studied by time-resolved potential electrometry. Surprisingly, photo dissociation of the uncharged car-
bon monoxide results in generation of a small-amplitude electric potential with the same sign as the physi-
ological charge separation during activity. The number of electrogenic events after CO photolysis depends on
the state of the enzyme. CO photolysis following immediately after activation by an enzymatic turnover,
showed a two-component potential development. A fast (~1.5 μs) phase was followed by slower potential
generation with a time constant varying from 8 μs at pH 7 to 250 μs at pH 10. The amplitude of the fast
phase was independent of the time of incubation after enzyme activation, whereas the slower phase van-
ished with a time constant of ~25 min. CO photolysis from enzyme that had not undergone a prior single
turnover showed the fast phase, but the amplitude of the slow phase was reduced to 10–30%. The amplitude
of the fast phase corresponds to charge movement of 0.83 Å perpendicular to the membrane dielectric, and is
independent of the time after enzyme activation. Thus it can be used as an internal ruler for normalization of
the electrogenic responses of CcO. The slow phase was absent in the K354M mutant with a blocked proton-
conducting K channel. We propose that CO photolysis increases the pK of the K354 residue, which results in
its partial protonation, and generation of electric potential.
© 2011 Elsevier B.V. All rights reserved.
Cytochrome c oxidase (CcO) is the terminal enzyme of the electron
transfer chain located in the inner membrane of mitochondria, and in
the plasma membrane of many species of aerobic bacteria. There are
two parallel mechanisms by which CcO catalysis is coupled to the gen-
erationofelectrochemical protongradient. Half of the energy conserva-
tion is the result of the vectorial nature of the oxygen reduction
chrome c at the positively charged outer side of the membrane (P side)
reduction of each molecule of oxygen to water is coupled to transloca-
tion of four electrical charges across the membrane. In addition, CcO
uses the energy of this redox reaction to pump four additional charges
(protons) per oxygen from the N- to the P side of the membrane .
The detailed molecular mechanism of the proton pump is still not
known, and is subject to intense investigation [3–6].
It is well established that during the enzyme's catalytic cycle there
is a large number of different electron and proton transfer events
with a great variety of rates. Detailed understanding of the topology
of all these charge translocations should provide significant help in
understanding the molecular mechanism. Therefore, the method of
time-resolved potential electrometry [7,8], which provides the ability
for precise determination of relative charge displacements, is a pow-
erful tool for investigating the CcO mechanism and for creating a "di-
electric topography" map of the enzyme . However, the amplitudes
of the electrometric response give only a relative displacement without
tudes may be converted into a number of moved charges only when
there is an internal molecular “ruler” to which the displacement can be
normalized. For example, in photosynthetic reaction centers with well-
known 3D structure , the ultra-fast charge separation between the
chlorophyll special pair and the primary quinone corresponds to the
movement of a single charge over an exactly established distance, and
this phase of charge separation can be used for calibrating all other
charge translocation events in the sample [11,12].
In this paper we make an attempt to find such an internal molec-
ular “ruler” for the charge transfer reactions in CcO. We used the abil-
ity of potential electrometry to distinguish very small electrical
changes in the enzyme to show that a 1.5 μs phase of potential gener-
ation upon CO photolysis from fully reduced CcO is proportional to
the enzyme concentration and does not depend on experimental con-
ditions (such as pH, enzyme activation, different single point muta-
tions), and can thus be used as an internal calibration for other
charge movement events in CcO, for example, the R (fully reduced)
to OH(oxidized) transition upon addition of molecular oxygen.
Biochimica et Biophysica Acta 1817 (2012) 269–275
Abbreviations: CcO, type aa3cytochrome c oxidase; DM, n-dodecyl-β-Dmaltoside;
Eh, ambient redox potential; Em, midpoint redox potential
⁎ Corresponding author.
E-mail address: Ilya.Belevich@Helsinki.Fi (I. Belevich).
1Deceased 4th October, 2011.
0005-2728/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
Contents lists available at SciVerse ScienceDirect
Biochimica et Biophysica Acta
journal homepage: www.elsevier.com/locate/bbabio
2. Materials and methods
2.1. Sample preparation
Cytochrome c oxidasefrom Paracoccusdenitrificanswasisolated and
preparedasdescribed earlier [13,14].Reconstitutionof the enzymeinto
vesicles was achieved with Bio-Beads (Bio-Rad Laboratories) using a
method described before , except that the enzyme concentration
during the reconstitution was 6.7 μM. In addition to the wild type en-
zyme, two batches of mutant enzymes were used, D124N and K354M,
the classical D- and K-channel variants, respectively.
2.2. The electrometric setup
The setup was described in detail previously, for example in Refs.
[15,16]. In the measuring system a membrane, i.e. a phospholipid im-
pregnated Teflon mesh, divides the measuring cell into two compart-
ments and voltage over the membrane is measured with two Ag/AgCl
electrodes (World Precision Instruments), one in each compartment.
The vesicles with reconstituted CcO were fused with the measuring
membrane by adding 25 mM MgCl2 in 100 mM HEPES (pH 7.5).
After incubation for 2 h, the buffer was replaced with the experimen-
tal one with the desired pH. Buffers were used at 100 mM final con-
centration as follows (corresponding pH in parenthesis): MES (6),
MOPS (7), HEPES (7.5), TRIS (8 and 8.5), CHES (9 and 9.5), CAPS
(10 and 10.5) (all buffers from Sigma). Measurements were con-
ducted under a 100% CO atmosphere and CO photolysis was induced
by a laser flash (Brilliant YAG, Quantel, pulse energy=180 mJ,
λ=532 nm). When multiple averages were taken, laser flashes
were separated by a 4–5 s pause. The voltage reported by the system
is proportional to charge movement inside the enzyme perpendicular
to the membrane plane.
2.3. The measurement procedure
(63 mM,(0.13–0.27 U/μL,
Merck) in the presence of catalase (0.45 U/μL, Sigma), using hexaam-
mineruthenium (1 μM, Aldrich) as a redox mediator. Reduction was
followed as the extent of electric potential generation upon electron
backflow , and the sample was judged to be fully reduced when
all traces of backflow (maximal at the 3-electron reduction level of
the enzyme ) had disappeared. Second, the potential generation
by the enzyme was followed after addition of 100 μL of oxygen satu-
rated buffer directly toward the measuring membrane, followed by
an immediate laser flash, as described previously . Third, sodium
dithionite (Merck) was added to a final concentration of 1.5 mM to
both sides of the measuring membrane to reduce the sample and to
quickly remove excess oxygen. Time-resolved CO photolysis of fully
reduced CcO was then followed after different times of incubation.
The transmembrane leak of potential through the measuring mem-
brane was recorded on a second time scale and fitted. These results
were then used to compensate the leak especially for the D124N mu-
tant that has a rather slow oxidation rate.
2.4. Data analysis
Custom made Matlab (MathWorks) scripts and functions were
used for data treatment, fitting and presentation.
3.1. CO photolysis of the fully reduced enzyme
Fig. 1 shows the development of electric potential upon photodis-
sociation of carbon monoxide from CcO that was immediately re-
reduced by dithionite after oxidation by O2at pH 9.5. Each trace rep-
resents an average of ten flashes fired with 4-second intervals at dif-
ferent times after re-reduction. Upon the laser flash there is a fast
1.5 μs phase, which stays unchanged over the time of the experiment.
It is followed by a slow phase (180 μs at this particular pH), the ampli-
tude of which decreases with the time between enzyme re-reduction
and CO photolysis. The last phase reflects decay of the potential due
to CO recombination, the rate of which becomes faster with time
after enzyme activation. Note that CO recombination in the figure is
presented on a much slower time scale relative to the two preceding
phases (Fig. 1). The dependence of the amplitudes of the first and sec-
ond phases on the delay between reduction and CO photolysis,
obtained as a result of the data fitting process, is presented in Fig. 2.
The data shows that the first phase maintains its amplitude over the
measuring period, while the second phase fades out completely
with a time constant of ~25 min.
3.1.1. The fast phase
The 1.5 μs phase has a constant amplitude and rate from pH 7 to 10
10.5, but at such a high pH part of the enzyme population could already
be denatured, or the hemes destroyed . Between pH 7 and 10 the
amplitude is 0.83 mV on the average with a standard deviation of
0.13 mV. The average time constant is 1.5 μs (rate 650±196 ms−1).
To verify the nature of the potential generation upon CO photolysis
Fig. 1. Electric potential development upon CO photolysis of the fully reduced Paracoc-
cus denitrificans aa3oxidase at pH 9.5. Reaction starts with the laser flash at time 0.
Each curve represents individual time points (time scale minutes), where the first
trace is immediately after dithionite re-reduction following activation of the enzyme
by turnover. Inset: early part of reaction, up to 3 μs. Conditions: 100 mM CHES, pH
9.5, 1 μM hexamine–ruthenium, 0.45 U/μL catalase, 0.25 U/μL glucose oxidase, 63 mM
glucose, 1.5 mM sodium dithionite after 100 μL pulse of O2saturated buffer, under
100% CO atmosphere.
Fig. 2. Dependence of the amplitudes of the fast ~1.1 μs (●) and slow ~180 μs (▲)
phases of potential generation upon CO photolysis from the fully reduced CcO on the
time after enzyme activation at pH 9.5. Conditions as in Fig. 1.
M. Rintanen et al. / Biochimica et Biophysica Acta 1817 (2012) 269–275
we also performed measurements on mutant enzymes which have
blocks in either ofthetwo proton-conductingchannels. The amplitudes
of the first phase for the two studied mutants, D124N and K354M, have
exactly the same time constant as wild type enzyme. The amplitude of
(Supplementary material Fig. S2). There is a slight decrease in the am-
plitude of the K354M variant towards the alkaline end of the titration,
but this could be due to easier alkaline denaturation of this variant.
3.1.2. The slow phase
In wild type oxidase the slow phase grows to its maximum ampli-
tude in 1 to 3 min after dithionite addition (Fig. 2), and then de-
creases with a time constant of about 25 min. Fig. 3 presents the
amplitudes and rate constants of the second phase as a function of
pH in wild type CcO. The amplitude increases with pH, becoming larg-
er than the fast phase at the alkaline end of the titration. Conversely,
the rate is fast at neutral pH and slows down with pH almost linearly
on a logarithmic scale, dropping from an average of 155 ms−1at pH 7
to 2.7 ms−1at pH 10.5. It should be noted that discriminating be-
tween the fast and slow phase becomes more difficult as the slow
phase speeds up at neutral pH. The slow phase is present in the
D124N variant, but it is entirely absent in the K354M variant. The re-
sults for the slow phase in mutant enzymes are presented in the sup-
plementary Fig. S3. We also conducted control experiments where we
followed CO photolysis for CcO that had not been activated by reduc-
tion and reoxidation by oxygen. In these controls the amplitude of the
slow phase was only 10–30 per cent of that in the pre-activated case.
3.1.3. The decay of electric potential
Upon CO recombination the decay of potential is slower when the
amplitude of the signal is larger. In other words, when the slow phase
is present, CO recombination is slow and gets faster when the slow
phase disappears. This is presented in Fig. 4, where the amplitudes
(Fig. 4A) and the time constants (Fig. 4B) of the CO recombination
phase of wild type enzyme are compared at pH 7 (■) and 9.5 (▼) at
different time points after enzyme activation. The CO recombination
rate is quite different for the activated enzyme at these two pH values,
but after enzyme relaxation (and disappearance of the slow phase) it
became similar (τ~25 ms) to the earlier reported values for the fully
reduced enzyme [19,20].
3.2. Oxygen reaction
In the same samples where electric potential generation upon CO
photolysis was followed, we also recorded the development of elec-
tric potential during the oxidative half of the catalytic cycle triggered
by CO photolysis in the presence of O2. In fact, this measurement was
used as the enzyme activation procedure before sample re-reduction
with dithionite. The data on electrogenic reactions of fully reduced
CcO with oxygen, i.e. transition from state R (four electron reduced)
to state OH(oxidized), was obtained as in an earlier study . Here
we aimed at comparing amplitudes of this response with the ampli-
tude of the fast 1.5 μs phase, which does not depend on pH, time of
the experiment or the mutation type, but rather reflects the amount
of CO-photolysable enzyme fused to the measuring membrane.
Fig. 5 presents data on the potential generation in R to OHtransitions
at different pH values, normalized to the amplitude of the first
(1.5 μs) phase of CO photolysis from fully reduced enzyme, for the
wild type enzyme (●), and two mutants, K354M (▲) and D124N
(■). A Henderson–Hasselbalch one-proton titration curve with pKa
9.11 is drawn through the wild type data points (for comparison see
Ref. ). R2of the fit was 0.98. As can be noticed, the amplitudes
of the oxygen response of the K354M mutant enzyme fit nicely with
the wild type data after normalization of the amplitudes to the ampli-
tude of the 1.5 μs phase. Yet, when the measured absolute amplitudes
of the oxygen response are compared, the K354M values are smaller
and drop faster in alkaline conditions than those of wild type. Howev-
er, as can be seen from Fig. S2, the amplitude of the 1.5 μs phase in
K354M also has a slight tendency to decrease at alkaline pH, and
after normalization the ratio of voltage generated by oxygen reaction
Fig. 3. pH dependence of the slow phase amplitude (A) and rate constant (B) of poten-
tial generation upon CO photolysis by the wild type CcO. Panel A presents averages of
maximum amplitudes (mV) of the second phase. Solid line: Henderson–Hasselbalch fit
with pKa10.1. Panel B shows averages of corresponding rate constants of the slow
phase on a logarithmic scale. Solid line: Henderson–Hasselbalch fit with pKa7.7 and
Hill coefficient 0.71. Conditions: for every pH value 100 mM buffer (for details see Ma-
terials and methods), 1 μM hexamine–ruthenium, 1.5 mM sodium dithionite after
100 μL pulse of O2saturated buffer, 0.45 U/μL catalase, 0.13–0.27 U/μL glucose oxidase
and 63 mM glucose, under 100% CO atmosphere.
Fig. 4. Dependence of the amplitude (A) and rate constant (B) of potential dissipation
upon CO recombination in the wild type CcO on the time of the experiments at two pH
values—7(■) and 9.5 (▼). Time point 0 is immediately after dithionite re-reduction fol-
lowing oxygen activation of the enzyme. Conditions: as in Fig. 3 (Materials and
M. Rintanen et al. / Biochimica et Biophysica Acta 1817 (2012) 269–275
to the amplitude of the 1.5 μs phase is the same as in the wild type en-
zyme. In contrast, the D124N mutant behaves differently. At neutral
pH the normalized response is only approximately one half of that
in the wild type and the K354M variant, and virtually independent
of pH so that the amplitude matches that of the wild type enzyme
at high pH. Such behaviour is expected because the D-channel mutant
does not pump protons, and only catalyses charge translocation
through vectorial chemistry (see Introduction) [21–23].
3.3. Redox state dependence of the fast phase
We have so far reported on the existence of the fast 1.5 μs phase in
the fully reduced enzyme. To find out whether this phase depends on
the redox state, we performed analogous experiments at different
redox potentials. Fig. 6 shows traces upon CO photolysis for non-
activated enzyme at pH 8.6. The fast 1.5 μs phase is clearly absent
from the mixed valence enzyme (Eh=420 mV), but shows slower
charge separation with opposite polarity due to the well-known
proton transfer via the K-pathway coupled to backflow of electrons
from the binuclear site at high pH [19,24,25].
3.4. Spectrum of the fast phase
Optical absorption measurements were done to determine the na-
ture of the potential generation phases upon CO photolysis. Experi-
ments were conducted at pH 9.5 where both phases were clearly
observable by electrometry. However, freshly oxidized enzyme, re-
reduced by dithionite, did not show any significant changes of optical
absorbance with a time constant of 180 μs corresponding to the slow
phase of potential generation, but clearly showed changes with a time
constant of 1.5μs (Fig. 7). These optical changes are identical to those
described earlier at neutral pH [26,27]. Fig. 7A shows a kinetic trace at
613 nm, where this fast change is clearly seen as a decrease of absor-
bance immediately after the initial increase caused by the scission of
the carbon monoxide-heme a3bond by the laser flash. Fig. 7B shows
the kinetic spectrum of this phase, which has a maximum at
600 nm, a trough at 613 and intersects zero close to 605 nm.
4.1. Nature of the fast phase
Photolysis of CO from the reduced heme-copper oxidase is quite
different from the equivalent process in other heme-containing pro-
teins. The main difference is due to the location of the copper atom
(CuB) in very close proximity to the iron of heme a3. Such architecture
of the binuclear site results in very fastb1 ps [26,28] initial migration
of CO from Fea3
sociates from the CuBsite with a time constant of about 1.5 μs ,
which coincides well with the time constant of the fast phase of po-
tential generation obtained in our experiments. Thus, it would be nat-
ural to propose that the fast phase is the result of CO dissociation from
CuB. However, as we showed by the redox state dependence (Fig. 6),
the reduced and CO-ligated binuclear site is a necessary but insuffi-
cient condition for appearance of the fast phase. In the mixed valence
enzyme, where the binuclear site has the same electronic state as in
the fully reduced enzyme, but where the donor centers (CuAand
heme a) are oxidized, there was no potential generation correspond-
ing to the 1.5 μs phase.
The absorbance changes in the visible region also show a compo-
nent with a time constant of 1.5 μs. The optical spectrum of this
phase has a derivative shape that is characteristic of a blue absor-
bance band shift with a cross-over point at 605 nm, and maximum
and minimum at 600 and 613 nm, respectively. Such properties
would be expected for a blue shift of the α-band of low spin ferrous
heme a. However, Einarsdottir et al. , who showed the same
+upon the laser flash. In the next step, CO dis-
Fig. 5. pH dependence of ratio of the oxygen reaction to the fast phase of CO photolysis.
The ratio is calculated by dividing the amplitude of the electric potential generation
during the reaction of fully reduced CcO with oxygen by the amplitude of the fast
phase of CO photolysis reaction measured on the same sample. Data represents wild
type (●), K354M mutant (▲) and D124N (■). Henderson–Hasselbalch fit of wild
type data (solid line) has pKa9.11. Conditions: for every pH 100 mM buffer (for details
see Materials andmethods),1 μM hexamine–ruthenium,
0.13–0.27 U/μL glucose oxidase and 63 mM glucose, under 100% CO atmosphere. In
case of K354M 10 μM hexamine–ruthenium was used instead of 1 μM.
0.45 U/μL catalase,
Fig. 6. Potential generation upon CO photolysis in mixed-valence (solid line) and fully
reduced (dotted line) Paracoccus denitrificans aa3oxidase at pH 8.6. Reaction starts
with a laser flash at time 0. Conditions: 100 mM bicine (pH 8.6), 0.5 U/μL catalase,
0.2 U/μL glucose oxidase, 63 mM glucose, 1 μM hexamine–ruthenium, under 100% CO
atmosphere. Mixed-valence state was induced by adding 15 mM ferri/ferrocyanide
with 1:1 stoichiometry. In reduced case Ehb190 mV and in mixed-valence case
Fig. 7. Absorbance changes upon CO photolysis from fully reduced CO-ligated wild type
CcO. The time course at 613 nm is shown in panel A. The kinetic spectrum of the 1.5 μs
phase as a result of a global fit of the recorded multiwavelength data surface is presented
inpanel B. Conditions: 200 mM CHES, pH9.5, 0.05% DM, 10 U/ml glucose oxidase, 10 mM
glucose,30 U/mlcatalase,1 μMhexamine–ruthenium([Ru(NH3)6]2+/3+),under100%CO
M. Rintanen et al. / Biochimica et Biophysica Acta 1817 (2012) 269–275
blue shift with similar kinetics in enzyme from bovine heart mito-
chondria, found it to occur not only in fully reduced enzyme, but
also in the mixed-valence state with only slightly smaller amplitude.
It is clear, therefore, that the 1.5 μs optical shift differs from the 1.5 μs
charge separation in that the latter is not observed in the mixed-
On the basis of the information at hand, it seems that the 1.5 μs
charge separation phase is kinetically limited by the rate of CO disso-
ciation from CuB. However, the absence of this phase in the mixed-
valence enzyme is enigmatic because CO dissociation, as such, is not
affected by the redox states of heme a or CuA. The insensitivity of
this phase to variations in pH and to key mutations in proton transfer
pathways provides no support for net proton transfer as the cause,
and electron transfer is excluded in the fully reduced enzyme. The
1.5 μs phase seems to be a fully conserved feature among diverse
members of the vast heme-copper oxidase family (ba3 and caa3
from Thermus thermophilus, cbb3from Rhodobacter sphaeroides, data
not shown). We may speculate that dissociation of CO from CuBin-
duces a rearrangement of ligands to both CuBand heme a3. These
could be water molecules, whose arrangement may depend on the
redox state of heme a . Rearrangement due to CO dissociation
from CuBcould change the overall dipole moment of a cluster of
water molecules to mimic a 0.83 Å movement of an elementary
charge perpendicular to the membrane plane.
4.2. Amplitude of the fast phase
In our experiments we were able to sequentially measure the po-
tential generation from the same sample for two cases: CO photolysis,
and the reaction of CcO with oxygen under the same experimental
conditions. For the fully reduced enzyme the ratio of amplitudes of
the oxygen response to the fast phase of CO photolysis (125, see
Fig. 5) is very stable, even though the absolute values of the ampli-
tudes vary depending on conditions of the experiment and sample
preparation. From this ratio we can estimate the dielectric distance
of charge separation due to CO photolysis. For that purpose we have
to know the number of charges translocated during the reaction of
fully reduced enzyme with oxygen. This number can be estimated
based on the current knowledge of the enzyme structure and its loca-
tion in the membrane dielectric . Our earlier measurements 
showed that during the oxidative phase two pumped protons are
translocated across the whole membrane. In addition, the redox
chemistry results in vectorial electrogenic events: two “chemical”
protons are taken up to the binuclear center from the N side, and
are translocated through about 2/3 of the membrane dielectric
where they meet an electron delivered via 1/3 of the membrane di-
electric from CuAat the P side of the membrane. Altogether, these
events make the net number of translocated charges 2+2×2/3+1/
3=3.7. The measured potential is thus the result of translocation of
3.7 charges across 28 Å of the hydrophobic membrane core .
The response of CO photolysis is 125 times smaller, which would be
equivalent to translocation of one charge across 3.7×28/125=0.83 Å
in a direction perpendicular to the membrane. This tiny charge move-
ment can be recorded very well by our method and, most importantly,
it canbeusedas aninternalrulerfor thecalibrationof charge transloca-
tion events in the enzyme. Additionally, it is important to note that the
fast phaseof CO photolysis waspresentin both mutantenzymes tested,
applicability for this "dielectric ruler" we analyzed two well-
characterized mutants of CcO. It was earlier shown [21–23] that the
K354M mutation, which blocks the K channel, does not influence pro-
ton translocation during the oxidative phase of enzyme turnover. The
measurements presented in this work confirmed this finding: the
ratio of the amplitude of the electrometric oxygen response to the fast
phase of CO photolysis in the K354M mutant is the same as in the
wild type enzyme (i.e. 125, see Fig. 5). It should be stressed that the
absolute amplitude of the oxygen response at neutral pH for the
K354M mutant is only around 50 mV, which is about twice smaller
than that for the wild type enzyme. Thus, the use of absolute values of
potential generation for comparison of the K354M mutant with wild
type is not possible. At the same time, the mutation in the D-channel
(D124N) decreases this ratio to 54.5, which is equivalent to transloca-
tion of a single charge across a 54.5×0.83=45.2 Å barrier. Or, taking
into account the thickness of the membrane dielectric (28 Å), we find
the number of charges translocated by the D124N mutant during the
oxidative phase of the catalytic cycle to be 45.2/28=1.62. This number
is smallerby~2 charges thantheoneforthe wild type enzyme (3.7, see
above), and correlates well with the knowledge  that 2 protons are
pumped during the oxidative phase of the catalytic cycle in the wild
type enzyme, and that proton pumping is abolished in the D124N
4.3. pH dependence of the charge separation in the oxidative half of the
A comparison of the potential changes at neutral and alkaline pH
values was done earlier [2,14,34]. Previous results were based on
the total potential change associated with oxidation of CcO, and did
not take into account the decline of the amount of active enzyme at
alkaline pH. Here, in Fig. 5, we presented the response [14,34,35] nor-
malized to the fraction of the enzyme active in CO photolysis. We rec-
ognize that there are problems in the assumption that all the signals
decrease in exactly the same manner as the enzyme denatures. De-
spite that, the normalization of the oxygen responses with CO photol-
ysis from fully reduced CcO ties mutants with different response
amplitudes nicely together, which is especially noticeable in the
case with wild type and K354M mutant (see above). In a previous
study  the amplitudes of potential generation upon addition of
oxygen at alkaline pH tended to approach zero. However, normaliza-
tion with the amplitude of the CO photolysis effect changes the pic-
ture. The normalization shows that at alkaline pH the potential
approaches 40% of the signal obtained at neutral pH, which corre-
sponds to translocation of 3.7×0.4=1.48 charges. Loss of about 2
translocated charges at high pH is consistent with disappearance of
proton pumping at alkaline pH . The fit of the data by the Hender-
son–Hasselbalch equation (solid curve, Fig. 5) shows that proton
pumping disappears in the aa3oxidase from P. denitrificans with an
apparent pKaof ~9.1. The fact that the pH dependence is absent in
the non-pumping D124N mutant enzyme, where the amplitude is
close to that in wild type at very high pH, provides further support
for our conclusion.
4.4. The nature and origin of the slow phase
The relaxation of the oxidized enzyme to the non-active “resting”
form of the “fast” , “pulsed”  or active  forms of CcO is well
documented in the literature. Here we showed that the protein relax-
ation kinetics is clearly seen also for the reduced form of the enzyme.
This conclusion is based on the time dependence of the amplitude of
the slow phase of potential generation upon CO photolysis after en-
zyme turnover (Fig. 2), which diminishes with a time constant of
~25 min after a single turnover of CcO (oxygen injection and CO pho-
tolysis was followed by immediate enzyme reduction). As may be
noted from Fig. 2, in the first moments after the addition of dithionite
the amplitude of the slow phase seems to be growing with a time
constant of about 1 min. The source of this effect is not completely
clear, but it may be induced by a small oxygen contamination of the
system from, for example, the needle by which dithionite was
injected. Such a contamination could result in the electron backflow
reaction and accompanying proton transfer . At pH 9.5 the back-
flow reaction would result in potential generation of the opposite
sign from the slow phase, thus decreasing the slow phase signal. If
M. Rintanen et al. / Biochimica et Biophysica Acta 1817 (2012) 269–275
this explanation is true, we estimate the amount of oxygen-
contaminated protein to be about 4%.
What could be the possible cause of the slow phase of potential
generation upon CO photolysis? Transmembrane electron transfer
can be excluded since the enzyme is fully reduced. The strong pH de-
pendence argues in favour of movement of a proton within the pro-
tein core, perpendicular to the membrane plane. The amplitude
increases dramatically at alkaline pH, which means that the acceptor
group involved should have an alkaline pKain order to be unproto-
nated before excitation, and to be able to take the proton after the
laser flash. For practical reasons we could not go to pH values higher
than 10.5 to finish the titration. However, it is clear that the ampli-
tude is still increasing at the alkaline edge and even though we
could not exactly define the pKaof this transition, its low limit can
be estimated as ~10.1 (Fig. 3), but the true value is likely to be higher
(see below). The abolition of this phase in the K354M variant pro-
vides an important clue to its nature. Lysine 354 itself could be the
group responsible, in which case the phase is due to electrogenic pro-
ton uptake from the N phase to the lysine. The pH dependence of the
amplitude of the slow phase found in our investigation is indeed rath-
er similar to the pH dependence of the K354 protonation state for the
fully reduced enzyme obtained from electrostatic calculations .
The pH dependence of the rate of the slow phase (Fig. 3B) shows
that this rate is controlled by the degree of protonation of another
group with a pKaof ~7.7. In accordance with the scenario suggested
above, K354 would be in protonic equilibrium with the N phase via
the input amino acid of the K-pathway, E(II)-78 (Fig. 8). The observed
pKaof 7.7 for the group controlling the rate is indeed close to the pKa
value for E(II)-78 found in continuum electrostatic calculations of the
fully reduced CcO . It also correlates well with the value found for
the pH dependence of the alkaline-induced electron backflow ,
which is now established [19,25] to be coupled to proton release
through the K channel from a water molecule bound to the ferric a3
The following analysis supports the notion that the amplitude of
the second phase is linked to the protonation state of lysine 354.
The dielectric topography map  gives the value of the relative di-
electric depth of this residue as 0.186 when counted from the N
side of the membrane from where the proton is taken up according
to our model (Fig. 8). When taking into account the thickness of the
membrane dielectric and the dielectric ruler obtained in this work
(0.83 mV/Å), the expected maximal amplitude of potential genera-
tion should be 28 Å×0.186×0.83 mV/Å=4.3 mV when the lysine is
initially unprotonated and takes up one full proton. In our experi-
ments the largest amplitude was ~1.3 mV observed at pH 10.5
(Fig. 3), which is 30% of the estimated maximum, and higher pH
values were not attempted due to enzyme denaturation. Continuum
electrostatic calculations with fully reduced enzyme  predict
that K354 is ~75% protonated at pH 10.5, so that maximally 25% can
be further protonated, which is in remarkable agreement with the ex-
The emerging picture is thus that the slow electrometric phase fol-
lowing CO photolysis from fully reduced enzyme is due to (fractional)
proton uptake via the K-pathway into K354, from the N side of the
membrane. The most plausible reason for this is evidently an increase
in the pKaof K354 caused by the dissociation of CO from the binuclear
site. However, considering that the distance from K354 to the binuclear
site is nearly20 Å, suchaneffectisby nomeans obvious. Perturbance of
the structure of water molecules within the K-pathway is expected to
change the pKaof K354 [38,39]. Moreover, Qin et al.  have shown
crystallographic evidence for redox-and ligand-dependent changes in
the water structure at the end of the K-pathway, close to the binuclear
site. It seems feasible, therefore, that dissociation of CO from the active
site changes the water structure along the K-pathway in such a way as
to cause an increase in the pKaof the lysine.
Another possible reason for the slow electrometric phase is a "flip"
of the protonated side chain of K354 "upwards" from the position in
the crystal structure (Fig. 8). However, it may be less evident why
such a reorientation would show the observed pH dependence. At
any rate, such a side chain reorganisation could occur in conjunction
with proton uptake, although the above analysis does not require it.
4.5. The effect of enzyme activation
The molecular basis for the fact that proton translocation in the re-
ductive phase of the catalytic cycle requires "activation" of the enzyme
by prior reduction and reoxidation by O2[22,33] is not understood.
Here, we showed that such "activation" also affects the state of the
fully reduced enzyme, in that the slow electrometric phase was much
diminished in amplitude if the enzyme was not pre-activated, and
that it disappeared completely after a sufficiently long incubation in
the reduced state prior to the laser pulse (Fig. 2). In view of the discus-
sion above, it would seem plausible that "activation" involves the reor-
ganisation of water molecules in the K-pathway, or filling strategic
positions in that pathway with water (see Ref. ). In fact, none of
the available crystal structures show a continuous water array in the
K-pathway, in contrast to the very well developed water file in the
lower parts of the D-pathway (see e.g. Ref. ).
Supplementary materials related to this article can be found on-
line at doi:10.1016/j.bbabio.2011.11.005.
This work was supported by Biocentrum Helsinki, the Sigrid Jusé-
lius Foundation and the Academy of Finland. We are grateful to Prof.
Mårten Wikström for the critical comments and invaluable help in
the writing and preparation of the manuscript.
Fig. 8. Structure of CcO subunits I and II based on PDB ID: 3HB3 . CuAand CuB,
hemes a and a3are highlighted together with K-pathway residues K354 and E78. The
figure was prepared using VMD software .
M. Rintanen et al. / Biochimica et Biophysica Acta 1817 (2012) 269–275
References Download full-text
 P. Mitchell, Coupling of phosphorylation to electron and hydrogen transfer by a
chemi-osmotic type of mechanism, Nature 191 (1961) 144–148.
 M.K. Wikström, Proton pump coupled to cytochrome c oxidase in mitochondria,
Nature 266 (1977) 271–273.
 P.E. Siegbahn, M.R. Blomberg, Proton pumping mechanism in cytochrome c oxi-
dase, J. Phys. Chem. A 112 (2008) 12772–12780.
 M. Wikström, M.I. Verkhovsky, Mechanism and energetics of proton translocation
by the respiratory heme-copper oxidases, Biochim. Biophys. Acta 1767 (2007)
 I. Belevich, M.I. Verkhovsky, Molecular mechanism of proton translocation by cy-
tochrome c oxidase, Antioxid. Redox Signal. 10 (2008) 1–29.
 G. Brändén, R.B. Gennis, P. Brzezinski, Transmembrane proton translocation by
cytochrome c oxidase, Biochim. Biophys. Acta 1757 (2006) 1052–1063.
 L.A. Drachev, A.D. Kaulen, V.P. Skulachev, Time resolution of the intermediate steps
in the bacteriorhodopsin-linked electrogenesis, FEBS Lett. 87 (1978) 161–167.
 M.I. Verkhovsky, J.E. Morgan, M.L. Verkhovskaya, M. Wikström, Translocation of
electrical charge during a single turnover of cytochrome c oxidase, Biochim. Bio-
phys. Acta 1318 (1997) 6–10.
 R. Sugitani, E.S. Medvedev, A.A. Stuchebrukhov, Theoretical and computational
analysis of the membrane potential generated by cytochrome c oxidase upon sin-
gle electron injection into the enzyme, Biochim. Biophys. Acta 1777 (2008)
 J. Deisenhofer, O. Epp, K. Miki, R. Huber, H. Michel, X-ray structure analysis of a
membrane protein complex. Electron density map at 3 Å resolution and a
model of the chromophores of the photosynthetic reaction center from Rhodop-
seudomonas viridis, J. Mol. Biol. 180 (1984) 385–398.
 S.K. Chamorovsky, D.A. Cherepanov, C.S. Chamorovsky, A.Y. Semenov, Correlation
of electron transfer rate in photosynthetic reaction centers with intraprotein di-
electric properties, Biochim. Biophys. Acta 1767 (2007) 441–448.
 A. Semenov, D. Cherepanov, M. Mamedov, Electrogenic reactions and dielectric
properties of photosystem II, Photosynth. Res. 98 (2008) 121–130.
 S. Riistama, L. Laakkonen, M. Wikström, M.I. Verkhovsky, A. Puustinen, The calci-
um binding site in cytochrome aa3from Paracoccus denitrificans, Biochemistry 38
 C. Ribacka, M.I. Verkhovsky, I. Belevich, D.A. Bloch, A. Puustinen, M. Wikström, An
elementary reaction step of the proton pump is revealed by mutation of
tryptophan-164 to phenylalanine in cytochrome c oxidase from Paracoccus deni-
trificans, Biochemistry 44 (2005) 16502–16512.
 M.I. Verkhovsky, A. Tuukkanen, C. Backgren, A. Puustinen, M. Wikström, Charge
translocation coupled to electron injection into oxidized cytochrome c oxidase
from Paracoccus denitrificans, Biochemistry 40 (2001) 7077–7083.
 D.A. Bloch, A. Jasaitis, M.I. Verkhovsky, Elevated proton leak of the intermediate
OHin cytochrome c oxidase, Biophys. J. 96 (2009) 4733–4742.
tion of cytochrome c oxidase with dioxygen, Biochemistry 38 (1999) 2697–2706.
 P.M. Callahan, G.T. Babcock, Origin of the cytochrome a absorption red shift: a pH-
dependent interaction between its heme a formyl and protein in cytochrome ox-
idase, Biochemistry 22 (1983) 452–461.
 I. Belevich, A. Tuukkanen, M. Wikström, M.I. Verkhovsky, Proton-coupled electron
equilibrium in soluble and membrane-bound cytochrome c oxidase from Para-
coccus denitrificans, Biochemistry 45 (2006) 4000–4006.
 I. Szundi, J. Ray, A. Pawate, R.B. Gennis, Ó. Einarsdóttir, Flash-photolysis of fully
reduced and mixed-valence CO-bound Rhodobacter sphaeroides cytochrome c ox-
idase: heme spectral shifts, Biochemistry 46 (2007) 12568–12578.
 P. Ädelroth, R.B. Gennis, P. Brzezinski, Role of the pathway through K(I-362) in
proton transfer in cytochrome c oxidase from R. sphaeroides, Biochemistry 37
 D. Bloch, I. Belevich, A. Jasaitis, C. Ribacka, A. Puustinen, M.I. Verkhovsky, M. Wik-
ström, The catalytic cycle of cytochrome c oxidase is not the sum of its two halves,
Proc. Natl. Acad. Sci. U. S. A. 101 (2004) 529–533.
 A.A. Konstantinov, S. Siletsky, D. Mitchell, A. Kaulen, R.B. Gennis, The roles of the
two proton input channels in cytochrome c oxidase from Rhodobacter sphaeroides
probed by the effects of site-directed mutations on time-resolved electrogenic
intraprotein proton transfer, Proc. Natl. Acad. Sci. U. S. A. 94 (1997) 9085–9090.
 S. Hallen, P. Brzezinski, B.G. Malmström, Internal electron transfer in cytochrome
c oxidase is coupled to the protonation of a group close to the bimetallic site, Bio-
chemistry 33 (1994) 1467–1472.
 Brändén, A. Namslauer, O. Hansson, R. Aasa, P. Brzezinski, Water-hydroxide ex-
change reactions at the catalytic site of heme-copper oxidases, Biochemistry 42
 O. Einarsdottir, R.B. Dyer, D.D. Lemon, P.M. Killough, S.M. Hubig, S.J. Atherton, J.J.
Lopez-Garriga, G. Palmer, W.H. Woodruff, Photodissociation and recombination
of carbonmonoxy cytochrome oxidase: dynamics from picoseconds to kilose-
conds, Biochemistry 32 (1993) 12013–12024.
 A. Namslauer, M. Branden, P. Brzezinski, The rate of internal heme–heme electron
transfer in cytochrome C oxidase, Biochemistry 41 (2002) 10369–10374.
 P.O. Stoutland, J.-C. Lambry, J.-L. Martin, W.H. Woodruff, Femtosecond dynamics of re-
duced cytochrome oxidase and its CO derivative, J. Phys. Chem. 95 (1991) 6406–6408.
 R.B. Dyer, O. Einarsdóttir, P.M. Killough, J.J. López-Garriga, W.H. Woodruff, Tran-
sient binding of photodissociated CO to CuB+of eukaryotic cytochrome oxidase
at ambient temperature. Direct evidence from time-resolved infrared spectrosco-
py, J. Am. Chem. Soc. 111 (1989) 7657–7659.
 M. Wikström, M.I. Verkhovsky, G. Hummer, Water-gated mechanism of proton
translocation by cytochrome c oxidase, Biochim. Biophys. Acta 1604 (2003) 61–65.
 I.V. Leontyev, A.A. Stuchebrukhov, Dielectric relaxation of cytochrome c oxidase: com-
parison of the microscopic and continuum models, J. Chem. Phys. 130 (2009) 085103.
 N. Kucerka, S. Tristram-Nagle, J.F. Nagle, Closer look at structure of fully hydrated
fluid phase DPPC bilayers, Biophys. J. 90 (2006) L83–L85.
 M.I. Verkhovsky, A. Jasaitis, M.L. Verkhovskaya, J.E. Morgan, M. Wikström, Proton
translocation by cytochrome c oxidase, Nature 400 (1999) 480–483.
 H. Lepp, P. Brzezinski, Internal charge transfer in cytochrome c oxidase at a limit-
ed proton supply: proton pumping ceases at high pH, Biochim. Biophys. Acta
1790 (2009) 552–557.
 M. Verkhovskaya, M. Verkhovsky, M. Wikström, pH dependence of proton trans-
location by Escherichia coli, J. Biol. Chem. 267 (1992) 14559–14562.
 A.J. Moody, C.E. Cooper, P.R. Rich, Characterisation of 'fast' and 'slow' forms of bo-
vine heart cytochrome-c oxidase, Biochim. Biophys. Acta 1059 (1991) 189–207.
 E. Antonini, M. Brunori, A. Colosimo, C. Greenwood, M.T. Wilson, Oxygen "pulsed"
cytochrome c oxidase: functional properties and catalytic relevance, Proc. Natl.
Acad. Sci. U. S. A. 74 (1977) 3128–3132.
 A. Tuukkanen, M.I. Verkhovsky, L. Laakkonen, M. Wikstrom, The K-pathway revis-
ited: a computational study on cytochrome c oxidase, Biochim. Biophys. Acta
1757 (2006) 1117–1121.
 M. Bränden, H. Sigurdson, A. Namslauer, R.B. Gennis, P. Ädelroth, P. Brzezinski, On
the role of the K-proton transfer pathway in cytochrome c oxidase, Proc. Natl.
Acad. Sci. U. S. A. 98 (2001) 5013–5018.
 L. Qin, J. Liu, D.A. Mills, D.A. Proshlyakov, C. Hiser, S. Ferguson-Miller, Redox-de-
pendent conformational changes in cytochrome a oxidase suggest a gating mech-
anism for proton uptake, Biochemistry 48 (2009) 5121–5130.
 V.R.I. Kaila, M.I. Verkhovsky, M. Wikstrom, Proton-coupled electron transfer in
cytochrome oxidase, Chem. Rev. 110 (2010) 7062–7081.
 J. Koepke, E. Olkhova, H. Angerer, H. Muller, G. Peng, H. Michel, High resolution
crystal structure of Paracoccus denitrificans cytochrome c oxidase: new insights
into the active site and the proton transfer pathways, Biochim. Biophys. Acta
1787 (2009) 635–645.
 W. Humphrey, A. Dalke, K. Schulten, VMD: visual molecular dynamics, J. Mol.
Graph. 14 (1996) 33–38.
M. Rintanen et al. / Biochimica et Biophysica Acta 1817 (2012) 269–275