Efficient electron transfer in a protein network lacking specific interactions

Article abstract of DOI:10.1021/ja205043f

In many biochemical processes, proteins need to bind partners amidst a sea of other molecules. Generally, partner selection is achieved by formation of a single-orientation complex with well-defined, short-range interactions. We describe a protein network that functions effectively in a metabolic electron transfer process but lacks such specific interactions. The soil bacterium Paracoccus denitrificans oxidizes a variety of compounds by channeling electrons into the main respiratory pathway. Upon conversion of methylamine by methylamine dehydrogenase, electrons are transported to the terminal oxidase to reduce molecular oxygen. Steady-state kinetic measurements and NMR experiments demonstrate a remarkable number of possibilities for the electron transfer, involving the cupredoxin amicyanin as well as four c-type cytochromes. The observed interactions appear to be governed exclusively by the electrostatic nature of each of the proteins. It is concluded that Paracoccus provides a pool of cytochromes for efficient electron transfer via weak, ill-defined interactions, in contrast with the view that functional biochemical interactions require well-defined molecular interactions. It is proposed that the lack of requirement for specificity in these interactions might facilitate the integration of new metabolic pathways.

Full text of DOI:10.1021/ja205043f

ARTICLE
pubs.acs.org/JACS
Efficient Electron Transfer in a Protein Network Lacking Specific
Interactions
Francesca Meschi,† Frank Wiertz,‡ Linda Klauss,‡ Anneloes Blok,‡ Bernd Ludwig,§ Angelo Merli,†
Hendrik A. Heering,‡ Gian Luigi Rossi,† and Marcellus Ubbink*,‡
†Department of Biochemistry and Molecular Biology, University of Parma, 43100 Parma, Italy
‡Institute of Chemistry, Leiden University, P.O. Box 9502, 2300 RA Leiden, The Netherlands
§Institute of Biochemistry, Molecular Genetics Group, and Cluster of Excellence Macromolecular Complexes, Goethe University,
D-60438 Frankfurt, Germany
S Supporting Information
b
ABSTRACT: In many biochemical processes, proteins need to
bind partners amidst a sea of other molecules. Generally,
partner selection is achieved by formation of a single-orienta-
tion complex with well-defined, short-range interactions. We
describe a protein network that functions effectively in a
metabolic electron transfer process but lacks such specific
interactions. The soil bacterium Paracoccus denitrificans oxidizes
a variety of compounds by channeling electrons into the main
respiratory pathway. Upon conversion of methylamine by
methylamine dehydrogenase, electrons are transported to the
terminal oxidase to reduce molecular oxygen. Steady-state
kinetic measurements and NMR experiments demonstrate a remarkable number of possibilities for the electron transfer, involving
the cupredoxin amicyanin as well as four c-type cytochromes. The observed interactions appear to be governed exclusively by the
electrostatic nature of each of the proteins. It is concluded that Paracoccus provides a pool of cytochromes for efficient electron
transfer via weak, ill-defined interactions, in contrast with the view that functional biochemical interactions require well-defined
molecular interactions. It is proposed that the lack of requirement for specificity in these interactions might facilitate the integration
of new metabolic pathways.
’ INTRODUCTION
electrons to a small blue copper protein, amicyanin (ami).1 From
there, electrons eventually reach the terminal oxidase cyto-
chrome aa3 (cyt. aa3), which reduces molecular oxygen to water.
In the periplasmic space, three soluble c-type cytochromes2,3 as
well as a membrane-anchored one,4 are also found, potentially all
able to catalyze the electron transfer (ET) from ami to the
oxidase. MADH, ami and cyt. c-551 were shown to crystallize in a
ternary complex,5 and on the basis of this result as well as kinetic
data,6 Davidson and co-workers proposed that ami is oxidized
within the ternary complex. ET from cyt. c-551 to the membrane-
bound oxidase was proposed to be mediated by cyt. c-550.3
Recently, we showed that ami can reduce cyt. c-551, but only after
dissociation from MADH, that is, in a ping-pong mechanism. We
could not detect the ternary complex in solution.7 The important
role of the cytochromes in ET is not supported by a genetic study
demonstrating that knockout of the genes encoding the three
soluble c-type cytochromes, cyt. c-550, c-551, and c-553, does not
prevent growth on MA.8 Apparently, these cytochromes are not
essential and various routes of ET are possible. We wondered
The general view of protein interactions is that specificity is
essential for biochemical processes to distinguish partners from
nonpartners. In specific complexes, the proteins assume a single,
well-defined orientation, stabilized by multiple short-range inter-
actions, like salt bridges, H-bonds, van der Waals contacts and
shielding of hydrophobic residues. In nonspecific complexes,
proteins can assume multiple orientations. Electrostatic attrac-
tion between patches of charged residues on the proteins can
stabilize nonspecific complexes and result in preorientation,
creating preferred docking sites. However, the Coulombic force
is a long-range interaction and on its own not sufficient to result
in a specific complex. Many proteins are highly charged and, thus,
such nonspecific complexes must be common, but these inter-
actions are normally expected to have no biological function.
Here, we report on a network of protein interactions that
functions effectively in the transport of electrons, without a
requirement for specificity.
The Gram-negative bacterium Paracoccus denitrificans can use
methylamine (MA) as sole source of carbon and energy. The
periplasmic enzyme MA dehydrogenase (MADH) converts this
substrate to formaldehyde and ammonia and transfers the
Received: June 1, 2011
Published: September 14, 2011
r
2011 American Chemical Society
16861
dx.doi.org/10.1021/ja205043f J. Am. Chem. Soc. 2011, 133, 16861–16867
|

Journal of the American Chemical Society
ARTICLE
how promiscuous the proteins responsible for ET from MA
to dioxygen are in their interactions. To address this question,
the oxidation rate constants of ami by all possible partners, as
well as most of the rates of ET among the cytochromes, were
determined with an in vitro steady state reconstitution assay.
The interaction of ami with all partners was also characterized
with NMR spectroscopy. The pattern of reactivities shows
that ET from ami involves a network of nonspecific protein
complexes, rather than a defined chain of interactions. We
discuss why the formation of specific complexes may be
functionally unnecessary or even unfavorable for small ET
proteins and speculate that having a pool of complementary
cytochromes may represent an adaptive advantage for the
bacterium.
Table 1. Reaction Parameters for the MA ET Networka
reaction
parameter
26 ( 2 sÀ1 b
9 ( 2b
MADH ami f ami + MADH
k1
k
3
Ami + MADH f MADH ami
À1
3
Ami f aa3
k2
k3
k4
k5
k6
k7
k8
k9
0.17 À 0.26b,c
0.0039 ((0.0003)b
13 ((0.7)
MADH f c-550
c-550 f aa3
app
Ami f c-550
Ami f c-551
c-551 f c-550
Ami f c-553
c-553 f c-550
MADH f c-552
c-552 f aa3
0.0032 ((0.0001)
>k7
3.1 ((0.2)
>k9
0.66 ((0.05)
not determined
6.1 ((0.2)
k10
k11
k12
k13
k14
k15
k16
’ RESULTS
app
Ami f c-552
Ami f c-551
c-551 f c-552
Ami f c-552
c-553 f c-552
c-551 f aa3
0.0051 ((0.0002)
>k14
Kinetics. To study the protein interactions involved in the
transfer of electrons from ami to the terminal oxidase cyt. aa3,
steady-state electron transfer reactions were measured with an
oxygraph. To an air-saturated solution containing MA in an
oxygraph cell, cyt. aa3 and MADH were added. No oxygen
consumption was detected, showing that MA-reduced MADH
cannot donate electrons directly to cyt. aa3 for reduction of
molecular oxygen. Addition of ami results in a constant decrease
of the oxygen concentration in time, indicating that ami catalyzes
the electron transfer from MADH to the oxidase, without the
need for any c-type cytochromes. We reported this surprising
finding recently in a previous paper.7 In the absence of either
MADH or cyt. aa3, ami is not capable of oxygen reduction on the
same time scale. The oxygen consumption depends nearly
linearly on the amicyanin concentration up to at least 100 μM.
The value of bimolecular rate constant k2 is listed in Table 1, along
with all other constants determined in the previous and in the
present study for the MA-network, which is represented in Figure 1.
To establish which of the c-type cytochromes can reduce cyt.
aa3, the steady-state oxidation of each of the four reduced c-type
cytochromes by cyt. aa3 and oxygen was measured using a
spectrophotometric assay. Cyt. c-550 and c-552 exhibit fast ET
to cyt. aa3, as expected from previous studies.9,10 A linear fit to the
concentration dependence (Figure S1) yields the rate constants k4
and k11, respectively. In contrast, cyt. c-551 and c-553 show no
detectable ET to cyt. aa3, in line with their reported inability to
become oxidized by intact spheroplasts of P. denitrificans.3
To determine whether the c-type cytochromes can catalyze ET
between ami and cyt. aa3, the oxygraph assay was repeated in the
presenceofboth ami and each ofthe c-type cytochromes. Figure 2
shows that cyt. c-551 and c-553 do not accelerate the reduction of
oxygen, but rather have a slight inhibitory effect. An acceleration
is not expected, because both cytochromes are unreactive toward
the oxidase. The slight inhibition may be related to complex
formation between ami and these cytochromes (see below).
Addition of either cytochrome c-550 or c-552 results in faster
ET to the cyt. aa3. Above a concentration of 5 μM, the increase in
oxygen consumption is linear with increasing cytochrome con-
1.07 ((0.06)
>k16
0.183 ((0.009)
no ET detected
no ET detected
2.9 ((0.7)
c-553 f aa3
MA
MADH ami H ami +
Kd
Kd
3
MADH (k1/k )
À1
MC550
MADH c-550 H c-550 + MADH
40 ((6)b
3
Aaa3
Ami aa3 H ami + aa3
Kd,app
12 ((1) (2:1)d
323 ((13)d
80 ((4)b,d
>1000d
3
AC550
Ami c-550 H ami + c-550
Kd
Kd
Kd
3
AC551
AC552
AC553
Ami c-551 H ami + c-551
3
Ami c-552 H ami + c-552
3
Ami c-553 H ami + c-553
Kd,app
58 ((3) (2:1)d
3
a The rate constants (k) are defined in Figure 1 and given in μMÀ1 sÀ1
,
except for k1. The dissociation constants (Kd) are given in μM. b Ref 7.
c Some variation in k2 was observed due to activity differences of the cyt.
aa3 in different experiments. d Determined from NMR titrations.
Figure 1. Schematic representation of the ET network for MA oxida-
tion. The arrows indicate electron flow, with the k symbols referring to
the reduction rate constants, except for k1 and kÀ1, which represent the
dissociation and association rate constants of ami and MADH, respec-
tively. The black and gray boxes denote negatively charged and dipolar
proteins, respectively (see Discussion).
constants are actually small, in the range of 103 MÀ1 À1, and
s
app
app
centration, yielding k5 and k12 for cyt. c-550 and c-552,
respectively, because the rate-limiting step is the ET from ami to
these cytochromes and not from the cytochromes to cyt. aa3. The
enhancement of the rate compared to that in the absence of these
cytochromes appears to be large in Figure 2, but it should be
noted that the concentration of the c-type cytochromes is up to
500-fold larger than the concentration of cyt. aa3. Both rate
about 50-fold lower than the rate for ET from ami directly to the
app
oxidase (k2). The value for k5 is close to what has been
reported for the equivalent reaction for Paracoccus versutus
proteins.11 The small rate constants for the reduction of the
cytochromes indicate that the interactions between ami and cyt.
c-550 and c-552 are unfavorable.
16862
dx.doi.org/10.1021/ja205043f |J. Am. Chem. Soc. 2011, 133, 16861–16867

Journal of the American Chemical Society
ARTICLE
Figure 2. Steady-state ET from MA to dioxygen. The rate at 0 μM
represents ET from ami to cyt. aa3. Addition of cyt. c-550 and c-552, but
not c-551 and c-553, enhances the rate. The solid lines represent linear
fits. Conditions: ami, 10 μM; MADH, 200 nM (protomer con-
centration); cyt. aa3, 40 nM; MA, 1 mM.
Figure 4. Binding curves for ami and cyt. aa3. The chemical shift
changes relative to those of free ami are plotted for the ratio of ami and
cyt. aa3. The solid lines represent a global fit to a 2:1 binding model with
a single apparent dissociation constant of 12 μM. The residue number
and observed amide nucleus are listed for each symbol. The error bars
represent estimated experimental errors.
than step 2, despite the unfavorable free energies of these
reactions (see below). Remarkably, all these rates are 3 orders
of magnitude higher than those for the interactions of ami with
cyt. c-550 and c-552, indicating that the interactions of ami with
both cyt. c-551 and cyt. c-553 are much more favorable.
NMR Spectroscopy. The interactions between ami and the
cytochromes were also studied using NMR spectroscopy. Per-
deuterated and 15N-labeled Zn(II)-ami was titrated into a
solution of cyt. aa3, resulting in chemical shift perturbations of
the amide resonances in a TROSY spectrum. The copper in ami
was substituted with Zn(II) to mimic the cupric form and avoid
line broadening by ET and the paramagnetic nature of Cu(II).
The binding curve (Figure 4) cannot be fitted well with a 1:1
binding model, but a model assuming a 2:1 stoichiometry
for ami/cyt. aa3 gives an excellent fit with a single, apparent
dissociation constant (Kd,appAaa3) of 12 ((1) μM.
No perturbations were observed for ami in the MADHÀami
complex upon addition of cyt. aa3, indicating that the interaction
between ami and cyt. aa3 is lost when ami is bound to MADH
(results not shown). Similarly, it has been observed that the
interaction between ami and cyt. c-551 is lost upon complex
formation of ami with MADH.7 The binding site on ami for cyt.
aa3 is well-defined and comprises the so-called hydrophobic
patch and surrounding positive residues, similar to the binding
site for MADH and cyt. c-551 (Figure 5).7,12
Figure 3. Steady-state ET from MA to dioxygen. The rate at 0 μM
represents ET from ami to cyt. aa3 and the solid symbols show the effect
of the addition of either cyt. c-550 or c-552 only (data are the same as in
Figure 2). The open circles show the effect of addition of cyt. c-550 or
c-552 in the presence of 100 nM cyt. c-551. The solid lines represent linear
fits yielding k7 and k14, respectively. For the cyt. c-550 + c-551 data set,
the two points at highest cyt. c-550 concentration were not used in the
fitting. Conditions: ami, 10 μM; MADH, 200 nM (protomer con-
centration); cytochrome aa3, 40 nM; MA, 1 mM.
The interaction of ami with the c-type cytochromes was
studied by titrating the cytochrome into a solution of 15N-labeled
Zn(II)Àami. Cyt. c-550, c-552, and c-551 were in the reduced
state, whereas cyt. c-553 was oxidized. The natural redox states
would be reduced for ami and oxidized for the cytochromes, but
the resulting ET would affect the NMR spectra negatively. Each
of the cytochromes shows an interaction with ami, albeit with
varying affinities. The binding curves are shown in Figure S3 and
the Kd values are reported in Table 1. For cyt. c-550 and cyt.
c-551,7 a 1:1 binding curve fits the data well, whereas for cyt. c-553,
a 2:1 (ami/cyt. c-553) model is required. The reason for this ratio
is unclear, but cyt. c-553 is larger (22 kDa) than the other
cytochromes. Its sequence has a single heme binding motif. Cyt.
We then wanted to establish whether cyt. c-551 or c-553 could
catalyze ET between ami and cyt. c-550 or c-552. We had
reported before that cyt. c-551 enhances the ET to cyt. c-5507
and this finding was reproduced in the current study. Further-
more, it was found that titrating either cyt. c-550 or c-552 into the
oxygraph assay in the presence of ami and a fixed amount of cyt.
c-551 (Figure 3) or c-553 (Figure S2) in all cases resulted in much
faster oxygen consumption than without cyt. c-551 or c-553.
Again, a linear increase in the rate is observed with increasing
concentration of cyt. c-550/c-552, suggesting that ET from cyt.
c-551/c-553 to cyt. c-550/c-552 (step 2) is rate limiting. This result
implies that ET from ami to cyt. c-551/c-553 (step 1) is faster
16863
dx.doi.org/10.1021/ja205043f |J. Am. Chem. Soc. 2011, 133, 16861–16867

Journal of the American Chemical Society
ARTICLE
Figure 5. Binding maps for ami. The Δδavg (see Materials and Methods) was plotted on a surface representation of ami (PDB entry 1AAN17). Δδavg
(ppm) > 0.1, red; Δδavg > 0.04 orange; Δδavg > 0.02 yellow; Δδavg e 0.02, blue, except for cyt. c-552, for which Δδavg e 0.02, orange, Δδavg e 0.01
yellow and Δδavg e 0.006, blue, were used; no data, gray. The data for cyt. c-551 have been taken from ref 7. In the bottom image, ami is rotated by 180°
around the vertical axis in the plane, compared to the top image.
c-552 induces perturbations that are too small to be fitted reliably,
suggesting a weak and dynamic interaction. The affinity for cyt.
c-550 is low, whereas the interaction with cyt. c-551 and cyt. c-553 is
in the normal range for transient electron transfer complexes.13,14 The
binding maps are shown in Figure 5. Clearly, ami uses the same
surface area for the interaction with all the cytochromes. This
finding suggests that ami does not bind to more than one partner
simultaneously, consistent with a ping-pong mechanism of ET.7
Cyclic Voltammetry. To determine the driving forces for ET
between ami and the c-type cytochromes, the midpoint potential
was determined for all proteins as a function of pH. The results
(Figure 6) show a very strong pH dependence for ami, in line
with published work,15 as a consequence of the change in copper
coordination at low pH due to protonation of His 95. Also the
data for the cytochromes can be fitted with protonation curves
(Table S1), except for cyt. c-551, for which the potential shows a
negligible pH dependence. At pH 7.5, used in the kinetic
experiments, the midpoint potential difference (cyt. c À ami)
Figure 6. Midpoint potentials (Em) for ami and c-type cytochromes as
a function of pH as determined by cyclic voltammetry. The solid lines
represent fits of the data to eq S5. The fitted parameters are listed in
Table S1. For cyt. c-551, no pK values could be resolved.
is À1, À45, +38, and À85 mV for cyt. c-550, c-551, c-552, and c-553,
respectively. Note that ET from ami to cyt. c-551 and, in
particular, cyt. c-553 is uphill.
’ DISCUSSION
ET from ami to cyt. c-550 and c-552 is probably too slow to be of
much importance physiologically, despite the fact that the driving
forces for these reductions are not unfavorable.
Our results indicate that electrons can flow from ami to cyt.
aa3 via several routes. Ami can form a transient complex with the
oxidase, with an apparent 2:1 stoichiometry and an apparent
affinity, at low ionic strength, in the low micromolar range. Ami
reduces cyt. aa3 with a rate that could well have physiological
relevance. These findings are in accord with the genetic study by
Van Spanning et al.8 that demonstrated that P. denitrificans can
still utilize MA as sole carbon and energy source in the absence of
cyt. c-550, c-551, and c-553, although those authors suggested ET
from ami to cyt. bc1 and then to either cyt. cbb3 or cyt. aa3.10,16
In contrast, ET to cyt. c-553 and c-551 is fast, even though
these reactions have positive free energy changes. The
reactivities of the c-type cytochromes toward cyt. aa3 are
opposite to those toward ami. Cyt. c-550 and c-552, which are
slowly reduced by ami, reduce the oxidase rapidly, whereas
cyt. c-551 and c-553, that are rapidly reduced by ami, do not
reduce cyt. aa3. Consequently, apart from direct ET from ami
to cyt. aa3, only three-step ET from ami via cyt. c-551/c-553
16864
dx.doi.org/10.1021/ja205043f |J. Am. Chem. Soc. 2011, 133, 16861–16867

Journal of the American Chemical Society
ARTICLE
Figure 7. The MA ET network. The structures of the proteins studied in this work are shown in surface representation, colored according to surface
charge. For MADH, cyt. c-551, and cyt. aa3, the scale ranges from À12 to +12 au, and for ami and cyt. c-550 and c-552 from À4 to +4 au. Negative charges
are in red, positive in blue. For cyt. c-553, no structure is available, but the sequence (UniProt entry P29967) indicates that it is a highly negative protein.
The green areas mark the sites for strong electronic coupling to the redox center. Cyt. c-552 and cyt. aa3 are depicted as integral membrane proteins. The
black solid arrows indicate fast ET to or from ami. The dotted lines represent slow ET from ami, the green lines depict rapid ET between cytochromes,
and the thin lines with a cross indicate that no ET was observed. PDB entries: MADH, 2BBK;20 ami, 1AAN;17 cyt. c-550, 155C;21 cyt. c-551, 2MTA;5 cyt.
c-552, 1QL4;22 cyt. aa3, 3MK7.23 Surface representations were produced using DeepView (Swiss-PdbViewer, http://spdbv.vital-it.ch/).
and c-550/c-552 to cyt. aa3 is likely to contribute signifi-
cantly to the flow of electrons (Figure 1).
Information, Table S2 and Figure S4). In MA-cultured cells, the
amount of cyt. c-553 is much lower than of MADH, ami, cyt.
c-550, and cyt. c-551, which are found in approximately equal
quantities. This estimation suggests that the ET flow through the
amiÀcyt. c-551Àcyt. c-550Àcyt. aa3 route is dominant. How-
ever, the other interactions are functional and kinetically possible
and thus can be expected to occur. The concept of a strictly defined
linear pathway amidst a multitude of nonfunctional interactions
is not applicable here. The relatively moderate reduction rates
and binding affinities and the finding that the fastest reactions are
uphill suggest that the network consists of protein interactions
that have not been optimized by evolution. The binding site on
ami is similar for all cytochromes and can therefore not be highly
complementary to any one partner. Nevertheless, ET from ami
to dioxygen is sufficiently fast to allow P. denitrificans to grow on
MA. An analysis of the electrostatic properties of the proteins
supports the view of an ET network dominated by electrostatic
interactions. In Figure 7, the proteins are shown in surface
representation, colored by charge. MADH, cyt. c-551, and the
hydrophilic domain of subunit II of cyt. aa3 are strongly
negatively charged (red). Ami, cyt. c-550, and c-552 are dipolar,
with the patch reactive in ET (green) in the positively charged
(blue) hemisphere of the protein. For cyt. c-553, no structure is
available, but the sequence contains nearly twice more negative
than positive residues. Ami reacts fast with proteins that have a
negative reaction patch and slowly with cyt. c-550 and c-552.
Similarly, the negative c-type cytochromes react rapidly with
those with positive reactive patches and not with the other
negative proteins. Thus, electrostatic attraction and repulsion
between the partners alone is sufficient to explain the observed
pattern of reactivity. The rate of ET depends on the driving force,
the reorganization energy, and the electronic coupling between
the redox centers.24 The driving force and reorganization energy
are not significantly influenced by complex formation, but the
electronic coupling depends strongly on the interaction between
the proteins. In large ET complexes, a specific state is required, in
which the redox centers are brought close enough to allow for
sufficient electron coupling and, thus, fast ET.25 For complexes of
small electron transfer proteins like ami and the c-type cyto-
chromes, it is possible to achieve very efficient electronic coupling by
It might appear as if the uphill reactions from ami to cyt. c-551/
c-553 cannot be relevant for the oxidation of MA under physio-
logical conditions. However, our experiments with the recon-
stituted redox chain show the opposite. The overall driving force
for the reaction of MA and dioxygen to formaldehyde, ammonia,
and water is very large (≈1 V). All the proteins in the redox chain
are catalysts and, therefore, cannot alter the free energy change of
this reaction. The uphill reaction from ami to cyt. c-551 or c-553
could only represent a kinetic barrier. However, a short distance
between the copper and heme cofactors can be achieved in these
small proteins, enabling an efficient electronic coupling and, thus,
fast ET from ami to these cytochromes. In combination with
rapid (downhill) ET from cyt. c-551/c-553 to cyt. c-550/c-552,
the role of cyt. c-551/c-553 as ET mediators can still be effective,
as also pointed out by Dutton and co-workers for other systems.18
Our finding that cyt. c-551 is the most reactive of the
cytochromes toward ami is in agreement with earlier studies.2
Compelling evidence indicates that it is free ami, not the ami-
MADH complex, that transfers electrons to cyt. c-551.7 However,
it is clear that cyt. c-553 can replace cyt. c-551. Also cyt. c-550 and
c-552 exhibit very similar properties in the ET reactions studied
here. At least for the growth on MA, there appears to be
redundancy in the available ET proteins, a finding that explains
why single knockouts of the genes of the c-type cytochromes have a
moderate16 or no effect8,10 on the growth of P. denitrificans on MA.
However, nonredundant roles may exist for these cytochromes for
ET from other donors. For example, genetic studies have shown
that only cyt. c-551 can act as a partner for methanol dehydro-
genase.19 In this case, cyt. c-553 cannot act as a substitute.
The results of the current study suggest that ami can interact
functionally with several proteins. ET from MADH to ami
involves a specific, well-defined interaction, whereas ET from
ami to cyt. aa3 is best described by a network, rather than a chain,
of protein interactions. The relative concentrations of the
proteins are also important in determining the contribution of
each branch of the ET network. On the basis of the protein yields
from P. denitrificans cells, a crude estimation can be made about
the relative concentrations and the electron fluxes (Supporting
16865
dx.doi.org/10.1021/ja205043f |J. Am. Chem. Soc. 2011, 133, 16861–16867

Journal of the American Chemical Society
ARTICLE
Table 2. Potential Monoheme c-Type Cytochromes in
P. denitrificans Pd1222 (http://genome.jgi-psf.org/parde/
parde.home.html)
proteins eluting in different fractions. The soluble domain of cyt. c-552
was prepared as published.9 The production of isotope-labeled Zn-
(II)Àami followed the reported procedure.7
Activity Assays. Kinetic assays were performed at 20 °C in 1 mM
MA hydrochloride, 10 mM potassium phosphate (pH 7.5), 50 mM KCl,
1 mM EDTA, and 0.1%, dodecyl maltoside. Oxygen consumption was
measured with a Hansatech Instruments oxygraph equipped with a Clark
electrode. The oxygraph was calibrated with sodium dithionite and the
reaction volume was 400 μL. The rate of oxygen consumption was
determined by measuring the slope of concentration decrease. A Varian
Cary 50 UVÀvis spectrophotometer was used to measure the steady-
state oxidation rate of the cytochromes by cyt. aa3. Cytochromes were
reduced with sodium ascorbate and applied to a PD-10 desalting column
(GE Healthcare) to remove the ascorbate. The reaction volume was 100
μL and the reaction was initiated by the addition of cyt. aa3. Errors were
set to the SD of duplicate experiments or estimated from the experi-
mental measurements and propagated. Errors of the protein concentra-
tions were estimated to be 5%. The models for the kinetic analysis are
described in the Supporting Information.
size (aa)
genetic
genome locus
name
mature protein
neighborhood
0021
1808
1937
2307
2489
2544
2995
3238
4150
4700
c-553
c-552
c-550
c1
198
148
134
425
86
XDH
Cyt. c reductase subunits
Nitrogen metabolism
Cobyrinate synthesis
Methanol oxidation
c-55x
131
154
130
133
131
c-551
SoxX
Sulfur oxidation
Acriflavin resistance
bringing the redox centers in close proximity. The favorable
electrostatic interactions can ensure that such positioning occurs
in a large fraction of the encounters of two proteins.26,27 A well-
defined single orientation, as observed in specific proteinÀprotein
complexes, is not required in small ET protein complexes and may
even be unfavorable, because it could result in lower turnover
rates.25,28À30 Thus, a network of electrostatic interactions is
sufficient to transfer electrons quickly and effectively down the
free energy gradient.
NMR Titrations. NMR experiments were performed at 300 K on a
Bruker Avance DMX 600 MHz NMR spectrometer equipped with a
TCI-Z-GRAD ATM cryoprobe. All solutions contained 20 mM Hepes
(pH 7.6) and 6% D2O. For samples containing cyt. aa3, 0.1%, dodecyl
maltoside was added. TROSY spectra were recorded during the inverse
titration of a 70 μM cyt. aa3 sample with a 0.9 mM stock solution of
2
15NÀ HÀZn(II)Àami. For the c-type cytochrome titration experi-
1
ments, a series of 15NÀ H HSQC and 1D spectra of a 100 μM
15NÀZn(II)Àami sample with increasing concentrations of cyto-
chromes was recorded. All cytochromes had been reduced with sodium
ascorbate, except for cyt. c-553 which had been oxidized with K3Fe-
(CN)6. Data processing and analysis of the chemical shift perturbations,
including the definition of Δδavg, followed the published procedure,7
except that a correction factor of 2 for the cytochrome concentration was
introduced for fitting the binding curves of ami with cyt. aa3 and cyt. c-
553, to account for a 2:1 binding ratio.
This observation poses the question whether there could be an
evolutionary advantage to produce several c-type cytochromes
with similar properties, capable of forming electrostatically favored,
but nonspecific, complexes with other proteins. The genome of
P. denitrificans strain Pd1222 has been sequenced. Table 2 lists
the proteins likely to be monoheme c-type cytochromes. The
genes are spread over the genome, with separate promoters and
most are located close to possible partner redox enzymes. Cyt.
c-551 and c-553 are induced simultaneously, probably by
formaldehyde,31 a common intermediate in the metabolism of
various C1 compounds. Cyt. c-550 appears to be constitutively
expressed.31 In our cultures grown on MA, we found three
soluble cytochromes, so it is assumed that the other soluble c-type
cytochromes are expressed only under different conditions.
We speculate that such a pool of ET proteins makes it possible
to ‘plug-in’ new metabolic routes. When the genes for enzymes
capable of oxidizing a particular substrate are acquired via lateral
gene transfer, as appears to be very common among bacteria,32
the minimal requirement to benefit from such an acquisition is
the need to transfer the electrons to a terminal oxidase. The pool
of c-type cytochromes ensures that the electrons will be trans-
ferred one way or another. We note that the entire set of genes for
MA growth is encoded by a single stretch of DNA (mau operon),
including the gene for ami and the genes required for the
processing of MADH and its TTQ-cofactor.33À37 This set might
simply have been acquired from another bacterium. Such a ‘plug-
and-grow’ mechanism would be highly beneficial for a nonmotile
soil bacterium like P. denitrificans that may be confronted with
strongly varying conditions, requiring metabolic versatility.
Cyclic Voltammetry. The reduction potentials of ami, cyt. c-550,
cyt. c-551, cyt. c-552 (soluble domain), and cyt. c-553 were determined
by cyclic voltammetry. The proteins were diluted to 69, 26, 16, 63, and
14 μM, respectively, in 90 mM of buffer (citric acid, MES, MOPS, TAPS,
or CAPS, titrated with NaOH to the desired pH between 4.0 and 10.6).
The working electrode was a 3.1 mm2 gold electrode (CH-instruments),
polished with suspensions of 1 μm, 0.3 μm, and 50 nm alumina on
microcloth (Buehler), sonicated in water, subsequently incubated for at
least 2 h in 1 mM aqueous mercaptohexanol (Sigma), and rinsed
thoroughly with water. The working electrode was mounted on a glass
microliter cell as described by Hagen,38 modified to accommodate the
CHI working electrode, and equipped with a water jacket connected to a
refrigerated circulating water bath (VWR) at constant temperature of
24 °C. The reference electrode was a saturated calomel electrode (SCE,
Radiometer REF401), and the counter electrode was a Pt wire. The
electrodes were connected to an Autolab PGstat12 potentiostat (Metrohm),
controlled by GPES 4.9 software. A 20 μL droplet of buffered protein
solution was suspended between the three electrodes, the cell headspace
was flushed with Ar, and staircase cyclic voltammograms were recorded
with a step size of 2 mV at a scan rate of 10 mV/s (20 mV/s for ami, 5 mV/s
for cyt. c-550), where the responses are (nearly) ideally reversible. The
midpoint potentials were determined by averaging the oxidative and
reductive peak positions of four cycles, and converted to the standard
hydrogen electrode scale by adding the SCE potential of 245 mV at 24 °C.39
’ MATERIALS AND METHODS
’ ASSOCIATED CONTENT
Protein Preparation. The purification of MADH, ami, cyt. c-550,
cyt. c-551, and cyt. aa3 from P. denitrificans has been reported before.7
Cyt. c-553 was obtained with the same protocol as cyt. c-551, with the
S
Supporting Information. Kinetic models, details of the
b
cyclic voltammetry, additional kinetic data and NMR binding
16866
dx.doi.org/10.1021/ja205043f |J. Am. Chem. Soc. 2011, 133, 16861–16867

Journal of the American Chemical Society
ARTICLE
curves. This material is available free of charge via the Internet at
http://pubs.acs.org.
(24) Marcus, R. A.; Sutin, N. Biochim. Biophys. Acta 1985,
811, 265–322.
(25) Bashir, Q.; Volkov, A. N.; Ullmann, G. M.; Ubbink, M. J. Am.
Chem. Soc. 2010, 132, 241–247.
(26) Ubbink, M. FEBS Lett. 2009, 583, 1060–1066.
(27) Schreiber, G.; Haran, G.; Zhou, H. X. Chem. Rev. 2009,
109, 839–860.
(28) Xu, X. F.; Reinle, W. G.; Hannemann, F.; Konarev, P. V.;
Svergun, D. I.; Bernhardt, R.; Ubbink, M. J. Am. Chem. Soc. 2008,
130, 6395–6403.
’ AUTHOR INFORMATION
Corresponding Author
m.ubbink@chem.leidenuniv.nl
’ ACKNOWLEDGMENT
(29) Xiong, P.; Nocek, J. M.; Griffin, A. K. K.; Wang, J. Y.; Hoffman,
F.W., H.A.H. and M.U. were supported by The Netherlands
Organization for Scientific Research (NWO), VIDI grant
700.53.424 (F.W. and H.A.H.) and VICI grant 700.58.441
(M.U.). The Parma group was supported by the Ministry of
University and Research, MIUR, Grant PRIN 20074TJ3ZB_004.
B.L. acknowledges the support of DFG (SFB472 and CEF-MC,
Project EXC 115).
B. M. J. Am. Chem. Soc. 2009, 131, 6938–6939.
(30) Liang, Z. X.; Jiang, M.; Ning, Q.; Hoffman, B. M. J. Biol. Inorg.
Chem. 2002, 7, 580–588.
(31) Baker, S. C.; Ferguson, S. J.; Ludwig, B.; Page, M. D.; Richter,
O. M. H.; van Spanning, R. J. M. Microbiol. Mol. Biol. Rev. 1998,
62, 1046–1078.
(32) Treangen, T. J.; Rocha, E. P. C. PLoS Genet. 2011, 7, e1001284.
(33) Chistoserdov, A. Y.; Boyd, J.; Mathews, F. S.; Lidstrom, M. E.
Biochem. Biophys. Res. Commun. 1992, 184, 1181–1189.
(34) van der Palen, C. J. N. M.; Slotboom, D. J.; Jongejan, L.;
Reijnders, W. N. M.; Harms, N.; Duine, J. A.; van Spanning, R. J. M. Eur.
J. Biochem. 1995, 230, 860–871.
(35) Delorme, C.; Huisman, T. T.; Reijnders, W. N. M.; Chan, Y. L.;
Harms, N.; Stouthamer, A. H.; van Spanning, R. J. M. Microbiology 1997,
143, 793–801.
(36) van der Palen, C. J. N. M.; Reijnders, W. N. M.; deVries, S.;
Duine, J. A.; van Spanning, R. J. M. Antonie van Leeuwenhoek 1997,
72, 219–228.
’ REFERENCES
(1) Husain, M.; Davidson, V. L. J. Biol. Chem. 1985, 260, 4626–4629.
(2) Husain, M.; Davidson, V. L. J. Biol. Chem. 1986, 261, 8577–8580.
(3) Davidson, V. L.; Kumar, M. A. FEBS Lett. 1989, 245, 271–273.
(4) Turba, A.; Jetzek, M.; Ludwig, B. Eur. J. Biochem. 1995, 231, 259–265.
(5) Chen, L.; Durley, R. C.; Mathews, F. S.; Davidson, V. L. Science
1994, 264, 86–90.
(6) Davidson, V. L.; Jones, L. H. J. Biol. Chem. 1995, 270, 23941–
23943.
(37) van Spanning, R. J. M.; Wansell, C. W.; Reijnders, W. N. M.;
Oltmann, L. F.; Stouthamer, A. H. FEBS Lett. 1990, 275, 217–220.
(38) Hagen, W. R. Eur. J. Biochem. 1989, 182, 523–530.
(39) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Funda-
mentals and Applications; Wiley & Sons: New York, 2001.
(7) Meschi, F.; Wiertz, F.; Klauss, L.; Cavalieri, C.; Blok, A.; Ludwig,
B.; Heering, H. A.; Merli, A.; Rossi, G. L.; Ubbink, M. J. Am. Chem. Soc.
2010, 132, 14537–14545.
(8) van Spanning, R. J. M.; Wansell, C. W.; Reijnders, W. N. M.;
Harms, N.; Ras, J.; Oltmann, L. F.; Stouthamer, A. H. J. Bacteriol. 1991,
173, 6962–6970.
(9) Reincke, B.; Thony-Meyer, L.; Dannehl, C.; Odenwald, A.;
Aidim, M.; Witt, H.; Ruterjans, H.; Ludwig, B. Biochim. Biophys. Acta,
Bioenerg. 1999, 1411, 114–120.
(10) Otten, M. F.; van der Oost, J.; Reijnders, W. N. M.; Westerhoff,
H. V.; Ludwig, B.; van Spanning, R. J. M. J. Bacteriol. 2001,
183, 7017–7026.
(11) Ubbink, M.; Hunt, N. I.; Hill, H. A.; Canters, G. W. Eur. J.
Biochem. 1994, 222, 561–571.
(12) Chen, L.; Durley, R.; Poliks, B. J.; Hamada, K.; Chen, Z.;
Mathews, F. S.; Davidson, V. L.; Satow, Y.; Huizinga, E.; Vellieux, F. M.
Biochemistry 1992, 31, 4959–4964.
(13) Crowley, P. B.; Ubbink, M. Acc. Chem. Res. 2003, 36, 723–730.
(14) Prudencio, M.; Ubbink, M. J. Mol. Recogn. 2004, 17, 524–539.
(15) Zhu, Z. Y.; Cunane, L. M.; Chen, Z. W.; Durley, R. E.; Mathews,
F. S.; Davidson, V. L. Biochemistry 1998, 37, 17128–17136.
(16) van Spanning, R. J. M.; Wansell, C.; Harms, N.; Oltmann, L. F.;
Stouthamer, A. H. J. Bacteriol. 1990, 172, 986–996.
(17) Durley, R.; Chen, L. Y.; Mathews, F. S.; Davidson, V. L. Protein
Sci. 1993, 2, 739–752.
(18) Page, C. C.; Moser, C. C.; Chen, X. X.; Dutton, P. L. Nature
1999, 402, 47–52.
(19) van Spanning, R. J. M.; Wansell, C. W.; de Boer, T.; Hazelaar,
M. J.; Anazawa, H.; Harms, N.; Oltmann, L. F.; Stouthamer, A. H.
J. Bacteriol. 1991, 173, 6948–6961.
(20) Chen, L. Y.; Doi, N.; Durley, R. C. E.; Chistoserdov, A. Y.;
Lidstrom, M. E.; Davidson, V. L.; Mathews, F. S. J. Mol. Biol. 1998,
276, 131–149.
(21) Timkovich, R.; Dickerson, R. E. J. Biol. Chem. 1976, 251, 4033–
4046.
(22) Harrenga, A.; Reincke, B.; Ruterjans, H.; Ludwig, B.; Michel, H.
J. Mol. Biol. 2000, 295, 667–678.
(23) Koepke, J.; Olkhova, E.; Angerer, H.; Muller, H.; Peng, G. H.;
Michel, H. Biochim. Biophys. Acta, Bioenerg. 2009, 1787, 635–645.
16867
dx.doi.org/10.1021/ja205043f |J. Am. Chem. Soc. 2011, 133, 16861–16867