Multistep, eight-electron oxidation catalyzed by the cofactorless oxidase, PqqC: Identification of chemical intermediates and their dependence on molecular oxygen

Article abstract of DOI:10.1021/bi4003315

The final step of the biosynthesis of prokaryotic cofactor PQQ is catalyzed by PqqC, a cofactorless oxidase that brings about a ring closure and overall eight-electron oxidation of its substrate. Time-dependent acid quenching and subsequent high-performance liquid chromatography separation and mass spectrometric analyses of reaction mixtures were performed to correlate the structures of intermediates with previously observed UV-visible signatures. The reaction is composed of four stepwise oxidations: three steps use O₂ as the two-electron acceptor, and the fourth uses hydrogen peroxide (H ₂O₂). The chemical nature of the intermediates, the stoichiometry of the reaction, and their dependence on the oxygen concentration indicate that the third oxidation uses the product, H₂O₂, from the preceding step to produce water. The last oxidation step can also be studied separately and is a reaction between O₂ and PQQH₂ trapped in the active site. This oxidation is approximately 10 times slower than the reoxidation of PQQH₂ in solution. From the order of the four oxidation steps and their sensitivity to O₂ concentration, we propose a progressive closure of the active site as the enzyme proceeds through its catalytic cycle.

Full text of DOI:10.1021/bi4003315

Article
pubs.acs.org/biochemistry
Multistep, Eight-Electron Oxidation Catalyzed by the Cofactorless
Oxidase, PqqC: Identification of Chemical Intermediates and
Their Dependence on Molecular Oxygen
Florence Bonnot,†,§ Anthony T. Iavarone,†,§ and Judith P. Klinman*,†,‡,§
†Department of Chemistry, ‡Department of Molecular and Cell Biology, and §California Institute for Quantitative Biosciences (QB3),
University of California, Berkeley, California 94720, United States
ABSTRACT: The final step of the biosynthesis of prokaryotic cofactor PQQ is
catalyzed by PqqC, a cofactorless oxidase that brings about a ring closure and
overall eight-electron oxidation of its substrate. Time-dependent acid quenching
and subsequent high-performance liquid chromatography separation and mass
spectrometric analyses of reaction mixtures were performed to correlate the
structures of intermediates with previously observed UV−visible signatures. The
reaction is composed of four stepwise oxidations: three steps use O2 as the two-
electron acceptor, and the fourth uses hydrogen peroxide (H2O2). The chemical
nature of the intermediates, the stoichiometry of the reaction, and their
dependence on the oxygen concentration indicate that the third oxidation uses
the product, H2O2, from the preceding step to produce water. The last oxidation
step can also be studied separately and is a reaction between O2 and PQQH2
trapped in the active site. This oxidation is approximately 10 times slower than
the reoxidation of PQQH2 in solution. From the order of the four oxidation steps and their sensitivity to O2 concentration, we
propose a progressive closure of the active site as the enzyme proceeds through its catalytic cycle.
yrroloquinoline quinone [4,5-dihydro-4,5-dioxo-1H-
Ppyrrolo[2,3-f]quinoline-2,7,9-tricarboxylic acid (PQQ)] is
an aromatic, tricyclic o-quinone that serves as a cofactor for a large
number of prokaryotic dehydrogenases found predominantly in
Gram-negative bacteria.1−3 PQQ belongs to the quinone family of
enzymatic cofactors and has remarkable antioxidant properties.4−6
Like all identified quinone cofactors, PQQ is derived from a
ribosomally encoded precursor.5 However, unlike the other
described quinone cofactors, topaquinone (TPQ),7 tryptophan
tryptophylquinone (TTQ),8 lysyl tyrosylquinone (LTQ),9 and
cysteine tryptophylquinone (CTQ),10 which are generated in situ
within their cognate proteins, PQQ is peptide-derived and is
freely dissociable from its site of catalytic function.5
The biosynthesis of PQQ is a complex process and is
catalyzed by gene products pqqA−F in Klebsiella pneumoniae.11
All of the carbon and nitrogen atoms in PQQ derive from a
conserved tyrosine and glutamate located near the C-terminus
Figure 1. Image of WT PqqC without PQQ (green) (Protein Data
Bank entry 10TV) in the open conformation and WT PqqC with
of peptide substrate PqqA.12 The final step of PQQ formation
is catalyzed by PqqC (Figure 1) and involves a ring closure and
eight-electron oxidation of AHQQ [3a-(2-amino-2-carboxyethyl)-
4,5-dioxo-4,5,6,7,8,9-hexahydroquinoline-7,9-dicarboxylic acid]
(structure 1, Scheme 1).13,14 As shown in Scheme 1, AHQQ
must undergo a two-electron oxidative ring closure (A) followed
by three additional two-electron oxidation steps (BI−BIII). X-ray
studies of a mutant form of PqqC support ring closure as the
initial step in AHQQ oxidation,15,16 while the order of reactions
BI−BIII is currently unknown. The stoichiometry of the reac-
tion has been well-characterized, and the reaction occurs via the
production of only 2 molar equiv of hydrogen peroxide (H2O2),
PQQ (black) (Protein Data Bank entry 10TW) in the closed
conformation. PQQ is colored orange, and the O2 or H2O2 present in
the active site is colored purple. A subset of the active site residues
surrounding the PQQ are labeled in dark red. Note how the
conformational change brings R157, R179 and Y175 into the active
site.
Received: March 13, 2013
Revised: May 28, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/bi4003315 | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
a
Scheme 1. Reaction of WT PqqC with AHQQ
a
The species observed in acid quench experiments at 5% O2 are boxed in red. While these are illustrated as anions, mass spectrometric analyses were
of the acidified species. Note that species detected anaerobically are also seen at the lowest O2 level (5%) studied.
implicating bound hydrogen peroxide as the electron acceptor in
either step BI or BII;13 the biochemical detection of the final
products, H2O2 and PQQ during a single turnover of PqqC
further indicates that O2 is the electron acceptor in the final
product-forming step (BIII).13 The absence of any requirement for
an external cofactor or metal raises very intriguing questions regard-
ing how O2 is activated in this metal and cofactor free active site.
In this work, we sought to identify the substrate-derived
intermediates formed during a single turnover of PqqC catalysis
and to understand the impact of different oxygen concentrations
on their distribution. The results allow us to determine the inter-
mediates that accumulate during the reaction as a function of
O2 concentration and to establish the order of the different
oxidations. X-ray crystallographic studies of PqqC show two
conformations for PqqC; the free enzyme is open, with the
active site solvent-exposed, whereas the addition of PQQ leads
to significant reorganization and a more closed active site, Figure 1.13
The dependence of the various reaction steps on oxygen con-
centration, the order of the different oxidations, and the inter-
mediates that accumulate during the reaction all suggest that
the enzyme undergoes a progressive closure at its active site in
the course of the conversion of AHQQ to PQQ.
Methylobacterium extorquens AM1 as described previously.14
The concentration of AHQQ was determined spectrophoto-
metrically at pH 7 by averaging the concentrations determined
at three wavelengths with the following extinction coefficients:
15.7 mM−1 cm−1 at 222 nm, 8.26 mM−1 cm−1 at 274 nm, and
2.01 mM−1 cm−1 at 532 nm.
Expression and Purification of PqqC. Plasmids were
constructed in pET29 with kanamycin resistance and a
C-terminal His tag. PqqC was expressed and purified as
described previously.13 A single band was seen on a sodium
dodecyl sulfate−polyacrylamide gel electrophoresis gel, and
Ni-NTA column chromatography provided >95% pure enzyme.
The expression for PqqC without a His tag was identical, but
after the dialysis, the PqqC His tag was cut using a thrombin
kit from Novagen (catalog no. 69022-3). The protein was
incubated with the thrombin overnight at 4 °C, and the tag
was removed with the Ni-NTA column and the thrombin
via filtration after being mixed with streptavidin agarose. PqqC,
without the His tag, exhibits the same level of activity with
AHQQ as PqqC with the His tag.
Single-Turnover Kinetics Monitored by UV−Vis Spec-
troscopy. The enzyme (50 μM) was mixed with substrate,
AHQQ (20 μM), in 100 mM potassium phosphate buffer (pH 8.0)
in a total volume of 100 μL. The mixture was transferred to a
microcuvette, and UV−vis spectra were recorded within
10 s of mixing. The temperature of the reaction mixture was
maintained at 20 °C. Spectra were acquired at predetermined
time points and the data exported to SPECFIT/32 for analysis.
A baseline was obtained under identical conditions with enzyme
but without AHQQ. Despite the baseline subtraction, the
absorbance below 300 nm was very noisy and has been omitted
METHODS
■
Chemical and Molecular Biological Reagents. Buffers,
salts, general reagents, and culture media were obtained from
Sigma-Aldrich and Fisher Scientific and were of the highest
available purity. Nickel-nitrilotriacetic acid (Ni-NTA) resin
was purchased from Qiagen. PQQ was purchased from Fluka.
AHQQ was purified from pqqC mutant strain EMS12 of
B
dx.doi.org/10.1021/bi4003315 | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
in the data presentation. Anaerobic reactions were performed
under the same conditions as described above. The enzyme and
substrate were made anaerobic by three vacuum/nitrogen cycles
on the freeze solution and brought into an anaerobic chamber
(Plas Laboratories 855-AC glovebox). Protein samples were
transferred into a microcuvette covered with a rubber septum,
and the substrate was loaded into a gastight syringe before being
removed from the anaerobic chamber. The substrate was
injected through the septum and the solution thoroughly mixed
immediately prior to the spectra being recorded.
This reduced quinol was then bound to 125 μM PqqC enzyme
anaerobically in 100 mM KPi (pH 8.0) in a glovebox (Plas
Laboratories 855-AC glovebox). A microcuvette fitted with a
rubber stopper containing 80 μL of 100 mM KPi (pH 8.0) was
flushed for 20 min under a nitrogen/oxygen flow (the ratio of
oxygen was determined using an oxygen electrode). The protein
was transferred into a microcuvette fitted with a rubber stopper,
and the substrate was loaded into a syringe and the needle
inserted through the stopper into the top of the cuvette. The
cuvette with the syringe attached was brought to the spectro-
photometer, and the substrate was mixed thoroughly with the
enzyme before spectra were recorded. The spectral traces
were analyzed with SPECFIT. The data were fit to a single
exponential to find the rate of oxidation of quinol to quinone.
Acid Quench Experiment. The enzyme (100 μM) was
mixed with substrate, AHQQ (40 μM), in 100 mM potassium
phosphate buffer (pH 8.0) in a total volume of 80 μL. The reac-
tion was quenched by the addition of 4 μL of concentrated HCl.
The reaction mixture was centrifuged for 5 min at 12000 rpm,
and the supernatant was then loaded onto a C18 column equili-
brated with 5% acetonitrile with 0.1% TFA. Reaction inter-
mediates were eluted using a gradient of 5 to 20% acetonitrile
over a 20 min period. The absorbance was monitored between
250 and 700 nm using a diode array detector. Fractions were
collected and analyzed by mass spectrometry as described below.
Mass Spectrometry. Mass spectrometry analysis was per-
formed using an LTQ Orbitrap XL mass spectrometer equipped
with an electrospray ionization (ESI) source (Thermo Fisher
Scientific, Waltham, MA) that was connected in-line with an
Agilent (Santa Clara, CA) 1200 series autosampler and pump.
Sample solutions contained in polypropylene autosampler vials
sealed with septa caps (Wheaton, Millville, NJ) were loaded into
the autosampler compartment prior to analysis. The injected
sample (15 μL) was pumped to the ESI probe of the mass
spectrometer using a flow of acetonitrile (50 μL/min). The
connection between the autosampler and the ESI probe of the
mass spectrometer was made using PEEK tubing [0.005 in.
inner diameter × 1/16 in. outer diameter (Agilent)]. Mass spectra
were recorded in positive ion mode using the Orbitrap mass
analyzer, in profile format, with a resolution setting of 100000
(at m/z 400). Mass spectra were processed using Xcalibur version
2.0.7 SP1 (Thermo).
RESULTS
■
Single-Turnover Kinetics of Wild-Type PqqC at Variable
O2 Concentrations, Monitored by UV−Vis Monitoring.
Reactions were performed with excess enzyme (2.5 equiv) to
ensure that all of the reactions that were monitored occurred
from the E−S complex. Spectra at various time points during
the WT PqqC reaction are shown in Figure 2. At all dioxygen
concentrations between 0 and 1.14 mM, qualitative assessment
shows a decay of substrate at 532 nm, concomitant with the
appearance of a chromophore with a λmax of ∼318 nm. This
species subsequently decays, and a direct precursor−product
relationship with a new chromophore with a λmax at 344 nm
can be seen by the isosbestic point at 332 nm. The final
chromophore represents enzyme-bound PQQ (Figure 2).
Singular-value decomposition analysis yielded each individual
spectral component, and the data were fit to an irreversible, two-
consecutive-exponential model, which yielded observed rate
constants that are dependent on O2 concentration (Figure 3A,B).
Compilation of the results was achieved via the global fitting
procedure with SPECFIT (as described previously15). The
concentration of dioxygen in the reaction mixtures clearly
affects the kinetics. The formation of the reduced intermediate
at 318 nm yields rate constants ranging from 0.2 to 1.5 min−1
at oxygen concentrations of 0−1.14 mM (Figure 3A), while the
formation of the 344 nm species (PQQ) is slower and shows
rate constants of 0−0.2 min−1 in 0−1.14 mM O2 (Figure 3B).
The apparent Kd values for O2 during these two processes are
also distinct:
Determination of the Kd for PQQH2. Binding of PQQH2
to PqqC was measured fluorometrically. Assays were performed
in 100 mM potassium phosphate buffer (pH 8.0) at room
temperature. A fixed concentration of PQQH2 was employed
(100 nM), and the concentration of enzyme was varied. Samples
were incubated for 5 min, and fluorescence emission was
recorded at 333 nm (slit width of 5 nm) using an excitation
wavelength at 282 nm (slit width of 5 nm). Data were fit to a
quadratic binding isotherm (eq 1)
k1app
k2app
AHQQ ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯→ 318 nm ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯→ 344 nm
Kdapp(O2)=0.51 0.16mM
Kdapp(O2)=0.05 0.02mM
Although the 318 nm intermediate represents a quinol,15
there are four such species possible in the course of the overall
reaction (Scheme 1). Identification of this intermediate was
accomplished using acid quench techniques as described below.
Identification of the Reduced Quinol Intermediates at
Varying O2 Concentrations by Acid Quench. The reaction
intermediates formed under the different experimental condi-
tions were determined by a combination of high-performance
liquid chromatography (HPLC) separation and mass spec-
trometry. The UV−vis spectra, retention times, and masses for
the starting material and oxidized and reduced products, AHQQ,
PQQ, and PQQH2, respectively, at pH 2 were obtained by
comparison with those of pure samples. Under the conditions
established, AHQQ elutes at 11.5 min, occurs at m/z 339.08,
and shows λmax values of 275 and 510 nm; PQQ elutes at
16.9 min, occurs at m/z 331.02, and shows λmax values of
ΔF = ΔFM([E]T + [L]T + Kd
− {[([E]T + [L]T + Kd)2 − 4[E]T [L]T ]0.5}/(2[E]T))
(1)
where ΔF equals the change in fluorescence between a given
sample and the free ligand, ΔFM equals the change in fluore-
scence at saturation, [E]T equals the total concentration of the
enzyme in each sample, [L]T equals the total ligand concentra-
tion, and Kd is the dissociation constant of the enzyme−ligand
complex. Data were fit using Kaleidagraph (Synergy Software).
The determination of the Kd (PQQ-PqqC) for the wild-type
(WT) enzyme at different concentrations of O2 was performed
15
fluorometrically as previously described in a closed cuvette.
Quinol Oxidation Assays. PQQ (100 μM) was anaerobi-
cally reduced using 1.25 equiv of dithiothreitol (125 μM).
C
dx.doi.org/10.1021/bi4003315 | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
Figure 3. WT PqqC with AHQQ (at concentrations of 50 and 20 μM,
respectively) as a function of O2 concentration in 0.1 M phosphate
(pH 8) at 20 °C followed by UV−vis monitoring and fit with SPECFIT.
(A) k1 describes the formation of the 318 nm species. (B) k2 describes
app
app
Figure 2. Reaction of WT PqqC with AHQQ (at concentrations of
50 and 20 μM, respectively) as a function of time in 0.1 M phosphate
(pH 8) at 20 °C followed by UV−vis monitoring over 20 min. (A)
At 4% O2, k1 = 0.27 min−1 and k2 = 0.08 min−1. (B) At 21% O2, k1 =
0.75 min−1 and k2 = 0.13 min−1. (C) At 100% O2, k1 = 1.4 min−1 and
k2 = 0.2 min−1.
the formation of the 344 nm species. (C) Comparison of k2 observed
app
during the reaction of WT PqqC and AHQQ (red) and the kinetics of
PQQH2 reoxidation (black). For the latter, PQQH2 is formed by
reduction of PQQ by DTT and mixed with WT PqqC (30 μM PqqC
and 20 μM PQQH2) in different concentrations of O2 in 0.1 M
phosphate (pH 8) at 20 °C followed by UV−vis monitoring at 344 nm.
247 and 349 nm with a shoulder at 270 nm; and PQQH2 elutes
at 20.1 min, occurs at m/z 333.03, and shows a λmax value of
320 nm (Figure 4A).
At 21% O2 (240 μM O2), the rapid decay of substrate is
concomitant with the appearance of a chromophore with a λmax
at 303 nm (m/z 335.05) at an elution time of 17.5 min. This
initial intermediate subsequently decays, and a second inter-
mediate with the same mass, elution time, and spectrum as
PQQ forms (Figure 4B). Additionally, once PQQ begins to
accumulate, a small amount of PQQH2 with a λmax of 320 nm
(m/z 333.03), eluting at 20.1 min, is observed (this can be seen
in the 3 min reaction shown in Figure 4B). The intermediate
occurring at m/z 335.05 may represent the quinol or the
quinone. To identify which intermediate was accumulating
during the reaction of AHQQ with PqqC, the intermediate was
mixed with phenylhydrazine, a derivatization agent known to
react readily with quinones.17 The reaction of phenylhydrazine
with PQQ and PQQH2 has previously been studied. The
reaction with PQQ leads to the formation of a chromophore
D
dx.doi.org/10.1021/bi4003315 | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
Figure 4. HPLC traces for the reactant, intermediates, and product under a variety of conditions. (A) HPLC traces monitored at 275 nm of a
standard mixture (AHQQ, PQQ, and PQQH2). The elution times were as follows: 11.4 min for AHQQ, 16.9 min for PQQ, and 20.1 min for
PQQH2. (B) HPLC traces monitored at 275 nm for WT PqqC reaction with AHQQ (100 and 40 μM, respectively) in 0.1 M phosphate (pH 8) at
20 °C in 20% O2 quenched with HCl at different times (30 s, 90 s, 3 min, and 10 min), centrifuged, and loaded onto HPLC. The elution times were
as follows: 11.4 min for AHQQ, 16.9 min for PQQ, 17.5 min for the intermediate, and 20.1 min for PQQH2. (C) HPLC traces monitored at 275 nm
for WT PqqC reaction with AHQQ (100 and 40 μM, respectively) in 0.1 M phosphate (pH 8) at 20 °C in 100% O2 quenched with HCl at different
times (40 s, 70 s, 2 min, and 4 min), centrifuged, and loaded onto HPLC. The elution times were as follows: 11.4 min for AHQQ, 16.9 min for
PQQ, 17.5 min for intermediate, and 20.1 min for PQQH2. (D) HPLC traces monitored at 275 nm before 12.5 min (blue) and at 340 nm after 12.5
min (black) of WT PqqC reaction with AHQQ (100 and 40 μM, respectively) in 0.1 M phosphate (pH 8) at 20 °C in 5% O2 quenched with HCl at
different times (30 s, 1 min, 2 min, and 10 min), centrifuged, and loaded onto HPLC. The elution times were as follows: 11.4 min for AHQQ, 16.9
min for PQQ, 17.2 for an anaerobic intermediate, 17.5 min for an intermediate, and 20.1 min for PQQH2. (E) HPLC traces monitored at 275 nm for
WT PqqC reaction with AHQQ (100 and 40 μM, respectively) in 0.1 M phosphate (pH 8) at 20 °C in the absence of O2 quenched with HCl at
different times (10, 30, and 60 min), centrifuged, and loaded onto HPLC. The elution times were as follows: 11.4 min for AHQQ and 17.2 min for
an anaerobic intermediate.
with a λmax of 315 nm, with the absorbance intensity being
proportional to the concentration of PQQ, whereas
incubation of phenylhydrazine with PQQH2 leads to no
apparent reaction.17 When this same reaction was performed
with the intermediate at m/z 335.05, no spectral changes were
observed. This indicates that the intermediate with a λmax at
E
dx.doi.org/10.1021/bi4003315 | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
303 nm and m/z 335.05 is a quinol and not a quinone species
(the quinol of BII in Scheme 1).
Reoxidation of Exogenous PQQH2 as a Function of O2
Concentration. The last step of the reaction catalyzed by PqqC
is one in which the reduced product, PQQH2, is converted to
PQQ. It is possible to probe and compare the disappearance
of exogenously added PQQH2 at different O2 concentrations.
Estimation of the Kd for added PQQH2 was first conducted by
following the protein fluorescence as a function of added PQQH2
(excitation at 282 nm, emission at 333 nm, under anaerobic
conditions). WT PqqC was found to bind PQQH2 less tightly
than PQQ (Kd = 50 nM for PQQH2, and Kd = 2.0 nM for
PQQ15). However, under the conditions used in the PQQH2
oxidation assay (20 μM PQQH2 with 25 μM enzyme), all the
PQQH2 is predicted to be bound to the enzyme. Concentrated
PQQH2 was first incubated with enzyme anaerobically to ensure
that PQQH2 is fully bound prior to the addition of O2 and then
diluted with a buffer with a known oxygen concentration in a
closed cuvette. Quinols are known to reoxidize to quinones in
buffer in air;20 in the presence of PqqC, the rate of PQQH2
reoxidation is found to be ∼8 times slower than in buffer.
Although the enzymatic reaction was conducted over a range of
O2 concentrations, a saturating concentration of O2 was never
achieved (Figure 3C). Below 200 μM O2, the rate of oxidation
At 100% O2 (1.14 mM O2), the data are similar to what is
observed for the reaction in 21% O2, though the reaction proceeds
more rapidly and without any detectable PQQH2 (Figure 4C). At
high O2 concentrations, the only intermediate that accumulates
during the reaction is the quinol BII shown in Scheme 1.
At 5% O2 (57 μM O2), the accumulation of PQQH2 is much
more pronounced than at 21% O2 (Figure 4D). Almost all the
substrate is PQQH2 at 3 min of reaction. Indeed, the rate of
PQQH2 reoxidation appears to be the most affected by the
change in O2 concentration. The cumulative data, thus, strongly
point toward an oxidation of bound PQQH2 by O2 and not
H2O2. In the beginning of the reaction at 5% O2, a very small
amount of an intermediate with a λmax at 347 nm (m/z 321.07)
can also be seen eluting at 17.2 min. This intermediate is
observed in the anaerobic reaction between PqqC and AHQQ
(Figure 4E), with a rate of formation of ∼0.2 min−1 (see below)
and can be assigned to quinol A. The low level of quinol A is
consistent with comparable rates for its formation and decay.
The reaction intermediates detected by acid quench at 57 μM O2
are boxed in red in Scheme 1. In contrast to these data, only a
single intermediate was detected during the UV−vis single-turnover
kinetics, most likely because the spectra of bound quinols A, BII,
and PQQH2 are too similar to be distinguished from one another.
Comparison of UV−Vis and Acid Quench under Anaerobic
Conditions. Under anaerobic conditions, the initial steps of the
PQQ reaction were probed using the established UV−vis and
acid quench techniques. The anaerobic reaction intermediates
have been previously characterized by UV−vis spectroscopy and
display the following absorption maxima.13
of PQQH2 is very close to the measured k2 , consistent with a
app
similar rate-determining process for the oxidation of either
AHQQ or added PQQH2 in this O2 concentration range
(Figure 2). In contrast, above 200 μM O2, where the rate of the
oxidation of bound PQQH2 becomes more rapid, the enzymatic
oxidation of the substrate AHQQ is limited by another step.
These results are fully compatible with the changes in reaction
intermediates observed using acid quench methodology at
different O2 concentrations.
We have previously reported that the PqqC−PQQH2
complex could be oxidized under anaerobic conditions with
exogenously added H2O2, as well as molecular oxygen.18,19
More recent studies show that this consumption of H2O2
occurs in the absence of PQQH2 and leads to a time-dependent
evolution of O2 as measured on an O2 electrode. Because the
earlier and current studies use His-tagged PqqC, we prepared
PqqC without its His tag and re-examined reoxidation of
PQQH2 by H2O2 using UV−vis monitoring, finding a rate
E‐AHQQ → 344 nm → 318 nm
E−(quinoid)
E−(quinol)
In the WT enzyme, the rate of transformation of E−quinoid
to E−quinol is very fast (i.e., not observable). The reaction
spectra in this study indicate formation of the species with a
λmax of 318 nm at a rate of 0.2 min−1. Quenching of the
anaerobic reaction by acid and subsequent HPLC separation
and mass spectral analysis of the reaction products shows a
single intermediate with an elution time of 17.2 min, a λmax of
347 nm, and an ion at m/z 321.07 (Figure 4E). This species
appears concomitant with the decay of the substrate (AHQQ),
and occurs at the same rate as the appearance of the λmax = 318
nm species seen via UV−vis spectroscopy.
Neither the mass (m/z) nor λmax of the acid quench-derived
anaerobic intermediate matched the values expected from
the UV−vis spectra for the protonated form of quinol A in
Scheme 1. The mass (m/z) is 18 Da smaller than expected,
and the UV−vis spectrum exhibits a red-shifted absorbance.
We propose that the acidic conditions of the quench catalyze a
second reaction, in which the initial cyclization step of AHQQ
is followed by an intramolecular reaction between the carbo-
xylate at C9 and nitrogen of the five-membered quinoid/quinol
ring to form a stable lactam product:
(0.04
0.005 min−1) that is <10% of what was previously
reported.19 Apparently, the His tag itself is capable of a reaction
with H2O2 to liberate O2, possibly because of the binding of a
trace redox-active metal to the tag itself. Much of the earlier
observation of oxidation of PQQH2 by exogenous H2O2 was,
thus, likely due to the O2 liberated via a side reaction. In the
study presented here, conducted in the absence of added H2O2,
the initial level of H2O2 during the reaction of AHQQ with
PqqC is very low. At later times, as the product peroxide
accumulates, this may undergo some oxidation via reaction at
the His tag. However, the oxidation of bound PQQH2 by pro-
duct peroxide can be neglected. Re-evaluation of the ratio for con-
sumption of O2 to production of H2O2 using PqqC without His
tag leads to a result identical to the previously reported result.13
As an alternative way of examining the possibility of the satura-
tion of the enzyme by O2 in the presence of bound product, the
Kd for the interaction of PQQ with PqqC was determined as a
function of increasing O2 concentration. Consistent with the
kinetic measurements for PQQH2 oxidation, the Kd for PQQ
increases in a linear fashion with O2 concentration, reaching a
value that is approximately 10 times larger at >1 mM O2 relative
to that under the anaerobic condition (Figure 5). O2 appears to
This structure is consistent with both the observed mass and
red-shifted UV−vis spectrum of the anaerobic intermediate.
F
dx.doi.org/10.1021/bi4003315 | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
modification). The latter is identical to the first quinol (A),
formed under anaerobic conditions.
The fact that the accumulation of the intermediates varies when
the reaction is conducted at different O2 concentrations indicates
that, in fact, the UV−vis traces at the three experimental oxygen
concentrations represent different processes. At low O2 con-
centrations (≤5%), k1 represents the sum of the formation of
app
quinol A, quinol BII, and PQQH2. At 21% O2, quinol A no longer
accumulates, a small amount of PQQH2 is observed, and the
majority of the product is quinol BII. The maximal level of
PQQH2 observed during the catalysis is <8% of the level of PQQ
formed at the end of the reaction (calculated using the extinction
coefficients of PQQH2 at 320 nm and PQQ at 275 nm and the
HPLC traces at these wavelengths at 3 and 10 min for the
determination of the amount of PQQH2 and PQQ, respectively).
At this O2 concentration, while the formation of PQQH2 is
still detectable, its contribution to k1 is quite small. At high
app
O2 concentrations, k1 is a direct observation of the formation of
app
Figure 5. Kd values of the WT PqqC−PQQ complex determined by
fluorescence in 0.1 M phosphate (pH 8) at different dioxygen
concentrations. Both the λmax and intensity for the emission are
different for bound and free PQQ.14
quinol BII, which is the only intermediate observed at 100% O2.
One very interesting observation from these comparative analyses
is the rate of appearance of quinol (k1 ) at high O2 concentra-
app
tions is more rapid than the rate of anaerobic formation of
quinol A, which appears with a rate constant of 0.2 s−1. This
means that the anaerobic portion of the reaction must be affected
by the presence of O2, accelerating the formation of quinol A,
whose formation would otherwise be rate-limiting at the elevated
O2 concentrations. This result implicates an interaction (binding)
between O2 and PqqC that alters a portion of the reaction that is
chemically independent of O2, namely, the conversion of AHQQ
to quinol A. The presence of an O2 binding pocket near the
bound substrate or product has previously been inferred from
the X-ray crystal structure of PqqC in the presence of PQQ17
(Figure 1) and is further implicated from the data in Figure 5.
The earlier determination of the stoichiometry of oxygen
consumption in the net PqqC reaction had indicated that one
of the four steps in Scheme 1 must involve the consumption of
the product H2O2, formed in a previous step. The resulting
prediction of an O2-independent step is, in fact, confirmed and
identified in this study. The only species seen to accumulate at
all O2 concentrations is quinol BII. According to the acid
quench experiments, the second kinetic trace observed by UV−
be unable to saturate the enzyme in the concentration range
studied (from 0 to 1.2 mM), though the sensitivity of the binding
of PQQ to O2 levels is consistent with a binding site for the gas
at the active site, as implicated from the X-ray structure
determined in the presence of both PQQ and O2 (Figure 1).
DISCUSSION
■
As shown in Scheme 1, AHQQ must undergo a two-electron
oxidative ring closure (A), followed by three additional two-
electron oxidation steps (BI−BIII) using two molecules of O2
and one molecule of H2O2.13 While the first of the four steps in
this single-turnover process has been identified as a ring closure
(A)17 and the last involves the oxidation of the reduced product
(BIII), the chemical steps in BI and BII may occur either as
shown or in the reverse order. The net reaction was previously
monitored solely by UV−vis spectroscopy at a single oxygen
concentration (air). The expectation of similar absorbance
properties for the quinoid and quinol intermediates in steps A
through BI−BIII made identification of the species that
accumulated during the reaction impossible. In this study, we
have turned to acid quench−mass spectrometry experiments
that, when combined with UV−vis measurements across a
range of O2 concentrations, provide considerable insight into
the nature of the accumulated intermediates and, by inference,
the nature of the rate-determining steps.
vis spectroscopy (k2 ) represents the disappearance of quinol
app
BII at high O2 concentrations, together with a small contribu-
tion from the disappearance of quinol A (5% O2) and PQQH2
(5 and 21% O2). Once the accumulated levels of quinol A and
PQQH2 approach zero, the disappearance of BII has become
independent of O2. This could have meant that the
disappearance of BII with O2 had a very low Km but was
intrinsically slow. However, as we argue below, quinol BII is
assigned to the species that reacts with hydrogen peroxide
rather than O2.
By UV−vis spectroscopy, the rates of reaction of PqqC with
a range of O2 levels can be accommodated by two kinetic
traces, k1 (representing the appearance of the quinol species at
app
318 nm) and k2 (representing the disappearance of quinol at
app
Ideally, these studies would have allowed a direct monitoring
of the disappearance of each quinol in Scheme 1 as a function
of O2 concentration, to distinguish its reaction with O2 from
that with H2O2. In the case of the first oxidation (of quinol A),
there is no H2O2 present and this reaction must proceed
with O2. Further, the fact that the amount of quinol A observed
318 nm and the appearance of quinone at 344 nm) (Figure 2).
These kinetic traces (k1 and k2 ) at each O2 concentration are,
app
app
in fact, the sum of different rates according to the acid quench
experiments. Comparison of m/z values for the isolated species
in air versus 5 and 100% O2 shows that the quinol of BII
(Scheme 1) is the only intermediate that accumulates at all
measured O2 concentrations. In air, the reduced product,
PQQH2, is also seen, whereas 5% O2 leads to both the reduced
product and a small amount of quinol A (i.e., the substrate
subsequent to ring closure but lacking any oxidative
at 5% O2 is already very low makes its contribution to k2
app
negligible at all O2 levels. Quinol BI was never observed in
these studies, making it impossible to correlate its disappear-
ance with a varying O2 concentration. We were, however,
G
dx.doi.org/10.1021/bi4003315 | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
able to follow the disappearance of PQQH2 as a function of O2
concentration. We accomplished this by reducing PQQ
with DTT, mixing the resulting PQQH2 (20 μM) with PqqC
(30 μM), and monitoring its reoxidation at 344 nm (Figure 3C,
black line). It can be seen that the O2 dependence of the direct
oxidation of the bound PQQH2 is very different from the O2
peroxide in the proximity of PQQ. This is consistent with the
approximately 10-fold tighter binding of PQQ than PQQH2 to
PqqC. Taken together, the aggregate data strongly support a
progressive closure of the active site during catalysis: at each step,
the active site becomes more insulated from solvent.
In all the cofactor-independent oxidases studied, the
substrate plays the key role in the activation of O2,20,21 with
the abstraction of a proton from the organic substrate by a
catalytic base initiating the oxidative process. In some cases, the
presence of a specific O2 pocket has been shown, ideally located
above the substrate.22 While none of these enzymes has, thus
far, indicated any direct activation of O2, we note that the active
site of PqqC contains positively charged and H-bonding
residues near the ascribed O2 pocket; these side chains may
play an important electrostatic role in stabilizing the superoxide
anion during those steps that use O2 as the immediate oxidant.
We now address a possible physiological relevance of the
observed slow oxidation of PqqC-bound PQQH2. As noted,
the oxidation of PQQH2 is, in fact, slower on the enzyme than
in buffer (cf. ref 18). With regard to in vivo function, enzyme-
bound PQQH2 may, thus, be the anticipated product. A low
concentration of O2 in the bacterial species that make PQQ
would lead to an accumulation of enzyme-bound PQQH2 rather
than PQQ. In the context of the very low (nanomolar) Kd value
for binding of PQQ to PqqC, the accumulation of the 25-fold
weaker binding-reduced product may be an evolved function to
ensure production of sufficient levels of the free cofactor. This
feature may be a particularly important accompaniment to the
postulated progressive closure of the enzyme as it proceeds
through its catalytic cycle. For the future, it should be particularly
informative to obtain and contrast an X-ray structure for PqqC in
complex with its reduced rather than oxidized product.
dependence seen in k2 and fails to show saturation under any
app
condition (Figure 3C). This property will be related to the
protein structure in the ensuing discussion. From a kinetic
perspective, the data show a much faster rate for PQQH2
oxidation above 21% O2 than for the oxidation of quinol BII,
consistent with the absence of PQQH2 by acid quench at
elevated O2 concentrations:
Limited step for k2app at >0.2 mM O2
E‐quinol BII ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯→ E − PQQ
⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯→E‐PQQH2⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯→
k
k
BIII
BII
Using the measured values for kBIII as a function of O2
concentration, it was possible to estimate the rate for the
disappearance of BII at all O2 concentrations according to eq 2:
k2 = kBIIkBIII[O2]/kBII + kBIII[O2]
app
(2)
The analysis indicates that all of the change in k2app as a function
of O2 concentration arises from the oxidation of bound PQQH2,
and that kBII is fully independent of O2 concentration.
The unique dependence of each step of the PqqC reaction
sequence on O2 or H2O2 suggests a deliberate property for each
of the oxidation steps. In particular, the following questions are
raised: (i) Why does the enzyme import three molecules of
dioxygen and not two or four? (ii) Why does the reoxidation of
BII occur using H2O2? (iii) Why is the oxidation of enzyme-
bound PQQH2 so slow at the ambient O2 concentration?
Crystallographic studies of PqqC have shown two conformations
for PqqC: an open form with the active site exposed to solvent
and a closed form that occurs when PQQ is bound (Figure 1).
The closure of the active site of PqqC must occur either during
the binding of AHQQ, along the reaction path, or via a
combination of both processes. The dependence of the reaction
sequence on O2 suggests that this closure takes place during each
of the chemical steps. For the two initial oxidation steps, O2 can
easily access the active site and rapidly oxidize the bound quinols;
only the formation of quinol A emerges as a potentially slow step
at >50 μM O2. After these two oxidation steps, the active site
appears to have become quite closed. This provides a rationale
for the use of bound peroxide during the third oxidation step. It
appears that PqqC encounters a “log jam” at its BII intermediate,
preventing either egress of the product H2O2 from the previous
step or the binding of O2. This enforces a reductive step that
uses hydrogen peroxide rather than O2 in an intrinsically slower
process. The oxidation of bound PQQH2 offers even more of a
challenge, with regard to access to O2. The nonsaturating nature
of O2 during this step indicates a process that is first-order in O2
up to 1.14 mM. In this case, the product (H2O) from the
previous reaction may be impeding access and/or binding of the
O2 required for the final step. Importantly, however, a previous
study of the net PqqC stoichiometry showed that only one
molecule of H2O2 was released to the buffer directly. The
detection of the second mole of peroxide product required acid
denaturation of PqqC. For reasons not yet understood, the
oxygen-derived product from the last reaction binds tightly,
despite the fact that the O2 reaction is fully second-order. This
suggests that only upon formation and oxidation of PQQH2 does
the enzyme enter into its fully closed form, trapping bound
CONCLUSION
■
This study has demonstrated the isolation and identification of
reaction intermediates during the final step of PQQ biogenesis,
catalyzed by PqqC. The comparison of acid quench-derived
species with UV−vis traces as a function of O2 concentration
shows that oxidant utilization is distinct in each of the four
catalytic steps. One of the four steps has been shown to use
H2O2 rather than molecular dioxygen, while a second step
undergoes a first-order reaction with O2. These properties are
attributed to a stepwise closure of the enzyme from an open to
fully closed form that is not complete until the final chemical
step of the reaction. The very slow rate of formation of enzyme-
bound PQQ from its precursor PQQH2 at the ambient O2
concentration suggests that PQQH2 is the physiologically
relevant product of this biosynthetic pathway.
AUTHOR INFORMATION
Corresponding Author
*E-mail: klinman@berkeley.edu. Telephone: (510) 642-2668.
■
Funding
Financial support was provided by National Institutes of Health
Grants GM025765 to J.P.K. and 1S10RR022393-01 to A.T.I.
Notes
The authors declare no competing financial interest.
ABBREVIATIONS
■
AHQQ, 3a-(2-amino-2-carboxyethyl)-4,5-dioxo-4,5,6,7,8,9-hex-
ahydroquinoline-7,9-dicarboxylic acid; PQQ, pyrroloquinoline
quinone; PQQH2, dihydroquinone PQQ; WT, wild type.
H
dx.doi.org/10.1021/bi4003315 | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
REFERENCES
Article
(20) Fetzner, S., and Steiner, R. A. (2010) Cofactor-independent
oxidases and oxygenases. Appl. Microbiol. Biotechnol. 86, 791−804.
(21) Gabison, L., Chopard, C., Colloc’h, N., Peyrot, F., Castro, B., El
■
(1) Anthony, C. (2001) Pyrroloquinoline quinone (PQQ) and
quinoprotein enzymes. Antioxid. Redox Signaling 3, 757−774.
(2) Arakawa, K., Sugino, F., Kodama, K., Ishii, T., and Kinashi, H.
(2005) Cyclization mechanism for the synthesis of macrocyclic
antibiotic lankacidin in Streptomyces rochei. Chem. Biol. 12, 249−256.
(3) Goodwin, P. M., and Anthony, C. (1998) The biochemistry,
physiology and genetics of PQQ and PQQ-containing enzymes. Adv.
Microb. Physiol. 40, 1−80.
(4) Rucker, R., Chowanadisal, W., and Nakano, M. (2009) Potential
physiological importance of pyrroloquinoline quinone. Alternative
Medicine Review: A Journal of Clinical Therapeutic 14, 268−277.
(5) Arnison, P. G., Bibb, P. G., Bierbaum, M. G., et al. (2013)
Ribosomally synthesized and post-translationally modified peptide
nature products: Overview and recommendations for a universal
nomenclature. Nat. Prod. Rep. 30, 106−160.
́
Hajji, M., Altarsha, M., Monard, G., Chiadmi, M., and Prange, T.
(2011) X-ray, ESR, and quantum mechanics studies unravel a spin well
in the cofactorless urate oxidase. Proteins 79, 1964−1976.
(22) Colloc’h, N., Gabison, L., Monard, G., Altarsha, M., Chiadmi,
M., Marassio, G., Sopkova-de Oliveira Santos, J., El Hajji, M., Castro,
́
B., Abraini, J. H., and Prange, T. (2008) Oxygen pressurized X-ray
crystallography: Probing the dioxygen binding site in cofactorless urate
oxidase and implications for its catalytic mechanism. Biophys. J. 95,
2415−2422.
(6) He, K., Nukada, H., Urakami, T., and Murphy, M. P. (2003)
Antioxidant and pro-oxidant properties of pyrroloquinoline quinone
(PQQ): Implications for its function in biological systems. Biochem.
Pharmacol. 65, 67−74.
(7) Janes, S. M., Mu, D., and Wemmer, D. (1990) A new redox
cofactor in eukaryotic enzymes-6-hydroxydopa at the active site of
bovine serum amine oxidase. Science 248, 981−987.
(8) McIntire, W. S., Wemmer, D. E., Christoserdov, A., et al. (1991)
A new cofactor in a prokaryotic enzyme tryptophan tryptophylquinone
as the redox prosthetic group in methylamine dehydrogenase. Science
252, 817−824.
(9) Wang, S., Mure, M., Medzihradszky, K. F., and Burlingame, A. L.
(1996) A crosslinked cofactor in lysyl oxidase: Redox function for
amino acid side chains. Science 273, 1078−1084.
(10) Datta, S., Mori, Y., Takagi, K., Kawaguchi, K., Chen, Z. W., et al.
(2001) Structure of a quinoprotein in amine dehydrogenase with an
uncommon redox cofactor and highly unusual crosslinking. Proc. Natl.
Acad. Sci. U.S.A. 98, 14268−14273.
(11) Shen, Y. Q., Bonnot, F., Imsand, E. M., RoseFigura, J. M.,
Sjolander, K., and Klinman, J. P. (2012) Distribution and properties of
̈
the genes encoding the biosynthesis of the bacterial cofactor,
pyrroloquinoline quinone. Biochemistry 51, 2265−2275.
(12) Houck, D. R., Hanners, J. L., and Unkefer, C. J. (1991)
Biosynthesis of pyrroloquinoline quinone. 2. Biosynthetic assembly
from glutamate and tyrosine. J. Am. Chem. Soc. 113, 3162−3166.
(13) Magnusson, O. T., Toyama, H., Saeki, M., Rojas, A., Reed, J. C.,
Liddington, R. C., Klinman, J. P., and Schwarzenbacher, R. (2004)
Quinone biogenesis: Structure and mechanism of PqqC, the final
catalyst in the production of pyrroloquinoline quinone. Proc. Natl.
Acad. Sci. U.S.A. 101, 7913−7918.
(14) Magnusson, O. T., Toyama, H., Saeki, M., Schwarzenbacher, R.,
and Klinman, J. P. (2004) The structure of a biosynthetic intermediate
of pyrroloquinoline quinone (PQQ) and elucidation of the final step
of PQQ biosynthesis. J. Am. Chem. Soc. 126, 5342−5343.
(15) Magnusson, O. T., RoseFigura, J. M., Toyama, H.,
Schwarzenbacher, R., and Klinman, J. P. (2007) Pyrroloquinoline
quinone biogenesis: Characterization of PqqC and its H84N and
H84A active site variants. Biochemistry 46, 7174−7186.
(16) Puehringer, S., RoseFigura, J. M., Metlitzky, M., Toyama, H.,
Klinman, J. P., and Schwarzenbacher, R. (2010) Structural studies of
mutant forms of the PQQ-forming enzyme PqqC in the presence of
product and substrate. Proteins 78, 2554−2562.
(17) Mure, M., Nii, K., Inoue, T., Itoh, S., and Ohshino, Y. (1990)
The reduction of coenzyme PQQ with hydrazines. J. Chem. Soc., Perkin
Trans. 2, 315−320.
(18) Itoh, S., Ohshiro, Y., and Agawa, T. (1986) Reaction of reduced
PQQ (PQQH2) and molecular oxygen. Bull. Chem. Soc. Jpn. 59, 1911−
1914.
(19) RoseFigura, J. M., Puehringer, S., Schwarzenbacher, R., Toyama,
H., and Klinman, J. P. (2011) Characterization of a protein-generated
O2 binding pocket in PqqC, a cofactorless oxidase catalyzing the final
step in PQQ production. Biochemistry 50, 1556−1566.
I
dx.doi.org/10.1021/bi4003315 | Biochemistry XXXX, XXX, XXX−XXX