---
pmid: '35380039'
title: Pre-Steady-State Reactivity of Peptidylglycine Monooxygenase Implicates Ascorbate
  in Substrate Triggering of the Active Conformer.
authors:
- Welch EF
- Rush KW
- Arias RJ
- Blackburn NJ
journal: Biochemistry
year: '2022'
full_text_available: true
full_text_extraction_method: xml
pmcid: PMC9064607
doi: 10.1021/acs.biochem.2c00080
pubmed_publication_types:
- Journal Article
- Research Support, N.I.H., Extramural
publication_type: PRIMARY_RESEARCH
---

# Pre-Steady-State Reactivity of Peptidylglycine Monooxygenase Implicates Ascorbate in Substrate Triggering of the Active Conformer.
**Authors:** Welch EF, Rush KW, Arias RJ, Blackburn NJ
**Journal:** Biochemistry (2022)
**DOI:** [10.1021/acs.biochem.2c00080](https://doi.org/10.1021/acs.biochem.2c00080)
**PMC:** [PMC9064607](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9064607/)

## Abstract

1. Biochemistry. 2022 Apr 19;61(8):665-677. doi: 10.1021/acs.biochem.2c00080.
Epub  2022 Apr 5.

Pre-Steady-State Reactivity of Peptidylglycine Monooxygenase Implicates
Ascorbate in Substrate Triggering of the Active Conformer.

Welch EF(1)(2), Rush KW(1)(3), Arias RJ(1), Blackburn NJ(1).

Author information:
(1)Department of Chemical Physiology and Biochemistry, Oregon Health and Science
University, 3181 SW Sam Jackson Park Road, Portland, Oregon 97239, United
States.
(2)Department of Biomedical Engineering, Oregon Health and Science University,
3181 SW Sam Jackson Park Road, Portland, Oregon 97239, United States.
(3)Department of Chemistry, Reed College, 3203 SE Woodstock Blvd, Portland,
Oregon 97202, United States.

Peptidylglycine monooxygenase (PHM) is essential for the posttranslational
amidation of neuroendocrine peptides. An important aspect of the PHM mechanism
is the complete coupling of oxygen reduction to substrate hydroxylation, which
implies no oxygen reactivity of the fully reduced enzyme in the absence of
peptidyl substrates. As part of studies aimed at investigating this feature of
the PHM mechanism, we explored pre-steady-state kinetics using chemical quench
(CQ) and rapid freeze-quench (RFQ) studies of the fully reduced ascorbate-free
PHM enzyme. First, we confirmed the absence of Cu(I)-enzyme oxidation by O2 at
catalytic rates in the absence of peptidyl substrate. Next, we investigated
reactivity in the presence of the substrate dansyl-YVG. Surprisingly, when
ascorbate-free di-Cu(I) PHM was shot against oxygenated buffer containing the
dansyl-YVG substrate, <15% of the expected product was formed. Substoichiometric
reactivity was confirmed by stopped-flow and RFQ EPR spectroscopy. Product
generation reached a maximum of 70% by the addition of increasing amounts of the
ascorbate cosubstrate in a process that was not the result of multiple
turnovers. FTIR spectroscopy of the Cu(I)-CO reaction chemistry was then used to
show that increasing ascorbate concentrations correlated with a
substrate-induced Cu(I)M-CO species characteristic of an altered conformation.
We conclude that ascorbate and peptidyl substrate work together to induce a
transition from an inactive to an active conformation and suggest that the
latter may represent the "closed" conformation (Cu-Cu of ∼4 Å) recently observed
for both PHM and its sister enzyme DBM by crystallography.

DOI: 10.1021/acs.biochem.2c00080
PMCID: PMC9064607
PMID: 35380039 [Indexed for MEDLINE]

Conflict of interest statement: The authors declare no competing financial
interest.

## Full Text

IntroductionMononuclear copper
monooxygenases constitute a family of metalloenzymes
important to both the health- and energy-related sciences. A prime
example is dopamine β-monooxygenase (DBM),1,2 which
catalyzes a critical step in catecholamine biosynthesis, as does its
insect homologue tyramine β-monooxygenase (TBM).3 Similarly, peptidylglycine α-amidating monooxygenase
(PAM)4,5 is a bifunctional enzyme and the only known
enzyme to convert glycine-extended neuropeptide hormones into their
active C-terminally amidated forms. Lytic polysaccharide monooxygenases
(LPMOs) contain functional mononuclear copper centers, which catalyze
the chemically challenging oxidation at hexose C1 and/or C4, making
them important enzymes in biomass degradation.6 Finally, new investigations into the structure and function of particulate
methane monooxygenase (pMMO) denote the likelihood of a mononuclear
copper site as the catalytic center for alkane hydroxylation.7,8The catalytic core (PHMcc, residues 42–356) of PHM
catalyzes
the first step in peptide amidation via a copper-dependent hydroxylation
of the α-C atom of the glycine-extended propeptide C-terminus
(Figure 1). The reaction
involves the transfer of four electrons to O2, two from
the C–H bond and two from an external reductant. To achieve
this conversion, the enzyme stores the two external electrons on two
noncoupled copper atoms (CuH and CuM), denoted the H-site and the
M-site, respectively, which are reduced to Cu(I) by ascorbate. Powerful
insights into the structure and catalytic mechanism of this copper
monooxygenase have been achieved in previous crystallographic,9−13 spectroscopic,14−20 kinetic,21−27 and theoretical28−30 studies. Structurally, it is known that the copper
centers of WT PHMcc11,12 and its complexes with exogenous
ligands9,31 are mononuclear and separated by 11 Å
across a solvent-filled cleft (Figure 1). The M-site (CuM) is deemed responsible for initial
oxygen binding and the ensuing catalytic chemistry10 and is coordinated by H242, H244, and a weak, yet catalytically
essential interaction with the thioether of M314.14,32−34 The two reducing electrons provided to the M-site
are hypothesized to be supplied by the H-site (CuH) via long-range
ET.12,26,35−37 It has been reported that exogenous ligands (O2, CO,
peroxide, and azide) can bind to the catalytic M-site but are excluded
from the H-site by electronic or steric factors, which are incompletely
understood.9,10,31,37,38Figure 1Top: structure
of the catalytic core of the peptidylglycine monooxygenase
with bound substrate di-iodotyrosyl-glycine (pdb code 3OPM). The left
panel shows the protein fold architecture and the position of the
two copper centers, which are separated by 11 Å. The right panel
shows an expanded view of the metal centers and their coordinated
residues. Copper atoms are presumed to be in the oxidized Cu(II) state.
Bottom: reaction catalyzed by PHM.A number of mechanisms have been proposed to explain how the electron
is transferred across the 11 Å solvent-filled cleft that separates
the H- and M-subdomains. Among these, substrate-mediated ET12 and transfer across the ordered solvent26,27 have secured the most support but another possibility proposed on
the basis of crystallography2 and computational
modeling30 is that the enzyme undergoes
a conformational change to bring the copper sites closer together.
Various crystallographic studies have shown examples of DBM2 and PHM13 in both
open and closed conformations, although the biological relevance of
these conformations has not yet been established. In one study, two
conformations of DBM were observed in the same asymmetric unit. In
one monomer, the structure of the catalytic core aligns with the open
state of PHM, where the copper ions are separated by approximately
11 Å (13–14 Å in DBM), while in the second monomer,
it aligns with a closed structure of PHM (mutant variant H108A in
complex with citrate). In each of the closed structures, the copper
sites are partially occupied (DBM) or half-filled (PHM, H-site absent),
raising the question of whether the closed conformer is an artifact
of copper loss. However, this closed conformation offers an appealing
route for ET between the H- and M-sites and has been shown to be an
energetically feasible active state by QMMM calculations.30The likelihood of multiple conformational
states of the enzyme
is supported by spectroscopic studies. These studies focus on CuM
binding to CO and demonstrate a substrate-induced electronic perturbation
of the diatomic CO/O2 molecule.18 CO forms a complex with the ascorbate-reduced enzyme with a C≡O
stretching frequency (v(CO)) of 2093 cm–1. The decrease
in v(CO) from that of free CO (2143 cm–1) is consistent
with back-bonding from the filled d-electron manifold into the empty
π* CO antibonding orbitals and is assisted by electron-rich
imidazole ligands. In the presence of the peptidyl substrate, a red
shift in v(CO) is observed, with acetyl-YVG substrate dropping the
frequency to 2063 cm–1. The ∼30 cm–1 decrease in v(CO) suggests a substrate-induced increase in the electron-donating
power of the CuM electronic environment. Interestingly, both populations
(2093 and 2063 cm–1) remain present at high acetyl-YVG
concentration, suggesting two forms of the enzyme.Here, we
explore the pre-steady-state reaction of the fully reduced
enzyme in the absence of a reductant and as a function of ascorbate
concentration using chemical quench, (CQ) stopped-flow, rapid freeze-quench
(RFQ), and Fourier transform infrared (FTIR) approaches. We show that
the hydroxylation of the dansyl-YVG substrate while rapid only proceeds
to <15% of the expected product based on enzyme concentration.
However, this percentage increases to about 70% in the presence of
ascorbate in an ascorbate-dependent manner. Pulse-chase data suggest
that this is not the result of multiple turnovers but represents an
activation of the Cu(I)–enzyme by excess ascorbate. Additionally,
the effect appears to correlate with an increase in the species responsible
for the 2063 cm–1 infrared band. Taken together,
the data require a new hypothesis for substrate activation, one possibility
of which includes an ascorbate-induced conformational change to bring
the protein from an open to a closed conformation, facilitating electron
transfer between the copper sites.

Introduction

Mononuclear copper
monooxygenases constitute a family of metalloenzymes
important to both the health- and energy-related sciences. A prime
example is dopamine β-monooxygenase (DBM),1,2 which
catalyzes a critical step in catecholamine biosynthesis, as does its
insect homologue tyramine β-monooxygenase (TBM).3 Similarly, peptidylglycine α-amidating monooxygenase
(PAM)4,5 is a bifunctional enzyme and the only known
enzyme to convert glycine-extended neuropeptide hormones into their
active C-terminally amidated forms. Lytic polysaccharide monooxygenases
(LPMOs) contain functional mononuclear copper centers, which catalyze
the chemically challenging oxidation at hexose C1 and/or C4, making
them important enzymes in biomass degradation.6 Finally, new investigations into the structure and function of particulate
methane monooxygenase (pMMO) denote the likelihood of a mononuclear
copper site as the catalytic center for alkane hydroxylation.7,8

The catalytic core (PHMcc, residues 42–356) of PHM
catalyzes
the first step in peptide amidation via a copper-dependent hydroxylation
of the α-C atom of the glycine-extended propeptide C-terminus
(Figure 1). The reaction
involves the transfer of four electrons to O2, two from
the C–H bond and two from an external reductant. To achieve
this conversion, the enzyme stores the two external electrons on two
noncoupled copper atoms (CuH and CuM), denoted the H-site and the
M-site, respectively, which are reduced to Cu(I) by ascorbate. Powerful
insights into the structure and catalytic mechanism of this copper
monooxygenase have been achieved in previous crystallographic,9−13 spectroscopic,14−20 kinetic,21−27 and theoretical28−30 studies. Structurally, it is known that the copper
centers of WT PHMcc11,12 and its complexes with exogenous
ligands9,31 are mononuclear and separated by 11 Å
across a solvent-filled cleft (Figure 1). The M-site (CuM) is deemed responsible for initial
oxygen binding and the ensuing catalytic chemistry10 and is coordinated by H242, H244, and a weak, yet catalytically
essential interaction with the thioether of M314.14,32−34 The two reducing electrons provided to the M-site
are hypothesized to be supplied by the H-site (CuH) via long-range
ET.12,26,35−37 It has been reported that exogenous ligands (O2, CO,
peroxide, and azide) can bind to the catalytic M-site but are excluded
from the H-site by electronic or steric factors, which are incompletely
understood.9,10,31,37,38

Top: structure
of the catalytic core of the peptidylglycine monooxygenase
with bound substrate di-iodotyrosyl-glycine (pdb code 3OPM). The left
panel shows the protein fold architecture and the position of the
two copper centers, which are separated by 11 Å. The right panel
shows an expanded view of the metal centers and their coordinated
residues. Copper atoms are presumed to be in the oxidized Cu(II) state.
Bottom: reaction catalyzed by PHM.

A number of mechanisms have been proposed to explain how the electron
is transferred across the 11 Å solvent-filled cleft that separates
the H- and M-subdomains. Among these, substrate-mediated ET12 and transfer across the ordered solvent26,27 have secured the most support but another possibility proposed on
the basis of crystallography2 and computational
modeling30 is that the enzyme undergoes
a conformational change to bring the copper sites closer together.
Various crystallographic studies have shown examples of DBM2 and PHM13 in both
open and closed conformations, although the biological relevance of
these conformations has not yet been established. In one study, two
conformations of DBM were observed in the same asymmetric unit. In
one monomer, the structure of the catalytic core aligns with the open
state of PHM, where the copper ions are separated by approximately
11 Å (13–14 Å in DBM), while in the second monomer,
it aligns with a closed structure of PHM (mutant variant H108A in
complex with citrate). In each of the closed structures, the copper
sites are partially occupied (DBM) or half-filled (PHM, H-site absent),
raising the question of whether the closed conformer is an artifact
of copper loss. However, this closed conformation offers an appealing
route for ET between the H- and M-sites and has been shown to be an
energetically feasible active state by QMMM calculations.30

The likelihood of multiple conformational
states of the enzyme
is supported by spectroscopic studies. These studies focus on CuM
binding to CO and demonstrate a substrate-induced electronic perturbation
of the diatomic CO/O2 molecule.18 CO forms a complex with the ascorbate-reduced enzyme with a C≡O
stretching frequency (v(CO)) of 2093 cm–1. The decrease
in v(CO) from that of free CO (2143 cm–1) is consistent
with back-bonding from the filled d-electron manifold into the empty
π* CO antibonding orbitals and is assisted by electron-rich
imidazole ligands. In the presence of the peptidyl substrate, a red
shift in v(CO) is observed, with acetyl-YVG substrate dropping the
frequency to 2063 cm–1. The ∼30 cm–1 decrease in v(CO) suggests a substrate-induced increase in the electron-donating
power of the CuM electronic environment. Interestingly, both populations
(2093 and 2063 cm–1) remain present at high acetyl-YVG
concentration, suggesting two forms of the enzyme.

Here, we
explore the pre-steady-state reaction of the fully reduced
enzyme in the absence of a reductant and as a function of ascorbate
concentration using chemical quench, (CQ) stopped-flow, rapid freeze-quench
(RFQ), and Fourier transform infrared (FTIR) approaches. We show that
the hydroxylation of the dansyl-YVG substrate while rapid only proceeds
to <15% of the expected product based on enzyme concentration.
However, this percentage increases to about 70% in the presence of
ascorbate in an ascorbate-dependent manner. Pulse-chase data suggest
that this is not the result of multiple turnovers but represents an
activation of the Cu(I)–enzyme by excess ascorbate. Additionally,
the effect appears to correlate with an increase in the species responsible
for the 2063 cm–1 infrared band. Taken together,
the data require a new hypothesis for substrate activation, one possibility
of which includes an ascorbate-induced conformational change to bring
the protein from an open to a closed conformation, facilitating electron
transfer between the copper sites.

Materials and MethodsChemicalsBuffer components and sodium ascorbate were
purchased from Sigma-Aldrich with purities of >99%. The substrate
acetyl-Tyr-Val-Gly (Ac-YVG) was purchased from Peptide International.
The substrate dansyl-Tyr-Val-Gly (Dns-YVG) was purchased from Chi
Scientific.Cell Production and Protein PurificationPHM WT was
produced from transfected CHO cells in an Accusyst Minimax bioreactor
and reconstituted to 2 Cu(II) atoms per monomer. All procedures and
analytical methods were as described previously.18,34,39 Protein concentration was determined using
OD280(1%) = 0.980 with a Cary 50 spectrophotometer. After
purification, the proteins were stored at −80 °C. Unless
otherwise stated, reactions were carried out in 50 mM sodium phosphate
buffer pH 7.5. Reactions were typically carried out at pH 7.5 since
the single-turnover rates were not too fast to measure at this pH.Copper Reconstitution and AnalysisThe protein was
reconstituted by the addition of 2.5 molar equivalents of copper sulfate
by a syringe pump over the course of an hour. Excess copper was removed
by dialysis in 3 L of pH 7.5 sodium phosphate buffer overnight. Reconstitution
was verified using a Perkin-Elmer Optima 2000 DV inductively coupled
plasma optical emission spectrometer (ICP-OES) and was determined
to be 2.0 ± 0.2 Cu(II) per protein molecule (Table S1).Generation of Ascorbate-Free Fully Reduced
PHMThe
fully reduced PHM was generated by the addition of 2.5 reducing equivalents
(1.25 equiv per copper) of buffered sodium ascorbate as a highly concentrated
small volume with mixing to a solution of anaerobic oxidized enzyme
in a Vacuum Atmospheres glovebox. A color change was observed immediately
from cerulean blue to pale straw. The solution was then desalted anaerobically
using two passes in 7 K MWCO Zeba spin desalting columns (Thermo Fisher).
The specification of these columns quotes a value of 5–10%
of low-molecular-weight solute remaining after each passage, implying
a negligible amount of ascorbate remaining after two passages. ICP-OES
measurement of copper concentration after this protocol showed that
little or no copper loss occurred during ascorbate reduction and removal.
Representative data are listed in Table S1.Oxygen Reactivity of Fully Reduced Ascorbate-Free PHM in the
Absence of Peptidyl SubstrateSix hundred microliters of
a 500 μM solution of oxidized PHM (1 mM in Cu(II)) was split
into three portions of 200 μL each. Sample 1 was transferred
directly to an EPR tube for determination of the EPR-detectable Cu(II)
in fully oxidized PHM. Samples 2 and 3 were reduced anaerobically
with ascorbate and made ascorbate-free as described above. Sample
2 was transferred anaerobically to an EPR tube for determination of
the EPR-detectable signal in the fully reduced anaerobic ascorbate-free
enzyme. Sample 3 was exposed to air with shaking for 120 s and then
transferred to an EPR tube for Cu(II) quantitation. EPR spectra of
all samples were measured, and the concentration of EPR-detectable
Cu(II) was determined via double integration versus a 300 μM
Cu(II)–EDTA standard measured under identical conditions. Spectral
analysis was performed using GRAMS AI spectroscopy software (Thermo).
EPR spectra were measured on a Bruker Elexsys E500 spectrometer equipped
with a superX microwave bridge and a dual-mode cavity with a helium
flow cryostat (ESR900, Oxford Instrument, Inc.). The following experimental
conditions were used: frequency 9.63 GHz, temperature 100 K, microwave
power 20 mW, gain 10 dB, modulation amplitude 10 G, and sweep time
84 s.Fully Reduced Ascorbate-Free Single-Turnover KineticsThe pre-steady-state reaction (single-turnover) of the fully reduced
ascorbate-free PHM was interrogated using a BioLogic microvolume quench-flow
instrument (QFM-4000). Six hundred microliters of fully reduced ascorbate-free
PHM was preincubated with 600 μM of dansyl-YVG, preloaded into
a syringe, and sealed with a septum. Sodium phosphate buffer (20 mM)
was saturated with oxygen by bubbling with pure humidified oxygen
gas and then loaded into a syringe sealed with a septum. The fully
reduced ascorbate-free PHM + dansyl was shot against an equal volume
of oxygenated buffer at varied aging times: 5, 20, 40, 60, 80, 100,
150, 200, 250, 300, and 500 ms. The reaction mixture was quenched
at each aging time with an equal volume of 8% trifluoroacetic acid
(TFA) and then diluted offline with 100 mM MES buffer resulting in
a final concentration of 100 μM PHM + 100 μM Dns-(substrate
+ product). In experiments with added ascorbate, the quenched, diluted
reaction mixture also contained 100 μM, 500 μM, 2 mM,
or 6 mM ascorbate. Samples were analyzed via high-performance liquid
chromatography (HPLC) with a Varian Pro Star solvent delivery module
equipped with a Varian Pro Star model 410 autosampler (250 μL
syringe, 100 μL sample loop) on a 250 mm × 4.6 mm Varian
Microsorb-MV 100-5 C18 column as previously reported.35 Dansyl fluorescence was monitored with a Waters 474 scanning
fluorescence detector at λex = 365 nm and λem = 558 nm. This fluorescence was used to monitor the Dns-YVG
substrate concentration and hydroxylated product (Dns-YVG-OH). The
product and substrate were resolved via an isocratic method using
25% solvent B (0.1% TFA in acetonitrile) and 75% solvent A (0.1% TFA
in water). Substrate concentrations were determined via standard curves
of known concentrations of Dns-YVG in 0.1% TFA in water.Ascorbate Dependence
of the Pre-Steady-State ReactionThe oxidized, 600 μM
PHM WT was preincubated with 600 μM
Dns-YVG. The PHM and Dns-YVG mixture were gently infused with pure
oxygen in a 5 mL septum-fitted Wheaton vial for approximately 30 min
at 23 °C. When removing the vial from the oxygen infuser, a small
amount of oxygen overpressure was allowed to remain in the vial. Anaerobic
ascorbate was prepared from powder in a Vacuum Atmospheres anaerobic
glovebox using anaerobic buffer (50 mM NaP, pH 7.5) and was prepared
in 200 μM, 1 mM, 4 mM, and 12 mM concentrations. Ascorbate mixtures
were preloaded into syringes within the anaerobic chamber and sealed
with septa when removing from the chamber. Ascorbate dependence was
assessed by shooting the 600 μM PHM + Dns-YVG mixture against
an equal volume of anaerobic ascorbate, with a set aging time of 300
ms. The samples were quenched with 8% TFA (4% final concentration)
in water and analyzed via HPLC as described above.Pre-Steady-State
Kinetics at High Ascorbate ConcentrationPre-steady-state
kinetics were performed using the QFM-4000 system
as mentioned above but with the following adjustments: 600 μM
PHM and 600 μM Dns-YVG were preincubated together and infused
with pure oxygen at 23 °C. Anaerobic ascorbate was prepared in
the anaerobic chamber, preloaded into a syringe, and sealed with a
septum. Single-turnover kinetic curves were obtained by shooting the
600 μM PHM + Dns-YVG mixture against anaerobic ascorbate, resulting
in a final concentration of 6 mM at varied aging times: 5, 20, 40,
60, 80, 100, 150, 200, 250, 300, and 500 ms. The samples were quenched
with 8% TFA (4% final concentration) in water and analyzed via HPLC
as described above.“Pulse-Chase” Kinetics with
Acetyl-YVG and Dansyl-YVGTo determine if the enzyme was
undergoing multiple turnovers, a
pulse-chase experiment was performed using the QFM. Oxidized, 600
μM PHM WT was preincubated with 600 μM Dns-YVG. The PHM
and Dns-YVG mixture were gently infused with pure oxygen in a 5 mL
septum-fitted Wheaton vial for approximately 30 min at 23 °C.
When removing the vial from the oxygen infuser, a small amount of
oxygen overpressure was allowed to remain in the vial. Anaerobic ascorbate
+ acetyl-YVG (Ac-YVG) was prepared from powder in a Vacuum Atmospheres
anaerobic glovebox using anaerobic buffer (20 mM NaP pH 7.5) at 10
mM ascorbate plus 3 mM Ac-YVG. The ascorbate mixture was preloaded
into syringes within the anaerobic chamber and sealed with septa when
removed from the chamber. Repeated turnover was then assessed by shooting
the 600 μM PHM + Dns-YVG mixture against an equal volume of
anaerobic ascorbate Ac-YVG mixture, with aging times of 0, 300, and
1000 ms performed in triplicate. The mixture was then quenched with
an equal volume of 8% trifluoroacetic acid (TFA) and then diluted
with 20 mM NaP buffer resulting in a final concentration of 100 μM
PHM + 100 μM Dns-(substrate + product) + 1.67 mM ascorbate +
500 μM Act-YVG. Samples were analyzed by HPLC with a Varian
Pro Star solvent delivery module equipped with a Varian Pro Star model
410 autosampler (250 μL syringe, 100 μL sample loop) on
a 250 mm × 4.6 mm Varian Microsorb-MV 100-5 C18 column as described
above and previously reported.35Fully
Reduced Ascorbate-Free Stopped-Flow KineticsAscorbate-free
reduced PHM WT was prepared in the anaerobic chamber
as described above. The protein was prepared in a pH 7.5 buffer system
containing equal volumes of 50 mM each of MES, HEPES, and sodium phosphate.
WT PHM (500 μM) was shot against 5 mM Ac-YVG + 20 mM sodium
azide. Data were collected for 500 ms at 390 nm, with readings taken
every 5 ms in an Applied Photosystems SX20 stopped-flow module.Fully Reduced Ascorbate-Free Rapid Freeze-Quench EPRRapidly
frozen EPR samples were prepared using a Quench-Flow-3 (KinTek)
apparatus configured for RFQ by the manufacturer. Reduced, 200 μM
ascorbate-free protein samples (400 μM in Cu(I)) in anaerobic
buffer and 200 μM acetyl-YVG substrate in fully oxygenated buffer
([O2] ≈ 1.1 mM before 1:1 dilution on mixing) were
loaded onto the instrument and rapidly mixed at the desired time point.
A universal buffer was used comprised of equal volumes of 50 mM MES,
HEPES, and CHES adjusted to pH 7.5. Mixtures were quenched in liquid
ethane at −170 °C. The resulting snow was then packed
into 707-SQ-250M EPR tubes (Wilmad LabGlass). Oxidation was monitored
in a Bruker E500 X-Band EPR spectrometer equipped with a superX microwave
bridge and a dual-mode cavity with a helium flow cryostat (ESR900,
Oxford Instrument, Inc.). The experimental conditions were frequency
9.63 GHz, temperature 100 K, microwave power 20 mW, gain 10 dB, modulation
amplitude 10 G, and sweep time 84 s.Fourier Transform Infrared
Spectroscopy (FTIR)PHM
(500 μM) in 20 mM pH 7.5 NaP buffer was deoxygenated and saturated
with humidified CO by streaming across the surface of the solution
with pure CO gas for 5 min with gentle agitation. Once carbonylated,
Dns-YVG was added to a final concentration of 2.5 mM, and then a small
aliquot of 2 M anaerobic buffered sodium ascorbate was added to achieve
final concentrations of ascorbate between 0 and 25 mM. The mixture
was then carbonylated again for 5 min. One hundred microliters of
solution was then added by a syringe into the IR cell composed of
two CaF2 windows separated by a 50 μm Teflon spacer.
FTIR data were recorded on a Bruker Tensor 27 FTIR spectrophotometer
continuously purged with CO2-free dry air as previously
described.18 One thousand scans were collected
for both protein sample and buffer blank from 2250 to 1900 cm–1 at a resolution of 2 cm–1. Spectral
analysis including subtraction of the buffer blank was performed using
GRAMS AI spectroscopy software (Thermo).Ratios of CO bands
at 2063 and 2093 cm–1 were obtained from the peak
heights rather than peak integrals. The rational for this less rigorous
approach stems from several factors; the cell was taken apart and
reassembled between measurements taken over several days, which introduced
small changes in path length and fill quality. Slight differences
in protein and copper concentrations were introduced due to evaporation
during carbonylation. Finally, different calcium fluoride windows
were used across the experiment, which may have very slightly different
optical properties.

Materials and Methods

ChemicalsBuffer components and sodium ascorbate were
purchased from Sigma-Aldrich with purities of >99%. The substrate
acetyl-Tyr-Val-Gly (Ac-YVG) was purchased from Peptide International.
The substrate dansyl-Tyr-Val-Gly (Dns-YVG) was purchased from Chi
Scientific.

Buffer components and sodium ascorbate were
purchased from Sigma-Aldrich with purities of >99%. The substrate
acetyl-Tyr-Val-Gly (Ac-YVG) was purchased from Peptide International.
The substrate dansyl-Tyr-Val-Gly (Dns-YVG) was purchased from Chi
Scientific.

Cell Production and Protein PurificationPHM WT was
produced from transfected CHO cells in an Accusyst Minimax bioreactor
and reconstituted to 2 Cu(II) atoms per monomer. All procedures and
analytical methods were as described previously.18,34,39 Protein concentration was determined using
OD280(1%) = 0.980 with a Cary 50 spectrophotometer. After
purification, the proteins were stored at −80 °C. Unless
otherwise stated, reactions were carried out in 50 mM sodium phosphate
buffer pH 7.5. Reactions were typically carried out at pH 7.5 since
the single-turnover rates were not too fast to measure at this pH.

Cell Production and Protein Purification

PHM WT was
produced from transfected CHO cells in an Accusyst Minimax bioreactor
and reconstituted to 2 Cu(II) atoms per monomer. All procedures and
analytical methods were as described previously.18,34,39 Protein concentration was determined using
OD280(1%) = 0.980 with a Cary 50 spectrophotometer. After
purification, the proteins were stored at −80 °C. Unless
otherwise stated, reactions were carried out in 50 mM sodium phosphate
buffer pH 7.5. Reactions were typically carried out at pH 7.5 since
the single-turnover rates were not too fast to measure at this pH.

Copper Reconstitution and AnalysisThe protein was
reconstituted by the addition of 2.5 molar equivalents of copper sulfate
by a syringe pump over the course of an hour. Excess copper was removed
by dialysis in 3 L of pH 7.5 sodium phosphate buffer overnight. Reconstitution
was verified using a Perkin-Elmer Optima 2000 DV inductively coupled
plasma optical emission spectrometer (ICP-OES) and was determined
to be 2.0 ± 0.2 Cu(II) per protein molecule (Table S1).

Copper Reconstitution and Analysis

The protein was
reconstituted by the addition of 2.5 molar equivalents of copper sulfate
by a syringe pump over the course of an hour. Excess copper was removed
by dialysis in 3 L of pH 7.5 sodium phosphate buffer overnight. Reconstitution
was verified using a Perkin-Elmer Optima 2000 DV inductively coupled
plasma optical emission spectrometer (ICP-OES) and was determined
to be 2.0 ± 0.2 Cu(II) per protein molecule (Table S1).

Generation of Ascorbate-Free Fully Reduced
PHMThe
fully reduced PHM was generated by the addition of 2.5 reducing equivalents
(1.25 equiv per copper) of buffered sodium ascorbate as a highly concentrated
small volume with mixing to a solution of anaerobic oxidized enzyme
in a Vacuum Atmospheres glovebox. A color change was observed immediately
from cerulean blue to pale straw. The solution was then desalted anaerobically
using two passes in 7 K MWCO Zeba spin desalting columns (Thermo Fisher).
The specification of these columns quotes a value of 5–10%
of low-molecular-weight solute remaining after each passage, implying
a negligible amount of ascorbate remaining after two passages. ICP-OES
measurement of copper concentration after this protocol showed that
little or no copper loss occurred during ascorbate reduction and removal.
Representative data are listed in Table S1.

Generation of Ascorbate-Free Fully Reduced
PHM

The
fully reduced PHM was generated by the addition of 2.5 reducing equivalents
(1.25 equiv per copper) of buffered sodium ascorbate as a highly concentrated
small volume with mixing to a solution of anaerobic oxidized enzyme
in a Vacuum Atmospheres glovebox. A color change was observed immediately
from cerulean blue to pale straw. The solution was then desalted anaerobically
using two passes in 7 K MWCO Zeba spin desalting columns (Thermo Fisher).
The specification of these columns quotes a value of 5–10%
of low-molecular-weight solute remaining after each passage, implying
a negligible amount of ascorbate remaining after two passages. ICP-OES
measurement of copper concentration after this protocol showed that
little or no copper loss occurred during ascorbate reduction and removal.
Representative data are listed in Table S1.

Oxygen Reactivity of Fully Reduced Ascorbate-Free PHM in the
Absence of Peptidyl SubstrateSix hundred microliters of
a 500 μM solution of oxidized PHM (1 mM in Cu(II)) was split
into three portions of 200 μL each. Sample 1 was transferred
directly to an EPR tube for determination of the EPR-detectable Cu(II)
in fully oxidized PHM. Samples 2 and 3 were reduced anaerobically
with ascorbate and made ascorbate-free as described above. Sample
2 was transferred anaerobically to an EPR tube for determination of
the EPR-detectable signal in the fully reduced anaerobic ascorbate-free
enzyme. Sample 3 was exposed to air with shaking for 120 s and then
transferred to an EPR tube for Cu(II) quantitation. EPR spectra of
all samples were measured, and the concentration of EPR-detectable
Cu(II) was determined via double integration versus a 300 μM
Cu(II)–EDTA standard measured under identical conditions. Spectral
analysis was performed using GRAMS AI spectroscopy software (Thermo).
EPR spectra were measured on a Bruker Elexsys E500 spectrometer equipped
with a superX microwave bridge and a dual-mode cavity with a helium
flow cryostat (ESR900, Oxford Instrument, Inc.). The following experimental
conditions were used: frequency 9.63 GHz, temperature 100 K, microwave
power 20 mW, gain 10 dB, modulation amplitude 10 G, and sweep time
84 s.

Oxygen Reactivity of Fully Reduced Ascorbate-Free PHM in the
Absence of Peptidyl Substrate

Six hundred microliters of
a 500 μM solution of oxidized PHM (1 mM in Cu(II)) was split
into three portions of 200 μL each. Sample 1 was transferred
directly to an EPR tube for determination of the EPR-detectable Cu(II)
in fully oxidized PHM. Samples 2 and 3 were reduced anaerobically
with ascorbate and made ascorbate-free as described above. Sample
2 was transferred anaerobically to an EPR tube for determination of
the EPR-detectable signal in the fully reduced anaerobic ascorbate-free
enzyme. Sample 3 was exposed to air with shaking for 120 s and then
transferred to an EPR tube for Cu(II) quantitation. EPR spectra of
all samples were measured, and the concentration of EPR-detectable
Cu(II) was determined via double integration versus a 300 μM
Cu(II)–EDTA standard measured under identical conditions. Spectral
analysis was performed using GRAMS AI spectroscopy software (Thermo).
EPR spectra were measured on a Bruker Elexsys E500 spectrometer equipped
with a superX microwave bridge and a dual-mode cavity with a helium
flow cryostat (ESR900, Oxford Instrument, Inc.). The following experimental
conditions were used: frequency 9.63 GHz, temperature 100 K, microwave
power 20 mW, gain 10 dB, modulation amplitude 10 G, and sweep time
84 s.

Fully Reduced Ascorbate-Free Single-Turnover KineticsThe pre-steady-state reaction (single-turnover) of the fully reduced
ascorbate-free PHM was interrogated using a BioLogic microvolume quench-flow
instrument (QFM-4000). Six hundred microliters of fully reduced ascorbate-free
PHM was preincubated with 600 μM of dansyl-YVG, preloaded into
a syringe, and sealed with a septum. Sodium phosphate buffer (20 mM)
was saturated with oxygen by bubbling with pure humidified oxygen
gas and then loaded into a syringe sealed with a septum. The fully
reduced ascorbate-free PHM + dansyl was shot against an equal volume
of oxygenated buffer at varied aging times: 5, 20, 40, 60, 80, 100,
150, 200, 250, 300, and 500 ms. The reaction mixture was quenched
at each aging time with an equal volume of 8% trifluoroacetic acid
(TFA) and then diluted offline with 100 mM MES buffer resulting in
a final concentration of 100 μM PHM + 100 μM Dns-(substrate
+ product). In experiments with added ascorbate, the quenched, diluted
reaction mixture also contained 100 μM, 500 μM, 2 mM,
or 6 mM ascorbate. Samples were analyzed via high-performance liquid
chromatography (HPLC) with a Varian Pro Star solvent delivery module
equipped with a Varian Pro Star model 410 autosampler (250 μL
syringe, 100 μL sample loop) on a 250 mm × 4.6 mm Varian
Microsorb-MV 100-5 C18 column as previously reported.35 Dansyl fluorescence was monitored with a Waters 474 scanning
fluorescence detector at λex = 365 nm and λem = 558 nm. This fluorescence was used to monitor the Dns-YVG
substrate concentration and hydroxylated product (Dns-YVG-OH). The
product and substrate were resolved via an isocratic method using
25% solvent B (0.1% TFA in acetonitrile) and 75% solvent A (0.1% TFA
in water). Substrate concentrations were determined via standard curves
of known concentrations of Dns-YVG in 0.1% TFA in water.

Fully Reduced Ascorbate-Free Single-Turnover Kinetics

The pre-steady-state reaction (single-turnover) of the fully reduced
ascorbate-free PHM was interrogated using a BioLogic microvolume quench-flow
instrument (QFM-4000). Six hundred microliters of fully reduced ascorbate-free
PHM was preincubated with 600 μM of dansyl-YVG, preloaded into
a syringe, and sealed with a septum. Sodium phosphate buffer (20 mM)
was saturated with oxygen by bubbling with pure humidified oxygen
gas and then loaded into a syringe sealed with a septum. The fully
reduced ascorbate-free PHM + dansyl was shot against an equal volume
of oxygenated buffer at varied aging times: 5, 20, 40, 60, 80, 100,
150, 200, 250, 300, and 500 ms. The reaction mixture was quenched
at each aging time with an equal volume of 8% trifluoroacetic acid
(TFA) and then diluted offline with 100 mM MES buffer resulting in
a final concentration of 100 μM PHM + 100 μM Dns-(substrate
+ product). In experiments with added ascorbate, the quenched, diluted
reaction mixture also contained 100 μM, 500 μM, 2 mM,
or 6 mM ascorbate. Samples were analyzed via high-performance liquid
chromatography (HPLC) with a Varian Pro Star solvent delivery module
equipped with a Varian Pro Star model 410 autosampler (250 μL
syringe, 100 μL sample loop) on a 250 mm × 4.6 mm Varian
Microsorb-MV 100-5 C18 column as previously reported.35 Dansyl fluorescence was monitored with a Waters 474 scanning
fluorescence detector at λex = 365 nm and λem = 558 nm. This fluorescence was used to monitor the Dns-YVG
substrate concentration and hydroxylated product (Dns-YVG-OH). The
product and substrate were resolved via an isocratic method using
25% solvent B (0.1% TFA in acetonitrile) and 75% solvent A (0.1% TFA
in water). Substrate concentrations were determined via standard curves
of known concentrations of Dns-YVG in 0.1% TFA in water.

Ascorbate Dependence
of the Pre-Steady-State ReactionThe oxidized, 600 μM
PHM WT was preincubated with 600 μM
Dns-YVG. The PHM and Dns-YVG mixture were gently infused with pure
oxygen in a 5 mL septum-fitted Wheaton vial for approximately 30 min
at 23 °C. When removing the vial from the oxygen infuser, a small
amount of oxygen overpressure was allowed to remain in the vial. Anaerobic
ascorbate was prepared from powder in a Vacuum Atmospheres anaerobic
glovebox using anaerobic buffer (50 mM NaP, pH 7.5) and was prepared
in 200 μM, 1 mM, 4 mM, and 12 mM concentrations. Ascorbate mixtures
were preloaded into syringes within the anaerobic chamber and sealed
with septa when removing from the chamber. Ascorbate dependence was
assessed by shooting the 600 μM PHM + Dns-YVG mixture against
an equal volume of anaerobic ascorbate, with a set aging time of 300
ms. The samples were quenched with 8% TFA (4% final concentration)
in water and analyzed via HPLC as described above.

Ascorbate Dependence
of the Pre-Steady-State Reaction

The oxidized, 600 μM
PHM WT was preincubated with 600 μM
Dns-YVG. The PHM and Dns-YVG mixture were gently infused with pure
oxygen in a 5 mL septum-fitted Wheaton vial for approximately 30 min
at 23 °C. When removing the vial from the oxygen infuser, a small
amount of oxygen overpressure was allowed to remain in the vial. Anaerobic
ascorbate was prepared from powder in a Vacuum Atmospheres anaerobic
glovebox using anaerobic buffer (50 mM NaP, pH 7.5) and was prepared
in 200 μM, 1 mM, 4 mM, and 12 mM concentrations. Ascorbate mixtures
were preloaded into syringes within the anaerobic chamber and sealed
with septa when removing from the chamber. Ascorbate dependence was
assessed by shooting the 600 μM PHM + Dns-YVG mixture against
an equal volume of anaerobic ascorbate, with a set aging time of 300
ms. The samples were quenched with 8% TFA (4% final concentration)
in water and analyzed via HPLC as described above.

Pre-Steady-State
Kinetics at High Ascorbate ConcentrationPre-steady-state
kinetics were performed using the QFM-4000 system
as mentioned above but with the following adjustments: 600 μM
PHM and 600 μM Dns-YVG were preincubated together and infused
with pure oxygen at 23 °C. Anaerobic ascorbate was prepared in
the anaerobic chamber, preloaded into a syringe, and sealed with a
septum. Single-turnover kinetic curves were obtained by shooting the
600 μM PHM + Dns-YVG mixture against anaerobic ascorbate, resulting
in a final concentration of 6 mM at varied aging times: 5, 20, 40,
60, 80, 100, 150, 200, 250, 300, and 500 ms. The samples were quenched
with 8% TFA (4% final concentration) in water and analyzed via HPLC
as described above.

Pre-Steady-State
Kinetics at High Ascorbate Concentration

Pre-steady-state
kinetics were performed using the QFM-4000 system
as mentioned above but with the following adjustments: 600 μM
PHM and 600 μM Dns-YVG were preincubated together and infused
with pure oxygen at 23 °C. Anaerobic ascorbate was prepared in
the anaerobic chamber, preloaded into a syringe, and sealed with a
septum. Single-turnover kinetic curves were obtained by shooting the
600 μM PHM + Dns-YVG mixture against anaerobic ascorbate, resulting
in a final concentration of 6 mM at varied aging times: 5, 20, 40,
60, 80, 100, 150, 200, 250, 300, and 500 ms. The samples were quenched
with 8% TFA (4% final concentration) in water and analyzed via HPLC
as described above.

“Pulse-Chase” Kinetics with
Acetyl-YVG and Dansyl-YVGTo determine if the enzyme was
undergoing multiple turnovers, a
pulse-chase experiment was performed using the QFM. Oxidized, 600
μM PHM WT was preincubated with 600 μM Dns-YVG. The PHM
and Dns-YVG mixture were gently infused with pure oxygen in a 5 mL
septum-fitted Wheaton vial for approximately 30 min at 23 °C.
When removing the vial from the oxygen infuser, a small amount of
oxygen overpressure was allowed to remain in the vial. Anaerobic ascorbate
+ acetyl-YVG (Ac-YVG) was prepared from powder in a Vacuum Atmospheres
anaerobic glovebox using anaerobic buffer (20 mM NaP pH 7.5) at 10
mM ascorbate plus 3 mM Ac-YVG. The ascorbate mixture was preloaded
into syringes within the anaerobic chamber and sealed with septa when
removed from the chamber. Repeated turnover was then assessed by shooting
the 600 μM PHM + Dns-YVG mixture against an equal volume of
anaerobic ascorbate Ac-YVG mixture, with aging times of 0, 300, and
1000 ms performed in triplicate. The mixture was then quenched with
an equal volume of 8% trifluoroacetic acid (TFA) and then diluted
with 20 mM NaP buffer resulting in a final concentration of 100 μM
PHM + 100 μM Dns-(substrate + product) + 1.67 mM ascorbate +
500 μM Act-YVG. Samples were analyzed by HPLC with a Varian
Pro Star solvent delivery module equipped with a Varian Pro Star model
410 autosampler (250 μL syringe, 100 μL sample loop) on
a 250 mm × 4.6 mm Varian Microsorb-MV 100-5 C18 column as described
above and previously reported.35

“Pulse-Chase” Kinetics with
Acetyl-YVG and Dansyl-YVG

To determine if the enzyme was
undergoing multiple turnovers, a
pulse-chase experiment was performed using the QFM. Oxidized, 600
μM PHM WT was preincubated with 600 μM Dns-YVG. The PHM
and Dns-YVG mixture were gently infused with pure oxygen in a 5 mL
septum-fitted Wheaton vial for approximately 30 min at 23 °C.
When removing the vial from the oxygen infuser, a small amount of
oxygen overpressure was allowed to remain in the vial. Anaerobic ascorbate
+ acetyl-YVG (Ac-YVG) was prepared from powder in a Vacuum Atmospheres
anaerobic glovebox using anaerobic buffer (20 mM NaP pH 7.5) at 10
mM ascorbate plus 3 mM Ac-YVG. The ascorbate mixture was preloaded
into syringes within the anaerobic chamber and sealed with septa when
removed from the chamber. Repeated turnover was then assessed by shooting
the 600 μM PHM + Dns-YVG mixture against an equal volume of
anaerobic ascorbate Ac-YVG mixture, with aging times of 0, 300, and
1000 ms performed in triplicate. The mixture was then quenched with
an equal volume of 8% trifluoroacetic acid (TFA) and then diluted
with 20 mM NaP buffer resulting in a final concentration of 100 μM
PHM + 100 μM Dns-(substrate + product) + 1.67 mM ascorbate +
500 μM Act-YVG. Samples were analyzed by HPLC with a Varian
Pro Star solvent delivery module equipped with a Varian Pro Star model
410 autosampler (250 μL syringe, 100 μL sample loop) on
a 250 mm × 4.6 mm Varian Microsorb-MV 100-5 C18 column as described
above and previously reported.35

Fully
Reduced Ascorbate-Free Stopped-Flow KineticsAscorbate-free
reduced PHM WT was prepared in the anaerobic chamber
as described above. The protein was prepared in a pH 7.5 buffer system
containing equal volumes of 50 mM each of MES, HEPES, and sodium phosphate.
WT PHM (500 μM) was shot against 5 mM Ac-YVG + 20 mM sodium
azide. Data were collected for 500 ms at 390 nm, with readings taken
every 5 ms in an Applied Photosystems SX20 stopped-flow module.

Fully
Reduced Ascorbate-Free Stopped-Flow Kinetics

Ascorbate-free
reduced PHM WT was prepared in the anaerobic chamber
as described above. The protein was prepared in a pH 7.5 buffer system
containing equal volumes of 50 mM each of MES, HEPES, and sodium phosphate.
WT PHM (500 μM) was shot against 5 mM Ac-YVG + 20 mM sodium
azide. Data were collected for 500 ms at 390 nm, with readings taken
every 5 ms in an Applied Photosystems SX20 stopped-flow module.

Fully Reduced Ascorbate-Free Rapid Freeze-Quench EPRRapidly
frozen EPR samples were prepared using a Quench-Flow-3 (KinTek)
apparatus configured for RFQ by the manufacturer. Reduced, 200 μM
ascorbate-free protein samples (400 μM in Cu(I)) in anaerobic
buffer and 200 μM acetyl-YVG substrate in fully oxygenated buffer
([O2] ≈ 1.1 mM before 1:1 dilution on mixing) were
loaded onto the instrument and rapidly mixed at the desired time point.
A universal buffer was used comprised of equal volumes of 50 mM MES,
HEPES, and CHES adjusted to pH 7.5. Mixtures were quenched in liquid
ethane at −170 °C. The resulting snow was then packed
into 707-SQ-250M EPR tubes (Wilmad LabGlass). Oxidation was monitored
in a Bruker E500 X-Band EPR spectrometer equipped with a superX microwave
bridge and a dual-mode cavity with a helium flow cryostat (ESR900,
Oxford Instrument, Inc.). The experimental conditions were frequency
9.63 GHz, temperature 100 K, microwave power 20 mW, gain 10 dB, modulation
amplitude 10 G, and sweep time 84 s.

Fully Reduced Ascorbate-Free Rapid Freeze-Quench EPR

Rapidly
frozen EPR samples were prepared using a Quench-Flow-3 (KinTek)
apparatus configured for RFQ by the manufacturer. Reduced, 200 μM
ascorbate-free protein samples (400 μM in Cu(I)) in anaerobic
buffer and 200 μM acetyl-YVG substrate in fully oxygenated buffer
([O2] ≈ 1.1 mM before 1:1 dilution on mixing) were
loaded onto the instrument and rapidly mixed at the desired time point.
A universal buffer was used comprised of equal volumes of 50 mM MES,
HEPES, and CHES adjusted to pH 7.5. Mixtures were quenched in liquid
ethane at −170 °C. The resulting snow was then packed
into 707-SQ-250M EPR tubes (Wilmad LabGlass). Oxidation was monitored
in a Bruker E500 X-Band EPR spectrometer equipped with a superX microwave
bridge and a dual-mode cavity with a helium flow cryostat (ESR900,
Oxford Instrument, Inc.). The experimental conditions were frequency
9.63 GHz, temperature 100 K, microwave power 20 mW, gain 10 dB, modulation
amplitude 10 G, and sweep time 84 s.

Fourier Transform Infrared
Spectroscopy (FTIR)PHM
(500 μM) in 20 mM pH 7.5 NaP buffer was deoxygenated and saturated
with humidified CO by streaming across the surface of the solution
with pure CO gas for 5 min with gentle agitation. Once carbonylated,
Dns-YVG was added to a final concentration of 2.5 mM, and then a small
aliquot of 2 M anaerobic buffered sodium ascorbate was added to achieve
final concentrations of ascorbate between 0 and 25 mM. The mixture
was then carbonylated again for 5 min. One hundred microliters of
solution was then added by a syringe into the IR cell composed of
two CaF2 windows separated by a 50 μm Teflon spacer.
FTIR data were recorded on a Bruker Tensor 27 FTIR spectrophotometer
continuously purged with CO2-free dry air as previously
described.18 One thousand scans were collected
for both protein sample and buffer blank from 2250 to 1900 cm–1 at a resolution of 2 cm–1. Spectral
analysis including subtraction of the buffer blank was performed using
GRAMS AI spectroscopy software (Thermo).Ratios of CO bands
at 2063 and 2093 cm–1 were obtained from the peak
heights rather than peak integrals. The rational for this less rigorous
approach stems from several factors; the cell was taken apart and
reassembled between measurements taken over several days, which introduced
small changes in path length and fill quality. Slight differences
in protein and copper concentrations were introduced due to evaporation
during carbonylation. Finally, different calcium fluoride windows
were used across the experiment, which may have very slightly different
optical properties.

Fourier Transform Infrared
Spectroscopy (FTIR)

PHM
(500 μM) in 20 mM pH 7.5 NaP buffer was deoxygenated and saturated
with humidified CO by streaming across the surface of the solution
with pure CO gas for 5 min with gentle agitation. Once carbonylated,
Dns-YVG was added to a final concentration of 2.5 mM, and then a small
aliquot of 2 M anaerobic buffered sodium ascorbate was added to achieve
final concentrations of ascorbate between 0 and 25 mM. The mixture
was then carbonylated again for 5 min. One hundred microliters of
solution was then added by a syringe into the IR cell composed of
two CaF2 windows separated by a 50 μm Teflon spacer.
FTIR data were recorded on a Bruker Tensor 27 FTIR spectrophotometer
continuously purged with CO2-free dry air as previously
described.18 One thousand scans were collected
for both protein sample and buffer blank from 2250 to 1900 cm–1 at a resolution of 2 cm–1. Spectral
analysis including subtraction of the buffer blank was performed using
GRAMS AI spectroscopy software (Thermo).

Ratios of CO bands
at 2063 and 2093 cm–1 were obtained from the peak
heights rather than peak integrals. The rational for this less rigorous
approach stems from several factors; the cell was taken apart and
reassembled between measurements taken over several days, which introduced
small changes in path length and fill quality. Slight differences
in protein and copper concentrations were introduced due to evaporation
during carbonylation. Finally, different calcium fluoride windows
were used across the experiment, which may have very slightly different
optical properties.

ResultsPHM Is Unreactive to Oxygen
in the Absence of SubstrateConsensus mechanisms for PHM
and DBM are predicated on the initial
binding of oxygen to the reduced Cu(I)M-site. To probe this chemistry,
we explored the sensitivity of the Cu(I)M-site to oxygen in the absence
of substrate. Figure 2 shows EPR spectra of PHM as a function of time exposed to oxygen.
A sample of the fully oxidized enzyme was divided into three aliquots.
One aliquot was transferred to an EPR tube and measured without further
modifications. A second aliquot was reduced anaerobically with 2 molar
equivalents of ascorbate, and the excess ascorbate was removed by
two passages through a spin desalting column and the EPR spectrum
was measured. The third sample was reduced and ascorbate-depleted
as described but then exposed to oxygen for 2 min, and the EPR spectrum
was determined once again. The spectra were double-integrated and
compared to the doubly integrated EPR spectrum of the first aliquot,
which had not been reduced. The results are shown in Figure 2. Ascorbate reduction and removal
resulted as expected in a low integral (3.5% EPR-detectable). However,
exposure to oxygen for 2 min generated only 11.3% of the fully oxidized
Cu(II) signal, indicating oxidation on a time scale much slower than
the catalytic rate. While the absence of an EPR spectrum could be
attributable to a spin-coupled cupric superoxo species, such an entity
should be labile leading to superoxide dissociation and disproportionation,
ultimately leading to an oxidized Cu(II)M-center.Figure 2Top: EPR spectra of WT
PHM mixed with oxygenated buffer at various
time points; the purple trace shows fully oxidized, the black trace
shows the ascorbate-reduced anaerobic enzyme, and the red trace shows
a 120 s reoxidized enzyme. Bottom: quantitation via double integration
(colors are the same as for spectra) and calibration against a standard
solution of 300 μM Cu(II)–EDTA (blue). EPR conditions:
frequency 9.63 GHz, T = 100 K, microwave power 20
mW, gain 10 dB, modulation amplitude 10 G, and sweep time 84 s.The low reactivity of the fully reduced ascorbate-free
enzyme with
oxygen is not unexpected since a notable feature of these enzymes
is the tight coupling of oxygen reduction to product formation24,25,34 as well as evidence that suggests
that substrate binding precedes oxygen binding.40 These data indicate that oxygen reactivity is contingent
on other factors, where the reactivity and/or redox potential of the
Cu(I)M-center is influenced by substrate binding. Additionally, the
data demonstrate that in the absence of peptidyl substrate, exogenous
ascorbate is not required to maintain the reduced state of the enzyme,
even over a relatively long duration of oxygen exposure. Interestingly,
while the fully oxidized enzyme clearly shows resolved components
due to the two chemically inequivalent copper centers, the slow, noncatalytic
reoxidation appears to favor one of these sites, as predicted by previous
redox potential measurements.41Reduction
Stoichiometry Is Dependent on Exogenous AscorbateTo further
our understanding of the PHM reaction mechanism, we
explored the pre-steady-state reactivity of the fully reduced ascorbate-free
enzyme with oxygen and peptidyl substrate. Under these conditions,
peptidyl substrate binding to the dicopper(I) enzyme should activate
the catalysis and lead to rapid generation of the α-hydroxyglycyl
product. The fully reduced, ascorbate-free PHM was shot against the
stoichiometric dansyl-YVG substrate dissolved in buffer saturated
with O2 (1.1 mM) in the absence of an exogenous ascorbate
using a BioLogic QFM-4000 quench-flow microvolume mixture instrument.
Here, the substrate is the limiting reagent, so the reaction should
terminate when all substrates have been converted to products. Reactions
were acid-quenched at successive time points with TFA, and both product
and substrate were quantified via HPLC using fluorescence of the associated
dansyl group. Under the canonical reaction mechanism, where ascorbate
is solely the physiological electron source, this prereduced PHM should
be fully primed with reducing equivalents and should be capable of
reacting stoichiometrically with the substrate to reach complete product
formation. Figure 3a shows the concentration of substrate and product over the course
of the reaction, indicating that this reaction under ascorbate-free
conditions while fast is substoichiometric, achieving
only approximately 10–15% of the expected product with a kobs = 150 s–1. Despite being
fully reduced, the enzyme shows only limited reactivity with substrate
and oxygen. The incomplete reactivity under these conditions is also
interesting given PHM’s high affinity for dansyl-YVG35 (KM = 5 μM, KD,reduced enzyme = 22 μM), suggesting
that while the substrate must be binding to the enzyme in the presence
of oxygen, turnover is not occurring.Figure 3Pre-steady-state reactivity of the fully
reduced ascorbate-free
PHM. (a) Time evolution of the product in the absence of excess ascorbate, kobs = 150 s–1: blue trace
represents the consumption of dansyl-YVG substrate and orange trace
is the rate of generation of the dansyl-α-hydroxyglycine-YVG
product. (b) Time evolution of substrate consumption (blue) and product
generation (orange) in the presence of excess ascorbate (6 mM), kobs = 36 s–1. (c) Bar graph
showing an increase in the mole fraction of product per mole of enzyme
as a function of added ascorbate. (d) Ascorbate dependence of product
formation fitted to a ligand-binding event with KD = 117 μM.We then explored the role of ascorbate as a possible reaction initiator
by titrating the reactivity with increasing amounts of ascorbate.
The addition of ascorbate resulted in an increase in the stoichiometry
of product formation in a concentration-dependent manner (Figure 3b,c), reaching a
final stoichiometry of ∼70% at an ascorbate concentration of
2 mM. Figure 3b shows
that in the presence of excess exogenous ascorbate the reaction proceeds
as expected with the rapid conversion of substrate to product in a
manner consistent with first-order kinetics, with a pseudo-rate constant kobs = 36 s–1 that is over
4-fold slower than for the ascorbate-free enzyme. This suggests that
in the presence of ascorbate some other process is rate-limiting.
To address whether the inability to generate 100% product under saturating
conditions of ascorbate was due to copper loss, we measured the Cu
concentration after ascorbate reduction and removal (Table S1). Results show that in some samples, copper loss
equivalent to 0.3 Cu/protein was detected, which could account for
the 70% ceiling on product since the loss of one of the two coppers
per enzyme eliminates the activity due to both coppers
in that molecule.The ascorbate dependency of the pre-steady-state
reaction is shown
in Figure 3c, while Figure 3d plots the mole
ratio of product formation as a function of ascorbate concentration.
Ascorbate is seen to influence the stoichiometry of product generation
via a process that involves a binding event with a KD of 117 μM.Ascorbate Dependency Is
Not the Result of Multiple TurnoversA plausible explanation
for the observed behavior of ascorbate
dependency of product stoichiometry is the hypothesis that a large
population of the enzyme is unreactive, perhaps due to aggregation
or some other feature of protein preparation. In this scenario, only
a small population of the enzyme is functional but can undergo multiple
turnovers under the conditions of the chemical quench experiment when
excess reductant is supplied. To assess this possibility, a pulse-chase
experiment was performed. Here, the oxidized enzyme treated with 1
equiv of dansyl-YVG in buffer saturated with O2 was shot
against excess ascorbate containing a 5-fold molar excess of the unlabeled
acetyl-YVG (Ac-YVG). Since the enzyme is preincubated with the fluorescent-detectable
dansyl-YVG, the expectation in the case of the fully reactive enzyme
is rapid pre-steady-state turnover to generate the dansyl-labeled
product in near-stoichiometric amounts. In the case of a bulk-unreactive
enzyme, the expectation is an approximate 5-fold reduction in dansyl-labeled
product as the unlabeled substrate will compete with the labeled substrate
for enzyme binding in the second and subsequent turnovers. Put another
way, the acetyl-labeled substrate should act as a competitive inhibitor
of dansylated product production. (This experiment works because dansyl-
and acetyl- labeled substrates have similar affinity for the enzyme). Figure 4 shows that while
the rate of the reaction was marginally decreased by the unlabeled
acetyl peptide, its reaction progress resembles the uninhibited reaction,
consistent with full enzyme reactivity. These data together demonstrate
that the enzyme requires activation to achieve full reactivity by
a process that is dependent in some way on ascorbate binding. This
is not due to increased reductive activity since in the absence of
ascorbate the enzyme is fully reduced.Figure 4Product generated from
the reaction of dansyl-YVG with the equimolar
fully reduced enzyme chased with a 5-fold excess of unlabeled Ac-YVG.
Blue circles represent the product generated with equimolar dansyl-YVG
and the fully reduced enzyme; purple diamonds represent the product
generated when the same experiment is conducted in the presence of
a 5-fold molar excess of Ac-YVG. The black dashed line represents
the expected dansyl product after chasing with Ac-YVG if multiple
turnovers occur.Ascorbate Reduction Kinetics
Indicate an Ascorbate Binding SiteAs a final probe on the
effects of exogenous ascorbate on PHM reactivity,
stopped-flow was employed to determine the rate of reduction as a
function of ascorbate. To accomplish this, azide was used as a reporter
for the oxidized state of the enzyme. Azide is useful in these experiments
as it binds only to the oxidized form of the enzyme and has a characteristic
390 nm UV–vis signal32 (Figure S1, Supporting Information). Additionally,
the formation of the PHM Cu(II)–azido complex is not rate-limiting,
being formed within the time of mixing (Figure S2, Supporting Information). Elimination of this signal by
ascorbate therefore serves as a proxy for the rate of reduction. Figure 5 shows the rate of
reduction as a function of ascorbate concentration. The reduction
obeys saturation kinetics with KD = 1120
μM and kobs = 146 s–1 and indicates that ascorbate binds to the enzyme to accomplish reduction.
This KD is expected to reflect the affinity
of ascorbate for the oxidized enzyme during the reduction
of the di-Cu(II) to di-Cu(I) states and should be similar to the apparent Km determined from the dependence of product
stoichiometry on ascorbate if the latter is due merely to multiple
turnovers. However, as shown in Figure 3d, the integrated peak signal for the product plotted
versus ascorbate concentration yields a KD of 117 μM, an order of magnitude lower than that of the reduction
event. This further supports a role for ascorbate in activation of
the fully reduced enzyme over and above a simple role as a reducing
agent.Figure 5Ascorbate dependence of the rate of PHM reduction determined by
stopped-flow using the Cu(II) PHM–azido adduct (λmax = 390 nm) as a reporter for the concentration of Cu(II)–enzyme
remaining at each time point. Data are fit to a pre-equilibrium binding
event with KD = 1.1 mM.Rapid Freeze-Quench EPR of the Pre-Steady-State ReactionThe above experiments demonstrate that product formation is substoichiometric
without exogenous ascorbate but do not definitively identify the redox
state of the enzyme during or after the pre-steady-state reaction.
One possible yet unlikely explanation for the apparent ascorbate dependency
is the oxidation of the enzyme without product formation, initiated
by substrate and oxygen binding. Here, excess ascorbate would rereduce
the enzyme to allow turnover and apparent ascorbate dependency. Such
reactivity does occur in other copper monooxygenase such as the LPMOs.42 To address this possibility, rapid freeze-quench
(RFQ) EPR was employed. Using this technique, the redox state of the
enzyme can be assessed independent of product formation under single-turnover
conditions. Figure 6 shows the results of shooting fully reduced ascorbate-free enzyme
against oxygenated buffer containing 1 mole equivalent of substrate
relative to the enzyme concentration. The results (Figure 6, bottom panel, and Table S2) show that in the absence of excess
ascorbate negligible amounts of the enzyme become EPR-detectable after
300 ms, a time point at which product formation has ceased. This result
is remarkable as consensus mechanisms predict that in the absence
of reductant, turnover should generate two atoms of EPR-detectable
Cu(II) for each product molecule produced. In the present experiment,
if we allow for 20% product production, we anticipate 60 μM
EPR-detectable Cu(II) compared with the experimentally determined
amount of <3 μM. While this result should be interpreted
with caution, it may suggest that an EPR-detectable entity such as
a di-Cu(II) spin-coupled species is formed during or after the product
is formed (vide infra). We note that a spin-coupled Cu(II)-superoxo
species is less consistent with the data as any cupric superoxide
formed is expected to have been converted into hydroxylated product
after 300 ms.Figure 6Rapid freeze-quench-derived EPR spectra of the products
of the
pre-steady-state PHM reaction. EPR spectra were generated by rapidly
mixing fully reduced ascorbate-free PHM (400 mM in Cu(I)) with the
oxygenated buffer containing equimolar dansyl-YVG and freezing in
liquid ethane (−120 °C). The top panel shows the actual
spectra, and the bottom panel shows the quantitation of the Cu(II)
signal by reference to a standard Cu(II)–EDTA sample: red,
13 ms; green, 25 ms; blue, 100 ms; orange, 300 ms; and purple, 250
μM Cu(II)–EDTA standard. To avoid errors from integration
of noise peaks, intensities were determined from peak-to-peak heights
at g⊥ relative to that of a 250 μM solution
of Cu(II)–EDTA taken through an identical cycle of mixing,
freezing, and packing. Errors in concentrations determined from the
peak heights are estimated from triplicate measurements of a Cu(II)–EDTA
standard to be ≅ 15% and are mainly due to differences in the
packing density of the samples. EPR conditions: frequency 9.63 GHz, T = 100 K, microwave power 20 mW, gain 10 dB, modulation
amplitude 10 G, and sweep time 84 s.Cu(II)–Azido-Detected Stopped-Flow Spectrometry of the
Pre-Steady-State ReactionAs discussed above, the oxidized
di-Cu(II) PHM reacts with azide to generate an azide-to-copper charge-transfer
complex with λmax = 390 nm. The formation of the
azido complex is contingent on the oxidation of the enzyme and therefore
serves as a reporter for the pre-steady state enzyme turnover from
Cu(I) to Cu(II) states. Further, its rate of formation is extremely
rapid, the complex being fully formed in the time of mixing in the
stopped-flow (Figure S2, Supporting Information).
The reaction of fully reduced, ascorbate-free enzyme with Ac-YVG and
oxygen was followed using stopped-flow spectrometry monitoring the
rate of appearance of the 390 nm absorption. Here, the fully reduced
enzyme loaded with Ac-YVG was shot against buffer saturated with oxygen
and containing 20 mM sodium azide. Figure 7 shows that in the absence of excess ascorbate
less than 30% of the fully oxidized Cu(II)–azido complex is
formed based on the expected absorbance for full copper oxidation
(red line, Figure 7). Although this amount is larger than that based on product formation
or RFQ EPR, it reinforces our findings that PHM has limited reactivity
with substrate in the absence of excess ascorbate but may also suggest
that product stoichiometry is influenced by other factors such as
azide concentration.Figure 7Pre-steady-state reactivity of the fully reduced ascorbate-free
PHM determined by measuring the rate of appearance of the 390 nm Cu(II)–azido
adduct at 390 nm using stopped-flow spectrophotometry. 500 μM
of WT PHM was shot against 5 mM Ac-YVG + 20 mM sodium azide. Data
were collected for 500 ms, with readings taken every 5 ms. The top
trace shows the generation of the 390 nm absorbance as a function
of time in seconds. The bottom trace shows the rate curve fitted to
a single exponential with kobs= 72 s–1. The red dashed line represents the absorbance expected
for complete reaction where all copper is oxidized to Cu(II) determined
from a separate experiment where fully oxidized Cu(II) PHM was shot
against buffer containing sodium azide under identical conditions.FTIR Indicates a Shift in Enzyme Population
in Response to Ascorbate
TitrationTo determine the specific effect of ascorbate on
PHM reactivity, we looked into insights previously garnered from FTIR.
Carbon monoxide (CO) FTIR has been used extensively to probe metal
oxygen-binding active sites, where it can give remarkably detailed
information about structure and oxygen binding.18,37,43−45 Subtle changes to the
binding site are reflected in shifts in the IR stretching frequency
of the carbon monoxide. Furthermore, carbon monoxide has been shown
to bind solely to the M-site in PHM and therefore it functions as
a probe of M-site configuration. It has been shown previously that
varied concentrations of substrates tune two separate populations
of stretching frequencies.18 In the absence
of substrate, the spectra are dominated by a large 2093 cm–1 stretching frequency with a minor 2063 cm–1 shoulder.
Addition of substrate causes an increase in the 2063 cm–1 species with an attenuation of 2093 cm–1 in a
concentration-dependent manner. Increasing substrate concentration
was, however, not sufficient to eliminate the 2093 cm–1 stretch nor did a complete absence of substrate abrogate the 2063
cm–1 signal. These data suggested a pair of interconvertible
states mediated by substrate binding, but the significance of these
states to reactivity as well as their identity was unclear.Peptidyl substrate’s ability to partition the enzyme population
echoes ascorbate’s ability to induce a catalytically competent
state, particularly since the CO binding experiments are performed
in the presence of exogenous ascorbate. To probe that relationship
further, we performed FTIR experiments to test if ascorbate induces
an alternative enzyme configuration in a manner resembling substrate.
WT PHM was exposed to carbon monoxide in the presence of substrate
and varied concentrations of ascorbate. Figure 8 (top) shows that increasing the ascorbate
concentration causes the ratio of 2063 to 2093 cm–1 peaks to increase in a concentration-dependent manner, reaching
a maximum at 15 mM exogenous ascorbate, while Figure 8 (bottom) depicts two representative spectra
at 1 and 15 mM exogenous ascorbate, respectively. Curiously, excess
ascorbate beyond 15 mM resulted in a decrease in the ratio of the
2063 to 2093 cm–1 signals, but these concentrations
are well above the ascorbate concentration (2 mM) required for maximum
product stoichiometry. This experiment demonstrates that ascorbate,
in the presence of substrate and CO, is sufficient to induce an alternate
conformer of PHM.Figure 8Evolution of the substrate-induced 2063 cm–1 M-site
Cu(I)–CO signal as a function of the added ascorbate. The top
trace shows the ratio of 2063 to 2093 cm–1 peaks
plotted against ascorbate concentration. The bottom trace shows the
representative spectra at 1 mM (orange) and 15 mM (green).This explicit connection between the 2063 cm–1 signal and ascorbate concentration provides insights into the role
of ascorbate in catalysis. Exogenous ascorbate is here shown to both
bias the population of enzyme to the 2063 cm–1 form
and increase overall reactivity. This suggests that substrate and
ascorbate work combinatorially to induce the 2063 cm–1 form of the enzyme and presents strong evidence that this form is
the active species.The identity of this active state remains
unclear. Changes to the
hydrogen-bonding network surrounding the CuM site have been proposed
as one mechanism of substrate activation.13,18 Another possible interpretation is a large-scale conformational
change. As discussed in more detail below, ascorbate binding could
potentially induce an open-to-closed conformational transition where
the Cu site separation decreases from 11 to 4–5 Å, similar
to that observed in DBM. Once in proximity, electron transfer is facile
between the two sites. There are several pieces of evidence that support
this interpretation. First is the scale of the change: a 30 cm–1 change suggests a relatively large increase in the
electron-donating power of the CuM site, larger than changes caused
by mutating nearby ligands to perturb the hydrogen-bonding network.
Second, the red-shifted IR frequency is reminiscent of that observed
in the dinuclear arthropodal hemocyanin (2043 cm–1) and attributed to a semibridging species.44 In the context of PHM, such a bridging species is not possible unless
the copper sites were in much closer proximity, as in the case of
hemocyanin, which has an intracopper distance of 4 Å. Third is
the observation of a closed conformer in crystal structures of the
PHM H108A mutant in the presence of citrate,13 indicating that under certain conditions the interdomain closure
is favored. Indeed, the ability for citrate to help induce a closed
conformer is evocative, as both citrate and ascorbate are hydroxy
acids. The binding of the citrate is interesting as well since it
binds in a manner to bridge the CuM and CuH sites and may provide
stability for the closed conformer (Figure 9).Figure 9Closed and open structures of DBM and PHM. Left,
two conformers
of DBM with Cu atom (bronze) locations modeled (PDB file 4EZL). Middle:
alignment of the protein fold of the closed crystal structure of H108A
in the presence of citrate (PDB file 6ALA), with the open structure
of the oxidized PHM (PDB file 1PHM). PHM structures were aligned on
the N-terminal (H) subdomain with an RMSD value of 0.41. The H108A-citrate
and the oxidized native PHM main chains are depicted by pink and purple
ribbons, respectively. To emphasize the hinge motion, the first β-strands
of the M-domains are colored green (closed: H108A-citrate) and dark
blue (open: oxidized PHM). Right: enlarged view of the metal binding
site in the H108A-citrate PHM structure. The single copper atom of
the closed H108A-citrate structure, which coordinates M-site residues
H242 and H244 and H-site residue H107, is shown in pink, while the
citrate molecule is shown as a green carbon backbone with oxygen atoms
in red. The positions of the two copper sites in the open PHM structure
are shown as transparent slate spheres.

PHM Is Unreactive to Oxygen
in the Absence of SubstrateConsensus mechanisms for PHM
and DBM are predicated on the initial
binding of oxygen to the reduced Cu(I)M-site. To probe this chemistry,
we explored the sensitivity of the Cu(I)M-site to oxygen in the absence
of substrate. Figure 2 shows EPR spectra of PHM as a function of time exposed to oxygen.
A sample of the fully oxidized enzyme was divided into three aliquots.
One aliquot was transferred to an EPR tube and measured without further
modifications. A second aliquot was reduced anaerobically with 2 molar
equivalents of ascorbate, and the excess ascorbate was removed by
two passages through a spin desalting column and the EPR spectrum
was measured. The third sample was reduced and ascorbate-depleted
as described but then exposed to oxygen for 2 min, and the EPR spectrum
was determined once again. The spectra were double-integrated and
compared to the doubly integrated EPR spectrum of the first aliquot,
which had not been reduced. The results are shown in Figure 2. Ascorbate reduction and removal
resulted as expected in a low integral (3.5% EPR-detectable). However,
exposure to oxygen for 2 min generated only 11.3% of the fully oxidized
Cu(II) signal, indicating oxidation on a time scale much slower than
the catalytic rate. While the absence of an EPR spectrum could be
attributable to a spin-coupled cupric superoxo species, such an entity
should be labile leading to superoxide dissociation and disproportionation,
ultimately leading to an oxidized Cu(II)M-center.Figure 2Top: EPR spectra of WT
PHM mixed with oxygenated buffer at various
time points; the purple trace shows fully oxidized, the black trace
shows the ascorbate-reduced anaerobic enzyme, and the red trace shows
a 120 s reoxidized enzyme. Bottom: quantitation via double integration
(colors are the same as for spectra) and calibration against a standard
solution of 300 μM Cu(II)–EDTA (blue). EPR conditions:
frequency 9.63 GHz, T = 100 K, microwave power 20
mW, gain 10 dB, modulation amplitude 10 G, and sweep time 84 s.The low reactivity of the fully reduced ascorbate-free
enzyme with
oxygen is not unexpected since a notable feature of these enzymes
is the tight coupling of oxygen reduction to product formation24,25,34 as well as evidence that suggests
that substrate binding precedes oxygen binding.40 These data indicate that oxygen reactivity is contingent
on other factors, where the reactivity and/or redox potential of the
Cu(I)M-center is influenced by substrate binding. Additionally, the
data demonstrate that in the absence of peptidyl substrate, exogenous
ascorbate is not required to maintain the reduced state of the enzyme,
even over a relatively long duration of oxygen exposure. Interestingly,
while the fully oxidized enzyme clearly shows resolved components
due to the two chemically inequivalent copper centers, the slow, noncatalytic
reoxidation appears to favor one of these sites, as predicted by previous
redox potential measurements.41

PHM Is Unreactive to Oxygen
in the Absence of Substrate

Consensus mechanisms for PHM
and DBM are predicated on the initial
binding of oxygen to the reduced Cu(I)M-site. To probe this chemistry,
we explored the sensitivity of the Cu(I)M-site to oxygen in the absence
of substrate. Figure 2 shows EPR spectra of PHM as a function of time exposed to oxygen.
A sample of the fully oxidized enzyme was divided into three aliquots.
One aliquot was transferred to an EPR tube and measured without further
modifications. A second aliquot was reduced anaerobically with 2 molar
equivalents of ascorbate, and the excess ascorbate was removed by
two passages through a spin desalting column and the EPR spectrum
was measured. The third sample was reduced and ascorbate-depleted
as described but then exposed to oxygen for 2 min, and the EPR spectrum
was determined once again. The spectra were double-integrated and
compared to the doubly integrated EPR spectrum of the first aliquot,
which had not been reduced. The results are shown in Figure 2. Ascorbate reduction and removal
resulted as expected in a low integral (3.5% EPR-detectable). However,
exposure to oxygen for 2 min generated only 11.3% of the fully oxidized
Cu(II) signal, indicating oxidation on a time scale much slower than
the catalytic rate. While the absence of an EPR spectrum could be
attributable to a spin-coupled cupric superoxo species, such an entity
should be labile leading to superoxide dissociation and disproportionation,
ultimately leading to an oxidized Cu(II)M-center.

Top: EPR spectra of WT
PHM mixed with oxygenated buffer at various
time points; the purple trace shows fully oxidized, the black trace
shows the ascorbate-reduced anaerobic enzyme, and the red trace shows
a 120 s reoxidized enzyme. Bottom: quantitation via double integration
(colors are the same as for spectra) and calibration against a standard
solution of 300 μM Cu(II)–EDTA (blue). EPR conditions:
frequency 9.63 GHz, T = 100 K, microwave power 20
mW, gain 10 dB, modulation amplitude 10 G, and sweep time 84 s.

The low reactivity of the fully reduced ascorbate-free
enzyme with
oxygen is not unexpected since a notable feature of these enzymes
is the tight coupling of oxygen reduction to product formation24,25,34 as well as evidence that suggests
that substrate binding precedes oxygen binding.40 These data indicate that oxygen reactivity is contingent
on other factors, where the reactivity and/or redox potential of the
Cu(I)M-center is influenced by substrate binding. Additionally, the
data demonstrate that in the absence of peptidyl substrate, exogenous
ascorbate is not required to maintain the reduced state of the enzyme,
even over a relatively long duration of oxygen exposure. Interestingly,
while the fully oxidized enzyme clearly shows resolved components
due to the two chemically inequivalent copper centers, the slow, noncatalytic
reoxidation appears to favor one of these sites, as predicted by previous
redox potential measurements.41

Reduction
Stoichiometry Is Dependent on Exogenous AscorbateTo further
our understanding of the PHM reaction mechanism, we
explored the pre-steady-state reactivity of the fully reduced ascorbate-free
enzyme with oxygen and peptidyl substrate. Under these conditions,
peptidyl substrate binding to the dicopper(I) enzyme should activate
the catalysis and lead to rapid generation of the α-hydroxyglycyl
product. The fully reduced, ascorbate-free PHM was shot against the
stoichiometric dansyl-YVG substrate dissolved in buffer saturated
with O2 (1.1 mM) in the absence of an exogenous ascorbate
using a BioLogic QFM-4000 quench-flow microvolume mixture instrument.
Here, the substrate is the limiting reagent, so the reaction should
terminate when all substrates have been converted to products. Reactions
were acid-quenched at successive time points with TFA, and both product
and substrate were quantified via HPLC using fluorescence of the associated
dansyl group. Under the canonical reaction mechanism, where ascorbate
is solely the physiological electron source, this prereduced PHM should
be fully primed with reducing equivalents and should be capable of
reacting stoichiometrically with the substrate to reach complete product
formation. Figure 3a shows the concentration of substrate and product over the course
of the reaction, indicating that this reaction under ascorbate-free
conditions while fast is substoichiometric, achieving
only approximately 10–15% of the expected product with a kobs = 150 s–1. Despite being
fully reduced, the enzyme shows only limited reactivity with substrate
and oxygen. The incomplete reactivity under these conditions is also
interesting given PHM’s high affinity for dansyl-YVG35 (KM = 5 μM, KD,reduced enzyme = 22 μM), suggesting
that while the substrate must be binding to the enzyme in the presence
of oxygen, turnover is not occurring.Figure 3Pre-steady-state reactivity of the fully
reduced ascorbate-free
PHM. (a) Time evolution of the product in the absence of excess ascorbate, kobs = 150 s–1: blue trace
represents the consumption of dansyl-YVG substrate and orange trace
is the rate of generation of the dansyl-α-hydroxyglycine-YVG
product. (b) Time evolution of substrate consumption (blue) and product
generation (orange) in the presence of excess ascorbate (6 mM), kobs = 36 s–1. (c) Bar graph
showing an increase in the mole fraction of product per mole of enzyme
as a function of added ascorbate. (d) Ascorbate dependence of product
formation fitted to a ligand-binding event with KD = 117 μM.We then explored the role of ascorbate as a possible reaction initiator
by titrating the reactivity with increasing amounts of ascorbate.
The addition of ascorbate resulted in an increase in the stoichiometry
of product formation in a concentration-dependent manner (Figure 3b,c), reaching a
final stoichiometry of ∼70% at an ascorbate concentration of
2 mM. Figure 3b shows
that in the presence of excess exogenous ascorbate the reaction proceeds
as expected with the rapid conversion of substrate to product in a
manner consistent with first-order kinetics, with a pseudo-rate constant kobs = 36 s–1 that is over
4-fold slower than for the ascorbate-free enzyme. This suggests that
in the presence of ascorbate some other process is rate-limiting.
To address whether the inability to generate 100% product under saturating
conditions of ascorbate was due to copper loss, we measured the Cu
concentration after ascorbate reduction and removal (Table S1). Results show that in some samples, copper loss
equivalent to 0.3 Cu/protein was detected, which could account for
the 70% ceiling on product since the loss of one of the two coppers
per enzyme eliminates the activity due to both coppers
in that molecule.The ascorbate dependency of the pre-steady-state
reaction is shown
in Figure 3c, while Figure 3d plots the mole
ratio of product formation as a function of ascorbate concentration.
Ascorbate is seen to influence the stoichiometry of product generation
via a process that involves a binding event with a KD of 117 μM.

Reduction
Stoichiometry Is Dependent on Exogenous Ascorbate

To further
our understanding of the PHM reaction mechanism, we
explored the pre-steady-state reactivity of the fully reduced ascorbate-free
enzyme with oxygen and peptidyl substrate. Under these conditions,
peptidyl substrate binding to the dicopper(I) enzyme should activate
the catalysis and lead to rapid generation of the α-hydroxyglycyl
product. The fully reduced, ascorbate-free PHM was shot against the
stoichiometric dansyl-YVG substrate dissolved in buffer saturated
with O2 (1.1 mM) in the absence of an exogenous ascorbate
using a BioLogic QFM-4000 quench-flow microvolume mixture instrument.
Here, the substrate is the limiting reagent, so the reaction should
terminate when all substrates have been converted to products. Reactions
were acid-quenched at successive time points with TFA, and both product
and substrate were quantified via HPLC using fluorescence of the associated
dansyl group. Under the canonical reaction mechanism, where ascorbate
is solely the physiological electron source, this prereduced PHM should
be fully primed with reducing equivalents and should be capable of
reacting stoichiometrically with the substrate to reach complete product
formation. Figure 3a shows the concentration of substrate and product over the course
of the reaction, indicating that this reaction under ascorbate-free
conditions while fast is substoichiometric, achieving
only approximately 10–15% of the expected product with a kobs = 150 s–1. Despite being
fully reduced, the enzyme shows only limited reactivity with substrate
and oxygen. The incomplete reactivity under these conditions is also
interesting given PHM’s high affinity for dansyl-YVG35 (KM = 5 μM, KD,reduced enzyme = 22 μM), suggesting
that while the substrate must be binding to the enzyme in the presence
of oxygen, turnover is not occurring.

Pre-steady-state reactivity of the fully
reduced ascorbate-free
PHM. (a) Time evolution of the product in the absence of excess ascorbate, kobs = 150 s–1: blue trace
represents the consumption of dansyl-YVG substrate and orange trace
is the rate of generation of the dansyl-α-hydroxyglycine-YVG
product. (b) Time evolution of substrate consumption (blue) and product
generation (orange) in the presence of excess ascorbate (6 mM), kobs = 36 s–1. (c) Bar graph
showing an increase in the mole fraction of product per mole of enzyme
as a function of added ascorbate. (d) Ascorbate dependence of product
formation fitted to a ligand-binding event with KD = 117 μM.

We then explored the role of ascorbate as a possible reaction initiator
by titrating the reactivity with increasing amounts of ascorbate.
The addition of ascorbate resulted in an increase in the stoichiometry
of product formation in a concentration-dependent manner (Figure 3b,c), reaching a
final stoichiometry of ∼70% at an ascorbate concentration of
2 mM. Figure 3b shows
that in the presence of excess exogenous ascorbate the reaction proceeds
as expected with the rapid conversion of substrate to product in a
manner consistent with first-order kinetics, with a pseudo-rate constant kobs = 36 s–1 that is over
4-fold slower than for the ascorbate-free enzyme. This suggests that
in the presence of ascorbate some other process is rate-limiting.
To address whether the inability to generate 100% product under saturating
conditions of ascorbate was due to copper loss, we measured the Cu
concentration after ascorbate reduction and removal (Table S1). Results show that in some samples, copper loss
equivalent to 0.3 Cu/protein was detected, which could account for
the 70% ceiling on product since the loss of one of the two coppers
per enzyme eliminates the activity due to both coppers
in that molecule.

The ascorbate dependency of the pre-steady-state
reaction is shown
in Figure 3c, while Figure 3d plots the mole
ratio of product formation as a function of ascorbate concentration.
Ascorbate is seen to influence the stoichiometry of product generation
via a process that involves a binding event with a KD of 117 μM.

Ascorbate Dependency Is
Not the Result of Multiple TurnoversA plausible explanation
for the observed behavior of ascorbate
dependency of product stoichiometry is the hypothesis that a large
population of the enzyme is unreactive, perhaps due to aggregation
or some other feature of protein preparation. In this scenario, only
a small population of the enzyme is functional but can undergo multiple
turnovers under the conditions of the chemical quench experiment when
excess reductant is supplied. To assess this possibility, a pulse-chase
experiment was performed. Here, the oxidized enzyme treated with 1
equiv of dansyl-YVG in buffer saturated with O2 was shot
against excess ascorbate containing a 5-fold molar excess of the unlabeled
acetyl-YVG (Ac-YVG). Since the enzyme is preincubated with the fluorescent-detectable
dansyl-YVG, the expectation in the case of the fully reactive enzyme
is rapid pre-steady-state turnover to generate the dansyl-labeled
product in near-stoichiometric amounts. In the case of a bulk-unreactive
enzyme, the expectation is an approximate 5-fold reduction in dansyl-labeled
product as the unlabeled substrate will compete with the labeled substrate
for enzyme binding in the second and subsequent turnovers. Put another
way, the acetyl-labeled substrate should act as a competitive inhibitor
of dansylated product production. (This experiment works because dansyl-
and acetyl- labeled substrates have similar affinity for the enzyme). Figure 4 shows that while
the rate of the reaction was marginally decreased by the unlabeled
acetyl peptide, its reaction progress resembles the uninhibited reaction,
consistent with full enzyme reactivity. These data together demonstrate
that the enzyme requires activation to achieve full reactivity by
a process that is dependent in some way on ascorbate binding. This
is not due to increased reductive activity since in the absence of
ascorbate the enzyme is fully reduced.Figure 4Product generated from
the reaction of dansyl-YVG with the equimolar
fully reduced enzyme chased with a 5-fold excess of unlabeled Ac-YVG.
Blue circles represent the product generated with equimolar dansyl-YVG
and the fully reduced enzyme; purple diamonds represent the product
generated when the same experiment is conducted in the presence of
a 5-fold molar excess of Ac-YVG. The black dashed line represents
the expected dansyl product after chasing with Ac-YVG if multiple
turnovers occur.

Ascorbate Dependency Is
Not the Result of Multiple Turnovers

A plausible explanation
for the observed behavior of ascorbate
dependency of product stoichiometry is the hypothesis that a large
population of the enzyme is unreactive, perhaps due to aggregation
or some other feature of protein preparation. In this scenario, only
a small population of the enzyme is functional but can undergo multiple
turnovers under the conditions of the chemical quench experiment when
excess reductant is supplied. To assess this possibility, a pulse-chase
experiment was performed. Here, the oxidized enzyme treated with 1
equiv of dansyl-YVG in buffer saturated with O2 was shot
against excess ascorbate containing a 5-fold molar excess of the unlabeled
acetyl-YVG (Ac-YVG). Since the enzyme is preincubated with the fluorescent-detectable
dansyl-YVG, the expectation in the case of the fully reactive enzyme
is rapid pre-steady-state turnover to generate the dansyl-labeled
product in near-stoichiometric amounts. In the case of a bulk-unreactive
enzyme, the expectation is an approximate 5-fold reduction in dansyl-labeled
product as the unlabeled substrate will compete with the labeled substrate
for enzyme binding in the second and subsequent turnovers. Put another
way, the acetyl-labeled substrate should act as a competitive inhibitor
of dansylated product production. (This experiment works because dansyl-
and acetyl- labeled substrates have similar affinity for the enzyme). Figure 4 shows that while
the rate of the reaction was marginally decreased by the unlabeled
acetyl peptide, its reaction progress resembles the uninhibited reaction,
consistent with full enzyme reactivity. These data together demonstrate
that the enzyme requires activation to achieve full reactivity by
a process that is dependent in some way on ascorbate binding. This
is not due to increased reductive activity since in the absence of
ascorbate the enzyme is fully reduced.

Product generated from
the reaction of dansyl-YVG with the equimolar
fully reduced enzyme chased with a 5-fold excess of unlabeled Ac-YVG.
Blue circles represent the product generated with equimolar dansyl-YVG
and the fully reduced enzyme; purple diamonds represent the product
generated when the same experiment is conducted in the presence of
a 5-fold molar excess of Ac-YVG. The black dashed line represents
the expected dansyl product after chasing with Ac-YVG if multiple
turnovers occur.

Ascorbate Reduction Kinetics
Indicate an Ascorbate Binding SiteAs a final probe on the
effects of exogenous ascorbate on PHM reactivity,
stopped-flow was employed to determine the rate of reduction as a
function of ascorbate. To accomplish this, azide was used as a reporter
for the oxidized state of the enzyme. Azide is useful in these experiments
as it binds only to the oxidized form of the enzyme and has a characteristic
390 nm UV–vis signal32 (Figure S1, Supporting Information). Additionally,
the formation of the PHM Cu(II)–azido complex is not rate-limiting,
being formed within the time of mixing (Figure S2, Supporting Information). Elimination of this signal by
ascorbate therefore serves as a proxy for the rate of reduction. Figure 5 shows the rate of
reduction as a function of ascorbate concentration. The reduction
obeys saturation kinetics with KD = 1120
μM and kobs = 146 s–1 and indicates that ascorbate binds to the enzyme to accomplish reduction.
This KD is expected to reflect the affinity
of ascorbate for the oxidized enzyme during the reduction
of the di-Cu(II) to di-Cu(I) states and should be similar to the apparent Km determined from the dependence of product
stoichiometry on ascorbate if the latter is due merely to multiple
turnovers. However, as shown in Figure 3d, the integrated peak signal for the product plotted
versus ascorbate concentration yields a KD of 117 μM, an order of magnitude lower than that of the reduction
event. This further supports a role for ascorbate in activation of
the fully reduced enzyme over and above a simple role as a reducing
agent.Figure 5Ascorbate dependence of the rate of PHM reduction determined by
stopped-flow using the Cu(II) PHM–azido adduct (λmax = 390 nm) as a reporter for the concentration of Cu(II)–enzyme
remaining at each time point. Data are fit to a pre-equilibrium binding
event with KD = 1.1 mM.

Ascorbate Reduction Kinetics
Indicate an Ascorbate Binding Site

As a final probe on the
effects of exogenous ascorbate on PHM reactivity,
stopped-flow was employed to determine the rate of reduction as a
function of ascorbate. To accomplish this, azide was used as a reporter
for the oxidized state of the enzyme. Azide is useful in these experiments
as it binds only to the oxidized form of the enzyme and has a characteristic
390 nm UV–vis signal32 (Figure S1, Supporting Information). Additionally,
the formation of the PHM Cu(II)–azido complex is not rate-limiting,
being formed within the time of mixing (Figure S2, Supporting Information). Elimination of this signal by
ascorbate therefore serves as a proxy for the rate of reduction. Figure 5 shows the rate of
reduction as a function of ascorbate concentration. The reduction
obeys saturation kinetics with KD = 1120
μM and kobs = 146 s–1 and indicates that ascorbate binds to the enzyme to accomplish reduction.
This KD is expected to reflect the affinity
of ascorbate for the oxidized enzyme during the reduction
of the di-Cu(II) to di-Cu(I) states and should be similar to the apparent Km determined from the dependence of product
stoichiometry on ascorbate if the latter is due merely to multiple
turnovers. However, as shown in Figure 3d, the integrated peak signal for the product plotted
versus ascorbate concentration yields a KD of 117 μM, an order of magnitude lower than that of the reduction
event. This further supports a role for ascorbate in activation of
the fully reduced enzyme over and above a simple role as a reducing
agent.

Ascorbate dependence of the rate of PHM reduction determined by
stopped-flow using the Cu(II) PHM–azido adduct (λmax = 390 nm) as a reporter for the concentration of Cu(II)–enzyme
remaining at each time point. Data are fit to a pre-equilibrium binding
event with KD = 1.1 mM.

Rapid Freeze-Quench EPR of the Pre-Steady-State ReactionThe above experiments demonstrate that product formation is substoichiometric
without exogenous ascorbate but do not definitively identify the redox
state of the enzyme during or after the pre-steady-state reaction.
One possible yet unlikely explanation for the apparent ascorbate dependency
is the oxidation of the enzyme without product formation, initiated
by substrate and oxygen binding. Here, excess ascorbate would rereduce
the enzyme to allow turnover and apparent ascorbate dependency. Such
reactivity does occur in other copper monooxygenase such as the LPMOs.42 To address this possibility, rapid freeze-quench
(RFQ) EPR was employed. Using this technique, the redox state of the
enzyme can be assessed independent of product formation under single-turnover
conditions. Figure 6 shows the results of shooting fully reduced ascorbate-free enzyme
against oxygenated buffer containing 1 mole equivalent of substrate
relative to the enzyme concentration. The results (Figure 6, bottom panel, and Table S2) show that in the absence of excess
ascorbate negligible amounts of the enzyme become EPR-detectable after
300 ms, a time point at which product formation has ceased. This result
is remarkable as consensus mechanisms predict that in the absence
of reductant, turnover should generate two atoms of EPR-detectable
Cu(II) for each product molecule produced. In the present experiment,
if we allow for 20% product production, we anticipate 60 μM
EPR-detectable Cu(II) compared with the experimentally determined
amount of <3 μM. While this result should be interpreted
with caution, it may suggest that an EPR-detectable entity such as
a di-Cu(II) spin-coupled species is formed during or after the product
is formed (vide infra). We note that a spin-coupled Cu(II)-superoxo
species is less consistent with the data as any cupric superoxide
formed is expected to have been converted into hydroxylated product
after 300 ms.Figure 6Rapid freeze-quench-derived EPR spectra of the products
of the
pre-steady-state PHM reaction. EPR spectra were generated by rapidly
mixing fully reduced ascorbate-free PHM (400 mM in Cu(I)) with the
oxygenated buffer containing equimolar dansyl-YVG and freezing in
liquid ethane (−120 °C). The top panel shows the actual
spectra, and the bottom panel shows the quantitation of the Cu(II)
signal by reference to a standard Cu(II)–EDTA sample: red,
13 ms; green, 25 ms; blue, 100 ms; orange, 300 ms; and purple, 250
μM Cu(II)–EDTA standard. To avoid errors from integration
of noise peaks, intensities were determined from peak-to-peak heights
at g⊥ relative to that of a 250 μM solution
of Cu(II)–EDTA taken through an identical cycle of mixing,
freezing, and packing. Errors in concentrations determined from the
peak heights are estimated from triplicate measurements of a Cu(II)–EDTA
standard to be ≅ 15% and are mainly due to differences in the
packing density of the samples. EPR conditions: frequency 9.63 GHz, T = 100 K, microwave power 20 mW, gain 10 dB, modulation
amplitude 10 G, and sweep time 84 s.

Rapid Freeze-Quench EPR of the Pre-Steady-State Reaction

The above experiments demonstrate that product formation is substoichiometric
without exogenous ascorbate but do not definitively identify the redox
state of the enzyme during or after the pre-steady-state reaction.
One possible yet unlikely explanation for the apparent ascorbate dependency
is the oxidation of the enzyme without product formation, initiated
by substrate and oxygen binding. Here, excess ascorbate would rereduce
the enzyme to allow turnover and apparent ascorbate dependency. Such
reactivity does occur in other copper monooxygenase such as the LPMOs.42 To address this possibility, rapid freeze-quench
(RFQ) EPR was employed. Using this technique, the redox state of the
enzyme can be assessed independent of product formation under single-turnover
conditions. Figure 6 shows the results of shooting fully reduced ascorbate-free enzyme
against oxygenated buffer containing 1 mole equivalent of substrate
relative to the enzyme concentration. The results (Figure 6, bottom panel, and Table S2) show that in the absence of excess
ascorbate negligible amounts of the enzyme become EPR-detectable after
300 ms, a time point at which product formation has ceased. This result
is remarkable as consensus mechanisms predict that in the absence
of reductant, turnover should generate two atoms of EPR-detectable
Cu(II) for each product molecule produced. In the present experiment,
if we allow for 20% product production, we anticipate 60 μM
EPR-detectable Cu(II) compared with the experimentally determined
amount of <3 μM. While this result should be interpreted
with caution, it may suggest that an EPR-detectable entity such as
a di-Cu(II) spin-coupled species is formed during or after the product
is formed (vide infra). We note that a spin-coupled Cu(II)-superoxo
species is less consistent with the data as any cupric superoxide
formed is expected to have been converted into hydroxylated product
after 300 ms.

Rapid freeze-quench-derived EPR spectra of the products
of the
pre-steady-state PHM reaction. EPR spectra were generated by rapidly
mixing fully reduced ascorbate-free PHM (400 mM in Cu(I)) with the
oxygenated buffer containing equimolar dansyl-YVG and freezing in
liquid ethane (−120 °C). The top panel shows the actual
spectra, and the bottom panel shows the quantitation of the Cu(II)
signal by reference to a standard Cu(II)–EDTA sample: red,
13 ms; green, 25 ms; blue, 100 ms; orange, 300 ms; and purple, 250
μM Cu(II)–EDTA standard. To avoid errors from integration
of noise peaks, intensities were determined from peak-to-peak heights
at g⊥ relative to that of a 250 μM solution
of Cu(II)–EDTA taken through an identical cycle of mixing,
freezing, and packing. Errors in concentrations determined from the
peak heights are estimated from triplicate measurements of a Cu(II)–EDTA
standard to be ≅ 15% and are mainly due to differences in the
packing density of the samples. EPR conditions: frequency 9.63 GHz, T = 100 K, microwave power 20 mW, gain 10 dB, modulation
amplitude 10 G, and sweep time 84 s.

Cu(II)–Azido-Detected Stopped-Flow Spectrometry of the
Pre-Steady-State ReactionAs discussed above, the oxidized
di-Cu(II) PHM reacts with azide to generate an azide-to-copper charge-transfer
complex with λmax = 390 nm. The formation of the
azido complex is contingent on the oxidation of the enzyme and therefore
serves as a reporter for the pre-steady state enzyme turnover from
Cu(I) to Cu(II) states. Further, its rate of formation is extremely
rapid, the complex being fully formed in the time of mixing in the
stopped-flow (Figure S2, Supporting Information).
The reaction of fully reduced, ascorbate-free enzyme with Ac-YVG and
oxygen was followed using stopped-flow spectrometry monitoring the
rate of appearance of the 390 nm absorption. Here, the fully reduced
enzyme loaded with Ac-YVG was shot against buffer saturated with oxygen
and containing 20 mM sodium azide. Figure 7 shows that in the absence of excess ascorbate
less than 30% of the fully oxidized Cu(II)–azido complex is
formed based on the expected absorbance for full copper oxidation
(red line, Figure 7). Although this amount is larger than that based on product formation
or RFQ EPR, it reinforces our findings that PHM has limited reactivity
with substrate in the absence of excess ascorbate but may also suggest
that product stoichiometry is influenced by other factors such as
azide concentration.Figure 7Pre-steady-state reactivity of the fully reduced ascorbate-free
PHM determined by measuring the rate of appearance of the 390 nm Cu(II)–azido
adduct at 390 nm using stopped-flow spectrophotometry. 500 μM
of WT PHM was shot against 5 mM Ac-YVG + 20 mM sodium azide. Data
were collected for 500 ms, with readings taken every 5 ms. The top
trace shows the generation of the 390 nm absorbance as a function
of time in seconds. The bottom trace shows the rate curve fitted to
a single exponential with kobs= 72 s–1. The red dashed line represents the absorbance expected
for complete reaction where all copper is oxidized to Cu(II) determined
from a separate experiment where fully oxidized Cu(II) PHM was shot
against buffer containing sodium azide under identical conditions.

Cu(II)–Azido-Detected Stopped-Flow Spectrometry of the
Pre-Steady-State Reaction

As discussed above, the oxidized
di-Cu(II) PHM reacts with azide to generate an azide-to-copper charge-transfer
complex with λmax = 390 nm. The formation of the
azido complex is contingent on the oxidation of the enzyme and therefore
serves as a reporter for the pre-steady state enzyme turnover from
Cu(I) to Cu(II) states. Further, its rate of formation is extremely
rapid, the complex being fully formed in the time of mixing in the
stopped-flow (Figure S2, Supporting Information).
The reaction of fully reduced, ascorbate-free enzyme with Ac-YVG and
oxygen was followed using stopped-flow spectrometry monitoring the
rate of appearance of the 390 nm absorption. Here, the fully reduced
enzyme loaded with Ac-YVG was shot against buffer saturated with oxygen
and containing 20 mM sodium azide. Figure 7 shows that in the absence of excess ascorbate
less than 30% of the fully oxidized Cu(II)–azido complex is
formed based on the expected absorbance for full copper oxidation
(red line, Figure 7). Although this amount is larger than that based on product formation
or RFQ EPR, it reinforces our findings that PHM has limited reactivity
with substrate in the absence of excess ascorbate but may also suggest
that product stoichiometry is influenced by other factors such as
azide concentration.

Pre-steady-state reactivity of the fully reduced ascorbate-free
PHM determined by measuring the rate of appearance of the 390 nm Cu(II)–azido
adduct at 390 nm using stopped-flow spectrophotometry. 500 μM
of WT PHM was shot against 5 mM Ac-YVG + 20 mM sodium azide. Data
were collected for 500 ms, with readings taken every 5 ms. The top
trace shows the generation of the 390 nm absorbance as a function
of time in seconds. The bottom trace shows the rate curve fitted to
a single exponential with kobs= 72 s–1. The red dashed line represents the absorbance expected
for complete reaction where all copper is oxidized to Cu(II) determined
from a separate experiment where fully oxidized Cu(II) PHM was shot
against buffer containing sodium azide under identical conditions.

FTIR Indicates a Shift in Enzyme Population
in Response to Ascorbate
TitrationTo determine the specific effect of ascorbate on
PHM reactivity, we looked into insights previously garnered from FTIR.
Carbon monoxide (CO) FTIR has been used extensively to probe metal
oxygen-binding active sites, where it can give remarkably detailed
information about structure and oxygen binding.18,37,43−45 Subtle changes to the
binding site are reflected in shifts in the IR stretching frequency
of the carbon monoxide. Furthermore, carbon monoxide has been shown
to bind solely to the M-site in PHM and therefore it functions as
a probe of M-site configuration. It has been shown previously that
varied concentrations of substrates tune two separate populations
of stretching frequencies.18 In the absence
of substrate, the spectra are dominated by a large 2093 cm–1 stretching frequency with a minor 2063 cm–1 shoulder.
Addition of substrate causes an increase in the 2063 cm–1 species with an attenuation of 2093 cm–1 in a
concentration-dependent manner. Increasing substrate concentration
was, however, not sufficient to eliminate the 2093 cm–1 stretch nor did a complete absence of substrate abrogate the 2063
cm–1 signal. These data suggested a pair of interconvertible
states mediated by substrate binding, but the significance of these
states to reactivity as well as their identity was unclear.Peptidyl substrate’s ability to partition the enzyme population
echoes ascorbate’s ability to induce a catalytically competent
state, particularly since the CO binding experiments are performed
in the presence of exogenous ascorbate. To probe that relationship
further, we performed FTIR experiments to test if ascorbate induces
an alternative enzyme configuration in a manner resembling substrate.
WT PHM was exposed to carbon monoxide in the presence of substrate
and varied concentrations of ascorbate. Figure 8 (top) shows that increasing the ascorbate
concentration causes the ratio of 2063 to 2093 cm–1 peaks to increase in a concentration-dependent manner, reaching
a maximum at 15 mM exogenous ascorbate, while Figure 8 (bottom) depicts two representative spectra
at 1 and 15 mM exogenous ascorbate, respectively. Curiously, excess
ascorbate beyond 15 mM resulted in a decrease in the ratio of the
2063 to 2093 cm–1 signals, but these concentrations
are well above the ascorbate concentration (2 mM) required for maximum
product stoichiometry. This experiment demonstrates that ascorbate,
in the presence of substrate and CO, is sufficient to induce an alternate
conformer of PHM.Figure 8Evolution of the substrate-induced 2063 cm–1 M-site
Cu(I)–CO signal as a function of the added ascorbate. The top
trace shows the ratio of 2063 to 2093 cm–1 peaks
plotted against ascorbate concentration. The bottom trace shows the
representative spectra at 1 mM (orange) and 15 mM (green).This explicit connection between the 2063 cm–1 signal and ascorbate concentration provides insights into the role
of ascorbate in catalysis. Exogenous ascorbate is here shown to both
bias the population of enzyme to the 2063 cm–1 form
and increase overall reactivity. This suggests that substrate and
ascorbate work combinatorially to induce the 2063 cm–1 form of the enzyme and presents strong evidence that this form is
the active species.The identity of this active state remains
unclear. Changes to the
hydrogen-bonding network surrounding the CuM site have been proposed
as one mechanism of substrate activation.13,18 Another possible interpretation is a large-scale conformational
change. As discussed in more detail below, ascorbate binding could
potentially induce an open-to-closed conformational transition where
the Cu site separation decreases from 11 to 4–5 Å, similar
to that observed in DBM. Once in proximity, electron transfer is facile
between the two sites. There are several pieces of evidence that support
this interpretation. First is the scale of the change: a 30 cm–1 change suggests a relatively large increase in the
electron-donating power of the CuM site, larger than changes caused
by mutating nearby ligands to perturb the hydrogen-bonding network.
Second, the red-shifted IR frequency is reminiscent of that observed
in the dinuclear arthropodal hemocyanin (2043 cm–1) and attributed to a semibridging species.44 In the context of PHM, such a bridging species is not possible unless
the copper sites were in much closer proximity, as in the case of
hemocyanin, which has an intracopper distance of 4 Å. Third is
the observation of a closed conformer in crystal structures of the
PHM H108A mutant in the presence of citrate,13 indicating that under certain conditions the interdomain closure
is favored. Indeed, the ability for citrate to help induce a closed
conformer is evocative, as both citrate and ascorbate are hydroxy
acids. The binding of the citrate is interesting as well since it
binds in a manner to bridge the CuM and CuH sites and may provide
stability for the closed conformer (Figure 9).Figure 9Closed and open structures of DBM and PHM. Left,
two conformers
of DBM with Cu atom (bronze) locations modeled (PDB file 4EZL). Middle:
alignment of the protein fold of the closed crystal structure of H108A
in the presence of citrate (PDB file 6ALA), with the open structure
of the oxidized PHM (PDB file 1PHM). PHM structures were aligned on
the N-terminal (H) subdomain with an RMSD value of 0.41. The H108A-citrate
and the oxidized native PHM main chains are depicted by pink and purple
ribbons, respectively. To emphasize the hinge motion, the first β-strands
of the M-domains are colored green (closed: H108A-citrate) and dark
blue (open: oxidized PHM). Right: enlarged view of the metal binding
site in the H108A-citrate PHM structure. The single copper atom of
the closed H108A-citrate structure, which coordinates M-site residues
H242 and H244 and H-site residue H107, is shown in pink, while the
citrate molecule is shown as a green carbon backbone with oxygen atoms
in red. The positions of the two copper sites in the open PHM structure
are shown as transparent slate spheres.

FTIR Indicates a Shift in Enzyme Population
in Response to Ascorbate
Titration

To determine the specific effect of ascorbate on
PHM reactivity, we looked into insights previously garnered from FTIR.
Carbon monoxide (CO) FTIR has been used extensively to probe metal
oxygen-binding active sites, where it can give remarkably detailed
information about structure and oxygen binding.18,37,43−45 Subtle changes to the
binding site are reflected in shifts in the IR stretching frequency
of the carbon monoxide. Furthermore, carbon monoxide has been shown
to bind solely to the M-site in PHM and therefore it functions as
a probe of M-site configuration. It has been shown previously that
varied concentrations of substrates tune two separate populations
of stretching frequencies.18 In the absence
of substrate, the spectra are dominated by a large 2093 cm–1 stretching frequency with a minor 2063 cm–1 shoulder.
Addition of substrate causes an increase in the 2063 cm–1 species with an attenuation of 2093 cm–1 in a
concentration-dependent manner. Increasing substrate concentration
was, however, not sufficient to eliminate the 2093 cm–1 stretch nor did a complete absence of substrate abrogate the 2063
cm–1 signal. These data suggested a pair of interconvertible
states mediated by substrate binding, but the significance of these
states to reactivity as well as their identity was unclear.

Peptidyl substrate’s ability to partition the enzyme population
echoes ascorbate’s ability to induce a catalytically competent
state, particularly since the CO binding experiments are performed
in the presence of exogenous ascorbate. To probe that relationship
further, we performed FTIR experiments to test if ascorbate induces
an alternative enzyme configuration in a manner resembling substrate.
WT PHM was exposed to carbon monoxide in the presence of substrate
and varied concentrations of ascorbate. Figure 8 (top) shows that increasing the ascorbate
concentration causes the ratio of 2063 to 2093 cm–1 peaks to increase in a concentration-dependent manner, reaching
a maximum at 15 mM exogenous ascorbate, while Figure 8 (bottom) depicts two representative spectra
at 1 and 15 mM exogenous ascorbate, respectively. Curiously, excess
ascorbate beyond 15 mM resulted in a decrease in the ratio of the
2063 to 2093 cm–1 signals, but these concentrations
are well above the ascorbate concentration (2 mM) required for maximum
product stoichiometry. This experiment demonstrates that ascorbate,
in the presence of substrate and CO, is sufficient to induce an alternate
conformer of PHM.

Evolution of the substrate-induced 2063 cm–1 M-site
Cu(I)–CO signal as a function of the added ascorbate. The top
trace shows the ratio of 2063 to 2093 cm–1 peaks
plotted against ascorbate concentration. The bottom trace shows the
representative spectra at 1 mM (orange) and 15 mM (green).

This explicit connection between the 2063 cm–1 signal and ascorbate concentration provides insights into the role
of ascorbate in catalysis. Exogenous ascorbate is here shown to both
bias the population of enzyme to the 2063 cm–1 form
and increase overall reactivity. This suggests that substrate and
ascorbate work combinatorially to induce the 2063 cm–1 form of the enzyme and presents strong evidence that this form is
the active species.

The identity of this active state remains
unclear. Changes to the
hydrogen-bonding network surrounding the CuM site have been proposed
as one mechanism of substrate activation.13,18 Another possible interpretation is a large-scale conformational
change. As discussed in more detail below, ascorbate binding could
potentially induce an open-to-closed conformational transition where
the Cu site separation decreases from 11 to 4–5 Å, similar
to that observed in DBM. Once in proximity, electron transfer is facile
between the two sites. There are several pieces of evidence that support
this interpretation. First is the scale of the change: a 30 cm–1 change suggests a relatively large increase in the
electron-donating power of the CuM site, larger than changes caused
by mutating nearby ligands to perturb the hydrogen-bonding network.
Second, the red-shifted IR frequency is reminiscent of that observed
in the dinuclear arthropodal hemocyanin (2043 cm–1) and attributed to a semibridging species.44 In the context of PHM, such a bridging species is not possible unless
the copper sites were in much closer proximity, as in the case of
hemocyanin, which has an intracopper distance of 4 Å. Third is
the observation of a closed conformer in crystal structures of the
PHM H108A mutant in the presence of citrate,13 indicating that under certain conditions the interdomain closure
is favored. Indeed, the ability for citrate to help induce a closed
conformer is evocative, as both citrate and ascorbate are hydroxy
acids. The binding of the citrate is interesting as well since it
binds in a manner to bridge the CuM and CuH sites and may provide
stability for the closed conformer (Figure 9).

Closed and open structures of DBM and PHM. Left,
two conformers
of DBM with Cu atom (bronze) locations modeled (PDB file 4EZL). Middle:
alignment of the protein fold of the closed crystal structure of H108A
in the presence of citrate (PDB file 6ALA), with the open structure
of the oxidized PHM (PDB file 1PHM). PHM structures were aligned on
the N-terminal (H) subdomain with an RMSD value of 0.41. The H108A-citrate
and the oxidized native PHM main chains are depicted by pink and purple
ribbons, respectively. To emphasize the hinge motion, the first β-strands
of the M-domains are colored green (closed: H108A-citrate) and dark
blue (open: oxidized PHM). Right: enlarged view of the metal binding
site in the H108A-citrate PHM structure. The single copper atom of
the closed H108A-citrate structure, which coordinates M-site residues
H242 and H244 and H-site residue H107, is shown in pink, while the
citrate molecule is shown as a green carbon backbone with oxygen atoms
in red. The positions of the two copper sites in the open PHM structure
are shown as transparent slate spheres.

DiscussionIn the present paper, we explored pre-steady-state
kinetics of
PHM as a function of ascorbate concentration using chemical quench
and rapid freeze-quench studies of the fully reduced ascorbate-free
enzyme. When ascorbate-free di-Cu(I) PHM was shot against oxygenated
buffer containing the dansyl-YVG substrate and the dansylated product
quantified by HPLC using fluorescence detection, <15% of the expected
product was formed. This result was recapitulated to varying degrees
by RFQ EPR spectroscopy and stopped-flow monitoring of the oxidized
Cu(II) species using the 390 nm Cu(II)–azido absorption as
readout. Product generation could be increased to a maximum of 70%
by the addition of increasing amounts of the ascorbate cosubstrate.
This phenomenon could be the result of multiple turnovers of a small
fraction of the “active” enzyme, with the remainder
of the sample in a nonfunctional state (case 1) or activation of the
fully reduced enzyme by excess ascorbate (case 2). To test which of
these scenarios was most probable, we performed a pulse-chase experiment,
wherein the enzyme was incubated with 1 equiv of dansyl-YVG in oxygenated
buffer and then shot against a 5-fold excess of Ac-YVG. Case 1 is
predicted to generate no more than 20% of the fluorescently labeled
product since after the first turnover the unlabeled Ac-YVG would
efficiently outcompete the labeled substrate, whereas case 2 is predicted
to generate higher amounts of the labeled product. The latter scenario
was observed with 50–60% of the dansylated product formed.
This allowed us to exclude multiple turnovers as an interpretation
of the data and to conclude that the cosubstrate ascorbate not only
reduces the Cu(II) centers but also induces a conformational change
to an active state. Further, the observed rate in the presence of
ascorbate is 4-fold slower than that of the ascorbate-free conformer,
indicating that ascorbate induces some process such as a conformational
change, which becomes rate-limiting. In previous work, we had documented
that the peptidyl substrate induces a downshift in the ν(CO)
of the carbon monoxide complex at the M-center from 2093 to 2063 cm–1 and had suggested that this species was a substrate-triggered
active conformer. Here, we found that this conformer was also induced
by increasing concentrations of ascorbate, confirming that ascorbate
does more than act as a reducing agent. The totality of the data led
to the hypothesis that ascorbate and peptidyl substrate work together
to induce a transition from the inactive to an active conformation.The present study demonstrates a correlation between the intensity
of the 2063 cm–1 CuM–CO band and the amount
of product formed during pre-steady-state turnover (since both increase
with increasing ascorbate), and one may cautiously infer that the
species responsible for the 2063 cm–1 band is derived
from a catalytically active form, where the CO ligand is bound in
the same configuration as the reactive oxygen molecule. Assigning
the chemical nature of the CO-bound species is therefore important
for understanding the catalytic mechanism and the mechanism of substrate
activation. The 30 cm–1 red shift of the CO frequency
implies a considerable degree of additional back-bonding into ligand
π* orbitals, which, in turn, weakens the CO bond via decreasing
the bond order, and similar electronic perturbation applied to a Cu(I)–O2 complex would result in significant activation of the metal–dioxygen
complex.Downshifted frequencies in Cu(I)–CO complexes
can potentially
arise from a number of sources such as (i) increase in the N-donor
(imidazole) coordination number, (ii) strong H-bonding between the
distal NH of the imidazole of a His ligand and a protein-derived negative
charge or dipole, (iii) interaction of the distal O of CO with a positive
charge or dipole, and (iv) a bridging or semibridging structure where
the CO can interact with a second metal ion. In model compounds, the
substitution of heterocyclic rings (imidazole, pyrazole) by electron-releasing
groups leads to decreased frequencies, while substitution by electron-withdrawing
groups leads to increased frequencies. Table S3 of the Supporting Information lists examples that validate these
trends. In a previous study,18 we showed
that the substitution of Met314 by His resulted in an 18 cm–1 downshift (2075 cm–1) in the absence of substrate,
but the addition of substrate led to a further downshift of 24 cm–1 (2051 cm–1). Therefore, simple
substitution of the weakly binding Met by His does not account for
the observed effect. Similarly, changes in the H-bonding networks,
which anchor the M-site coordinating His ligands, are unlikely to
generate sufficient perturbation. For example, in recent unpublished
work, we perturbed the H242–Q272 H bond via a Q272E mutation
and detected a mere 4 cm–1 downshift. A substrate-induced
interaction of the CO distal O with a positive dipole as documented
by Spiro and co-workers46 for heme carbonyls
could lead to significant downshifted frequencies, yet it is hard
to see how binding of either Ac-YVG or ascorbate could increase the
positive charge in the vicinity of the bound CO since both carry a
negative charge at the pH of our experiments. Indeed, the salt-bridge
formed between the peptide carboxylate and R240 would remove a source
of positive charge (protonated guanidinium) that could potentially
interact with a coordinated CO. We are therefore left with the final
possibility that a conformational change leads to the formation of
a semibridging CO complex where the CO can interact with a second
metal ion. It is worth noting that for the M314H Cu(I)–CO the
measured ν(CO) of 2051 cm–1 is one of the
lowest frequencies ever observed for a Cu(His)3 ligand
set (Table S3) and is in the range observed
for the CO complexes of hemocyanins 2063–2043 cm–1, which have the same His3 coordination44 and where the interaction of the O atom of the CO ligand
with the positive charge of the second Cu(I) was deemed the most likely
origin of the lowered frequency.Red-shifted IR frequencies
are also observed in the CuB–CO
complexes of cytochrome-c-oxidase47,48 between 2060
and 2036 cm–1. Like the hemocyanin case, Cu(I)–CO
frequencies at or below 2050 cm–1 appear to correlate
with the presence of a second positively charged metal ion in the
vicinity of the CO binding site. In the case of cytochrome oxidase,
the Cu(I)–CO species are formed by the photodissociation of
the CO ligand from the ferrous heme-a3 with subsequent
rebinding to copper. Recent X-ray crystallographic studies49 have captured an X-ray photodissociated state
in the crystal where the CO binds side-on to Cu(I) with Cu–C
and Cu–O distances of 2.5 and 2.3 Å, respectively. While
it is not known whether this state is representative of the photodissociated
CuB–CO complexes detected by IR, it is indicative of a plurality
of CO binding motifs that are accessible in dinuclear metal environments.
In the substrate-induced PHM CuM–CO complex, we may cautiously
infer that the interaction of the O atom with the positive charge
of a second copper atom is an attractive interpretation of the red
shift, which would imply that activation is the result of an open-to-closed
conformational transition.Our data provide experimental context
for recent computational
work, which has documented the energetic feasibility of the open-to-closed
conformational transition in catalysis. The new theoretical computations
that consider the closed conformer as the active species have suggested
a binuclear intermediate, which significantly lowers the calculated
activation energy for HAA from the peptide substrate.30 The new modeling shows that hydrogen atom abstraction by
a cupric superoxo from the reductant ascorbate is always more favorable
than from the peptide Cα–H, implying that in the presence
of excess ascorbate, any cupric superoxo that forms will always rapidly
form a cupric hydroperoxo and the ascorbyl radical. The study suggests
that the Cu(II)M hydroperoxo species (formed only in the presence
of ascorbate) drives the conversion of open-to-closed conformer which
are separated by only 2 kcal mol–1, and forms a
μ-oxo-, μ-hydroxo-mixed-valence Cu(I)–Cu(II) entity
as the active intermediate. While experimental data are yet to support
such an intermediate, there is clearly sufficient evidence to warrant
re-examination of the canonical mechanism. The extremely small amount
of EPR-detectable Cu(II) observed in our RFQ pre-steady-state experiment
offers the intriguing possibility that the closed conformer may persist
after product formation as a starting point for subsequent catalytic
cycles.An open-to-closed mechanism is also more consistent
with a number
of experimental observations that do not fit well with the consensus
mechanism. The latter posits that the initial attack of O2 on the reduced CuM center generates an electrophilic Cu(II)–superoxo
intermediate capable of H atom abstraction from the peptidylglycine
Cα to form a substrate radical. This mechanism fails
to explain (i) how substrate binding converts the redox-inactive Cu(I)M-site
to an active state apparently associated with the 2063 cm–1 Cu(I)M–CO species; (ii) the requirement for methionine at
position 314 and the inactivity of the M314H variant18 since Cu(I)His3 systems are known to be reactive
toward O2 in inorganic model compounds;50 and (iii) how substrate hydroxylation can be driven by
hydrogen peroxide from the oxidized Cu(II)–enzyme (peroxide
shunt) in a slow catalytic reaction in which the 18O of
labeled peroxide is scrambled 60–70% with ambient 16O2 in air.35 However, an open-to-closed
mechanism rationalizes all of these observations and can be discussed
in the context of the H108A-citrate structure shown in Figure 9. We suggest that the peptide
and ascorbate together induce the open-to-closed transition in a rate-limiting
step to form the active “binuclear” state. This is reactive
to both oxygen in the Cu(I) state and hydrogen peroxide in the Cu(II)
state. An interesting feature of the H108-citrate structure is the
lengthening of the CuM to M314 Cu–S bond from 2.3 to 5.6 Å.
This movement requires a labile fluxional Cu–S interaction
that is provided by the Cu(I)–thioether bond as recently demonstrated
in our studies of the SMet and SeMet complexes of a PHM CusF model32,33 but conversely would be frustrated in the M314H variant due to the
stronger sigma donor interaction of the Cu(I)–N(H314) bond.The peroxide shunt chemistry is also easily interpreted within
the binuclear formalism. From work published nearly two decades ago,35 we showed that the reaction was catalytic (at
least 40 turnovers) and was 100% coupled to peroxide consumption.
The conclusion that a mononuclear Cu(II)–peroxo species must
be a reactive intermediate was negated by the observation that 60%
of the 18O label was scrambled with ambient 16O2 in air when H218O2 was used as the source of peroxide, implying the formation
of an intermediate that was in equilibrium with ambient dioxygen at
some stage of the catalytic mechanism. While difficult to
rationalize within the context of the “open” PHM conformation,
a binuclear intermediate formed within the closed conformer can readily
explain the data due to the well-known equilibrium between di-Cu(II)–peroxo
and di-Cu(I)–O2 species.Our studies show
that ascorbate appears to play a dual role in
PHM catalysis both as a reductant and as an activator but the stoichiometry
and binding site(s) remain obscure. While ascorbate is technically
able to supply two electrons, earlier work on DBM established that
it functioned as a 1-electron reductant generating the semidehydroascorbate
radical as the primary product, which subsequently disproportionates
to ascorbate and dehydroascorbate.51 Thus,
we anticipate that during the reductive cycle, ascorbate must bind
and its oxidized radical must dissociate twice to generate the fully
reduced dicopper(I) enzyme. This process appears to have a KD of around 1.1 mM. On the other hand, the activation
of catalysis occurs with a lower KD (117
μM), suggesting a different mode/site of ascorbate interaction
with the enzyme. Yet, another study reported inactivation by ascorbate with a KD of 5 mM,52 whereas crystallography failed to identify any
ascorbate binding site.4 Interestingly,
the ascorbate dependence of the 2063 cm–1 CO conformer
shows a maximum between 10 and 15 mM ascorbate suggestive of both
activation and subsequent inactivation as ascorbate concentrations
increase. The data suggest complex activation/inactivation behavior
for the ascorbate cosubstrate over a concentration range of 0.1–10
mM, and it is possible that this may relate to the physiological regulation
of hormone production in the vesicle where the ascorbate levels also
span this concentration range.53Notwithstanding the growing evidence for the open-to-closed mechanism,
an alternative mechanism for substrate triggering in the open-only
formalism may be suggested based on the calculations of Cowley and
Solomon, who concluded that the formation of the Cu(II)–superoxide
at the M-site was energetically favorable only if the dioxygen displaced
a hydroxide ligand from the Cu(I) center.29 While Cu(I)–OH complexes are rare, crystal structures of
the reduced enzyme do show water/hydroxide density near the CuM center,
which appears to be a coordinated solvent species. It is therefore
possible that substrate and/or ascorbate binding to the Cu(I) state
deprotonates the coordinated water to a hydroxide, which destabilizes
the Cu(I)M-site and induces reaction with oxygen. However, the fact
that the catalytic rate is inversely proportional to pH in the 4–8
pH region is inconsistent with M-site water deprotonation as the driver
of substrate triggering. We continue to search for conditions where
closed conformers can be formed and studied as potential intermediates.

In the present paper, we explored pre-steady-state
kinetics of
PHM as a function of ascorbate concentration using chemical quench
and rapid freeze-quench studies of the fully reduced ascorbate-free
enzyme. When ascorbate-free di-Cu(I) PHM was shot against oxygenated
buffer containing the dansyl-YVG substrate and the dansylated product
quantified by HPLC using fluorescence detection, <15% of the expected
product was formed. This result was recapitulated to varying degrees
by RFQ EPR spectroscopy and stopped-flow monitoring of the oxidized
Cu(II) species using the 390 nm Cu(II)–azido absorption as
readout. Product generation could be increased to a maximum of 70%
by the addition of increasing amounts of the ascorbate cosubstrate.
This phenomenon could be the result of multiple turnovers of a small
fraction of the “active” enzyme, with the remainder
of the sample in a nonfunctional state (case 1) or activation of the
fully reduced enzyme by excess ascorbate (case 2). To test which of
these scenarios was most probable, we performed a pulse-chase experiment,
wherein the enzyme was incubated with 1 equiv of dansyl-YVG in oxygenated
buffer and then shot against a 5-fold excess of Ac-YVG. Case 1 is
predicted to generate no more than 20% of the fluorescently labeled
product since after the first turnover the unlabeled Ac-YVG would
efficiently outcompete the labeled substrate, whereas case 2 is predicted
to generate higher amounts of the labeled product. The latter scenario
was observed with 50–60% of the dansylated product formed.
This allowed us to exclude multiple turnovers as an interpretation
of the data and to conclude that the cosubstrate ascorbate not only
reduces the Cu(II) centers but also induces a conformational change
to an active state. Further, the observed rate in the presence of
ascorbate is 4-fold slower than that of the ascorbate-free conformer,
indicating that ascorbate induces some process such as a conformational
change, which becomes rate-limiting. In previous work, we had documented
that the peptidyl substrate induces a downshift in the ν(CO)
of the carbon monoxide complex at the M-center from 2093 to 2063 cm–1 and had suggested that this species was a substrate-triggered
active conformer. Here, we found that this conformer was also induced
by increasing concentrations of ascorbate, confirming that ascorbate
does more than act as a reducing agent. The totality of the data led
to the hypothesis that ascorbate and peptidyl substrate work together
to induce a transition from the inactive to an active conformation.

The present study demonstrates a correlation between the intensity
of the 2063 cm–1 CuM–CO band and the amount
of product formed during pre-steady-state turnover (since both increase
with increasing ascorbate), and one may cautiously infer that the
species responsible for the 2063 cm–1 band is derived
from a catalytically active form, where the CO ligand is bound in
the same configuration as the reactive oxygen molecule. Assigning
the chemical nature of the CO-bound species is therefore important
for understanding the catalytic mechanism and the mechanism of substrate
activation. The 30 cm–1 red shift of the CO frequency
implies a considerable degree of additional back-bonding into ligand
π* orbitals, which, in turn, weakens the CO bond via decreasing
the bond order, and similar electronic perturbation applied to a Cu(I)–O2 complex would result in significant activation of the metal–dioxygen
complex.

Downshifted frequencies in Cu(I)–CO complexes
can potentially
arise from a number of sources such as (i) increase in the N-donor
(imidazole) coordination number, (ii) strong H-bonding between the
distal NH of the imidazole of a His ligand and a protein-derived negative
charge or dipole, (iii) interaction of the distal O of CO with a positive
charge or dipole, and (iv) a bridging or semibridging structure where
the CO can interact with a second metal ion. In model compounds, the
substitution of heterocyclic rings (imidazole, pyrazole) by electron-releasing
groups leads to decreased frequencies, while substitution by electron-withdrawing
groups leads to increased frequencies. Table S3 of the Supporting Information lists examples that validate these
trends. In a previous study,18 we showed
that the substitution of Met314 by His resulted in an 18 cm–1 downshift (2075 cm–1) in the absence of substrate,
but the addition of substrate led to a further downshift of 24 cm–1 (2051 cm–1). Therefore, simple
substitution of the weakly binding Met by His does not account for
the observed effect. Similarly, changes in the H-bonding networks,
which anchor the M-site coordinating His ligands, are unlikely to
generate sufficient perturbation. For example, in recent unpublished
work, we perturbed the H242–Q272 H bond via a Q272E mutation
and detected a mere 4 cm–1 downshift. A substrate-induced
interaction of the CO distal O with a positive dipole as documented
by Spiro and co-workers46 for heme carbonyls
could lead to significant downshifted frequencies, yet it is hard
to see how binding of either Ac-YVG or ascorbate could increase the
positive charge in the vicinity of the bound CO since both carry a
negative charge at the pH of our experiments. Indeed, the salt-bridge
formed between the peptide carboxylate and R240 would remove a source
of positive charge (protonated guanidinium) that could potentially
interact with a coordinated CO. We are therefore left with the final
possibility that a conformational change leads to the formation of
a semibridging CO complex where the CO can interact with a second
metal ion. It is worth noting that for the M314H Cu(I)–CO the
measured ν(CO) of 2051 cm–1 is one of the
lowest frequencies ever observed for a Cu(His)3 ligand
set (Table S3) and is in the range observed
for the CO complexes of hemocyanins 2063–2043 cm–1, which have the same His3 coordination44 and where the interaction of the O atom of the CO ligand
with the positive charge of the second Cu(I) was deemed the most likely
origin of the lowered frequency.

Red-shifted IR frequencies
are also observed in the CuB–CO
complexes of cytochrome-c-oxidase47,48 between 2060
and 2036 cm–1. Like the hemocyanin case, Cu(I)–CO
frequencies at or below 2050 cm–1 appear to correlate
with the presence of a second positively charged metal ion in the
vicinity of the CO binding site. In the case of cytochrome oxidase,
the Cu(I)–CO species are formed by the photodissociation of
the CO ligand from the ferrous heme-a3 with subsequent
rebinding to copper. Recent X-ray crystallographic studies49 have captured an X-ray photodissociated state
in the crystal where the CO binds side-on to Cu(I) with Cu–C
and Cu–O distances of 2.5 and 2.3 Å, respectively. While
it is not known whether this state is representative of the photodissociated
CuB–CO complexes detected by IR, it is indicative of a plurality
of CO binding motifs that are accessible in dinuclear metal environments.
In the substrate-induced PHM CuM–CO complex, we may cautiously
infer that the interaction of the O atom with the positive charge
of a second copper atom is an attractive interpretation of the red
shift, which would imply that activation is the result of an open-to-closed
conformational transition.

Our data provide experimental context
for recent computational
work, which has documented the energetic feasibility of the open-to-closed
conformational transition in catalysis. The new theoretical computations
that consider the closed conformer as the active species have suggested
a binuclear intermediate, which significantly lowers the calculated
activation energy for HAA from the peptide substrate.30 The new modeling shows that hydrogen atom abstraction by
a cupric superoxo from the reductant ascorbate is always more favorable
than from the peptide Cα–H, implying that in the presence
of excess ascorbate, any cupric superoxo that forms will always rapidly
form a cupric hydroperoxo and the ascorbyl radical. The study suggests
that the Cu(II)M hydroperoxo species (formed only in the presence
of ascorbate) drives the conversion of open-to-closed conformer which
are separated by only 2 kcal mol–1, and forms a
μ-oxo-, μ-hydroxo-mixed-valence Cu(I)–Cu(II) entity
as the active intermediate. While experimental data are yet to support
such an intermediate, there is clearly sufficient evidence to warrant
re-examination of the canonical mechanism. The extremely small amount
of EPR-detectable Cu(II) observed in our RFQ pre-steady-state experiment
offers the intriguing possibility that the closed conformer may persist
after product formation as a starting point for subsequent catalytic
cycles.

An open-to-closed mechanism is also more consistent
with a number
of experimental observations that do not fit well with the consensus
mechanism. The latter posits that the initial attack of O2 on the reduced CuM center generates an electrophilic Cu(II)–superoxo
intermediate capable of H atom abstraction from the peptidylglycine
Cα to form a substrate radical. This mechanism fails
to explain (i) how substrate binding converts the redox-inactive Cu(I)M-site
to an active state apparently associated with the 2063 cm–1 Cu(I)M–CO species; (ii) the requirement for methionine at
position 314 and the inactivity of the M314H variant18 since Cu(I)His3 systems are known to be reactive
toward O2 in inorganic model compounds;50 and (iii) how substrate hydroxylation can be driven by
hydrogen peroxide from the oxidized Cu(II)–enzyme (peroxide
shunt) in a slow catalytic reaction in which the 18O of
labeled peroxide is scrambled 60–70% with ambient 16O2 in air.35 However, an open-to-closed
mechanism rationalizes all of these observations and can be discussed
in the context of the H108A-citrate structure shown in Figure 9. We suggest that the peptide
and ascorbate together induce the open-to-closed transition in a rate-limiting
step to form the active “binuclear” state. This is reactive
to both oxygen in the Cu(I) state and hydrogen peroxide in the Cu(II)
state. An interesting feature of the H108-citrate structure is the
lengthening of the CuM to M314 Cu–S bond from 2.3 to 5.6 Å.
This movement requires a labile fluxional Cu–S interaction
that is provided by the Cu(I)–thioether bond as recently demonstrated
in our studies of the SMet and SeMet complexes of a PHM CusF model32,33 but conversely would be frustrated in the M314H variant due to the
stronger sigma donor interaction of the Cu(I)–N(H314) bond.

The peroxide shunt chemistry is also easily interpreted within
the binuclear formalism. From work published nearly two decades ago,35 we showed that the reaction was catalytic (at
least 40 turnovers) and was 100% coupled to peroxide consumption.
The conclusion that a mononuclear Cu(II)–peroxo species must
be a reactive intermediate was negated by the observation that 60%
of the 18O label was scrambled with ambient 16O2 in air when H218O2 was used as the source of peroxide, implying the formation
of an intermediate that was in equilibrium with ambient dioxygen at
some stage of the catalytic mechanism. While difficult to
rationalize within the context of the “open” PHM conformation,
a binuclear intermediate formed within the closed conformer can readily
explain the data due to the well-known equilibrium between di-Cu(II)–peroxo
and di-Cu(I)–O2 species.

Our studies show
that ascorbate appears to play a dual role in
PHM catalysis both as a reductant and as an activator but the stoichiometry
and binding site(s) remain obscure. While ascorbate is technically
able to supply two electrons, earlier work on DBM established that
it functioned as a 1-electron reductant generating the semidehydroascorbate
radical as the primary product, which subsequently disproportionates
to ascorbate and dehydroascorbate.51 Thus,
we anticipate that during the reductive cycle, ascorbate must bind
and its oxidized radical must dissociate twice to generate the fully
reduced dicopper(I) enzyme. This process appears to have a KD of around 1.1 mM. On the other hand, the activation
of catalysis occurs with a lower KD (117
μM), suggesting a different mode/site of ascorbate interaction
with the enzyme. Yet, another study reported inactivation by ascorbate with a KD of 5 mM,52 whereas crystallography failed to identify any
ascorbate binding site.4 Interestingly,
the ascorbate dependence of the 2063 cm–1 CO conformer
shows a maximum between 10 and 15 mM ascorbate suggestive of both
activation and subsequent inactivation as ascorbate concentrations
increase. The data suggest complex activation/inactivation behavior
for the ascorbate cosubstrate over a concentration range of 0.1–10
mM, and it is possible that this may relate to the physiological regulation
of hormone production in the vesicle where the ascorbate levels also
span this concentration range.53

Notwithstanding the growing evidence for the open-to-closed mechanism,
an alternative mechanism for substrate triggering in the open-only
formalism may be suggested based on the calculations of Cowley and
Solomon, who concluded that the formation of the Cu(II)–superoxide
at the M-site was energetically favorable only if the dioxygen displaced
a hydroxide ligand from the Cu(I) center.29 While Cu(I)–OH complexes are rare, crystal structures of
the reduced enzyme do show water/hydroxide density near the CuM center,
which appears to be a coordinated solvent species. It is therefore
possible that substrate and/or ascorbate binding to the Cu(I) state
deprotonates the coordinated water to a hydroxide, which destabilizes
the Cu(I)M-site and induces reaction with oxygen. However, the fact
that the catalytic rate is inversely proportional to pH in the 4–8
pH region is inconsistent with M-site water deprotonation as the driver
of substrate triggering. We continue to search for conditions where
closed conformers can be formed and studied as potential intermediates.