Evidence for only oxygenative cleavage of aldehydes to alk(a/e)nes and formate by cyanobacterial aldehyde decarbonylases

Article abstract of DOI:10.1021/bi300912n

Cyanobacterial aldehyde decarbonylases (ADs) catalyze the conversion of C_n fatty aldehydes to formate (HCO₂^-) and the corresponding C_n-1 alk(a/e)nes. Previous studies of the Nostoc punctiforme (Np) AD produced in Escherichia coli (Ec) showed that this apparently hydrolytic reaction is actually a cryptically redox oxygenation process, in which one O-atom is incorporated from O₂ into formate and a protein-based reducing system (NADPH, ferredoxin, and ferredoxin reductase; N/F/FR) provides all four electrons needed for the complete reduction of O ₂. Two subsequent publications by Marsh and co-workers [Das, et al. (2011) Angew. Chem. Int. Ed.50, 7148-7152; Eser, et al. (2011) Biochemistry50, 10743-10750] reported that their Ec-expressed Np and Prochlorococcus marinus (Pm) AD preparations transform aldehydes to the same products more rapidly by an O₂-independent, truly hydrolytic process, which they suggested proceeded by transient substrate reduction with obligatory participation by the reducing system (they used a chemical system, NADH and phenazine methosulfate; N/PMS). To resolve this discrepancy, we re-examined our preparations of both AD orthologues by a combination of (i) activity assays in the presence and absence of O₂ and (ii) ¹⁸O₂ and H₂ ¹⁸O isotope-tracer experiments with direct mass-spectrometric detection of the HCO₂^- product. For multiple combinations of the AD orthologue (Np and Pm), reducing system (protein-based and chemical), and substrate (n-heptanal and n-octadecanal), our preparations strictly require O₂ for activity and do not support detectable hydrolytic formate production, despite having catalytic activities similar to or greater than those reported by Marsh and co-workers. Our results, especially of the ¹⁸O-tracer experiments, suggest that the activity observed by Marsh and co-workers could have arisen from contaminating O₂ in their assays. The definitive reaffirmation of the oxygenative nature of the reaction implies that the enzyme, initially designated as aldehyde decarbonylase when the C1-derived coproduct was thought to be carbon monoxide rather than formate, should be redesignated as aldehyde-deformylating oxygenase (ADO).

Full text of DOI:10.1021/bi300912n

Article
pubs.acs.org/biochemistry
Evidence for Only Oxygenative Cleavage of Aldehydes to Alk(a/e)nes
and Formate by Cyanobacterial Aldehyde Decarbonylases
Ning Li,† Wei-chen Chang,‡ Douglas M. Warui,‡ Squire J. Booker,*,†,‡ Carsten Krebs,*,†,‡
and J. Martin Bollinger, Jr.*,†,‡
†
‡
Departments of Biochemistry and Molecular Biology and Chemistry, The Pennsylvania State University, University Park,
Pennsylvania 16802, United States
S
* Supporting Information
ABSTRACT: Cyanobacterial aldehyde decarbonylases (ADs)
catalyze the conversion of Cn fatty aldehydes to formate
−
(HCO2 ) and the corresponding Cn‑1 alk(a/e)nes. Previous
studies of the Nostoc punctiforme (Np) AD produced in
Escherichia coli (Ec) showed that this apparently hydrolytic
reaction is actually a cryptically redox oxygenation process, in
which one O-atom is incorporated from O2 into formate and a
protein-based reducing system (NADPH, ferredoxin, and
ferredoxin reductase; N/F/FR) provides all four electrons
needed for the complete reduction of O2. Two subsequent publications by Marsh and co-workers [Das, et al. (2011) Angew.
Chem. Int. Ed. 50, 7148−7152; Eser, et al. (2011) Biochemistry 50, 10743−10750] reported that their Ec-expressed Np and
Prochlorococcus marinus (Pm) AD preparations transform aldehydes to the same products more rapidly by an O2-independent,
truly hydrolytic process, which they suggested proceeded by transient substrate reduction with obligatory participation by the
reducing system (they used a chemical system, NADH and phenazine methosulfate; N/PMS). To resolve this discrepancy, we re-
examined our preparations of both AD orthologues by a combination of (i) activity assays in the presence and absence of O2 and
−
(ii) 18O2 and H218O isotope-tracer experiments with direct mass-spectrometric detection of the HCO2 product. For multiple
combinations of the AD orthologue (Np and Pm), reducing system (protein-based and chemical), and substrate (n-heptanal and
n-octadecanal), our preparations strictly require O2 for activity and do not support detectable hydrolytic formate production,
despite having catalytic activities similar to or greater than those reported by Marsh and co-workers. Our results, especially of the
18O-tracer experiments, suggest that the activity observed by Marsh and co-workers could have arisen from contaminating O2 in
their assays. The definitive reaffirmation of the oxygenative nature of the reaction implies that the enzyme, initially designated as
aldehyde decarbonylase when the C1-derived coproduct was thought to be carbon monoxide rather than formate, should be
redesignated as aldehyde-deformylating oxygenase (ADO).
the enzymes’ similarity to other ferritin-like diiron-carboxylate
oxidases and oxygenases1 and the dependence of in vitro activity
on the presence of a reducing system (NADPH, ferredoxin, and
ferredoxin reductase; N/F/FR) analogous to those employed
by such oxidases/oxygenases,17,18 hinted that the enzyme could
be an oxygenase. We subsequently confirmed this possibility by
showing that O2 is also required for activity and that the O-
atom incorporated into the HCO2− product originates from O2,
implying a reaction stoichiometry of four electrons from the
reducing system per aldehyde cleaved and establishing the
reaction as an unusual oxygenation process (Scheme 1A).19 This
conclusion, which is definitively reaffirmed by the results
presented below, implies that the enzyme is more aptly
designated as aldehyde-deformylating oxygenase (ADO), a
designation that we hereby adopt.
recently discovered orthologous group of ferritin-like
Anon-heme dimetal-carboxylate enzymes from cyanobac-
teria catalyzes the second of two enzymatic steps through which
fatty acids linked to acyl carrier protein (ACP) are converted to
diesel-fuel alk(a/e)nes.1 This pathway has been touted as a
potential basis for a sunlight-driven, carbon-neutral bioprocess
to renewable, fungible (able to be burned in existing engines)
fuels.1−15 Schirmer, et al. identified the cyanobacterial genes
encoding these enzymes and demonstrated the ability of the
Nostoc punctiforme (Np) orthologue, isolated after over-
expression in Escherichia coli (Ec), to convert n-octadecanal
(which is produced by the first enzyme in the pathway, acyl-
ACP reductase) to heptadecane in vitro.1 They suggested that
the other product might be carbon monoxide and therefore
named the enzyme aldehyde decarbonylase (AD).1 We
subsequently showed that the C1-derived coproduct is actually
16
−
formate (HCO2 ). Conversion of a Cn aldehyde to the
corresponding Cn‑1 alk(a/e)ne and formate is an apparently
redox-neutral, formally hydrolytic outcome, but the structure of
the Prochlorococcus marinus (Pm) orthologue, which revealed
Received: July 9, 2012
Revised: September 3, 2012
Published: September 4, 2012
© 2012 American Chemical Society
7908
dx.doi.org/10.1021/bi300912n | Biochemistry 2012, 51, 7908−7916

Biochemistry
Article
concentrations and product yields depicted in figures were
generally very modest (tens of μM) and similar to our product
yields in O2-depleted reactions. The low product yields and
inconsistency with our published observations combined to
raise concerns about the Marsh conclusion of O2-independent
turnover by the ADOs.
Scheme 1. Predicted Origins of the Oxygen Atoms in the
Formate Co-Product Generated in the Oxygenative (A) and
Hydrolytic (B) Reactions Purportedly Catalyzed by
a
Cyanobacterial ADOs
Isotope-tracer experiments to assess the origin of the O-atom
incorporated into the formate product represent the ultimate
arbiter for the nature of the ADO reaction (compare Scheme 1,
reactions A and B). For example, in proving that the Np ADO
can cleave its substrate by an oxygenative process, we carried
out the reaction under an atmosphere of 18O2, converted the
formate to its 2-nitrophenylhydrazide (2NPH) derivative
(which extrudes one of the two O-atoms of formate), and
showed that the formyl-2NPH derivative had ∼34% 18O
(compared to a content of 50% expected for the production of
entirely HC16O18O−).19 Although shortcomings in the analysis,
specifically the use of a chemical-coupling procedure that
removes one of the two atoms being isotopically traced and
partially exchanges the other with solvent (as a result of
abortive derivatization events), prevented a firm conclusion as
to whether any hydrolytic cleavage occurs, the conclusion that
oxygenative formate production occurs was unambiguous.19
Analogously, following their initial claim of hydrolytic activity,20
Marsh and co-workers carried out the reaction in H218O with
the intent to verify the proposed solvent origin of the
incorporated O-atom in their reactions.21 In this case,
interpretation was further complicated by the potential for
exchange of the substrate carbonyl O-atom with solvent 18O.
Theoretically, complete solvent exchange prior to enzyme-
mediated formate production by the putative hydrolytic
pathway would have resulted in the amide carbonyl in the
formyl-2NPH analyte having precisely the same 18O/16O
isotopic composition as the solvent [which, unfortunately,
they did not report, thereby preventing critical evaluation of
their conclusions21]. By contrast, in concluding that the
observed result [61−64% 18O in the formyl-2NPH carbonyl
(note that this is almost identical to the percentage of solvent-
derived O-atoms in our prior analysis)] was consistent with a
solvent origin for the incorporated O-atom, they invoked very
slow exchange of the substrate carbonyl O-atom with solvent
(40% in 2 h), which they reported to have observed
experimentally but presented without documentation.21
Importantly, if one were to invoke the opposite assumption
of fast solvent exchange, as is commonly observed for aldehyde
carbonyl oxygen atoms,27 then their result would imply
precisely the opposite conclusion, that the incorporated O-
atom was derived from residual atmospheric 16O2. It is unclear
why Marsh and co-workers did not preincubate the aldehyde
substrate in the H218O solvent to permit complete exchange
prior to initiation of the ADO reaction and thereby eliminate
this ambiguity. In our view, the combination of (i) the
shortcomings in the execution and reporting of the crucial
isotope-tracer experiment (the omission of solvent isotopic
composition, the failure to pre-exchange the substrate with the
solvent, the surprising and undocumented claim of very slow
carbonyl-oxygen solvent exchange, and, most importantly, the
use of an analytical procedure that removes one of the two O-
atoms being traced and partially exchanges the other) and (ii)
our prior, unequivocal demonstration of formate production by
the oxygenative pathway cast doubt on the claim of hydrolytic
activity and made it imperative to seek more definitive analysis.
a
Non-enzymatic exchange of the carbonyl O-atom of the aldehyde
substrate with solvent is depicted on the left.
Shortly after our studies were published, Marsh and co-
workers reported that their preparations of the Np and Pm
ADOs catalyze cleavage of aldehydes to alk(a/e)nes and
formate in the presence of a reducing system but in the absence
of O2 (Scheme 1B).20,21 Their work raised the stunning
possibility that the same enzyme might catalyze two
fundamentally different reactions, the first involving reductive
activation of O2 and its subsequent attack on the aldehyde
carbonyl in a manner distinct from, but clearly related to,
mechanisms of other well-studied diiron oxidases/oxygenases,
and the second, for which they posited a mechanism involving
transient one-electron reduction of the substrate carbonyl by
the Fe2II/II cofactor and formation of an organometallic formyl-
FeII intermediate, bearing little resemblance to any reaction
previously attributed to a member of this well-studied enzyme
family.20 Apart from this provocative central claim, Marsh and
co-workers reported several other interesting observations.
First, the chemical reducing system that they employed, NADH
and phenazine methosulfate (N/PMS), reportedly supported
greater activity than the N/F/FR protein-based reducing
system used by both Schirmer et al. and us,16,19 suggesting
1
that the extremely modest in vitro turnover rates achieved in
previous studies could have resulted from sluggish electron
delivery to the ADO cofactor by the heterologous (spinach)
ferredoxin. Second, they reported that the phenazine of the
reducing system binds tightly to the enzyme, an observation
hinting that heterocyclic redox cofactors might have a role in
the reaction as it occurs in vivo.20 Finally, they noted that linear
aldehydes as short as C7 are good substrates,20 an observation
potentially resolving the challenge to mechanistic analysis
presented by the very low solubility of longer-chain fatty
aldehyde substrates.
The central claim of the Marsh work, aldehyde cleavage in
the absence of O2, is inconsistent with our own published
observations. Most importantly, we consistently detected much
less product (<20 μM) when O2 was intentionally removed
prior to constitution of the reactions than when reactions were
constituted with air-saturated solutions, which, under optimized
conditions, yielded >100 μM products.19 Product yields in the
O2-depleted reactions were invariably similar to the levels of
residual O2 typically remaining after our standard deoxygena-
tion procedure,22−26 consistent with a stoichiometric require-
ment for O2. By contrast, Marsh and co-workers reported that,
in their experiments, inclusion of O2 actually diminished product
yield.20 It is noteworthy that, in both of their studies, enzyme
7909
dx.doi.org/10.1021/bi300912n | Biochemistry 2012, 51, 7908−7916

Biochemistry
Article
We reasoned that mass-spectrometric (MS) analysis of the
formate product itself, rather than its 2NPH derivative, would
provide for a very sensitive isotopic probe for the proposed
hydrolytic activity. Following complete exchange of the
substrate O-atom with H218O solvent, any hydrolytic cleavage
would produce HC18O18O−, whereas any oxygenative activity
occurring in a reaction intended to be O2-free but containing
residual atmospheric 16O2 would produce HC18O16O−. Thus,
the absence or presence of a signal for the M+4 product (where
M is the mass of the formate with the natural abundance of
oxygen isotopes) in a reaction carried out in H218O should be
rigorously dispositive regarding the existence of hydrolytic
activity. It appears that direct MS analysis of formate has, in the
past, proven difficult, undoubtedly owing to its prevalence in
the environment. However, our previous studies showed that,
Flash column chromatography was performed on silica gel
(230−400 mesh, grade 60) obtained from Sorbent Technol-
1
ogies. Products were characterized by H and 13C nuclear
magnetic resonance (NMR) spectroscopy (see the Supporting
Information). NMR spectra were recorded on a Bruker 300,
360, or 400 MHz spectrometer at the nuclear magnetic
resonance facility of the Department of Chemistry, The
Pennsylvania State University. Chemical shifts (δ in ppm)
were determined from the known signals of solvent (CDCl3 or
d6-DMSO), and coupling constants are given in Hertz (Hz).
Preparation of Proteins. The DNA sequence that encodes
Pm ADO was codon-optimized for overexpression in Ec,
synthesized, and inserted into the NdeI and BamHI restriction
sites of expression vector pET-28a by GeneArt (Regensburg,
Germany). This plasmid construct, which places the gene under
the control of a T7 promoter, allows for overproduction of the
protein containing an N-terminal His6-tag separated from its
native start codon by a spacer of 10 amino acids. The resulting
plasmid, designated pPmADOwt, was used to transform Ec
BL21 (DE3) (Invitrogen; Carlsbad, CA) for protein
production. Protein overproduction and purification followed
procedures similar to those reported by Schirmer, et al.,1 with
minor differences, as noted. Protein expression was carried out
in shake flasks at 37 °C in Luria−Bertani (LB) medium
supplemented with 50 μg/mL kanamycin and was induced at
an OD600 nm of 0.8 by addition of isopropyl β-D-1-
thiogalactopyranoside (IPTG) to a final concentration of 0.25
mM. The culture was shaken at 37 °C for 4 additional hours
before being harvested by centrifugation at 6,000 rpm for 20
min. Cell paste was frozen in liquid nitrogen and stored at −80
°C until it was used. All purification steps were performed
according to the procedure reported previously.16,19 Np ADO
was prepared as described in previous work.16,19 The contents
of various redox-active, first-row transition metals in the
preparations of Np and Pm ADOs used in this work were
determined by inductively coupled plasma atomic emission
spectroscopy (ICP-AES) by Mr. Henry Gong at the Penn State
Materials Research Institute. For the Np ADO, the results were
Co, not detected; Cu, not detected; Fe, 0.82 equiv; Mn, 0.06
equiv; and Ni, 0.05 equiv. For the Pm ADO, the results were
Co, not detected; Cu, 0.01 equiv; Fe, 0.75 equiv; Mn, 0.07
equiv; and Ni, 0.06 equiv. Subsequent work in our group has
shown that growth of the overexpressing strain in minimal
medium supplemented with iron permits further enrichment in
Fe and affords more active protein, consistent with the active
cofactor being a diiron cluster. These procedures and results
will be published separately in a later paper.
2
with the substrate labeled with 13C or H at C1, the MS signal
arising from ADO-generated formate is resolved from that
arising from the environmental formate.16 In the present case,
mass resolution was anticipated to be enhanced even further by
the addition of 2 or 4 amu from the 18O in the HC18O16O− or
HC18O18O− product, respectively.
In this work, we report results of the 18O-tracer experiments
with direct MS analysis of the formate product that we have
forecast above, in combination with simpler determinations of
product yields in O2-replete and O2-depleted reactions, which
show unequivocally that our ADO preparations simply do not
support detectable O2-independent, hydrolytic cleavage,
irrespective of the ADO orthologue (Np or Pm), the reducing
system (N/F/FR or N/PMS), or the substrate (n-heptanal or
n-octadecanal) employed. In addition, aided by the important
technical advances of Marsh and co-workers,21 including the
N/PMS reducing system (which we improved upon slightly by
use of a further modified phenazine) and the more soluble,
shorter-chain substrate (n-heptanal) on which they reported,
we demonstrate turnover numbers (∼ 0.3 s−1) approaching the
100−101 s−1 regime typical of diiron oxidases/oxy-
genases,25,28−33 thus setting the stage for transient-state kinetic
and spectroscopic dissection of the reaction mechanism.
EXPERIMENTAL PROCEDURES
■
Materials. Phenazine methosulfate (PMS; >90%), 2-nitro-
phenylhydrazine (97%), 1-methoxy-5-methylphenazinium
methylsulfate (MeOPMS; >95%), pyridine (>99.9%), spinach
ferredoxin, spinach ferredoxin reductase, NADH (>97%), and
NADPH (>97%) were purchased from Sigma-Aldrich. 1-(3-
Dimethylaminopropyl)-3-ethylcarbodiimide (98%) and pyridi-
nium dichlorochromate were purchased from Alfa Aesar. 18O2
gas (99% 18O) was purchased from ICON Isotopes. H218O
(>97% H218O), 1-[13C]-stearic acid, n-1-[13C]-octanoic acid,
[13C]-formate, and B2H3 (98% 2H, 1 M in THF) were
purchased from Cambridge Isotope Laboratories.
Aldehyde Cleavage Reactions. All ADO reactions were
carried out in 100 mM HEPES buffer, pH 7.5, containing 10%
glycerol. Standard reactions with the n-octadecanal substrate
and the N/PMS reducing system contained 0.1 mM ADO, 0.5
mM n-1-[13C]-octadecanal, 0.3 mM PMS, 6 mM NADH, and
0.2% triton x-100. Reactions with the n-octadecanal substrate
and the N/F/FR reducing system contained 0.1 mM ADO, 0.5
mM n-1-[13C]-octadecanal, 2 mM NADPH, 100 μg/mL
ferredoxin, 100 mU/mL spinach ferredoxin reductase, and
0.2% triton x-100. Reactions with the n-heptanal substrate
contained 0.1 mM ADO, 16 mM n-1-[2H]-heptanal, 0.3 mM
MeOPMS, and 6 mM NADH (with no detergent).
Synthesis of Substrates. n-1-[13C]-Octadecanal was
synthesized as described by Schirmer, et al.1 1-[13C]-Stearic
acid (1.0 g, 3.5 mmol) was reduced with BH3 (1 M in THF
solution; 4.2 mL, 4.2 mmol) in THF (40 mL) and subsequently
treated with pyridinium dichlorochromate (1.98 g, 5.3 mmol)
in methylene chloride (40 mL). n-1-[13C]-Octanal was
prepared in an analogous manner from n-1-[13C]-octanoic
acid. n-1-[2H]-Heptanal was synthesized in identical fashion
from n-1-heptanoic acid using B2H3 as reductant (1 M B2H3
solution in THF). Analytical thin layer chromatography (TLC)
was carried out on precoated TLC aluminum plates (silica gel,
grade 60, F254, 0.25 mm layer thickness) from EMD Chemicals.
Analysis for Formate by Liquid Chromatography/
Mass Spectrometry (LC/MS). Reaction samples for direct
formate analysis were filtered through Amicon spin filters with a
molecular weight cutoff of 10 kDa and stored at 4 °C before
7910
dx.doi.org/10.1021/bi300912n | Biochemistry 2012, 51, 7908−7916

Biochemistry
Article
LC/MS analysis. A 3-μL aliquot of each sample was injected
onto an Agilent QQQ 6410 LC/MS spectrometer equipped
with an Agilent HILIC analytical column. The isocratic mobile
phase contained 80/20 (v/v) acetonitrile/20 mM aqueous
ammonium acetate. The flow rate was 0.1 mL/min. The
negative-ion detection mode was used with single ion
the Experimental Procedures section, and Figure S1 of the
Supporting Information shows that the peak area in the single-
ion chromatogram at m/z = 46, corresponding to the H13CO2
−
analyte, increases with the concentration of H13CO2 injected.
−
Thus, the formate product from a n-1-[13C]-aldehyde substrate
can readily be detected at tens of μM concentrations by this
assay. Comparison of peak areas obtained by analysis of ADO
reactions to the standard curve generated from Figure S1
(Supporting Information) provides for quantification of the
formate product, but, in the many experiments performed as
part of this study, we found that the experimental uncertainty
associated with quantification of the product by the direct
method is somewhat larger than that associated with the
2NPH-derivatization method. We therefore carried out both
procedures in this study: the direct analysis for purposes of
detection and isotopic analysis of the formate, and the 2NPH-
derivatization procedure for its accurate quantification (when
required).
−
monitoring (SIM) at m/z values of 45 (1H12C16O2 ), 46
(2H12C16O2 or 1H13C16O2−), 48 (2H12C16O18O− or
−
−
1
−
1H13C16O18O−), and 50 (2H12C18O2 or H13C18O2 ). The
sample preparation and analysis for formate by derivatization to
formyl-2NPH were the same as described in previous work.16
Analysis for Hexane by GC/MS. A solution containing 0.1
mM ADO, 4 mM n-heptanal, 0.3 mM MeOPMS, 6 mM
NADH, 50 mM HEPES buffer (pH 7.5), and 10% glycerol in a
final volume of 700 μL was allowed to react in a sealed vessel
for 5 min. Gas from the head space of the sample was analyzed
on a Shimadzu GCMS-QP2010S interfaced with a Shimadzu-
5MS 30 m column (ID, 0.25 narrow bore; film, 0.25 μm). The
inlet and oven temperature were set to 250 and 40 °C,
respectively. Upon injection, the oven temperature was held at
40 °C for 3 min and then ramped up to 120 °C at a rate of 10
°C/min. Total ion chromatograms were generated by scanning
the range m/z = 30−3,000.
Demonstration of the O2 Requirement for Formate
Production by the Np and Pm ADOs. Detecting both the
2NPH derivative of formate and its alkane co-product, we
previously demonstrated that the activity of the Np ADO
supported by the N/F/FR reducing system requires the
presence of O2. When the first paper by Marsh and co-workers
reporting O2-independent ADO activity (of the Pm orthologue
supported by the N/PMS reducing system) was published,20
we formulated three most likely explanations for the apparent
discrepancy between our observations and theirs: (1) different
catalytic capabilities of the different ADO orthologues; (2)
different catalytic capabilities supported by the different
reducing systems; and (3) O2 contamination in their assays.
With the direct LC/MS assay in hand, we simultaneously
evaluated possibilities 1 and 2 by testing for formate production
from n-1-[13C]-octadecanal, in the presence and absence of O2,
by both orthologues in the presence of Marsh’s N/PMS
reducing system (Figure 1). Following a 1-h incubation in an
18O2 and H218O Isotope-Tracer Experiments. Standard
reaction solutions (see above) were prepared in an MBraun
anoxic chamber from solutions rendered O2-free on a vacuum-
argon manifold. Samples were exposed to either 16O2 or 18O2
(∼720 Torr) as described previously34 and incubated for
sufficient time for the reaction to reach completion. The
resultant solution was either filtered anaerobically (for direct
formate analysis) or derivatized according to the standard
procedure (see above) before storage or analysis. In the H218O
isotope tracer experiments, the reaction solution was prepared
with O2-free H218O, and the aldehyde substrate was allowed
several minutes to undergo complete exchange of its carbonyl
oxygen atom with solvent before the reaction was initiated. The
final 18O content of the reaction solution was ∼85%.
Determination of the Rate of Exchange of the
Aldehyde Carbonyl of Octanal with 18O from Solvent.
Reaction mixtures containing 15 mM n-1-[13C]-octanal in 0.75
mL of H218O with 0.06 mL of d6-dimethylsulfoxide (d6-DMSO)
added as internal standard were prepared and immediately
subjected to 13C NMR analysis on a Bruker 850 MHz
spectrometer at 298 K. A time-dependent shift of the carbonyl
13C signal from δ = 211.053 to δ = 211.005 (upfield shift by
10.2 Hz) signifies exchange of 16O with 18O.35,36
Determination of the NADH/Formate Reaction Stoi-
chiometry. Reaction solutions contained 0.1 mM ADO, 16
mM n-1-[2H]-heptanal, 0.3 mM MeOPMS, and varying
[NADH]. They were allowed to react in air for 10 min at
room temperature. They were then subject to the 2NPH
derivatization reaction and LC/MS analysis.
Figure 1. LC/MS detection of formate produced enzymatically by Pm
ADO and Np ADO. Three identical standard reaction solutions with
the N/PMS reducing system and n-1-[13C]-octadecanal substrate were
prepared anareobically for both Np ADO (A) and Pm ADO (B),
respectively. They were filtered and measured by LC/MS right away
(black line) or after a 1-h incubation at 22 °C in the absence (gray
line) or presence of O2 (red line). An authentic H13COO− solution
was analzyed similarly as a standard (blue dotted line). The traces are
the single-ion-monitoring (SIM) LC/MS chromatograms at the
indicated values of m/z.
RESULTS
■
Development of a Direct LC/MS Assay for Formate. As
noted above, the most rigorously dispositive difference between
the reaction pathway reported by Marsh and co-workers20,21
and that previously demonstrated by us19 is the origin (H2O or
O2, respectively) of the new O-atom incorporated into the
formate product (Scheme 1). To assess this origin, a direct LC/
MS assay for formate, rather than its 2NPH derivative, was
sought. The conditions of the LC/MS analysis are provided in
O2-depleted solution, neither reaction gave a m/z = 46 peak for
H13CO2− (gray traces in panels A and B) that was significantly
enhanced relative to the corresponding peak from an identical
sample quenched at the shortest possible reaction time (black
traces). By contrast, the same 1-h incubation in the presence of
ambient O2 resulted in an intense m/z = 46 peak (red traces) at
−
the same retention time as that for the authentic H13CO2
7911
dx.doi.org/10.1021/bi300912n | Biochemistry 2012, 51, 7908−7916

Biochemistry
Article
standard (dashed blue trace in panel A). These results establish
that our preparations of both ADO orthologues require O2 to
produce formate, weighing against explanations 1 and 2. As
they were both carried out in the presence of natural abundance
O2 (0.2% 18O), it is not surprising that neither reaction gave a
Testing Short-chain Aldehydes as ADO Substrates.
The small peaks at m/z = 46 in the red traces of Figure 2,
panels C and D, potentially suggest the production of a small
but detectable quantity of formate by a truly hydrolytic
pathway, which would result in the incorporation of an O-atom
from H216O rather than 18O2. Alternatively, these peaks could
reflect contamination of these reactions by environmental 16O2
(equivalent to explanation 3 for the discrepancy between our
results and those of Marsh and co-workers). To distinguish
between these two possible interpretations and provide for the
most sensitive detection of even trace hydrolytic formate
production, we sought to carry out reactions in H218O and
1
significant peak at m/z = 48 corresponding to H13C18O16O−
(rear traces in both panels).
Assessment of the Origin of the New O-Atom in
Formate by the Use of 18O2. Our previous work employing
the 2NPH-derivatization assay established that the O-atom
incorporated into the formate product upon aldehyde cleavage
by the Np ADO supported by N/F/FR reducing system arises
from O2 in the majority of reaction events. To assess the origin
of the formate O-atom with other combinations of the ADO
orthologue and reducing system and to verify more
quantitatively its origin in the Np ADO−N/F/FR reaction,
we applied the direct LC-MS formate analysis to reactions
carried out with 18O2 (Figure 2). The O2-depleted reactions of
analyze for the presence of any 1H13C18O2 , which would result
−
from exchange of the aldehyde carbonyl O-atom of the
substrate with solvent27 prior to C1−C2-bond cleavage and
incorporation of the second O-atom from solvent in the ADO-
mediated hydrolytic event (Scheme 1B). Proper interpretation
of the results of such an experiment requires knowledge of the
extent of the carbonyl-solvent exchange occurring prior to
turnover. We therefore sought to employ a substrate that would
be amenable to the direct determination of the carbonyl-solvent
exchange kinetics by 13C NMR spectroscopy. The report by
Marsh and co-workers that linear aldehydes as short as n-
heptanal are substrates for the ADOs20,21 inspired us to test
various saturated linear aldehydes (n-heptanal, n-octanal, and n-
decanal), and, indeed, all were found to be active ADO
substrates (Table S1, Supporting Information).
To verify that the catalytic activity of the ADO operating on
even the shortest substrate, n-heptanal, has the same general
characteristics as ADO-mediated cleavage of the more
physiologically relevant n-octadecanal, we carried out reactions
with the Np ADO and the chemical reducing system [in this
case, 1-methoxy-5- methylphenazinium methylsulfate
(MeOPMS) was used in place of the PMS used by Marsh
and co-workers because the former compound reportedly
exhibits greater photolytic stability,37 a property desirable for
mechanistic analysis by stopped-flow absorption experiments]
at two different ADO concentrations and also with serial
omission of a single reaction component to define the
requirements for activity (Figure 3). The complete enzyme
reactions rapidly generated formate (black and gray symbols
and fit lines), and, although only a very short (if any) steady
state was observed (the reason for this characteristic is not
Figure 2. LC/MS detection of formate produced in 18O2-tracer
experiments by Np ADO and Pm ADO. Standard reaction solutions
with the N/F/FR (A and B) or N/PMS (C and D) reducing system
and n-1-[13C]-octadecanal substrate were prepared anaerobically for
both Np ADO (A and C) and Pm ADO (B and D). Solutions were
incubated at 22 °C for 1 h either anaerobically (black line) or after the
introduction of 18O2 (red line), filtered, and assayed directly for
formate by LC/MS. The traces are the single-ion-monitoring (SIM)
LC/MS chromatograms at the indicated values of m/z.
both Np ADO (panels A and C) and Pm ADO (panels B and
1
−
D) gave negligible peaks at m/z = 46 for H13C16O2 (black
traces), regardless of whether the N/F/FR reducing system (A
and B) or the N/PMS reducing system (C and D) was
employed. This observation is consistent with the results of
Figure 1. For all four combinations, the reactions to which 18O2
was subsequently added after the removal of atmospheric O2
1
−
(red traces) gave peaks at m/z = 48 for H13C18O16O2 that
were much more intense than the corresponding peaks at m/z
= 46 for H13C16O2 . These results demonstrate that, for all
four combinations of ADO orthologue and reducing system, O2
rather than H2O is the source of the O-atom incorporated into
the formate product in the vast majority of reaction events. The
minor peaks at m/z = 46 in panels C and D reflect a minor
atmospheric 16O2 contaminant in these reactions rather than a
trace level of hydrolytic formate production, as established
below.
1
−
Figure 3. Time dependence of and requirements for the Np ADO-
catalyzed production of formate. Standard reaction solutions with the
n-1-[2H]-heptanal substrate were prepared as described in Exper-
imental Procedures, and, as indicated, the [Np ADO] was varied, or a
single reaction component was omitted. Aliquots were removed after
varied times of incubation at 22 °C, derivatized, and analyzed for
formyl-2NPH by LC/MS.
7912
dx.doi.org/10.1021/bi300912n | Biochemistry 2012, 51, 7908−7916

Biochemistry
Article
known), the initial rates of formate production showed the
expected correlation to ADO concentration (compare black to
gray trace). Fits of the equation for an exponential burst and
linear rise phases to the progress curves gave initial rates (v/
[ADO]) of 0.27 0.03 s−1, much greater than any previously
reported ADO turnover number.20 This result hints at the value
of the short substrates in mechanistic analysis. Upon omission
of the ADO enzyme (purple), NADH (green), MeOPMS
(red), or O2 (blue), minimal formate was produced.
Importantly, after a 4 min unproductive incubation in the
absence of O2, opening of the sealed reaction vessel to the air
(blue arrow) led to the commencement of formate production,
further confirming the requirement for O2. The data in Figure 3
suggest that turnover of the shorter, more soluble substrate, n-
heptanal, has the same requirements as turnover of the
physiologically relevant longer fatty aldehydes.
Quantification of n-Hexane Produced by the ADO-
Catalyzed Conversion of n-Heptanal. As additional
evidence for the mechanistic correspondence between ADO-
mediated cleavage of n-heptanal and turnover of the longer
fatty aldehydes, the stoichiometry of n-hexane to formate was
determined. GC/MS analysis of the headspace from triplicate,
sealed ADO reactions (described in the Methods section)
permitted both the detection of the n-hexane co-product and,
by integration of the peak of the total-ion chromatogram and
comparison to an external standard curve, its quantification.
The triplicate reactions produced 322 35 μM formate and
362 40 μM hexane (mean standard deviation), verifying
the expected 1:1 reaction stoichiometry of the two coproducts.
Measurement of the Exchange Rate of the Aldehyde
Carbonyl of Octanal with H218O. The published inter-
pretation of the solvent-18O-tracer experiment carried out by
Marsh and co-workers in an effort to verify the expected solvent
origin of the O-atom incorporated into the formate product
relied on their report that the substrate carbonyl oxygen
exchanges with solvent very slowly on the time scale of the
ADO reaction.21 Indeed, the assumption of rapid exchange
would have led them to exactly the opposite conclusion: that
the origin of the incorporated O-atom must have been
contaminating O2. With the knowledge that short-chain
aldehydes are viable ADO substrates, we used 13C NMR
spectroscopy to directly determine the rate of carbonyl-solvent
exchange (Figure 4). Incubation of the specifically labeled
substrate in H218O resulted in a time-dependent change in the
chemical shift of the 1-13C nucleus from 211.053 ppm to
211.005 ppm, reflecting exchange of the 16O in the carbonyl
group for 18O.35,36 This shift was nearly complete after 6 min
(top spectrum) and had a half-life of approximately 1 min
(second spectrum from the bottom). These spectra demon-
strate that preincubation of a short-chain n-aldehyde substrate
in H218O for ≥6 min permits the substrate to attain the O-
isotopic composition of the solvent.
Figure 4. Determination of the kinetics of exchange of the carbonyl
oxygen of n-[13C]-octanal with solvent by 13C NMR-spectroscopy. n-1-
[13C]-Octanal (natural abundance of oxygen, >99% 16O) was dissolved
in 700 μL of H218O with 60 μL of d6-DMSO as the internal standard.
Spectra were required at varying times of incubation at ambient
temperature (∼22 °C). The exchange of 16O from the C1 position of
the aldehyde with solvent 18O results in the indicated change in the
chemical shift.
Figure 5. LC/MS 18O2- and H218O-isotope-tracer experiments to
determine the origin of the O-atom incorporated into formate by Np
ADO and Pm ADO. Standard reaction solutions (see Experimental
Procedures) of 500 μL total volume with the n-1-[2H]-heptanal
substrate and the N/MeOPMS reducing system were prepared
anaerobically with H216O (A) or H218O (B). All reactions were
incubated for 5 min at 22 °C, one set in the absence of O2 (black line),
the second under ∼1 atm of natural-abundance (>99% 16O) O2 (red
line), and the third under ∼1 atm of 18O2 (>99% isotopic enrichment)
(blue line). The reactions were filtered and subjected to direct LC/MS
analysis for formate. The traces are the single-ion-monitoring (SIM)
LC/MS chromatograms at the indicated values of m/z.
Test for Trace Hydrolytic Cleavage of n-Heptanal by
Analysis of Reactions in H218O. With the kinetics of
carbonyl-solvent exchange elucidated, the definitive test for
hydrolytic formate production forecast above could now be
carried out (Figure 5). The substrate employed in this
experiment, n-1-[2H]-heptanal, was selected as the synthetically
simplest isotopic perturbation that would afford mass-
resolution of enzymatically produced formate from the
environmental formate contaminant (our previous work
established that the C1 hydrogen of the aldehyde substrate is
fully retained in the formate product, giving the formate an m/z
of 46, equivalent to that produced from a 1-[13C]-labeled
aldehyde16). The n-1-[2H]-heptanal substrate was preincubated
in either natural O-abundance H2O (panel A) or 85% 18O H2O
(panel B) for several minutes to ensure that the carbonyl O-
atom would achieve the isotopic composition of the solvent,
and the reaction solution was then rendered complete by the
addition of the Np ADO and the N/MeOPMS chemical
reducing system. In either solvent, negligible formate was
7913
dx.doi.org/10.1021/bi300912n | Biochemistry 2012, 51, 7908−7916

Biochemistry
Article
detected in reactions from which O2 had been removed and not
added back (black traces in both panels). In H216O solvent
under 16O2 gas (panel A, red traces), an intense peak at m/z =
46 was detected (2H12C16O2 ), and no significant peak at either
−
−
m/z = 48 (2H12C18O16O−) or m/z = 50 (2H12C18O2 ) was
seen, as expected. In H216O solvent under 18O2 gas (panel A,
blue traces), the major peak shifted to m/z = 48
(2H12C18O16O−), reflecting incorporation of a single O-atom
from 18O2 into the formate product. Interestingly, a small but
significant peak at m/z = 50 reflects the production of a minor
quantity of formate with both O-atoms originating from O2
−
(2H12C18O2 ). This product could result from (1) an abortive
event that results in incorporation of one O-atom from O2 into
the aldehyde substrate without achieving C1−C2 fragmentation
and (2) subsequent, successful aldehyde cleavage occurring
Figure 6. NADH/formate stoichiometry of the Np ADO-catalyzed
production of formate from n-1-[2H]-heptanal. Reactions were carried
out at 22 °C in 100 mM HEPES buffer (pH 7.5) for 5 min. They
contained 0.3 mM MeOPMS, 0.2% triton x-100, 4 mM n-1-[2H]-
heptanal, the indicated concentration of Np ADO [0.033 mM (red
circles), 0.1 mM (blue circles), or 0.2 mM (green circles)], and varying
concentration of NADH (values shown on abscissa). The solutions
were subjected to the derivatization procedure and analyzed for
formyl-2NPH by LC/MS.
either without substrate release to solvent (upon which the 18
O
incorporated during the abortive event could exchange) or
before carbonyl-solvent exchange can occur. This rare double-
incorporation event may hold useful mechanistic clues. In
H218O solvent under 16O2 gas (panel B, red traces), the
intensities of the major peak at m/z = 48 and the minor peak at
2
2
−
m/z = 46, corresponding to H12C16O18O− and H12C16O2 ,
respectively, precisely mirror the solvent isotopic composition
of 85% 18O and 15% 16O. Most importantly, no significant peak
at m/z = 50, corresponding to the 2H12C18O2− that should have
been produced in 72% of hydrolytic events (the fraction
expected to result from 0.85 18O from exchange and 0.85 18O
from hydrolysis), could be detected (rear red trace in panel B).
Only when the reaction was carried out in both H218O and 18O2
did the m/z = 50 peak become prevalent (indeed,
predominant), and, under these conditions, the ratio of the
m/z = 50 and m/z = 48 peaks again reflected the solvent
18O/16O composition (blue traces in panel B), implying a vastly
predominant outcome of precisely one O-atom of formate from
solvent (by exchange) and precisely one from O2 (by
oxygenative cleavage). In other words, our preparations do
not support any detectable O2-independent, hydrolytic formate
production.
Test for a Stoichiometric or Catalytic Role of the
Reducing System. The final piece of evidence cited by Marsh
and co-workers in support of O2-independent, hydrolytic
aldehyde cleavage by ADO was their observation that NADH
is not consumed during the reaction.21 Turnover without
consumption of reducing equivalents (4 e−/turnover) would be
incompatible with the oxygenative pathway indicated by our
data (Scheme 1A). To assess the Marsh report of a catalytic
rather than stoichiometric role for the reducing system,21
reactions were run to completion (verified by examination of
different reaction times) with limiting and varying [NADH]
(Figure 6). The yield of formate was found to increase linearly
with increasing [NADH]. The slope of the line gives 0.25
0.03 formate/NADH. Given that NADH is a two-electron
donor, this experimental stoichiometry corresponds to 8 e−/
formate, twice the theoretical ratio of 4 e−/formate predicted
by Scheme 1A. This deviation most likely reflects the
uncoupling of NADH oxidation from formate production in
roughly half of the reaction events, a result that is not surprising
in view of the nonphysiological nature of the reducing system
and its inherent reactivity to O2 even in the absence of ADO.
the N/F/FR protein-based reducing system or the N/
(MeO)PMS chemical reducing system, together with the O-
isotope labeling pattern of the formate product, firmly establish
that our ADO preparations do not support any hydrolytic
aldehyde cleavage. Most definitively, we are unable to detect
any of the H12C18O2 that would necessarily be produced by
O2-independent, hydrolytic aldehyde cleavage in reactions run
in H218O. This absence of the doubly-18O-labeled formate
product is stark evidence that one oxygen atom of the formate
invariably derives from O2. Whereas it would remain formally
possible that Marsh and co-workers had obtained a different
enzyme form (e.g., possessing a different metallocofactor or
some post-translational modification) having distinct catalytic
capabilities (e.g., O2-independent, hydrolytic aldehyde cleav-
age), it appears that, immediately before the submission of this
manuscript, these authors also came to recognize that their
report of O2-independent ADO activity is in doubt.38 Thus, the
nature of the reaction is now firmly established as oxygenative
aldehyde cleavage to formate and alk(a/e)ne, providing
argument that the designation aldehyde-deformylating oxygen-
ase (ADO) is more appropriate than aldehyde decarbonylase
(AD).
2
−
Although the main conclusion of Marsh and co-workers is
incorrect, their work nevertheless provided valuable technical
tools from which we have profited in this study.20,21 For
example, we concur that the N/(MeO)PMS chemical reducing
system supports greater activity than the N/F/FR protein-
1
based system originally reported by Schirmer, et al. and
adopted by us in our first two studies.16,19 Indeed, the coupling
efficiency supported by this reducing system (∼ 50%) is
surprisingly high and much greater than the value that we
previously measured for the heterologous protein-based system.
As importantly, their discovery that ADO accepts much shorter,
more soluble aldehydes is an extremely useful advance. Our
past experience in dissecting the mechanisms of O2-utilizing
metalloenzymes by direct detection and characterization of
intermediate states has taught us that the most generally
informative approach involves rapid mixing of the complex of
the reduced enzyme form and its substrate(s) with O2 to
initiate rapid formation of the intermediate complexes.23 In
these experiments, it is desirable to saturate the enzyme by its
substrate. For the case of ADO, the physiologically relevant
DISCUSSION
■
The strict, stoichiometric O2 and NAD(P)H requirements of
both the Np and Pm ADO orthologues in the presence of either
7914
dx.doi.org/10.1021/bi300912n | Biochemistry 2012, 51, 7908−7916

Biochemistry
Article
long-chain n-aldehyde substrates are only sparingly soluble, and
we have experienced difficulty in achieving saturation at
accessible substrate concentrations. In addition, steady-state
turnover rates thus far achieved with these long substrates are
far too modest (<1 min−1) to have inspired hope that reactive
intermediates might be induced to accumulate. Shorter
aldehydes such as n-heptanal are more soluble and, indeed,
we find that n-heptanal supports an initial rate of ∼0.3 s−1.
Although 0.3 s−1 is still a relatively modest rate constant
compared to those exhibited by many oxygenases,25,28−33 it
approaches the realm of 100−102 s−1 that, in our experience,
signifies that the selected reaction conditions may support
intermediate accumulation.23
REFERENCES
■
(1) Schirmer, A., Rude, M. A., Li, X., Popova, E., and del Cardayre, S.
B. (2010) Microbial biosynthesis of alkanes. Science 329, 559−562.
(2) Ducat, D. C., Way, J. C., and Silver, P. A. (2011) Engineering
cyanobacteria to generate high-value products. Trends Biotechnol. 29,
95−103.
(3) Schirmer, A., Rude, M. A., and Brubaker, S. (2009) Methods and
compositions for producing hydrocarbons, Patent Application WO/
2009/140696, LS9.
(4) Service, R. F. (2009) ExxonMobil fuels Venter’s efforts to run
vehicles on algae-based oil. Science 325, 379.
(5) Reppas, N. B., and Ridley, C. P. (2010) Methods and
compositions for the recombinant biosynthesis of n-alkanes, Patent
7794969, Joule Unlimited Inc., United States.
(6) Keasling, J. D. (2010) Manufacturing molecules through
metabolic engineering. Science 330, 1355−1358.
(7) Radakovits, R., Jinkerson, R. E., Darzins, A., and Posewitz, M. C.
(2010) Genetic engineering of algae for enhanced biofuel production.
Eukaryotic Cell 9, 486−501.
ASSOCIATED CONTENT
■
S
* Supporting Information
Figures depicting LC/MS analysis of the 13C-labeled formate
1
standard, H- and 13C NMR spectra of n-1-[13C]-octadecanal,
n-1-[13C]-octanal, and n-1-[2H]-heptanal, and a table compar-
ing the ADO-catalyzed production of formate from n-heptanal,
n-octanal, and n-decanal. This material is available free of charge
via the Internet at http://pubs.acs.org.
(8) Du, J., Shao, Z. Y., and Zhao, H. M. (2011) Engineering microbial
factories for synthesis of value-added products. J. Ind. Microbiol.
Biotechnol. 38, 873−890.
(9) Hess, W. R. (2011) Cyanobacterial genomics for ecology and
biotechnology. Curr. Opin. Microbiol. 14, 608−614.
(10) Malcata, F. X. (2011) Microalgae and biofuels: a promising
partnership? Trends Biotechnol. 29, 542−549.
AUTHOR INFORMATION
Corresponding Author
■
(11) Patra, T., Manna, S., and Maiti, D. (2011) Metal-mediated
deformylation reactions: synthetic and biological avenues. Angew.
Chem.-Int. Ed. 50, 12140−12142.
*(S.J.B.) Department of Chemistry, 302 Chemistry Building,
University Park, PA 16802. Phone: 814-865-8793. Fax: 814-
865-2927. E-mail: sjb14@psu.edu. (C.K.) Department of
Chemistry, 332 Chemistry Building, University Park, PA
16802. Phone: 814-865-6089. Fax: 814-865-2927. E-mail:
ckrebs@psu.edu. (J.M.B.) Department of Chemistry, 336
Chemistry Building, University Park, PA 16802. Phone: 814-
863-5707. Fax: 814-865-2927. E-mail: jmb21@psu.edu.
(12) Robertson, D. E., Jacobson, S. A., Morgan, F., Berry, D., Church,
G. M., and Afeyan, N. B. (2011) A new dawn for industrial
photosynthesis. Photosynth. Res. 107, 269−277.
(13) Zhang, F. Z., Rodriguez, S., and Keasling, J. D. (2011) Metabolic
engineering of microbial pathways for advanced biofuels production.
Curr. Opin. Biotechnol. 22, 775−783.
(14) Hallenbeck, P. (2012) Microbial Technologies in Advanced
Biofuels Production, Springer, New York.
Funding
(15) Serrano-Ruiz, J. C., Ramos-Fernandez, E. V., and Sepulveda-
Escribano, A. (2012) From biodiesel and bioethanol to liquid
hydrocarbon fuels: new hydrotreating and advanced microbial
technologies. Energy Environ. Sci. 5, 5638−5652.
This work was supported by the National Science Foundation
(MCB-1122079 to C.K., S.J.B., and J.M.B.).
Notes
The authors declare the following competing financial
interest(s): The authors have a financial stake in the company,
LS9, incorporated, which is attempting to use the enzyme that
is the subject of this paper in a bioprocess to produce diesel fuel
in bacteria.
(16) Warui, D. M., Li, N., Nørgaard, H., Krebs, C., Bollinger, J. M.,
Jr., and Booker, S. J. (2011) Detection of formate, rather than carbon
monoxide, as the stoichiometric co-product in conversion of fatty
aldehydes to alkanes by a cyanobacterial aldehyde decarbonylase. J.
Am. Chem. Soc. 133, 3316−3319.
(17) Merkx, M., Kopp, D. A., Sazinsky, M. H., Blazyk, J. L., Muller, J.,
̈
ACKNOWLEDGMENTS
and Lippard, S. J. (2001) Dioxygen activation and methane
hydroxylation by soluble methane monooxygenase: a tale of two
irons and three proteins. Angew. Chem., Int. Ed. 40, 2782−2807.
(18) Fox, B. G., Lyle, K. S., and Rogge, C. E. (2004) Reactions of the
diiron enzyme stearoyl-acyl carrier protein desaturase. Acc. Chem. Res.
37, 421−429.
■
Technical assistance from Mr. Henry Gong in the ICP-AES
experiments and from Prof. Alan Benesi in the NMR
experiments is gratefully acknowledged.
ABBREVIATIONS USED
(19) Li, N., Nørgaard, H., Warui, D. M., Booker, S. J., Krebs, C., and
Bollinger, J. M., Jr. (2011) Conversion of fatty aldehydes to alka(e)nes
and formate by a cyanobacterial aldehyde decarbonylase: cryptic redox
by an unusual dimetal oxygenase. J. Am. Chem. Soc. 133, 6158−6161.
(20) Das, D., Eser, B. E., Han, J., Sciore, A., and Marsh, E. N. G.
(2011) Oxygen-independent decarbonylation of aldehydes by
cyanobacterial aldehyde decarbonylase: a new reaction of diiron
enzymes. Angew. Chem.-Int. Ed. 50, 7148−7152.
■
2NPH, 2-nitrophenylhydrazide; ACP, acyl carrier protein; AD,
aldehyde decarbonylase; ADO, aldehyde-deformylating oxy-
genase; amu, atomic mass unit; DMSO, dimethylsulfoxide; Ec,
Escherichia coli; GC, gas chromatography; IPTG, isopropyl-β-
D1-thiogalactopyranoside; LB, Luria−Bertani; LC, liquid
chromatography; MeOPMS, 1-methoxy-N-methylphenazine
methosulfate; MS, mass spectrometry; N/F/FR, reducing
system comprising NADPH, ferredoxin, and ferredoxin
reductase from spinach; NMR, nuclear magnetic resonance;
N/PMS, reducing system comprising NADH and N-methyl-
phenazinium methylsulfate; Np, Nostoc punctiforme; Pm,
Prochlorococcus marinus; PMS, phenazine methosulfate or N-
methyl-phenazinium methylsulfate; SIM, single ion monitoring
(21) Eser, B. E., Das, D., Han, J., Jones, P. R., and Marsh, E. N. G.
(2011) Oxygen-independent alkane formation by non-heme iron-
dependent cyanobacterial aldehyde decarbonylase: investigation of
kinetics and requirement for an external electron donor. Biochemistry
50, 10743−10750.
(22) Krebs, C., Price, J. C., Baldwin, J., Saleh, L., Green, M. T., and
Bollinger, J. M., Jr. (2005) Rapid freeze-quench 57Fe Mossbauer
̈
7915
dx.doi.org/10.1021/bi300912n | Biochemistry 2012, 51, 7908−7916

Biochemistry
Article
spectroscopy: monitoring changes of an iron-containing active site
during a biochemical reaction. Inorg. Chem. 44, 742−757.
(23) Bollinger, J. M., Jr., and Krebs, C. (2006) Stalking intermediates
in oxygen activation by iron enzymes: motivation and method. J. Inorg.
Biochem. 100, 586−605.
(24) Krebs, C., Galonic Fujimori, D., Walsh, C. T., and Bollinger, J.
́
M., Jr. (2007) Non-heme Fe(IV)-oxo intermediates. Acc. Chem. Res.
40, 484−492.
(25) Jiang, W., Hoffart, L. M., Krebs, C., and Bollinger, J. M., Jr.
(2007) A manganese(IV)/iron(IV) intermediate in assembly of the
manganese(IV)/iron(III) cofactor of Chlamydia trachomatis ribonu-
cleotide reductase. Biochemistry 46, 8709−8716.
(26) Li, N., Korboukh, V. K., Krebs, C., and Bollinger, J. M., Jr.
(2010) Four-electron oxidation of p-hydroxylaminobenzoate to p-
nitrobenzoate by a peroxodiferric complex in AurF from Streptomyces
thioluteus. Proc. Natl. Acad. Sci. U.S.A. 107, 15722−15727.
(27) Carey, F. A., and Sundberg, R. J. (2007) Advanced Organic
Chemistry: Part A Structure and Mechanisms, 5th ed., Springer, New
York.
(28) Baldwin, J., Krebs, C., Ley, B. A., Edmondson, D. E., Huynh, B.
H., and Bollinger, J. M., Jr. (2000) Mechanism of rapid electron
transfer during oxygen activation in the R2 subunit of Escherichia coli
ribonucleotide reductase. 1. Evidence for a transient tryptophan
radical. J. Am. Chem. Soc. 122, 12195−12206.
(29) Price, J. C., Barr, E. W., Tirupati, B., Bollinger, J. M., Jr., and
Krebs, C. (2003) The first direct characterization of a high-valent iron
intermediate in the reaction of an α-ketoglutarate-dependent
dioxygenase: a high-spin Fe(IV) complex in taurine/α-ketoglutarate
dioxygenase (TauD) from Escherichia coli. Biochemistry 42, 7497−
7508.
(30) Hoffart, L. M., Barr, E. W., Guyer, R. B., Bollinger, J. M., Jr., and
Krebs, C. (2006) A unified mechanism for the α-ketoglutarate-
dependent dioxygenases: evidence for a C-H-cleaving Fe(IV) complex
in prolyl-4-hydroxylase. Proc. Natl. Acad. Sci. U.S.A. 103, 14738−
14743.
(31) Xing, G., Diao, Y., Hoffart, L. M., Barr, E. W., Prabhu, K. S.,
Arner, R. J., Reddy, C. C., Krebs, C., and Bollinger, J. M., Jr. (2006)
Evidence for C-H cleavage by an iron-superoxide complex in the glycol
cleavage reaction catalyzed by myo-inositol oxygenase. Proc. Natl. Acad.
Sci. U.S.A. 103, 6130−6135.
́
(32) Galonic, D. P., Barr, E. W., Walsh, C. T., Bollinger, J. M., Jr., and
Krebs, C. (2007) Two interconverting Fe(IV) intermediates in
aliphatic chlorination by the halogenase CytC3. Nature Chem. Biol.
3, 113−116.
(33) Korboukh, V. K., Li, N., Barr, E. W., Bollinger, J. M., Jr., and
Krebs, C. (2009) A long-lived, substrate-hydroxylating peroxodiiron-
(III/III) intermediate in the amine oxygenase, AurF, from Streptomyces
thioluteus. J. Am. Chem. Soc. 131, 13608−13609.
(34) Matthews, M. L., Neumann, C. S., Miles, L. A., Grove, T. L.,
Booker, S. J., Krebs, C., Walsh, C. T., and Bollinger, J. M., Jr. (2009)
Substrate positioning controls the partition between halogenation and
hydroxylation in the aliphatic halogenase, SyrB2. Proc. Natl. Acad. Sci.
U.S.A. 106, 17723−17728.
(35) Jameson, C. J. (1977) Isotope shift in NMR. J. Chem. Phys. 66,
4983−4988.
(36) Risley, J. M., and Van Etten, R. L. (1979) O-18 Isotope shift
upon C-13 NMR spectra and its application to the study of oxygen-
exchange kinetics. J. Am. Chem. Soc. 101, 252−253.
(37) Hisada, R., and Yagi, T. (1977) 1-Methoxy-5-methylphenazi-
nium methyl sulfate. A photochemically stable electron mediator
between NADH and various electron acceptors. J. Biochem. 82, 1469−
1473.
(38) Eser, B. E., Das, D., Han, J., Jones, P. R., and Marsh, E. N. G.
(2012) Correction to oxygen-independent alkane formation by non-
heme iron-dependent cyanobacterial aldehyde decarbonylase: inves-
tigation of kinetics and requirement for an external electron donor.
Biochemistry 51, 5703.
7916
dx.doi.org/10.1021/bi300912n | Biochemistry 2012, 51, 7908−7916