Isofunctional enzymes PAD1 and UbiX catalyze formation of a novel cofactor required by ferulic acid decarboxylase and 4-hydroxy-3-polyprenylbenzoic acid decarboxylase

Article abstract of DOI:10.1021/cb5008103

The decarboxylation of antimicrobial aromatic acids such as phenylacrylic acid (cinnamic acid) and ferulic acid by yeast requires two enzymes described as phenylacrylic acid decarboxylase (PAD1) and ferulic acid decarboxylase (FDC). These enzymes are of interest for various biotechnological applications, such as the production of chemical feedstocks from lignin under mild conditions. However, the specific role of each protein in catalyzing the decarboxylation reaction remains unknown. To examine this, we have overexpressed and purified both PAD1 and FDC from E. coli. We demonstrate that PAD1 is a flavin mononucleotide (FMN)-containing protein. However, it does not function as a decarboxylase. Rather, PAD1 catalyzes the formation of a novel, diffusible cofactor required by FDC for decarboxylase activity. Coexpression of FDC and PAD1 results in the production of FDC with high levels cofactor bound. Holo-FDC catalyzes the decarboxylation of phenylacrylic acid, coumaric acid and ferulic acid with apparent k_cat ranging from 1.4-4.6 s^-1. The UV-visible and mass spectra of the cofactor indicate that it appears to be a novel, modified form of reduced FMN; however, its instability precluded determination of its structure. The E. coli enzymes UbiX and UbiD are related by sequence to PAD1 and FDC respectively and are involved in the decarboxylation of 4-hydroxy-3-octaprenylbenzoic acid, an intermediate in ubiquinone biosynthesis. We found that endogenous UbiX can also activate FDC. This implies that the same cofactor is required for decarboxylation of 4-hydroxy-3-polyprenylbenzoic acid by UbiD and suggests a wider role for this cofactor in metabolism.

Full text of DOI:10.1021/cb5008103

Articles
pubs.acs.org/acschemicalbiology
Isofunctional Enzymes PAD1 and UbiX Catalyze Formation of a Novel
Cofactor Required by Ferulic Acid Decarboxylase and 4‑Hydroxy-3-
polyprenylbenzoic Acid Decarboxylase
Fengming Lin,†,‡,∥ Kyle L. Ferguson,† David R. Boyer,† Xiaoxia Nina Lin,*,‡ and E. Neil G. Marsh*,†,§
‡
§
†Department of Chemistry, Department of Chemical Engineering, and Department of Biological Chemistry, University of
Michigan, Ann Arbor, Michigan 48109, United States
ABSTRACT: The decarboxylation of antimicrobial aromatic acids such as
phenylacrylic acid (cinnamic acid) and ferulic acid by yeast requires two enzymes
described as phenylacrylic acid decarboxylase (PAD1) and ferulic acid
decarboxylase (FDC). These enzymes are of interest for various biotechnological
applications, such as the production of chemical feedstocks from lignin under
mild conditions. However, the specific role of each protein in catalyzing the
decarboxylation reaction remains unknown. To examine this, we have
overexpressed and purified both PAD1 and FDC from E. coli. We demonstrate
that PAD1 is a flavin mononucleotide (FMN)-containing protein. However, it
does not function as a decarboxylase. Rather, PAD1 catalyzes the formation of a
novel, diffusible cofactor required by FDC for decarboxylase activity.
Coexpression of FDC and PAD1 results in the production of FDC with high
levels cofactor bound. Holo-FDC catalyzes the decarboxylation of phenylacrylic
acid, coumaric acid and ferulic acid with apparent kcat ranging from 1.4−4.6 s−1.
The UV-visible and mass spectra of the cofactor indicate that it appears to be a novel, modified form of reduced FMN; however,
its instability precluded determination of its structure. The E. coli enzymes UbiX and UbiD are related by sequence to PAD1 and
FDC respectively and are involved in the decarboxylation of 4-hydroxy-3-octaprenylbenzoic acid, an intermediate in ubiquinone
biosynthesis. We found that endogenous UbiX can also activate FDC. This implies that the same cofactor is required for
decarboxylation of 4-hydroxy-3-polyprenylbenzoic acid by UbiD and suggests a wider role for this cofactor in metabolism.
volatile biproducts pentadiene and styrene.9 Genetic analysis
nzyme-catalyzed decarboxylation constitutes an important
Eclass of biological reactions that are ubiquitous in both
established that two gene products are necessary for the
primary and secondary metabolism. However, decarboxylation
reactions are generally associated with high-energy barriers to
metabolism of cinnamic acid to styrene (Figure 1); these were
reaction because the transition state involves a buildup of
negative charge on the α-carbon. Nature has therefore evolved
decarboxylases that employ a wide variety of catalytic strategies
to facilitate decarboxylation using cofactors such as pyridoxal
phosphate and thiamin pyrophosphate that serve as electron
sinks and Lewis acidic metal ions.1−3
Decarboxylases are also of interest for their potential use as
industrial catalysts in the synthesis of optically pure
intermediates for organic synthesis under mild reaction
conditions.4 They have also been used in fermentation
processes to produce short-chain alcohols and in the
conversion of amino acids to chemical feedstocks such as
styrene and acrylamide.5−7 One such enzyme is ferulic acid
decarboxylase (FDC) from Saccharomyces cerevisiae that has
Figure 1. Decarboxylation reactions catalyzed by PAD1/FDC in yeast
and the related decarboxylation catalyzed by UbiX/UbiD in bacteria as
part of ubiquinone biosynthesis.
been used in conjunction with phenylalanine ammonia lyase to
engineer an E. coli strain capable of producing styrene from
endogenous produced phenylalanine.8
FDC is found in various strains of yeast that detoxify
antimicrobial compounds commonly used as food preservatives
and flavorings, such as sorbic acid and cinnamic acid
(phenylacrylic acid), by decarboxylation to yield CO2 and the
Received: October 7, 2014
Accepted: January 20, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/cb5008103
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Figure 2. Purification and initial characterization of FDC. (A) SDS-PAGE and native PAGE analysis of purified FDC. (B) LC-ESI-MS analysis of
purified FDC. (C) pH dependence of cinnamic acid decarboxylase activity; for assay details, see the main text. The buffers used were pH 6.0−8.0, 50
mM potassium phosphate buffer, 100 mM NaCl and pH 8.0−9.0, 50 mM Tris/Cl, 100 mM NaCl.
designated PAD1 and FDC for phenylacrylic acid decarboxylase
and ferulic acid decarboxylase, respectively.9−11 (Note that the
term PAD is also used in the literature as an abbreviation for
phenolic acid decarboxylase,4,7 leading to confusion between
the two enzymes. However, whereas phenolic acid decarbox-
ylase, which requires no cofactors for activity, catalyzes the
decarboxylation of variously substituted phenolic acrylic acids,
it does not catalyze the decarboxylation of phenylacrylic acid.)
Interestingly, the PAD1 and FDC genes were found to be
homologous to ubiX and ubiD respectively, two genes involved
in the biosynthesis of ubiquinone in a wide range of
bacteria12−14 where they are involved in the decarboxylation
of the aromatic intermediate 4-hydroxy-3-octaprenylbenzoic
acid to 2-octaprenylphenol (Figure 1). However, how FDC/
PAD1 and UbiD/UbiX function together to effect these
decarboxylation reactions remain unknown.
protein migrated with an apparent molecular weight of ∼480
kDa, suggesting that the native protein adopts an oligomeric
structure in solution. Interestingly, FDC purified from E. coli
was active in the absence of PAD1, consistent with previous
studies that found styrene production from engineered E. coli
cells did not require the introduction of the PAD1 gene.8 The
enzyme activity was optimal between pH 7.0 and 8.0 (Figure
2C) and decreased sharply below pH 7.0.
The specific activity of FDC at pH 8.0 as determined by the
GC-MS assay described in Methods was ∼0.54 μmol styrene·
min−1·mg−1 enzyme. Unexpectedly, it was found that dialyzing
FDC against buffer (membrane cutoff = 3.5 kDa) for 24 h
resulted in almost complete loss of activity (Figure 3),
suggesting that FDC required a low molecular weight cofactor
Sequence comparisons have suggested that both UbiX and
UbiD, and by extension PAD1 and FDC, may be flavin-
containing proteins,15,16 but the role of this redox cofactor in
the decarboxylation reaction remains unexplained. Here, we
report the initial biochemical characterization of PAD1 and
FDC. We provide evidence that PAD1 and FDC comprise a
two-component system in which PAD1 synthesizes a novel
cofactor used by FDC. This cofactor is probably a covalently
modified derivative of reduced flavin mononucleotide (FMN).
Our results suggest that UbiX and UbiD constitute a similar
two-component decarboxylase system and that the same flavin-
based cofactor is necessary for the biosynthesis of ubiquinone.
RESULTS AND DISCUSSION
■
Properties of S. cerevisiae FDC Recombinantly Ex-
pressed in E. coli. FDC containing an N-terminal 6-histidine
tag was successfully overexpressed in E. coli BL21 and purified
to homogeneity by standard methods utilizing affinity
chromatography on a nickel-NTA column (Figure 2A). The
molecular weight of the enzyme as determined by LC-ESI-MS
Figure 3. Effect of dialysis and addition of E. coli BL21 cell lysate on
FDC activity. The activity of FDC as purified from E. coli BL21 is
arbitrarily assigned as 100% and corresponds to a specific activity of
∼0.54 μmol styrene·min−1·mg−1 enzyme.
was 57898.15
0.5 Da (Figure 2B), in excellent agreement
with the predicted mass of 57897.8 Da. Analysis of FDC by
native gel electrophoresis (Figure 2A) indicated that the
B
DOI: 10.1021/cb5008103
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
for activity. Consistent with this, addition of a cell-free lysate of
E. coli BL21 was able to completely rescue activity (Figure 3).
However, if the cell-free lysate was first dialyzed against buffer
through a 3.5 kDa cutoff membrane, its ability to activate FDC
was lost. Attempts to activate FDC using commonly
encountered, commercially available biochemical cofactors,
including thiamine pyrophosphate, pyridoxal phosphate,
nucleotides, NADH, NADPH, FMN, and FAD, and various
metal ions were unsuccessful.
UbiX Substitutes for PAD1 in E. coli. Both FDC and
PAD1 are reported to be essential for the decarboxylation of
phenylacrylic acid in vivo in S. cerevisae;9 therefore, we
considered whether the homologous E. coli protein, UbiX,
was substituting for PAD1 in E. coli. To test this, we disrupted
the ubiX gene in E. coli BL21 to create an E. coli BL21/ΔubiX
strain. Expression and purification of FDC from the ΔubiX
strain resulted in protein completely lacking in decarboxylase
activity and cell lysates prepared from this strain were unable to
reactivate dialyzed FDC (Figure 4). However, the inactive
on a nickel-NTA column (Figure 5A). The truncated form of
PAD1 is designated tPAD1. The molecular weight of the
protein determined by LC-ESI-MS was 21808.6 Da (Figure
5B), in excellent agreement with the predicted molecular
weight of 21808.4 Da. The molecular weight of the native
protein was estimated as ∼300 kDa by gel filtration
chromatography (data not shown), indicating that it is an
oligomeric protein in solution.
Purified tPAD1 was yellow and the UV−visible spectrum
exhibited maxima at 384 and 458 nm, characteristic of oxidized
FMN (Figure 5C). The presence of FMN was confirmed by
LC-ESI-MS. FMN was released from tPAD1 by precipitation of
the protein with 1:1 dichloromethane/methanol, and subse-
quent chromatographic analysis of the supernatant identified a
molecule with (m+H+)/z of 457.1 Da (Figure 5D) matching an
authentic standard of FMN, Mr = 456. Interestingly, the
absorption maximum at 384 nm is more intense than that at
458 nm, (Figure 5C) whereas for oxidized FMN the opposite is
true. This suggested that tPAD1 may contain another
chromophore, although no other molecules were identified by
LC-ESI-MS. tPAD1 was unable to catalyze the conversion of
cinnamic acid to styrene, demonstrating that, contrary to initial
genetic analyses,10 the enzyme is not an isofunctional enzyme
with FDC.
The crystal structures of E. coli O157:H7 Pad1p16 and
Pseudomonas aeruginosa UbiX17 have been determined,
revealing them to be a dodecameric proteins containing
FMN. This is consistent with our characterization of PAD1
as an FMN-containing protein that exists as large oligomers in
solution and suggests a common structure and function for
these proteins. Interestingly, Pad1p and UbiX are structurally
related to the family of flavin-dependent peptidyl cysteine
decarboxylases. These include EpiD and MrsD, for which
structures have been determined,18,19 that are involved in the
biosynthesis of the lantibiotic peptides epidermin and
mersacidin, respectively,20 and Dpf that catalyzes the
decarboxylation of 4′-phospho-N-pantothenoyl-cysteine to 4′-
phospho-N-pantetheine in coenzyme A biosynthesis.21−23
However, there is currently no evidence to support either
Pad1p or PAD1 functioning as cysteinyl decarboxylases.
Instead, as discussed below, our results suggest these enzymes
are more likely involved in catalyzing the formation of a
modified form of FMN.
Role of tPAD1 in Activation of FDC. To examine whether
a direct interaction between FDC and PAD1 is necessary for
the activation of FDC, the following experiment was
performed. A dialysis cell was set up in which 1 μM apo-
FDC was introduced on one side of a 3.5 kDa cutoff dialysis
membrane in a buffer containing 6.7 mM cinnamic acid. On the
other side of the membrane purified tPAD1 and/or cell lysates
of either E. coli BL21 or E. coli BL21/ΔubiX were introduced.
At various times, the reaction was analyzed for the production
of styrene by GC-MS (Figure 6). Whereas E. coli ΔubiX cell
lysate was unable to activate FDC, addition of 10 μM purified
tPAD1 to the lysate resulted in the activation of FDC,
indicating that tPAD catalyzes the conversion of a small
molecule in the cell lysate to the cofactor for FDC.
Interestingly, in this experiment, purified tPAD alone was
quite effective at activating FDC in the absence of cell lysate.
This suggests that during purification some of the cofactor
remained bound and subsequently diffused from the enzyme.
The activation of FDC is not due to FMN diffusing from
tPAD1, as FMN itself does not activate FDC (data not shown).
Figure 4. Role of tPAD1 in activation of apo-FDC purified from E. coli
BL21/ΔubiX. Incubation of inactive, apo-FDC with E. coli BL21/
ΔubiX cell lysates and tPAD1 activates the enzyme. The activity of
FDC as purified from E. coli BL21 is arbitrarily assigned as 100%
(column 1 Figure 3).
enzyme could be activated by lysates prepared from either wild-
type E. coli BL21 or by lysates prepared from E. coli BL21/
ΔubiX when purified PAD1 (lacking the N-terminal mitochon-
drial targeting sequence, as discussed below) was also added
(Figure 4). These observations suggested that PAD1 and UbiX
might function to synthesize a low molecular weight cofactor
necessary for the activity of FDC.
Expression and Characterization of PAD1 in E. coli. To
obtain a better understanding of the role of PAD1 in activating
FDC, we overexpressed and purified this enzyme from E. coli.
Initial attempts to express the full-length PAD1 in E. coli
resulted in poor levels of expression and the recombinant
protein lacked the expected FMN cofactor. However, deletion
of the first 58 residues, which appear to encode a mitochondrial
targeting sequence, resulted in the expression of soluble protein
that could be purified by standard methods utilizing an N-
terminal-encoded 6-histidine tag and affinity chromatography
C
DOI: 10.1021/cb5008103
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Figure 5. Purification and initial characterization of tPAD1. (A) SDS-PAGE analysis of purified tPAD1. (B) ESI-MS of purified tPAD1. (C) UV−
visible spectrum of tPAD1 indicative of bound flavin cofactor(s). LS-MS of flavin cofactor isolated from tPAD1, demonstrating the protein contains
FMN.
Initial Characterization of the FDC Cofactor. The
experiments described above suggested that only a small
fraction of FDC isolated from E. coli BL21 contained the
cofactor. Therefore, we attempted to obtain FDC with a higher
fraction of cofactor bound by coexpressing FDC with tPAD1
lacking a his-tag in E. coli BL21 cells. The specific activity of
FDC purified from these cells was much higher: 4.3 0.2 μmol
styrene·min−1·mg−1 enzyme, or ∼8-fold higher than that of
preparations of FDC from E. coli lacking tPAD1. Furthermore,
the increased cofactor content resulted in an enzyme
preparation that was visibly a pale yellow color. The UV−
visible spectrum of the FDC-bound chromophore exhibited an
absorbance maximum at 342 nm characteristic of a reduced
flavin24,25 (Figure 7). However, whereas reduced flavins are
readily reoxidized under the aerobic conditions in which the
protein was handled,26 the chromophore associated with FDC
was stable and its spectrum remained unchanged when it was
released from the protein by denaturation with 1:1 dichloro-
methane/methanol. A preparation of the cofactor, obtained by
precipitation of the protein with 1:1 dichloromethane/
Figure 6. Interaction of tPAD1 and FDC is not required for activation
of FDC. 1 μM FDC in buffer containing 6.7 mM cinnamic acid was
separated from the other components of the experiment by a 3.5 kDa
cutoff dialysis membrane, and the production of styrene was
monitored as a function of time. The components added to the
○
other side of the membrane were ( ) E. coli BL21/ΔubiX cell lysate;
■
●
) E. coli BL21 cell lysate; ( ) purified tPAD1, 10 μM; ( ) E. coli
▲
(
BL21/ΔubiX cell lysate and tPAD1, 10 μM.
Surprisingly, we were unable to detect the presence of the
cofactor by LC-ESI-MS (see discussion below) in samples of
tPAD1. The reason for this is unclear. It is possible that this
reflects the low occupancy of the cofactor in the active site of
tPAD1, as a high concentration tPAD1 relative to FDC was
used in the dialysis experiments.
Figure 7. UV−visible spectrum of holo-FDC. The spectrum is
suggestive of a reduced flavin-like chromophore.
D
DOI: 10.1021/cb5008103
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
methanol, followed by removal of the organic solvent was
capable of activating apo-FDC, thereby confirming that the
yellow chromophore associated with FDC is indeed responsible
for the activity of the enzyme.
The protein-free cofactor was further subjected to LC-ESI-
MS-MS analysis. The major species in the primary MS
spectrum (Figure 8) was characterized a (M+H+)/z =
= 457.1 corresponding to FMN, providing further evidence that
the cofactor comprises a modified form of FMN.
Interestingly, the structures of E. coli UbiD,30 a homologous
enzyme to FDC, and an UbiD-like protein from Pseudomonas
aeruginosa15 have been solved; however, no cofactors are
present in the crystal structures. Although the E. coli enzyme is
a hexamer whereas the P. aeruginosa enzyme is dimeric, they
have very similar tertiary structures, and it is likely that FDC
adopts a similar tertiary structure. The UbiD enzymes comprise
three domains, and it is particularly intriguing that the central
domain possesses a split β-barrel fold characteristic of flavin-
reductases.15 The structures reveal a large cleft in the putative
flavin-binding domain that could potentially accommodate
FMN or a modified FMN derivative. We speculate that under
the conditions used to overexpress these enzymes for
crystallographic analysis, endogenous UbiX could not synthe-
size sufficient cofactor to saturate UbiD and that the cofactor
may also have been lost or degraded during purification. The
unavailability of the substrate of these enzymes, 4-hydroxy-3-
octaprenylbenzoic acid, prevented their assay to determine
whether they were active.
Based on the UV−visible spectrum and its molecular weight,
together the fact that PAD1 contains FMN and that FDC
resembles other flavin-binding enzymes, we propose that the
novel cofactor is likely a modified form of reduced FMN. The
fact that the cofactor is stable to air suggests that it is modified
at N5, as alkylation on nitrogen would stabilize the reduced
flavin ring against reoxidation. Determination of the cofactor’s
structure presents the next challenge to understanding the
reactions catalyzed by PAD1 and FDC. This will likely require
the production of isotopically labeled material to facilitate
NMR experiments or alternatively could be accomplished if the
X-ray structures of holo-FDC or holo-UbiD could be solved.
Steady State Kinetic Properties of FDC. The significant
improvement in the activity of FDC resulting from its
coexpression with tPAD1 allowed the decarboxylation of
cinnamic acid to be followed spectrophotometrically by the
decrease in absorbance at 304 nm. Although this limited the
upper concentration range of substrate that could be studied
before the substrate absorbance became too high, this assay,
being far less cumbersome than the GC-MS-based assay, made
it feasible to determine kcat and Km for cinnamic acid and the
structurally related compounds ferulic acid (trans-(4-hydroxy-3-
methoxy-phenyl)-acrylic acid) and p-coumaric acid (trans-(4-
hydroxyphenyl)-acrylic acid). The kinetic constants for these
substrates, at pH 7.0 and 25 °C in 100 mM potassium
phosphate buffer, are given in Table 1.
Figure 8. High resolution of ESI-MS spectrum of the novel cofactor
for FDC. The less abundant peaks at (m+1)/z = 643.2162 and (m+1)/
z = 625.2054 most likely represent the loss of one and two water
molecules from the most abundant species with (m+1)/z = 661.2279.
Inset: the same spectrum showing a wider spectral window.
661.2212
0.0132, with minor peaks of (m+1)/z =
643.2162 and (m+1)/z = 625.2054 that probably represent
the loss of one and two water molecules respectively from the
parent ion. From this a molecular weight of the cofactor was
deduced to be 660.2212
0.0132. The experimentally
determined molecular weight, together with the intensity of
the natural abundance 13C peak, is consistent with a molecule
possessing between 27−32 carbon atoms. We cautiously note
that we cannot definitively rule out, given its labile nature, that
this species might represent a fragment of the cofactor. The
major (m/z = 661) ion was subjected to fragmentation and
attempts were made to match the resulting MS-MS data with
various databases of small molecules including Chemical
Entities of Biological Interest (ChEBI),27 the Kyoto Encyclo-
pedia of Genes and Genomes (KEGG)28 metabolomics
database, the SciFinder (Chemical Abstracts Service) database,
and the METLIN29 metabolomics database. However, no
convincing matches to compounds in these databases were
found. Further attempts to determine the identity of the FDC
cofactor were hindered by the small amounts of material
available and the instability of the molecule, which appears to
lose activity upon release from the enzyme. Although the
cofactor retained activity under the mildly acid conditions used
in the LC analysis (0.1% formic acid), more strongly acidic (0.1
M HCl for 1 h) or basic (0.1 M NaOH for 1 h) conditions,
followed by neutralization, resulted in the loss of activity. LC-
MS analysis of acid-treated samples revealed a complex mixture
of products that could not be interpreted (data not shown).
However, LC-MS analysis of base-treated samples revealed,
among other fragments, the presence of a peak with (m+H+)/z
Although cinnamic acid proved to be the most active
substrate, FDC decarboxylated the phenolic substrates, ferulic
and coumaric acids, only slightly less efficiently. This suggests
that it may be active with a wider range of substituted
phenylacrylic acid derivatives, which may lend it to industrial
Table 1. Steady State Kinetic Parameters for FDC
a
substrate
kcat (app) (s−1
)
Km (μM)
kcat (app)/Km (s−1M−1
)
cinnamic acid
ferulic acid
4.6 0.2
3.8 0.3
1.5 0.1
180 24
180 41
110 26
25500 3500
21000 4500
13600 880
coumaric acid
a
kcat (app) values were calculated based on the concentration of FDC
protein. As discussed in the text the true kcat is likely significantly
higher due to incomplete cofactor occupancy.
E
DOI: 10.1021/cb5008103
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
uses. Although coexpression of FDC with tPAD1 results in a
large increase in enzyme activity, it appears that the cofactor
occupancy is still only 20−25% of active sites, assuming that the
extinction coefficient for the FDC cofactor is similar to that for
reduced FMN. Therefore, the true kcat values for FDC acting on
these substrates is most likely 4−5 fold higher than the values
reported here.
We note that ferulic acid decarboxylases, also described as
phenolic acid decarboxylases, have been described from a
variety of bacteria, for example, Pseudomonas fluoresencs and
Bacillus subtillis.31,32 However, these enzymes require no
organic cofactors and are dimers of subunit Mr = 20 kDa.
They catalyze the decarboxylation of phenolic acids, such as
coumaric and ferulic acids, through a mechanism involving
deprotonation of the phenolic group to generate a quinone-
methide intermediate. However, they do not catalyze the
decarboxylation of cinnamic acid, in which the phenyl ring is
unactivated. The two classes of FDC enzymes are therefore
mechanistically quite distinct from each other.
Conclusions. The function of PAD1 and FDC as enzymes
involved in the detoxification of aromatic carboxylic acids was
first elucidated through genetic analysis of yeast mutants
deficient in the metabolism of cinnamic acid and its
derivatives.9−11 Homologues of these genes are also present
in a range of bacteria that have been shown to decarboxylate
ferulic and coumaric acids.33 Although sequence similarities
with better-characterized enzymes from other organisms
allowed some predictions as to the structure and cofactor
requirements of these enzymes to be made, the function of
PAD1, which is widely distributed in microbes, has remained
confusing. Thus, E. coli strains overexpressing the PAD1 gene
are unable to decarboxylate phenylacrylic acid derivatives8 and
the homologue of PAD1 from the enterohemorrhagic E. coli
strain O157:H7, which has been overexpressed and purified,
also exhibited no decarboxylase activity with various phenyl-
acrylic acid derivatives.16 These observations made it apparent
that PAD1 is miss-named as a phenylacrylic acid decarboxylase;
however, its actual function remained obscure.
The experiments reported here, which to our knowledge
constitute the first time PAD1 and FDC have been biochemi-
cally characterized, have clarified the function of PAD1 by
showing that it synthesizes a cofactor required by FDC for
decarboxylase activity. Although the lability of the cofactor has
so far prevented us from determining its structure, based on its
UV−visible spectrum, its molecular weight, and the resem-
blance of PAD1 and FDC to other flavin-binding enzymes, we
propose that the cofactor is likely a new, modified form of
reduced FMN. In the absence of a structure, the mechanism by
the cofactor facilitates the decarboxylation of cinnamic acid
remains a mystery, as this type of decarboxylation reaction
appears to lack any obvious precedent in flavo-protein
chemistry.
zoic acid decarboxylase and that it uses the same cofactor as
FDC. The existence of Pad1p, distinct from UbiX, in some E.
coli strains is intriguing as it suggests further biological roles for
this novel cofactor.
METHODS
■
Plasmid and Strain Construction. A derivative strain of E. coli
BL21 in which the ubiX gene was deleted was obtained by P1 phage
transduction as previously described.35,36 The P1 phage lysate was
prepared from the Keio single gene knockout strain ΔubiX::kan and
used to transduce the ubiX knockout into E. coli BL21 Star (DE3) to
obtain a strain designated E. coli BL21/ΔubiX.
The genes encoding FDC (YDR539W) and PAD1 (YDR538W)
from S. cerevisiae were synthesized commercially by GenScript U.S.A.
Inc. (Piscataway, NJ, U.S.A.) and codon-optimized for expression in E.
coli. The synthetic genes FDC and PAD1 were subcloned into the
expression vector pET-28b utilizing the NdeI and BamHI restriction
sites to generate plasmids pFDC and pPAD1. This introduced a
vector-encoded N-terminal 6 histidine-tag to facilitate enzyme
purification. For protein expression, the plasmids were transformed
into the E. coli strains BL21 (DE3) (Invitrogen) and E. coli BL21/
ΔubiX.
To delete the N-terminal mitochondrial targeting peptide sequence
(residues 1−58) from PAD1 a truncated gene was constructed by PCR
using the following oligonucleotides as the forward (5′-
TAACATATGAAACGTATTGTGGTTGCGAT-3′) and reverse pri-
mers (5′-AGTGGATCCTTATTTTGATTTGATACCTTCCC-3′)
and pPAD1 as the template. After amplification, the PCR product
was digested with NdeI and BamHI and ligated with pET28b linearized
with NdeI and BamHI. The resulting plasmid, ptPAD1 was
transformed into E. coli BL21 (DE3) for expression of tPAD1.
For the coexpression of tPAD1 and FDC the two genes were
introduced in tandem into pET28b and expressed from a single vector-
encoded promoter. The tPAD1 gene was amplified by PCR using
forward primer (5′-ATTCCATGGGCAAACGTATTGTGGT-
TGCGAT-3′) and reverse primer (5′-AGTGGATCCTTATTTT-
GATTTGATACCTTCCC-3′) using pPAD1 as the template. After
amplification, the PCR product was digested with NcoI and BamHI
and ligated with pET28b linearized with NcoI and BamHI. The
resulting plasmid, ptPAD1nohis lacks an encoded N- terminal His tag.
The FDC gene was introduced behind the tPAD1 gene and the stop
codon deleted to allow for the incorporation of a C-terminal vector-
encoded 6 histidine tag. This was accomplished by PCR using the
forward primer (5′- ATTGAGCTCGAAGTATAAGAAGG-
AGATATATTCATGCGCAAACTGAACCCGG-3′) which intro-
duced a ribosome binding sequence, shown in bold, and the reverse
primer (5′- AATGTCGACTTTGTAACCGTAGCGTTTCCAGTT-
TTCATT-3′) and using pFDC as the template. The PCR product was
digested with SacI and SalI and ligated with ptPAD1nohis cut with
SacI and SalI to produce ptPAD1-FDC, which was transformed into E.
coli BL21 (DE3) to coexpress both tPAD1 and FDC genes.
Enzyme Expression and Purification. Protein expression was
performed by standard methods. E. coli cultures were grown at 32 °C
in LB medium supplemented with 50 μg/mL kanamycin to an OD600
of 0.6−0.8 and gene expression induced with 0.2 mM isopropyl-D-
thiogalactopyranoside (IPTG) overnight before harvesting the cells by
centrifugation.
PAD1 and FDC share sequence similarities with the E. coli
enzymes UbiX and UbiD, which are involved in decarbox-
ylation of 4-hydroxy-3-octaprenylbenzoic acid, an early step
ubiquinone biosynthesis. (We note that yeast, being eukaryotes,
synthesize ubiquinone by a slightly different pathway34 and that
PAD1 and FDC are not involved in ubiquinone biosynthesis.)
We have shown that UbiX can replace PAD1 in activating FDC
and it was previously shown that PAD1 restores ubiquinone
biosynthesis in a ΔubiX E. coli strain. Therefore, UbiX and
PAD1 appear to be isofunctional enzymes. This strongly
suggests that UbiD is, in fact, the 4-hydroxy-3-octaprenylben-
In a typical purification of FDC or PAD1, 7 g (damp weight) of
cells were resuspended in 50 mL binding buffer (20 mM phosphate
buffer, pH 8.0, 500 mM NaCl, 20 mM imidazole, 5% glycerol, 1 mM
THP). 0.5 mg mL−1 of lysozyme, 1 protease inhibitor tablet (Roche),
1 μL of DNase (Novagen), and 1 mM tris(hydroxylpropyl) phosphine
(THP) were added to the cell suspension and incubated on ice for 1 h
with gentle shaking. The cells were lysed by sonication at maximum
power using 2 s pulses separated by 8 s to prevent overheating for a
total time of 30 min. The supernatant was separated from cell debris
by centrifugation at 15000g at 4 °C for 30 min. Protein purification
̈
was performed using an AKTAexplorer chromatography system. The
supernatant was loaded onto a HisTrap column (GE Healthcare) and
F
DOI: 10.1021/cb5008103
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
the column was washed with buffer (20 mM phosphate buffer, pH 8.0,
containing 500 mM NaCl, 60 mM imidazole, 5% glycerol) at a flow
rate of 1.0 mL/min. The enzyme was eluted from the column with 500
mM imidazole in the same buffer. The resulting fractions were
analyzed by SDS-PAGE on a 15% gel. The fractions containing pure
protein were pooled and passed through PD-10 column (GE
Healthcare) for desalting into PBS buffer plus 5% glycerol.
using small molecular mode and analyzed using Agilent MassHunter
Qualitative Analysis software. MS/MS data was obtained by automatic
MS/MS acquisition. A continuous scan event consisting of one full MS
scan (50−2000 m/z) followed by three data-dependent MS/MS scans
were carried out. The three most intense ions from the initial MS scan
were selected individually for collision-induced dissociation at 50 eV.
GC-MS Enzyme Assay. Routine assays of FDC activity were
carried out in PBS buffer, pH 8.0, with a total volume of 500 μL, using
a saturating concentration of cinnamic acid, 6.7 mM, as the substrate;
the concentration of FDC was typically 1.0 μM. Assays utilizing
tPAD1 typically included this enzyme at 10 μM concentration. The
assay mixtures were incubated at 32 °C for varying lengths of time,
after which the reaction was quenched and the reaction products were
extracted by addition of 500 μL ethyl acetate; samples were vortexed
for 15 min and then centrifuged in a microfuge at 12000 rpm for 20
min to separate the organic and aqueous phases. The amount of
styrene produced was determined by GC-MS. For assays in which the
pH was varied, assays containing 6.7 mM cinnamic acid and 1 μM
FDC were incubated at 32 °C for 10 min. The styrene produced was
then extracted with ethyl acetate and the concentration of styrene was
determined by GC-MS. Control experiments were undertaken to
ensure that the substrate concentration remained saturating
throughout the pH range.
GC-MS analysis was performed using a Shimadzu QP-2010S GC-
MS instrument equipped with a quadrupole mass detector and a DB-5
column (Restek, 30m × 0.25 mm × 0.25 μm). The flow rate of the
helium carrier gas is constant at 1 mL/min and the inlet temperature
was maintained at 200 °C. The interface temperature was maintained
at 250 °C. Injections (10 μL) of the ethyl acetate extracted were made
in splitless mode. The oven temperature was held initially at 40 °C for
3.5 min, gradually increased to 90 °C at 14 °C/min, then gradually
increased from 90 to 315 °C at 20 °C/min, and, finally, maintained at
315 °C for 1 min. Chromatographic data were analyzed by GC-MS
PostRun analysis software. Enzymatic conversion of cinnamic acid to
styrene was quantified using a calibration plot of styrene standards of
known concentration.
AUTHOR INFORMATION
■
Corresponding Authors
*E-mail: nmarsh@umich.edu.
*E-mail: ninalin@umich.edu.
Present Address
∥State Key Laboratory of Bioelectronics, School of Biological
Science and Medical Engineering, Southeast University,
Nanjing 210096, China.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This research was supported in part by grants from the
National Science Foundation, CHE 1152055 (to E.N.G.M.)
and CBET 1336636 (to E.N.G.M. and X.N.L.). F.L. is a
recipient of a PISET fellowship from the University of
Michigan Energy Institute. We thank Dr. Stephen Brown of
the University of Michigan Metabolomics Core Services
supported by grant U24 DK097153 of the National Institutes
of Health (NIH) Common Funds Project to the University of
Michigan for LC-MS analyses of cofactor samples.
REFERENCES
■
(1) Cleland, W. W. (1999) Mechanisms of enzymatic oxidative
decarboxylation. Acc. Chem. Res. 32, 862−868.
(2) Jordan, F., and Patel, H. (2013) Catalysis in enzymatic
decarboxylations: Comparison of selected cofactor-dependent and
cofactor-independent examples. ACS Catal. 3, 1601−1617.
(3) Belcher, J., McLean, K. J., Matthews, S., Woodward, L. S., Fisher,
K., Rigby, S. E. J., Nelson, D. R., Potts, D., Baynham, M. T., Parker, D.
A., Leys, D., and Munro, A. W. (2014) Structure and biochemical
properties of the alkene producing cytochrome P450 OleTJE
(CYP152L1) from the Jeotgalicoccus sp. 8456 bacterium. J. Biol.
Chem. 289, 6535−6550.
(4) Kourist, R., Guterl, J.-K., Miyamoto, K., and Sieber, V. (2014)
Enzymatic decarboxylationAn emerging reaction for chemicals
production from renewable resources. ChemCatChem 6, 689−701.
(5) Straathof, A. J. J. (2014) Transformation of biomass into
commodity chemicals using enzymes or cells. Chem. Rev. 114, 1871−
1908.
(6) Shin, J. H., Kim, H. U., Kim, D. I., and Lee, S. Y. (2013)
Production of bulk chemicals via novel metabolic pathways in
microorganisms. Biotechnol. Adv. 31, 925−935.
(7) Jung, D.-H., Choi, W., Choi, K.-Y., Jung, E., Yun, H., Kazlauskas,
R. J., and Kim, B.-G. (2013) Bioconversion of p-coumaric acid to p-
hydroxystyrene using phenolic acid decarboxylase from B. amylolique-
faciens in biphasic reaction system. Appl. Microbiol. Biotechnol. 97,
1501−1511.
(8) McKenna, R., and Nielsen, D. R. (2011) Styrene biosynthesis
from glucose by engineered E. coli. Metab. Eng. 13, 544−554.
(9) Mukai, N., Masaki, K., Fujii, T., Kawamukai, M., and Iefuji, H.
(2010) PAD1 and FDC1 are essential for the decarboxylation of
phenylacrylic acids in Saccharomyces cerevisiae. J. Biosci. Bioeng. 109,
564−569.
(10) Clausen, M., Lamb, C. J., Megnet, R., and Doerner, P. W.
(1994) Pad1 encodes phenylacrylic acid decarboxylase which confers
resistance to cinnamic acid in Saccharomyces cerevisiae. Gene 142, 107−
112.
Spectroscopic Enzyme Assay. For high activity preparations of
FDC, that is, enzyme prepared from E. coli strains coexpressing
tPAD1, the decarboxylation of substrates could be followed by the
decrease in absorbance of the UV-active substrates. Typical assays
were performed in 100 mM potassium phosphate buffer, pH 7.0, at
room temperature (RT) and contained 250−500 nM FDC. The
substrate concentrations ranged between 0.05 mM and 1.5 mM; at
higher concentrations, the absorbance of the substrate became too
high. Decarboxylase activity was followed by monitoring the decrease
in absorbance at 304 nm for cinnamic acid εo = 1140 M−1cm−1; 334
nm for para-coumeric acid εo = 988 M−1 cm−1; and 344 nm for ferulic
acid εo = 1330 M−1 cm−1. Velocity data were fitted to the Michaelis−
Menten equation using Origin software.
LC-MS Analysis of Proteins and Small Molecules. The
molecular weights of proteins and small molecules were determined
using an Agilent 6520 LC-accurate-mass Q-TOF MS system. The
protein was passed through a desalting column and acidified with 0.1%
formic acid. The sample (5 μL) was injected onto a Poroshell 300SB-
C8 column equilibrated with 0.1% formic acid and 5% acetonitrile.
Proteins were eluted for 5 min with 95% water: 5% acetonitrile
followed by an increasing gradient of acetonitrile to 100% over 7 min
at a flow rate of 0.5 mL/min. Eluting proteins were detected at 280
nm. Mass data were obtained using intact protein mode and analyzed
using Agilent MassHunter Qualitative Analysis software. The raw data
was deconvoluted with respect to maximum entropy. For small
molecule analysis, proteins were precipitated with an equal volume of
1:1 dichloromethane/methanol in glass tubes and removed by
centrifugation at 3000g for 10 min. The organic solvent was removed
under a stream of nitrogen. The sample (20 μL) then was injected
onto a Zorbax 300SB-C18 column equilibrated with 0.1% formic acid
and 5% acetonitrile, followed by a 2 min washing with 95% water/5%
acetonitrile. Then, the gradient of acetonitrile was increased to 100%
over 10 min at a flow rate of 0.4 mL/min. Mass data were obtained
G
DOI: 10.1021/cb5008103
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
(11) Stratford, M., Plumridge, A., and Archer, D. B. (2007)
Decarboxylation of sorbic acid by spoilage yeasts is associated with
the PAD1 gene. Appl. Environ. Microbiol. 73, 6534−6542.
(12) Bentinger, M., Tekle, M., and Dallner, G. (2010) Coenzyme
QBiosynthesis and functions. Biochem. Biophys. Res. Commun. 396,
74−79.
(31) Huang, Z. X., Dostal, L., and Rosazza, J. P. N. (1994)
Purification and characterization of a ferulic acid decarboxylase from
Pseudomonas fluorescens. J. Bacteriol. 176, 5912−5918.
(32) Cavin, J. F., Dartois, V., and Divies, C. (1998) Gene cloning,
transcriptional analysis, purification, and characterization of phenolic
acid decarboxylase from Bacillus subtilis. Appl. Environ. Microbiol. 64,
1466−1471.
(33) Barthelmebs, L., Divies, C., and Cavin, J. F. (2000) Knockout of
the p-coumarate decarboxylase gene from Lactobacillus plantarum
reveals the existence of two other inducible enzymatic activities
involved in phenolic acid metabolism. Appl. Environ. Microbiol. 66,
3368−3375.
(34) Meganathan, R. (2001) Ubiquinone biosynthesis in micro-
organisms. FEMS Microbiol. Lett. 203, 131−139.
(35) Baba, T., Ara, T., Hasegawa, M., Takai, Y., Okumura, Y., Baba,
M., Datsenko, K. A., Tomita, M., Wanner, B. L., and Mori, H. (2006)
Construction of Escherichia coli K-12 in-frame, single-gene knockout
mutants: The Keio collection. Mol. Syst. Biol. 2, 2006.0008.
(36) Miller, J. H. (1992) A Short Course in Bacterial Genetics: A
Laboratory Manual and Handbook for Eschericia coli and Related
Bacteria; Cold Spring Harbor Laboratory Press, Plainview, NY.
(13) Gulmezian, M., Hyman, K. R., Marbois, B. N., Clarke, C. F., and
Javor, G. T. (2007) The role of UbiX in Escherichia coli coenzyme Q
biosynthesis. Arch. Biochem. Biophys. 467, 144−153.
(14) Zhang, H. T., and Javor, G. T. (2003) Regulation of the
isofunctional genes ubiD and ubiX of the ubiquinone biosynthetic
pathway of Escherichia coli. FEMS Microbiol. Lett. 223, 67−72.
(15) Jacewicz, A., Izumi, A., Brunner, K., Schnell, R., and Schneider,
G. (2013) Structural insights into the UbiD protein family from the
crystal structure of PA0254 from Pseudomonas aeruginosa. PLoS One 8,
e63161.
(16) Rangarajan, E. S., Li, Y. G., Iannuzzi, P., Tocili, A., Hung, L. W.,
Matte, A., and Cygler, M. (2004) Crystal structure of a dodecameric
FMN-dependent UbiX-like decarboxylase (Pad1) from Escherichia coli
O157:H7. Protein Sci. 13, 3006−3016.
(17) Kopec, J., Schnell, R., and Schneider, G. (2011) Structure of
PA4019, a putative aromatic acid decarboxylase from Pseudomonas
aeruginosa. Acta Crystallogr., Sect. F: Struct. Biol. Cryst. Commun. 67,
1184−1188.
(18) Blaesse, M., Kupke, T., Huber, R., and Steinbacher, S. (2000)
Crystal structure of the peptidyl-cysteine decarboxylase EpiD
complexed with a pentapeptide substrate. EMBO J. 19, 6299−6310.
(19) Blaesse, M., Kupke, T., Huber, R., and Steinbacher, S. (2003)
Structure of MrsD, an FAD-binding protein of the HFCD family. Acta
Crystallogr. Sect. D: Biol. Crystallogr. 59, 1414−1421.
(20) Majer, F., Schmid, D. G., Altena, K., Bierbaum, G., and Kupke,
T. (2002) The flavoprotein MrsD catalyzes the oxidative decarbox-
ylation reaction involved in formation of the peptidoglycan biosyn-
thesis inhibitor mersacidin. J. Bacteriol. 184, 1234−1243.
(21) Strauss, E., and Begley, T. P. (2001) Mechanistic studies on
phosphopantothenoylcysteine decarboxylase. J. Am. Chem. Soc. 123,
6449−6450.
(22) Strauss, E., Kinsland, C., Ge, Y., McLafferty, F. W., and Begley,
T. P. (2001) Phosphopantothenoylcysteine synthetase from Escher-
ichia coliIdentification and characterization of the last unidentified
coenzyme A biosynthetic enzyme in bacteria. J. Biol. Chem. 276,
13513−13516.
(23) Kupke, T. (2002) Molecular characterization of the 4′-
phosphopantothenoylcysteine synthetase domain of bacterial Dfp
flavoproteins. J. Biol. Chem. 277, 36137−36145.
(24) Ghisla, S., and Massey, V. (1989) Mechanisms of flavoprotein-
catalyzed reactions. Eur. J. Biochem. 181, 1−17.
(25) Fraaije, M. W., and Mattevi, A. (2000) Flavoenzymes: Diverse
catalysts with recurrent features. Trends Biochem. Sci. 25, 126−132.
(26) Massey, V. (1994) Activation of molecular-oxygen by flavins and
flavoproteins. J. Biol. Chem. 269, 22459−22462.
(27) de Matos, P., Alcantara, R., Dekker, A., Ennis, M., Hastings, J.,
Haug, K., Spiteri, I., Turner, S., and Steinbeck, C. (2010) Chemical
entities of biological interest: An update. Nucleic Acids Res. 38, D249−
D254.
(28) Kanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M.,
Itoh, M., Katayama, T., Kawashima, S., Okuda, S., Tokimatsu, T., and
Yamanishi, Y. (2008) KEGG for linking genomes to life and the
environment. Nucleic Acids Res. 36, D480−D484.
(29) Smith, C. A., O’Maille, G., Want, E. J., Qin, C., Trauger, S. A.,
Brandon, T. R., Custodio, D. E., Abagyan, R., and Siuzdak, G. (2005)
METLINA metabolite mass spectral database. Ther. Drug Monit. 27,
747−751.
(30) Zhou, W., Forouhar, F., Seetharaman, J., Fang, Y., Xiao, R.,
Cunningham, K., Ma, L. C., Chen, C. X., Acton, T. B., Montelione, G.
T., Hunt, J. F., and Tong, L. (2006) Crystal structure of 3-octaprenyl-
4-hydroxybenzoate decarboxylase (UbiD) from Escherichia coli,
Northeast Structural Genomics target ER459. Protein Data Bank,
DOI: 10.2210/pdb2idb/pdb.
H
DOI: 10.1021/cb5008103
ACS Chem. Biol. XXXX, XXX, XXX−XXX