Chemical intervention in bacterial lignin degradation pathways: Development of selective inhibitors for intradiol and extradiol catechol dioxygenases

Article abstract of DOI:10.1016/j.bioorg.2015.05.002

Bacterial lignin degradation could be used to generate aromatic chemicals from the renewable resource lignin, provided that the breakdown pathways can be manipulated. In this study, selective inhibitors of enzymatic steps in bacterial degradation pathways

Full text of DOI:10.1016/j.bioorg.2015.05.002

Bioorganic Chemistry 60 (2015) 102–109
Contents lists available at ScienceDirect
Bioorganic Chemistry
journal homepage: www.elsevier.com/locate/bioorg
Chemical intervention in bacterial lignin degradation pathways:
Development of selective inhibitors for intradiol and extradiol
catechol dioxygenases
⇑
Paul D. Sainsbury, Yelena Mineyeva, Zoe Mycroft, Timothy D.H. Bugg
Department of Chemistry, University of Warwick, Coventry CV4 7AL, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 23 March 2015
Available online 9 May 2015
Bacterial lignin degradation could be used to generate aromatic chemicals from the renewable resource
lignin, provided that the breakdown pathways can be manipulated. In this study, selective inhibitors of
enzymatic steps in bacterial degradation pathways were developed and tested for their effects upon lignin
degradation. Screening of a collection of hydroxamic acid metallo-oxygenase inhibitors against two catechol
dioxygenase enzymes, protocatechuate 3,4-dioxygenase (3,4-PCD) and 2,3-dihydroxyphenylpropionate
1,2-dioxygenase (MhpB), resulted in the identification of selective inhibitors D13 for 3,4-PCD (IC50
Keywords:
Lignin valorization
Catechol dioxygenase
Enzyme inhibition
Hydroxamic acid
Aldehyde dehydrogenase
Disulfiram
15 lM) and D3 for MhpB (IC50 110 lM). Application of D13 to Rhodococcus jostii RHA1 in minimal media
containing ferulic acid led to the appearance of metabolic precursor protocatechuic acid at low
concentration. Application of 1 mM disulfiram, an inhibitor of mammalian aldehyde dehydrogenase, to
R. jostii RHA1, gave rise to 4-carboxymuconolactone on the b-ketoadipate pathway, whereas in
Pseudomonas fluorescens Pf-5 disulfiram treatment gave rise to a metabolite found to be glycine betaine
aldehyde.
Ó 2015 Published by Elsevier Inc.
1. Introduction
lignin degradation: vanillic acid (2) has been observed as a lignin
breakdown product in R. jostii and P. putida [8]; lignin-degrading
Lignin is an aromatic heteropolymer found as 15–25% of ligno-
cellulose in plant cell walls, hence is a major component of plant
biomass. The aromatic content of lignin represents an attractive
raw material for conversion to renewable aromatic chemicals, pro-
vided that biocatalytic or chemocatalytic conversion routes can be
found for lignin valorization [1]. Despite research into fungal lignin
degradation since the early 1980s [2], there is currently no com-
mercial process for lignin valorization. We have recently identified
a number of soil bacteria with activity for lignin oxidation [3,4],
and we have identified Dyp-type peroxidase enzymes with activity
for lignin oxidation in Rhodococcus jostii RHA1 [5] and Pseudomonas
fluorescens Pf-5 [6].
Following the initial oxidation of the lignin polymer, the path-
ways for microbial degradation of lignin fragments are only partly
understood [7]. Identification of metabolites from lignin break-
down in R. jostii RHA1 and Pseudomonas putida mt-2 has led to
hypotheses for lignin degradation pathways in these bacteria
[3,7]. Several pieces of evidence indicate that the vanillic acid
degradation pathway, shown in Fig. 1, is involved in bacterial
bacteria isolated from environmental samples can utilise vanillic
acid from growth on minimal media [4]; gene deletion of vanillin
dehydrogenase in R. jostii RHA1 leads to accumulation of vanillin
(1) up to 96 mg/L when grown on minimal media containing wheat
straw lignocellulose [8].
The accumulation of vanillin (1) via gene deletion demonstrates
that intervention in bacterial lignin degradation pathways is a pos-
sible strategy for accumulation of aromatic bioproducts from lignin
[8]. However, genetic modification via gene deletion is only possi-
ble in well characterised genetically tractable bacteria, therefore
we wished to also investigate whether lignin degradation path-
ways could be intercepted via chemical intervention using selec-
tive chemical inhibitors of enzymes on the breakdown pathways.
Oxidative demethylation of vanillic acid (2) generates protocat-
echuic acid (3), which can be metabolized as shown in Fig. 1 either
via intradiol oxidative cleavage by protocatechuate 3,4-dioxygenase
via the b-ketoadipate pathway [7], or via extradiol oxidative
cleavage by protocatechuate 4,5-dioxygenase, a well-characterised
enzyme in Sphingobium SYK6 [9]. Oxidative cleavage of protocate-
chuic acid (3) is therefore
a key branch-point in aromatic
degradation pathways [7]. The ability to selectively block either
ortho- or meta-cleavage would be potentially valuable for pathway
⇑
Corresponding author. Fax: +44 02476 524112.
E-mail address: T.D.Bugg@warwick.ac.uk (T.D.H. Bugg).
http://dx.doi.org/10.1016/j.bioorg.2015.05.002
0045-2068/Ó 2015 Published by Elsevier Inc.

P.D. Sainsbury et al. / Bioorganic Chemistry 60 (2015) 102–109
103
Fig. 1. Vanillic acid degradation pathway, showing enzymes for chemical intervention. Intermediates: 1, vanillin; 2, vanillic acid; 3, protocatechuic acid; 4, ferulic acid; 5,
3-carboxymuconic acid; 6, 4-carboxymuconolactone.

104
P.D. Sainsbury et al. / Bioorganic Chemistry 60 (2015) 102–109
Fig. 2. Chemical structures of hydroxamic acid inhibitors, and disulfiram.
engineering approaches, hence selective inhibitors for either intra-
diol or extradiol catechol dioxygenases would be valuable tools to
investigate chemical intervention in bacterial lignin degradation
pathways. Both families of enzyme use a non-haem iron cofactor
for catalysis, but the intradiol dioxygenases use an iron (III) cofactor
ligated by two tyrosine and two histidine residues, whereas the
extradiol dioxygenases use an iron (II) cofactor, ligated by two his-
tidine and one glutamic acid residue [10].
pressure, and to the residue was added 1 M aqueous sodium
hydroxide solution (50 mL). The mixture was filtered through a
Buchner funnel and acidified with concentrated HCl to pH 4, pre-
cipitating the hydroxamic acid, which was recovered by filtration,
and then recrystallized from hot ethanol to give the product D12 as
pale yellow crystals (4.71 g, 84%).
Data for D12: mp 188–189 °C (lit 188–190 °C [17]); 1H NMR
(300 MHz, d6-DMSO) dH: 7.39 (1H, d, J = 15.0 Hz, CH = CHCO),
7.18 (1H, s, Ar H-2), 7.11 (1H, d, J = 8.0 Hz, Ar H-5), 7.05 (1H, d,
We have previously synthesised a collection of substituted
hydroxamic acid inhibitors for the non-haem iron-dependent car-
otenoid cleavage dioxygenases (see Fig. 2), whose active sites con-
J = 8.0 Hz, Ar H-6), 6.32 (1H, d, J = 15.0 Hz, CH = CHCO), 3.80 (3H,
s), 3.78 (3H, s); 13C NMR (75 MHz, d6-DMSO) dC: 163.7, 150.0,
148.8, 138.4, 127.6, 121.2, 116.6, 111.6, 110.0, 55.3, 55.2; HRMS
(ES+) 246.0737 (MNa+), calc. 246.0737 for C11H13NNaO4.
Data for D13, prepared as above from 4-methoxycinnamic acid
in 81% yield: mp 141–142 °C (lit 143–144 °C [13]); 1H NMR
(300 MHz, d6-DMSO) dH: 7.54 (2H, d, J = 8.0 Hz, Ar H-2), 7.44 (1H,
tain
a mononuclear iron cofactor, similar to the catechol
dioxygenases [11]. In this study we have used this inhibitor library
to identify selective inhibitors for Pseudomonas sp. protocatechuate
3,4-dioxygenase (3,4-PCD), an intradiol dioxygenase. In order to
investigate selectivity for inhibition of intradiol versus extradiol
catechol dioxygenases, we have also screened the collection
d, J = 15.0 Hz, CH = CHCO), 7.01 (2H, d, J = 8.0 Hz, Ar H-3), 6.47
against
Escherichia
coli
2,3-dihydroxyphenylpropionate
(1H, d, J = 15.0 Hz, CH = CHCO), 3.82 (3H, s); 13C NMR (75 MHz,
d6-DMSO) dC: 163.2, 160.3, 138.0, 129.0, 127.3, 116.4, 114.3,
55.0; HRMS (ES+) 216.0627 (MNa+), calc. 216.0631 for
1,2-dioxygenase (MhpB), a Class III extradiol dioxygenase previ-
ously overexpressed in our group [12], that is sequence-related
to protocatechuate 4,5-dioxygenase from Sphingobium SYK-6
[13], and for which mhpBÀ gene knockout strains of E. coli are avail-
able for microbiological testing [14]. We have also investigated
whether disulfiram, a commercially available inhibitor of aldehyde
dehydrogenase [15], could be used to inhibit vanillin dehydroge-
nase and hence generate vanillin as a bioproduct.
C10H11NNaO3.
2.3. Procedure for preparation of D12H, D13H
To a stirred solution of D12 (2.51 g, 11.25 mmol) in methanol
(20 mL) in a 3-necked flask was added 10% palladium/charcoal
(0.25 g). The solution was degassed and flushed twice with nitro-
gen gas, then exposed to hydrogen gas, and stirred under a balloon
of hydrogen for 16 h at room temperature. The reaction mixture
was filtered through Celite, and the solvent removed under
reduced pressure to give D12H as a pale yellow oil (2.11 g, 94%).
Data for D12H: 1H NMR (300 MHz, d6-DMSO) dH: 6.92 (1H, s, Ar
H-2), 6.79 (1H, d, J = 8.0 Hz, Ar H-5), 6.68 (1H, d, J = 8.0 Hz, Ar H-6),
3.72 (3H, s), 3.69 (3H, s), 2.73 (2H, t, J = 7.0 Hz, CH2CO), 2.22 (2H, t,
J = 7.0 Hz, CH2Ar); 13C NMR (75 MHz, d6-DMSO) dC: 168.4, 148.5,
147.0, 133.4, 119.9, 112.1, 111.7, 55.4, 55.3, 34.1, 30.5; HRMS
(ES+) 248.0895 (MNa+), calc. 248.0893 for C11H15NNaO4.
Data for D13H, prepared as above from D13 in 89% yield: mp
109–110 °C, 1H NMR (300 MHz, d6-DMSO) dH: 7.10 (2H, d,
J = 8.0 Hz, Ar H-2), 6.82 (2H, d, J = 8.0 Hz, Ar H-3), 3.71 (3H, s),
2.74 (2H, t, J = 7.0 Hz, CH2CO), 2.20 (2H, t, J = 7.0 Hz, CH2Ar); 13C
NMR (75 MHz, d6-DMSO) dC: 168.3, 161.9, 147.0, 129.1, 113.6,
51.1, 36.8, 34.1; HRMS (ES+) 218.0787 (MNa+), calc. 218.0788 for
2. Materials and methods
2.1. Materials
Hydroxamic acids D1–D13 were prepared as previously
described [11], also included for inhibitor screening were hydrox-
amic acids D20, D21 and D30–32 (structures shown in Supporting
Information Fig. S4), whose synthesis and biological evaluation is
described elsewhere [16]. E. coli 2,3-dihydroxyphenylpropionate
1,2-dioxygenase was purified as previously described [12].
Pseudomonas sp. protocatechuate 3,4-dioxygenase, disulfiram and
other chemicals and biochemicals were purchased from Sigma–
Aldrich Ltd.
2.2. Procedure for preparation of D12, D13
To a stirred solution of 3,4-dimethoxycinnamic acid (5.0 g,
24.1 mmol) in chloroform (50 mL) was added dropwise thionyl
chloride (3.2 g, 26.5 mmol, 1.1 equiv). The reaction mixture was
heated under reflux for 30 min, then cooled, and the solvent evap-
orated under reduced pressure to give the corresponding acyl chlo-
ride, which was dissolved in chloroform (50 mL). The solution of
acyl chloride was added slowly to a solution of hydroxylamine
hydrochloride (2.64 g, 38.0 mmol) and triethylamine (10.45 mL,
74.9 mmol) in chloroform (50 mL), stirred at 0 °C. The reaction
mixture was stirred at 0 °C for 20 min, and then stirred at room
temperature for 40 min. The solvent was removed under reduced
C10H13NNaO3.
2.4. Enzyme assays
Protocatechuate 3,4-dioxygenase was assayed as described by
Fujisawa and Uyeda [18] by measuring the decrease in absorbance
at 290 nm (e290 3.8 Â 103 MÀ1 cmÀ1) in a 1.0 mL total volume in a
quartz cuvette. To a solution of 380
50 mM Tris buffer pH 7.5 was added 15
l
M protocatechuic acid in
g Pseudomonas sp.
3,4-PCD (Sigma–Aldrich), and the change in absorbance at
l

P.D. Sainsbury et al. / Bioorganic Chemistry 60 (2015) 102–109
105
290 nm measured over a 2 min assay. The hydroxamic acid collec-
tion was tested for inhibition at 200 concentration.
water/0.1% trifluoroacetic acid (solvent A) and methanol/0.1% tri-
fluoroacetic acid (solvent B). The applied gradient was 5–25% B
over 15 min; 25% B for 8 min; and 25–70% B for 12 min, at a flow
rate of 1.0 mL/min. Glycine aldehyde betaine (retention time
2.6 min, m/z 102.1 M+, 120.0 hydrate M+) was detected via UV
absorbance at 220 nm, in positive ion mode, and was compared
with an authentic standard. Protocatechuic acid (retention time
15.6 min, m/z 153.3 MÀH) was detected via UV absorbance at
280 nm, in negative ion mode, and was compared with an authen-
tic standard.
lM
Compounds D12, D13 and D12H were then tested at 5–500
lM
inhibitor concentrations in order to measure IC50 values.
2,3-Dihydroxyphenylpropionate 1,2-dioxygenase (MhpB) was
assayed as previously described [12], by measuring the increase
in absorbance at 394 nm (e394 15.6 Â 103 MÀ1 cmÀ1), either in a
250 lL total volume in a 96-well microtitre plate, or in a 1.0 mL
quartz cuvette. To a solution of 2,3-dihydroxyphenylpropionic acid
(100 M, prepared as previously described [12]) in 50 mM potas-
sium phosphate buffer pH 8.0 was added 10 g E. coli MhpB (acti-
l
l
vated by sodium ascorbate and iron (II) sulphate as previously
described [12]), and the change in absorbance at 394 nm measured
over a 2 min assay. The hydroxamic acid collection was tested for
2.7. Preparation of authentic sample of 4-carboxymuconolactone
Pseudomonas sp. protocatechuate 3,4-dioxygenase (Sigma–
Aldrich, 50 lg) was added to a 0.4 mM solution of protocatechuic
inhibition at 250
lM concentration. Compounds D3 and D8 were
then tested at 5–500
l
M inhibitor concentrations in order to mea-
acid in 50 mM Tris acetate buffer pH 7.5 at 37 °C for 1 h. The
solution was then acidified by addition of 1 M HCl, and incubated
for 16 h at room temperature. Analysis by LC–MS (extracted ion
m/z 187.0) showed the presence of two peaks at retention times
8.5 min and 18.1 min, corresponding to 3-carboxymuconic acid
(5) and 4-carboxymuconolactone (6) respectively. Analysis by
GC–MS, after derivatisation by silylation (see above for method),
gave peaks at: 7.8 min, monosilylated 4-carboxymuconolactone
(m/z 243, M – CH3); 9.1 min, disilylated 4-carboxymuconolactone
(m/z 315, M – CH3); 9.5 min, trisilylated protocatechuic acid (m/z
371, M+); 9.6 min, trisilylated carboxymuconic acid (m/z 402, M+).
sure IC50 values.
2.5. Effects of enzyme inhibitors on bacterial fermentations
Cultures of R. jostii RHA1, P. putida mt-2 or P. fluorescens Pf-5
were grown at 30 °C in either Luria-Bertani broth or modified M9
minimal media (0.85 g/L Na2HPO4.12H2O, 1 g/L NH4Cl, 3 g/L
KH2PO4, 0.5 g/L NaCl, 0.24 g/L MgSO4, 0.1 mM CaCl2, 0.1 mM
MnCl2, 0.05 mM FeCl3) containing either 2.5% chopped wheat
straw or pine lignocellulose, or 10 mM ferulic acid/0.1% glucose,
and samples of culture media (1 mL) were withdrawn for analysis
after 24–168 h. The effects of compounds D13 and disulfiram were
3. Results
tested by addition of each compound at t = 0 h at 50 lM (D13) or 1
or 5 mM concentration (disulfiram), and analysed versus control
fermentations lacking inhibitor. Compound D3 was tested at
3.1. Screening of hydroxamic acid collection for inhibitors of
protocatechuate 3,4-dioxygenase
110–330
30 °C on CEM media [15] (5 g/L bactopeptone, 1.5 g/L
Na2HPO4.12H2O, 5 g/L NaCl, 3 g/L yeast extract, 2 g/L -proline)
lM concentration for effects upon E. coli K12 grown at
We investigated whether a selective inhibitor could be devel-
oped for protocatechuate 3,4-dioxygenase, the first enzyme of
the b-ketoadipate pathway, using a collection of substituted
hydroxamic acids, used previously to identify selective inhibitors
for non-haem iron-dependent carotenoid cleavage dioxygenases
[11]. Using commercially available Pseudomonas sp. protocatechu-
ate 3,4-dioxygenase (3,4-PCD), 20 hydroxamic acids were tested
L
containing 1.3 mM phenylpropionic acid. Mutated strain LW366
(mhpBÀ) was tested as a positive control, as previously described
[14].
2.6. Metabolite analysis
for enzyme inhibition at 200 lM concentration (see Supporting
Information Fig. S4), monitoring the disappearance of substrate
at 290 nm. Four compounds, namely D7, D10, D12 and D13, were
Each sample was centrifuged (5 min, 10,000 rpm microcen-
trifuge), and the supernatant was removed and cooled on ice.
Then it was treated with 50 lL of cooled CCl3COOH (1 mg/mL)
and left on ice for 3 min. The precipitate formed was removed by
centrifugation, and the final supernatant was taken and analysed
by reverse phase HPLC. The HPLC gradient was 20–30%
MeOH/H2O over 5 min, then 30–50% MeOH/H2O over 5–12 min,
and finally 50–80% MeOH/H2O over 12–25 min, flow rate
0.5 mL/min, detection at 310 nm.
found to show >50% inhibition at 200
in Table 1.
lM concentration, as shown
Further assays at 100 lM inhibitor concentration confirmed
that compounds D12 and D13 containing an unsubstituted hydrox-
amic acid moiety were the most potent inhibitors of 3,4-PCD. The
saturated analogues D12H and D13H were also synthesised from
the corresponding cinnamic acids as shown in Fig. 3. Compounds
D12 and D13 gave IC50 values of 20 and 15 lM respectively against
3,4-PCD, as shown in Table 1, whereas the hydrogenated analogues
Samples for GC–MS were extracted into dichloromethane
(1 mL), dried with MgSO4 and filtered, using a disposable syringe
filter Whatman 0.2
extracts were further derivatised by treatment with 200
O-bis(trimethylsilyl)acetamide and 10 L of chlorotrimethylsilane.
lm
PTFE W/GMF. Dichloromethane
D12H and D13H were found to be less effective inhibitors, an IC50
lL of N,
value of 160
lM was measured for D12H, and D13H showed only
l
20% inhibition at 200
l
M concentration.
The samples were left overnight and diluted 30-fold in
dichloromethane. GC–MS experiments were carried out on a
Varian 3800 machine, using
a
Varian Factor Four column,
3.2. Testing of compound D13 on R. jostii RHA1
30 m  0.25 mm  0.25 m, with electron impact ionisation (full
l
scan 70–400 m/z, ion trap). The oven temperature was kept at
75 °C for 0–1 min, then raised by 25 °C/min for 9 min to 300 °C,
and then maintained for 9 min.
Compound D13 was then tested for effects on lignin & aromatic
degradation in R. jostii RHA1. 50 lM D13 was added to cultures of
R. jostii in P1 media containing either 2.5% wheat straw lignocellu-
lose, or 10 mM ferulic acid/0.1% glucose, and samples of culture
media were analysed after 48–96 h by LC–MS. Protocatechuic acid
(3, Fig. 1), the substrate for 3,4-PCD, was detected at retention time
15.6 min in the sample taken after 96 h in media containing 10 mM
Samples for LC–MS were evaporated at reduced pressure,
re-suspended in methanol, then injected onto a C18 Zorbax
Eclipse plus (Agilent) reverse phase HPLC column on an Agilent
1200 Series system, and analysed by LC–MS using a Bruker
HTC-Ultra ESI mass spectrometer. The HPLC solvents were
ferulic acid/0.1% glucose and 50 lM D13, and was found to be

106
P.D. Sainsbury et al. / Bioorganic Chemistry 60 (2015) 102–109
Table 1
Inhibition of Pseudomonas sp. protocatechuate 3,4-dioxygenase by hydroxamic acids. ND, not determined. IC50 plots are shown in Supporting Information Fig. S3.
Compound
D7
Structure
% Inhibition @ 200
60
lM
IC50
ND
(lM)
D10
57
ND
20
D12
>90
60
5
D12H
D13
160 10
>90
20
15
5
D13H
ND
Fig. 3. Synthetic route for hydroxamic acid analogues. Reagents and yields: a, SOCl2, CHCl3; b, NH2OHÁHCl, NEt3, combined yields 81–84% c, H2, Pd/C, MeOH, yields 89–94%.
identical to an authentic sample of protocatechuic acid, as shown
in Fig. 4. The concentration of protocatechuic acid was estimated
to be 12 lM (1.8 mg/L) by comparison with authentic standards.
It is known that ferulic acid (4) is converted into vanillic acid (2)
in R. jostii RHA1 via b-oxidation [19]. Using media containing
wheat straw lignocellulose as carbon source, protocatechuic acid
was not detected, implying that its concentration in this case is
too low to detect by LC–MS.
inactive mhpB gene encoding 2,3-dihydroxyphenylpropionate
1,2-dioxygenase generate a red/purple colouration when grown
on a particular CEM medium containing phenylpropionic acid,
due to air oxidation of the catechol metabolite [14]. This coloura-
tion was verified using E. coli mhpBÀ mutant strain LW366 [14]
(see Supporting Information Fig. S2). However, treatment with
compound D3 at 110–330 lM concentrations in media containing
phenylpropionic acid failed to generate the same colouration, sug-
gesting that the compound is not taken up into E. coli cells.
3.3. Screening of hydroxamic acid collection for inhibitors of 2,3-
dihydroxyphenylpropionate 1,2-dioxygenase
3.4. Effects of disulfiram on bacterial lignin-degrading strains
The collection of hydroxamic acid inhibitors was also tested for
inhibition of an extradiol catechol dioxygenase, 2,3-dihydroxyphe-
nylpropionate 1,2-dioxygenase from E. coli, purified from an
overproducing strain [12]. Compounds were tested at 50–250 lM
concentrations, monitoring the formation of the yellow ring fission
product at 394 nm, as shown in Fig. 5 [12].
In this case, enzyme inhibition was observed at 250 lM concen-
tration with two compounds D3 and D8, as shown in Table 2.
Disulfiram is a commercially available inhibitor of mammalian
aldehyde dehydrogenase, whose mechanism proceeds via chemi-
cal modification of an active site cysteine residue, and disulfiram
is used as a treatment for alcohol abuse [15]. The oxidation of
vanillin (1) to vanillic acid (2) on the vanillic acid degradation
pathway (see Fig. 1) is catalysed by an aldehyde dehydrogenase
enzyme, therefore disulfiram might be a useful tool for chemical
intervention in this pathway.
Compound D3 was found to be the more effective inhibitor at
Disulfiram was found to inhibit growth of R. jostii RHA1 on min-
imal M9 media containing 1% wheat straw lignocellulose, however,
growth in rich Luria-Bertani broth containing 1% wheat straw lig-
nocellulose and 1 mM disulfiram was observed, and disulfiram
showed no growth inhibition of P. putida at 1 mM concentration
lower concentrations, and an IC50 value of 110
for D3.
lM was measured
For the E. coli phenylpropionic acid catabolic pathway, it has
been shown previously that E. coli mutant strains containing an

P.D. Sainsbury et al. / Bioorganic Chemistry 60 (2015) 102–109
107
Fig. 4. Detection of protocatechuic acid by LC–MS (extracted ion chromatogram, m/z 153) (a) in 96 h sample from R. jostii RHA1 grown in the presence of 10 mM ferulic acid/
0.1% glucose, treated with 50 lM D13; (b) authentic sample of protocatechuic acid; and (c) negative ion electrospray mass spectrum recorded for peak at retention time
15.6 min, showing the expected [MÀH]À ion at m/z 153.3.
7 days by GC–MS analysis in either organism. However, analysis
of the extract from R. jostii RHA1 after 24 h and 48 h showed the
presence of a new peak at 7.72 min with m/z 243 (see Supporting
Information Fig. S5), which was not present in controls lacking bac-
terial inoculum. In both cases the size of this peak was greatest
after 24 h fermentation, decreasing after 48 h and was not
observed thereafter.
Since the molecular weight of this metabolite corresponds to a
silylated oxidative ring cleavage product of protocatechuic acid,
and since R. jostii RHA1 is known to contain the b-ketoadipate
pathway for oxidative cleavage of protocatechuic acid, we sus-
pected that this metabolite could be either 3-carboxymuconic acid
(5) or 4-carboxymuconolactone (6, see Fig. 1). Authentic samples of
3-carboxymuconic acid and 4-carboxymuconolactone were gener-
ated as shown in Fig. 6. Enzymatic treatment of protocatechuic
acid with protocatechuate 3,4-dioxygenase gave a new GC–MS
peak for 3-carboxymuconic acid (5) at 9.6 min (trisilylated,
Fig. 5. Reaction catalysed by 2,3-dihydroxyphenylpropionate 1,2-dioxygenase.
in either media. Cultures of P. putida and R. jostii were then grown
in Luria-Bertani broth containing either milled wheat straw ligno-
cellulose or pine lignocellulose, at 1% (w/v) concentration, and
1 mM disulfiram, and cell culture supernatant was analysed for
metabolites. No accumulation of vanillin (1) was observed over
Table 2
Inhibition of E. coli 2,3-dihydroxyphenylpropionate 1,2-dioxygenase (MhpB) by hydroxamic acids. ND, not determined. IC50 plots are shown in Supporting Information Fig. S1.
Compound
D3
Structure
% Inhibition @ 250
>90
l
M
IC50
(lM)
110 10
D8
25
ND

108
P.D. Sainsbury et al. / Bioorganic Chemistry 60 (2015) 102–109
Fig. 6. Generation of 3-carboxymuconic acid (5) and 4-carboxymuconolactone (6).
Fig. 7. A. b-ketoadipate pathway involving intermediate 4-carboxymuconolactone (6). B. Choline catabolic pathway involving intermediate glycine betaine aldehyde (7).
m/z 402); further incubation of 5 under acidic conditions gave new
peaks corresponding to lactonisation product 4-carboxymucono-
lactone (6) at retention times 7.8 min (monosilylated, m/z 243)
and 9.1 min (disilylated, m/z 315), thereby confirming that the
observed metabolite was indeed 4-carboxymuconolactone.
bacterial lignin degradation. Using a collection of substituted
hydroxamic acids, members of which were shown previously to
inhibit non-haem iron-dependent carotenoid cleavage dioxyge-
nases [11] and Arabidopsis thaliana p-hydroxyphenylpropionate
dioxygenase [20], we have identified D13 as a selective inhibitor
Treatment of P. fluorescens Pf-5 with 100
lM disulfiram in min-
(IC50 16
lM) of protocatechuate 3,4-dioxygenase, and D3 as a
imal M9 media containing 1% wheat straw lignocellulose also did
not generate vanillin as observed by LC–MS, but was also found
to generate a new metabolite after 24 h fermentation by LC–MS
analysis, at RT 2.6 min (m/z 102, 120). High resolution mass spec-
trometry indicated that the peak at m/z 102.0913 was consistent
with molecular formula C5H12NO, and database searches revealed
a match with glycine betaine aldehyde (mass 102.09), which in
aqueous solution exists mainly as its hydrate (mass 120.10). An
authentic sample of glycine betaine aldehyde (7, see Fig. 7) was
found to match the LC–MS retention time and mass spectrum of
the unknown metabolite (see Supporting Information Fig. S6).
selective inhibitor (IC50 110
nylpropionate 1,2-dioxygenase (MhpB). There was no observable
inhibition of MhpB by D13, and only 30% inhibition of 3,4-PCD
by 200 lM D3 (see Supporting Information Fig. S4), indicating that
it is possible to inhibit either intradiol or extradiol catechol dioxy-
genase enzymes with some selectivity.
There are no published inhibitors for protocatechuate
3,4-dioxygenase, and the only previously published extradiol cate-
chol dioxygenase inhibitors are reaction intermediate analogues
prepared for E. coli MhpB that showed Ki values of 0.7–7.6 mM
[21]. Iron-chelating inhibitors have been prepared for other
non-haem iron-dependent enzymes such as mammalian p-hydro-
xyphenylpropionate dioxygenase [22] and 2-oxoglutarate dioxyge-
nases such as human PHD2 [23]. The selectivity of D12 and D13
may be due to the fact that there is no alkyl N-substituent on their
hydroxamic acid group, suggesting that access to the active site
l
M) of E. coli 2,3-dihydroxyphe-
4. Conclusions
In this work we have identified selective inhibitors for intradiol
and extradiol catechol dioxygenases, which catalyse key steps in

P.D. Sainsbury et al. / Bioorganic Chemistry 60 (2015) 102–109
109
iron (II) cofactor is sterically hindered, while the selectivity of D3
might be due to the presence of a chelating catechol group in the
inhibitor structure.
degradation pathways in organisms for which genetic tools are
not available.
We have investigated whether application of the catechol
dioxygenase inhibitors, or aldehyde dehydrogenase inhibitor disul-
firam, to bacterial lignin degraders could lead to the accumulation
of products from lignin breakdown. Small amounts of protocate-
chuic acid were observed upon treatment of R. jostii with inhibitor
D13, but only when the carbon source was ferulic acid, rather than
lignocellulose, and no substrate accumulation was observed for
inhibitor D3 in E. coli. We have observed previously that ferulic
acid is converted via b-oxidation into vanillic acid in R. jostii
RHA1 [8], and a cluster of genes responsible for metabolism of
ferulic acid in R. jostii has been identified [19]. In the case of
protocatechuic acid, it is possible that, upon accumulation of this
metabolite, R. jostii can convert this compound via decarboxylation
into catechol [24], preventing the accumulation of higher concen-
trations of protocatechuic acid. Hence although some intermediate
accumulation was observed, this strategy for bioproduct
accumulation may be limited both by uptake of the inhibitor into
bacterial cells, and by modulation of metabolic flux upon product
accumulation.
In the case of disulfiram, no accumulation of vanillin was
observed, in contrast to the gene deletion of vanillin dehydroge-
nase in R. jostii RHA1 [8], but two other metabolites were observed.
In R. jostii, accumulation of 4-carboxymuconolactone was observed
on the b-ketoadipate pathway, presumably due to inhibition
of the following bifunctional enzyme 4-carboxymuconolactone
decarboxylase/3-oxoadipate enol-lactone hydrolase (PcaL) [25],
as shown in Fig. 7A. Although there are no known inhibitors for
this enzyme, there is a conserved cysteine residue Cys-191 posi-
tioned close to the active site Asp-198 of the Ser-His-Asp catalytic
triad of 3-oxoadipate enol-lactone hydrolase (see Supporting
Information Fig. S7 for sequence alignment), which might be
susceptible to covalent modification by disulfiram.
In P. fluorescens Pf-5, treatment with disulfiram led to the
accumulation of a metabolite glycine betaine aldehyde unrelated
to lignin breakdown. This metabolite is found on a biosynthetic
pathway established in Arthrobacter pascens [25] for glycine
betaine, a bacterial osmoprotectant, from choline, as shown in
Fig. 7B. Oxidation of choline to glycine betaine aldehyde is
catalysed by choline oxidase, and further oxidation to glycine
betaine is catalysed by betaine aldehyde dehydrogenase [26].
Accumulation of glycine betaine aldehyde would therefore be
caused by inhibition of betaine aldehyde dehydrogenase by
disulfiram, and Zaldivar-Machorro et al. have reported that this
enzyme is inhibited by metabolites of disulfiram in Pseudomonas
aeruginosa, leading to growth inhibition [27].
Acknowledgments
This work was supported by a BBSRC PhD studentship (to P.D.S.)
and a grant from the Technology Strategy Board (Project Number
TP131141). We would like to thank Peter Harrison (University of
Warwick) for assistance with the hydroxamic acid inhibitor collec-
tion, and Goran Mahmoud (University of Warwick) for assistance
with GC–MS analysis.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bioorg.2015.05.
002.
References
[1] J. Zakzeski, P.C. Bruijnincx, A.L. Jongerius, B.M. Weckhuysen, Chem. Rev. 110
(2010) 3552–3599.
[2] R. Ten Have, P.J.M. Teunissen, Chem. Rev. 101 (2001) 3397–3413.
[3] M. Ahmad, C.R. Taylor, D. Pink, K. Burton, D. Eastwood, G.D. Bending, T.D.H.
Bugg, Mol. Biosyst. 6 (2010) 815–821.
[4] C.R. Taylor, E.M. Hardiman, M. Ahmad, P.D. Sainsbury, P.R. Norris, T.D.H. Bugg,
J. Appl. Microbiol. 113 (2012) 521–530.
[5] M. Ahmad, J.N. Roberts, E.M. Hardiman, R. Singh, L.D. Eltis, T.D.H. Bugg,
Biochemistry 50 (2011) 5096–5107.
[6] R. Rahmanpour, T.D.H. Bugg, Arch. Biochem. Biophys. 574 (2015) 93–98.
[7] T.D.H. Bugg, M. Ahmad, E.M. Hardiman, R. Rahmanpour, Nat. Prod. Rep. 28
(2011) 1883–1896.
[8] P.D. Sainsbury, E.M. Hardiman, M. Ahmad, H. Otani, N. Seghezzi, L.D. Eltis,
T.D.H. Bugg, ACS Chem. Biol. 8 (2013) 2151–2156.
[9] E. Masai, Y. Katayama, M. Fukuda, Biosci. Biotechnol. Biochem. 71 (2007) 1–15.
[10] T.D.H. Bugg, C.J. Winfield, Nat. Prod. Rep. 15 (1998) 513–530.
[11] M.J. Sergeant, J.-J. Li, C. Fox, N. Brookbank, D. Rea, T.D.H. Bugg, A.J. Thompson, J.
Biol. Chem. 284 (2009) 5257–5264.
[12] T.D.H. Bugg, Biochim. Biophys. Acta 1202 (1993) 258–264.
[13] E.L. Spence, M. Kawamukai, J. Sanvoisin, H. Braven, T.D.H. Bugg, J. Bacteriol.
178 (1996) 5249–5256.
[14] R.P. Burlingame, L. Wyman, P.J. Chapman, J. Bacteriol. 168 (1986) 55–64.
[15] J.J. Lipsky, M.L. Shen, S. Naylor, Chem. Biol. Interact. 130–132 (2001) 81–91.
[16] P.J. Harrison, S.A. Newgas, F. Descombes, S.A. Shepherd, A.J. Thompson, T.D.H.
Bugg, 2015 (submitted for publication).
[17] K. Tanaka, K. Matsuo, A. Nakanishi, T. Hatano, H. Izeki, Y. Ishida, W. Mori,
Chem. Pharm. Bull. 31 (1983) 2810–2819.
[18] H. Fujisawa, M. Uyeda, Eur. J. Biochem. 45 (1974) 223–231.
[19] H. Otani, Y.-E. Lee, I. Casabon, L.D. Eltis, J. Bacteriol. 196 (2014) 4293–4303.
[20] M.J. Sergeant, P.J. Harrison, R. Jenkins, G.R. Moran, T.D.H. Bugg, A.J. Thompson,
New J. Chem. 37 (2013) 3461–3465.
[21] C.J. Winfield, Z. Al-Mahrizy, M. Gravestock, T.D.H. Bugg, J. Chem. Soc. Perkin
Trans. 1 (2000) 3277–3289.
[22] S. Lindstedt, E. Holme, E.A. Lock, O. Hjalmarson, B. Strandvik, Lancet 340 (1992)
813–817.
[23] N.R. Rose, M.A. McDonough, O.N.F. King, A. Kawamura, C.J. Schofield, Chem.
Soc. Rev. 40 (2011) 4364–4397.
In summary, we have developed selective inhibitors for
intradiol and extradiol catechol dioxygenase enzymes, and we
have shown that inhibition of enzymes on aromatic degradation
pathways can generate metabolites from bacterial lignin
breakdown, however, uptake of enzyme inhibitors into bacterial
cells is a limitation of this approach. The inhibitors developed here
may be useful tools for investigation of bacterial aromatic
[24] H. Hara, L.D. Eltis, J.E. Davies, W.W. Mohn, J. Bacteriol. 189 (2007) 1641–
1647.
[25] D. Eulberg, S. Lakner, L.A. Golovleva, M. Schlömann, J. Bacteriol. 180 (1998)
1072–1081.
[26] K.L. Rozwadowski, G.G. Khachatourians, G. Selvaraj, J. Bacteriol. 173 (1991)
472–478.
[27] V.J. Zaldivar-Machorro, M. Lopez-Ortiz, P. Demare, I. Regla, R.A. Munoz-Clares,
Biochimie 93 (2011) 286–295.