Evidence from mechanistic probes for distinct hydroperoxide rearrangement mechanisms in the intradiol and extradiol catechol dioxygenases

Article abstract of DOI:10.1021/ja8029569

Three mechanistic probes were used to investigate whether the Criegee rearrangement step of catechol 1,2-dioxygenase (CatA) from Acinetobacter sp. proceeds via a direct 1,2-acyl migration, via homolytic O-O cleavage, or via a benzene oxide-oxepin rearrangement. Incubation of CatA with 3- chloroperoxybenzoic acid led to the formation of a 9:1 mixture of 2-chlorophenol and 3-chlorophenol, via a mechanism involving O-O homolytic cleavage. Incubation of CatA with 2-hydroperoxy-2-methylcyclohexanone led to formation of 5,6-diketoheptan-1-ol, also consistent with an O-O homolytic cleavage mechanism, and not consistent with a direct 1,2-acyl migration. No reaction product was isolated from incubation of CatA with 6-hydroxymethyl-6-methylcyclohexa-2,4- dienone, an analogue that is able to undergo the benzene oxide-oxepin rearrangement, but not able to undergo O-O homolytic cleavage. In contrast, incubation of extradiol dioxygenase MhpB from Escherichia coli with 6-hydroxymethyl-6-methylcyclohexa2,4-dienone led to the formation of a 2-tropolone ring expansion product, consistent with a direct 1,2-alkenyl migration for extradiol cleavage. Taken together, the results imply different mechanisms for the Criegee rearrangement steps of intradiol and extradiol catechol dioxygenases: a direct 1,2-alkenyl migration for extradiol cleavage and an O-O homolytic cleavage mechanism for intradiol cleavage.

Full text of DOI:10.1021/ja8029569

Published on Web 07/16/2008
Evidence from Mechanistic Probes for Distinct Hydroperoxide
Rearrangement Mechanisms in the Intradiol and Extradiol
Catechol Dioxygenases
Meite Xin and Timothy D. H. Bugg*
Department of Chemistry, UniVersity of Warwick, CoVentry CV4 7AL, U.K.
Received April 22, 2008; E-mail: T.D.Bugg@warwick.ac.uk
Abstract: Three mechanistic probes were used to investigate whether the Criegee rearrangement step of
catechol 1,2-dioxygenase (CatA) from Acinetobacter sp. proceeds via a direct 1,2-acyl migration, via
homolytic OsO cleavage, or via a benzene oxide-oxepin rearrangement. Incubation of CatA with
3-chloroperoxybenzoic acid led to the formation of a 9:1 mixture of 2-chlorophenol and 3-chlorophenol, via
a mechanism involving OsO homolytic cleavage. Incubation of CatA with 2-hydroperoxy-2-methylcyclo-
hexanone led to formation of 5,6-diketoheptan-1-ol, also consistent with an OsO homolytic cleavage
mechanism, and not consistent with a direct 1,2-acyl migration. No reaction product was isolated from
incubation of CatA with 6-hydroxymethyl-6-methylcyclohexa-2,4-dienone, an analogue that is able to undergo
the benzene oxide-oxepin rearrangement, but not able to undergo OsO homolytic cleavage. In contrast,
incubation of extradiol dioxygenase MhpB from Escherichia coli with 6-hydroxymethyl-6-methylcyclohexa-
2,4-dienone led to the formation of a 2-tropolone ring expansion product, consistent with a direct 1,2-
alkenyl migration for extradiol cleavage. Taken together, the results imply different mechanisms for the
Criegee rearrangement steps of intradiol and extradiol catechol dioxygenases: a direct 1,2-alkenyl migration
for extradiol cleavage and an OsO homolytic cleavage mechanism for intradiol cleavage.
Introduction
hydroperoxide undergoes different Criegee rearrangements in
the two classes of enzyme, as shown in Figure 1: the intradiol
dioxygenases catalyze a 1,2-acyl migration to form an anhydride
intermediate, while the extradiol dioxygenases catalyze a 1,2-
alkenyl migration to form a lactone intermediate.3
The nonheme iron-dependent catechol dioxygenases catalyze
the oxidative cleavage of catechol substrates as part of aromatic
degradation pathways found in soil bacteria.1,2 The intradiol
catechol dioxygenases, containing a mononuclear iron(III)
cofactor ligated by two tyrosine and two histidine residues,
generate a muconic acid product, whereas the extradiol catechol
dioxygenases, containing a mononuclear iron(II) cofactor ligated
by two histidine and one glutamic acid residue, generate a
2-hydroxymuconaldehyde product.
A long-standing question in the study of these enzymes has
been: what factors control the choice of intradiol vs extradiol
reaction specificity?3 The initial steps of substrate/dioxygen
activation are different: the iron(III) center of the intradiol
dioxygenases is believed to activate the substrate catecholate
dianion for reaction with dioxygen,2 as a high-spin iron(III)
catecholate complex with strong π-character,4 while the iron(II)
center of the extradiol dioxygenases is believed to activate
dioxygen as superoxide to react with a catecholate monoanion.5
In both cases, however, the same proximal hydroperoxide
intermediate is formed, and this intermediate has been observed
directly by protein crystallography in homoprotocatechuate 2,3-
dioxygenase (2,3-HPCD) from BreVibacterium fuscum.6 This
The 1,2-alkenyl migration of the extradiol dioxygenases is
assisted by acid-base catalysis by active-site histidine residues,7
whereas the active sites of the intradiol dioxygenases contain
no outer sphere acid-base residues. Extensive model chemistry
on catechol cleavage by iron complexes has shown that intradiol
cleavage can be catalyzed by a wide range of tripodal tetraden-
tate iron(III) complexes,8 while extradiol cleavage is only
observed in a small number of complexes with facial, tridentate
ligands;9,10 hence, the coordination chemistry of the iron center
clearly influences its reactivity, but exactly how the choice of
acyl vs alkenyl migration is controlled is not clear.
The 1,2-acyl migration rearrangement in the intradiol catechol
dioxygenase mechanism has always been assumed to proceed via
a Criegee rearrangement mechanism, although the acyl group is a
relatively poor migrating group for 1,2-migration reactions. It is
known that Baeyer-Villiger oxidation of 1,2-diketones proceeds
(7) Mendel, S.; Arndt, A.; Bugg, T. D. H. Biochemistry 2004, 43, 13390–
13396.
(1) Bugg, T. D. H. Tetrahedron 2003, 59, 7075–7101.
(2) Que, L., Jr.; Ho, R. Y. N. Chem. ReV. 1996, 96, 2607–2624.
(3) Bugg, T. D. H.; Lin, G. Chem. Commun. 2001, 941–952.
(4) Pau, M. Y. M.; Davis, M. I.; Orville, A. M.; Lipscomb, J. D.; Solomon,
E. I. J. Am. Chem. Soc. 2007, 129, 1944–1958.
(8) Yamahara, R.; Ogo, S.; Masuda, H.; Watanabe, Y. J. Inorg. Biochem.
2002, 88, 284–294.
(9) Lin, G.; Reid, G.; Bugg, T. D. H. J. Am. Chem. Soc. 2001, 123, 5030–
5039.
(5) Spence, E. L.; Langley, G. J.; Bugg, T. D. H. J. Am. Chem. Soc. 1996,
118, 8336–8343.
(10) Bruijnincx, P. C. A.; Lutz, M.; Spek, A. L.; Hagen, W. R.;
Weckhuysen, B. M.; van Koten, G.; Gebbink, R. J. M. K. J. Am. Chem.
Soc. 2007, 129, 2275–2286.
(6) Kovaleva, E. G.; Lipscomb, J. D. Science 2007, 316, 453–457.
9
10422 J. AM. CHEM. SOC. 2008, 130, 10422–10430
10.1021/ja8029569 CCC: $40.75
2008 American Chemical Society

Catechol Dioxygenase Reaction Mechanisms
A R T I C L E S
Figure 1. Mechanistic convergence and divergence of the extradiol and intradiol catechol dioxygenase mechanisms, via processing of a common proximal
hydroperoxide intermediate.
Figure 2. Three possible mechanisms for 1,2-acyl migration reaction during intradiol dioxygenase cleavage: (a) direct 1,2-acyl migration, (b) epoxide
formation, followed by benzene oxide-oxepin ring opening, and (c) OsO bond homolysis. In the case of path c, there is an alternative radical mechanism
involving ꢀ-scission to form an acyl radical; however, recyclization to form muconic anhydride would be needed, since muconic anhydride is a known
intermediate.1,2
via 1,2-acyl migration of a similar hydroperoxide intermediate;11
however, isotope labeling studies have shown that precise mech-
anism of this reaction may not proceed via a direct 1,2-shift.12 There
are two further mechanisms for 1,2-acyl migration, shown in Figure
2: a mechanism involving OsO bond homolysis, for which
precedent exists in reactions of alkyl hydroperoxides with iron
salts,13 and a mechanism involving a benzene oxide-oxepin ring
opening, for which we have previously found a chemical prece-
dent.14 Modeling of the intradiol cleavage reaction catalyzed by
protocatechuate 3,4-dioxygenase has been carried out using hybrid
DFT methods by Borowski and Siegbahn, who found that the
lowest energy pathway in their model proceeded via a Criegee
rearrangement, but that a pathway involving OsO homolytic
cleavage could under some circumstances compete with an
activation energy 2.2 kcal/mol higher.15 In this study, we designed
and investigated three mechanistic probes to distinguish between
these mechanisms, which we tested against catechol 1,2-dioxyge-
nase (CatA) from Acinetobacter sp.; ADP1, an enzyme whose
three-dimensional structure has been determined;16 and 2,3-
dihydroxyphenylpropionate 1,2-dioxygenase (MhpB) from Es-
cherichia coli, an extradiol catechol dioxygenase enzyme that we
have previously studied.5,7
Results
(11) Krow, G. R. Org. React. 1993, 43, 251–798.
(12) Cullis, P. M.; Arnold, J. R. P.; Clarke, M.; Howell, R.; De Mira, M.;
Naylor, M.; Nicholls, D. Chem. Commun. 1987, 1088–1089.
(13) Schreiber, S. L. J. Am. Chem. Soc. 1980, 102, 6163–6165.
(14) Eley, K. L.; Crowley, P. J.; Bugg, T. D. H. J. Org. Chem. 2001, 66,
2091–2097.
Mechanistic Probe 6-Hydroxymethyl-6-methylcyclohexa-2,4-
dienone. We previously synthesized carba analogues of the
proximal hydroperoxide reaction intermediate, based on sub-
stituted cyclohexane skeletons, which acted as inhibitors for
9
J. AM. CHEM. SOC. VOL. 130, NO. 31, 2008 10423

A R T I C L E S
Xin and Bugg
Figure 3. Anticipated reaction products from enzymatic processing of mechanistic probe 2 via 1,2-acyl migration, 1,2-alkenyl migration, or epoxide formation
mechanisms.
extradiol dioxygenase MhpB.17 In this study, we prepared the
corresponding carba analogue on the basis of a cyclohexa-2,4-
dienone skeleton. The 1,1-spiroepoxy cyclohexadienone (1) was
prepared by NaIO4 oxidation of salicyl alcohol, following
literature procedures.18 Treatment of 1 with commercially
available Aspergillus niger sp. epoxide hydrolase was found to
generate the corresponding diol (2), which, although rather
unstable, was observed as a discrete peak by reverse-phase
HPLC at a retention time 18.5 min and was characterized by
NMR spectroscopy and mass spectrometry.
Diol 2 represents an interesting mechanistic probe to inves-
tigate possible rearrangement mechanisms, since 1,2-alkenyl
migration (via a pinacol-type reaction) would generate a 1,2-
tropolone product, whereas 1,2-acyl migration would generate
a 1,3-tropolone product. Diol 2 is not able to undergo OsO
homolytic cleavage but should be able to undergo the benzene
oxide-oxepin rearrangement, which would yield an acyclic
carboxylic acid product (Figure 3).
Catechol 1,2-dioxygenase (CatA) from Acinetobacter sp.
ADP1 was expressed and purified as previously described.19
Incubation of diol 2 with purified CatA gave no new product
by HPLC analysis, even after overnight incubation with enzyme.
However, incubation of 2 with extradiol dioxygenase MhpB
from E. coli rapidly formed a bright pink reaction product.
Analysis of the silylated reaction product by GC-MS identified
the reaction product as 2-tropolone, with identical retention time
and fragmentation to authentic 2-tropolone. Complexes of
2-tropolone with iron salts are known to be brightly colored,20
accounting for the formation of the pink color observed. No
formation of 1,3-tropolone or acyclic reaction products was
detected by HPLC or GC-MS analysis.
triacetic acid,21 known to effect intradiol cleavage. However,
formation of 2-tropolone was detected upon incubation with
iron(II) and iron(III) complexes of 1,4,7-triazacyclononane in
methanol, conditions that are known to effect extradiol cleav-
age.9 The formation of 2-tropolone is consistent with the 1,2-
alkenyl migration reaction believed to occur in the extradiol
reaction mechanism, and the lack of product with CatA indicates
a difference in mechanism.
Mechanistic Probe 3-Chloroperoxybenzoic Acid. The second
mechanistic probe, designed to investigate a possible OsO bond
homolysis mechanism in CatA, is the stable peracid 3-chlorop-
eroxybenzoic acid (MCPBA). If OsO bond homolysis occurs
upon ligation to iron(III), then the resulting carboxyl radical
will rapidly decarboxylate22 to form a 3-chlorophenyl radical,
which upon recombination with an iron(IV)-oxo intermediate
would give 3-chlorophenol, as shown in Figure 4. Incubation
of MCPBA with Acinetobacter CatA for 10 min resulted in the
formation of two new reaction products by HPLC analysis,
shown to be 2-chlorophenol (RT 19.6 min) and 3-chlorophenol
(RT 20.5 min) in a 9:1 ratio. The identity of the reaction products
was confirmed by GC-MS analysis and comparison with
authentic standards. Extended incubation with CatA for 2 h
resulted in the same two products, but in an 8:2 ratio of
2-chlorophenol to 3-chlorophenol. The rates of product forma-
tion over 10 min were calculated from HPLC peak heights as
16.1 µmol/min/mg protein for 2-chlorophenol and 1.8 µmol/
min/mg protein for 3-chlorophenol, compared with a specific
activity of 8.5 µmol/min/mg protein for catechol. Thus, the
processing of MCPBA occurs 2-fold faster than kcat for catechol.
Incubation of MCPBA with E. coli MhpB resulted in the
formation of the same products, but with 3-chlorophenol as
the major product, in a 9:1 ratio of 3- and 2-chlorophenol.
The same products were observed upon treatment of MCPBA
with the iron(II) salt of 1,4,7-triazacyclononane, in a ratio
of 3:1, respectively, while no product formation was observed
with Fe(III)TACN or Fe(III)NTA. In these reactions, a new
species at m/z 281 was also detected by GC-MS, whose
fragmentation pattern is consistent with the structure of a
dimeric peroxy ester, shown in Figure 4.
Diol 2 was also incubated with iron complexes that are known
to catalyze catechol cleavage reactions. No reaction product was
observed upon incubation with the iron(III) complex of nitrilo-
(15) Borowski, T.; Siegbahn, P. E. M. J. Am. Chem. Soc. 2006, 128, 12941–
12953.
(16) Vetting, M. W.; Ohlendorf, D. H. Structure 2000, 8, 429–440.
(17) Winfield, C. J.; Al-Mahrizy, Z.; Gravestock, M.; Bugg, T. D. H.
J. Chem. Soc., Perkin Trans. 1 2000, 3277–3289.
(18) Adler, E.; Brasen, S.; Miyake, H. Acta Chem. Scand. 1971, 25, 2055–
2069.
(19) Strachan, P. D.; Freer, A. A.; Fewson, C. A. Biochem. J. 1998, 333,
741–747.
(21) Weller, M. G.; Weser, U. Inorg. Chim. Acta 1985, 107, 243–245.
(22) Newcomb, M. Tetrahedron 1993, 49, 1151–1176.
(20) Pietra, F. Chem. ReV. 1973, 73, 293–364.
9
10424 J. AM. CHEM. SOC. VOL. 130, NO. 31, 2008

Catechol Dioxygenase Reaction Mechanisms
A R T I C L E S
Figure 4. Processing of 3-chloroperoxybenzoic acid via OsO homolytic cleavage mechanism to form the observed 2- and 3-chlorophenol reaction products.
The proposed structure of a dimeric byproduct is also shown.
Figure 5. Anticipated reaction products from enzymatic processing of mechanistic probe 3 via 1,2-acyl migration or OsO bond homolysis mechanisms.
The formation of 2- and 3-chlorophenol from MCPBA by
intradiol dioxygenase CatA is consistent with the ability to
catalyze OsO bond homolysis. The formation of 2-chlorophenol
implies that an unusual isomerization has taken place. Since
no hydroxylated benzoic or perbenzoic acid products were
detected by GC-MS analysis, we believe that the mechanism
proceeds via a 3-chlorophenyl radical intermediate, which
isomerizes to a 2-chlorophenyl radical via an apparent 1,2-H
shift, as shown in Figure 4; however, the precise mechanism
of this 1,2-shift has yet to be determined. 2-Chlorophenol is
only seen as the major product in the incubation with CatA.
OsO bond homolysis products are also observed using extradiol
dioxygenase MhpB, and with complexes of iron(II) with
tridentate ligand TACN, but not with tetradentate ligand NTA,
indicating that OsO bond homolysis can be carried out by
iron(II) and (III) salts, but that the coordination state of the iron
center is important.
bonyl group adjacent to a hydroperoxide and is therefore able
to undergo 1,2-acyl migration. This compound therefore allows
us to compare the OsO bond homolysis vs acyl migration
reaction pathways with an intradiol catechol dioxygenase. The
possible reaction products arising from 1,2-acyl migration or
OsO homolytic cleavage are illustrated in Figure 5. 1,2-
Migration, followed by elimination, would generate an unsatur-
ated lactone (4). OsO homolytic cleavage would generate an
alkoxy radical, which would undergo CsC fragmentation of
an adjacent CsC bond. Oxygenation of the resulting radical
would generate either 6-ketoheptanoic acid (5) or 5,6-diketo-
heptan-1-ol (6), depending on which CsC bond cleaves.
Incubation of 2-hydroperoxy-2-methylcyclohexanone with
CatA for 30 min led to the formation of a strongly UV-absorbent
peak at retention time 4.4 min by HPLC analysis. At longer
reaction times, this polar reaction product disappeared, and a
new peak at retention time 59.1 min was formed, with λmax 323
nm, as shown in Figure 6. Inspection of the possible reaction
products in Figure 5 reveals only one that has a conjugated
system that would be expected to be UV absorbent: the enol
tautomer 7 of 5,6-diketoheptan-1-ol.
Incubation of 2-Hydroperoxy-2-methylcyclohexanone with
Intradiol Dioxygenase CatA. 2-Hydroperoxy-2-methylcyclo-
hexanone (3) is a stable hydroperoxide that can be prepared by
oxygenation of 2-methylcyclohexanone23 and contains a car-
To confirm the structure of this species, an independent
chemical synthesis was undertaken, as illustrated in Scheme 1.
(23) Cubbon, R. C. P.; Hewlett, C. J. Chem. Soc. C 1968, 24, 2978–2982.
9
J. AM. CHEM. SOC. VOL. 130, NO. 31, 2008 10425

A R T I C L E S
Xin and Bugg
Analysis of 6 by reverse-phase HPLC showed very similar
behavior to the CatA reaction product: in aqueous solution a
peak was initially observed at 4.4 min, followed by formation
of a peak at 59.1 min, as shown in Figure 6. Our interpretation
of the HPLC data is that, in both cases, cyclization of the free
alcohol (6) occurs rapidly in aqueous solution, as observed in
monosaccharide chemistry, to form the six-membered hemiketal,
which can eliminate to form a cyclic enol ether 10, as shown
in Scheme 1B. HRMS analysis of the peak at 59.1 min gives
molecular ion 127.0773, consistent with molecular formula
C7H10O2, consistent with the structure of 10.
The identical retention times and UV-vis spectra of the
synthetic sample confirm the identity of the product from
reaction with CatA. The formation of this product is consistent
with an OsO homolysis reaction pathway, while none of the
possible products arising from direct 1,2-acyl migration were
detected by HPLC or GC-MS analysis. The rate of conversion
of probe 3 by CatA was estimated from the rate of diasappear-
ance of 3 by HPLC as 44 µmol/min/mg of protein. Thus, 3 is
processed 5-fold faster than kcat for catechol (specific activity
8.5 µmol/min/mg of protein).
2-Hydroperoxy-2-methylcyclohexanone (3) was also incu-
bated with 1,4,7-triazacyclononane/FeCl2 in methanol. Under
these conditions, a different reaction product was observed by
HPLC analysis, which was identified by NMR and MS analysis
as 6-ketoheptanoic acid (5), another possible reaction product
arising from OsO homolytic cleavage (Figure 5). None of the
diketone 6 product was detected in this reaction, and none of
the methyl ester of 6-ketoheptanoic acid (5) was observed, which
could be formed by methanolysis of 4.
Figure 6. HPLC analysis of reaction products formed by processing
compound 3 with CatA after 10 min (A) and 30 min (C), compared with
synthetic standard 6 (B, D). (A) Incubation of compound 3 with Cat A (10
min, 220 nm). (B) Synthetic compound 6 (immediate injection, 220 nm).
(C) Incubation of compound 3 with CatA (30 min, 320 nm). (D) Synthetic
compound 6, incubated in pH 7.5 buffer (30 min, 320 nm).
Discussion
Three mechanistic probes were used to investigate three
possible mechanisms for the 1,2-acyl migration step of the
intradiol catechol dioxygenases. The first mechanism (path a,
Figure 2) involves a direct 1,2-acyl migration. Mechanistic
probes 2 and 3 are both able to undergo 1,2-rearrangements
via acyl migration, 2 via a pinacol rearrangement (migration
onto carbon), and 3 via a Criegee rearrangement (migration onto
oxygen), but none of the reaction products expected from 1,2-
acyl migration were observed in either case.
Scheme 1a
The second possible mechanism (path c, Figure 2) involves
OsO bond homolysis, generating an alkoxy radical, which then
undergoes a radical-mediated rearrangement. The transformation
of 3-chloroperoxybenzoic acid to a mixture of 2- and 3-chlo-
rophenol by CatA is consistent with a radical mechanism
initiated by OsO bond homolysis. However, only the CatA
products contained a majority of the 2-chlorophenol product,
arising from an apparent 1,2-aryl H shift. The mechanism of
this 1,2-shift reaction is intriguing and will be the subject of
future studies. The reaction of percarboxylic acids with heme
or nonheme iron centers has been studied previously, and it is
known that mechanisms involving either homolytic or hetero-
lytic cleavage of the OsO bond of the peracid are possible,
depending on the nature of the iron center.24 The same products
were isolated from incubations with extradiol dioxygenase
MhpB and from some model iron(II) and iron(III) complexes,
implying that OsO bond homolysis can be effected by iron(II)
or iron(III) complexes, as is known from the chemical litera-
ture.13 A recent article by Ray et al. shows that OsO bond
a (a) EtPh3P+I-, NaH, 34%. (b) MCPBA, CH2Cl2. (c) NaOH/H2O. (d)
Pyridinium chlorochromate, CH2Cl2. (e) Silica gel, yield 15% for steps b-e.
Wittig reaction of ethylphosphonium iodide with a THP-
protected aldehyde gave alkene 8 as a 2:1 mixture of Z/E
isomers. The alkene was epoxidized by MCPBA, then converted
to the corresponding diol using NaOH, followed by oxidation
of the diol to the diketone 9 with pyridinium chlorochromate.
Purification of the reaction product by silica column chroma-
tography led to the partial deprotection of the THP protecting
group, yielding the target alcohol 6 directly.
(24) Balasubramanian, P. N.; Bruice, T. C. J. Am. Chem. Soc. 1986, 108,
5495–5503.
9
10426 J. AM. CHEM. SOC. VOL. 130, NO. 31, 2008

Catechol Dioxygenase Reaction Mechanisms
A R T I C L E S
Figure 7. Proposed mechanism for intradiol dioxygenase CatA via OsO homolytic cleavage mechanism. It is not certain whether axial tyrosine-200
remains detached from the iron(III) center in the later steps of the mechanism or becomes reattached (as drawn).
homolysis in MCPBA is favored by complexation to iron(III),
whereas complexation by iron(II) favors OsO bond heterolysis,
consistent with our data.25
droxymethyl-6-methylcyclohexa-2,4-dienone (2) to be trans-
formed into an acyclic carboxylic acid product; however, no
such product was observed. Therefore, this mechanism seems
very unlikely.
The conversion of 2-hydroperoxy-2-methylcyclohexanone (3)
to 5,6-diketoheptan-1-ol (6) by CatA is also consistent with an
OsO bond homolysis pathway and not with 1,2-acyl migration.
Incubation of 2-hydroperoxy-2-methylcyclohexanone with FeCl2
and 1,4,7-triazacyclononane in methanol, conditions used for a
biomimetic extradiol cleavage reaction,9 resulted in the forma-
tion of 2-ketoheptanoic acid (5), which is the other expected
product arising from OsO bond homolysis (Figure 5), resulting
from cleavage of the acyl CsC bond, rather than the alkyl CsC
bond. It is interesting that different radical fragmentations occur
under different conditions. Cleavage of R-acylalkoxy radicals
to acyl radicals is precedented in the literature,26 whereas
cleavage to form a less stable primary carbon-centered radical
would appear to be a less favorable fragmentation (Figure 5).
Therefore, it is likely that the transient iron-oxo species present
in the CatA active site is positioned close to the site of the
primary radical that is formed, so that CsC cleavage is
concerted with CsO bond formation. The observation of a
unique OsO bond homolysis product using CatA is consistent
with an OsO homolysis mechanism in the CatA reaction
mechanism.
Taken together, these results imply that CatA is able to
catalyze OsO bond homolysis on mechanistic probes that are
structurally related to the hydroperoxide intermediate involved
in intradiol catechol cleavage, at reaction rates that are 2-5-
fold higher than kcat for the natural substrate catechol, but it is
unable to catalyze a direct 1,2-acyl migration. The results
therefore support a catalytic mechanism involving OsO bond
homolysis, as shown in Figure 7. This conclusion is apparently
at odds with previous theoretical studies by Lehnert et al., who
found a significant activation energy barrier for OsO homolytic
cleavage in a high-spin (TPA)FeIIIsOOR complex,27 whereas
the corresponding low-spin iron(III) complex is readily able to
undergo OsO homolytic cleavage.28 However, modeling studies
on the protocatechuate 3,4-dioxygenase reaction mechanism by
Borowski and Siegbahn revealed that ligand Tyr-408 is able to
assist OsO homolysis via single electron transfer; hence, the
activation energy for OsO homolytic cleavage was only 2.2
kcal/mol higher than that for the Criegee rearrangement path-
way.15 The latter study assumed a protonation of the proximal
oxygen atom by a bound water molecule,15 which will assist
the Criegee rearrangement pathway. Our experimental results
are not consistent with a direct Criegee rearrangement; therefore,
further modeling studies on these mechanisms might reveal the
effects that the precise geometry of the hydroperoxide complex
and the positioning of active-site residues have on the OsO
The third possible mechanism involves the formation of a
benzene oxide intermediate, followed by electrocyclic ring
opening (path b, Figure 2). If this mechanism were operating
in CatA, then one would expect the carba analogue 6-hy-
(25) Ray, K.; Lee, S. M.; Que, L., Jr. Inorg. Chim. Acta 2008, 361, 1066–
(27) Lehnert, N.; Ho, R. Y. N.; Que, L., Jr.; Solomon, E. I. J. Am. Chem.
Soc. 2001, 123, 12802–12816.
1069.
(26) Chatgilialoglu, C.; Crich, D.; Komatsu, M.; Ryu, I. Chem. ReV. 1999,
99, 1991–2069.
(28) MacFaul, P. A.; Ingold, K. U.; Wayner, D. D. M.; Que, L., Jr. J. Am.
Chem. Soc. 1997, 119, 10594–10598.
9
J. AM. CHEM. SOC. VOL. 130, NO. 31, 2008 10427

A R T I C L E S
Xin and Bugg
Figure 8. Possible role of arginine-221 in stabilizing the transition state
for OsO homolytic cleavage.
homolysis versus Criegee rearrangement pathways. Evidence
in favor of an OsO bond homolysis mechanism in model
intradiol cleavage reactions has been reported by Hitomi et al.,
via study of substituent effects and solvent dependence.29
Figure 9. Possible radical mechanism for formation of pyrone catechol
cleavage products from 2,4-ditert-butylcatechol.
Extradiol dioxygenase enzyme MhpB, in contrast, processed
6-hydroxymethyl-6-methylcyclohexa-2,4-dienone(2)into2-tropolo-
ne, via a pinacol rearrangement involving a 1,2-alkenyl migra-
tion. This is entirely consistent with the expected Criegee
rearrangement mechanism for extradiol dioxygenase cleavage,
which is believed to occur through 1,2-alkenyl migration,
assisted by acid-base catalysis.3,7 The different processing of
probe 2 by CatA and MhpB provides a clear difference in
behavior between an intradiol and an extradiol catechol dioxy-
genase. The ring expansion of 2 to a tropolone is of possible
relevance to natural product biosynthesis, since there are a
number of tropolone natural products whose biosynthesis may
proceed via such a ring expansion reaction.30
in the active site of heme peroxidases, which is known to assist
OsO bond cleavage and stabilize the iron (IV)-oxo species.32
Finally, we note that the existence of a radical-based OsO
cleavage mechanism that can operate on the proximal hydro-
peroxide intermediate offers a possible alternative mechanism
for the formation of pyrone products observed in several model
studies with facial tridentate iron(II) complexes, using 3,5-ditert-
butylcatechol as substrate.8 It has been proposed that the pyrone
products arise from loss of CO from the extradiol R-ketolactone
intermediate;33 however, such a decarbonylation reaction oc-
curring on a stable lactone would be mechanistically very
unusual. In contrast, acyl radicals are known to readily decar-
bonylate to form the corresponding alkyl radical.26 A possible
radical mechanism for pyrone formation is shown in Figure 9,
whereby, in the presence of the two bulky t-butyl substituents
and when ligated in a certain conformation, the alkoxy radical
formed upon OsO homolysis undergoes CsC homolytic
cleavage to form an acyl radical. Decarbonylation of the acyl
radical, followed by ring closure, could lead to the formation
of the observed pyrone product.
The implication of this work is that, having formed a proximal
hydroperoxide intermediate, the choice of intradiol vs extradiol
ring cleavage pathways is ultimately determined by a difference
in hydroperoxide rearrangement mechanism. The 1,2-alkenyl
migration of the extradiol dioxygenases requires stereoelectronic
positioning of the hydroperoxide group, relative to the cyclo-
hexadienone ring,3,6 and also requires acid-base catalysis by
active-site residues.7 The 1,2-acyl migration of the intradiol
dioxygenases appears to be mediated by OsO bond homolysis,
which might explain why there are no additional acid-base
residues in the intradiol dioxygenase active site. What factors
might promote OsO bond cleavage, versus other possible
reaction pathways? Higher Lewis acidity of the iron(III) center,
which is known to favor intradiol cleavage activity in model
systems,31 would assist bond cleavage, and/or stabilization of
the iron(IV) oxo species that is thereby formed. There is a
conserved arginine residue in the active site of intradiol
dioxygenases studied to date: Arg-221 in the catechol 1,2-
dioxygenase structure (PDB file 1DLT) is positioned 3.0-3.8
Å above the bound catechol substrate and 4.0 Å from the
iron(III) center.16 The close proximity of this conserved arginine
residue to the bound substrate suggests that it has a role in
catalysis, but its precise function is not known. In a catalytic
mechanism involving OsO homolytic cleavage, this arginine
residue might assist OsO bond homolysis by stabilization of
the anionic transition state and/or by stabilization of the transient
iron (IV)-oxo species formed, as illustrated in Figure 8. A
precedent for this proposal is the distal arginine residue found
In conclusion, these experiments indicate that the Criegee
rearrangements that operate in the intradiol and extradiol
catechol dioxygenases occur via quite different mechanisms.
The 1,2-alkenyl migration occurring in the extradiol dioxyge-
nases requires acid catalysis and a facial tridentate iron(II) center,
and it is assisted by stereoelectronic factors; whereas the
observed results suggest that the 1,2-acyl migration occurring
in the intradiol dioxygenases operates via a homolytic OsO
cleavage mechanism.
Experimental Section
Materials. The Acinetobacter sp. ADP1 catA gene was sub-
cloned from plasmid pBAC557 (gift of Dr. N. Ornston) using
restriction enzymes BamHI and SpeI into a pZErO vector (Invit-
rogen). 6-Hydroxymethyl-6-methylcyclohexa-2,4-dienone (2) was
prepared by the method of Adler et al.18 2-Hydroperoxy-2-
methylcyclohexanone (3) was prepared by the method of Cubbon
and Hewlett.23
Synthesis of 5,6-Diketoheptan-1-ol. 2-(Hept-5-enyloxy)-tet-
rahydro-2H-pyran (8). 5-(Tetrahydro-2H-2-yloxy)pentanal was
prepared by conversion of 1,5-pentanediol to the mono-THP ether,
(29) Hitomi, Y.; Yoshida, H.; Tanaka, T.; Funabiki, T. J. Mol. Catal. A:
Chem. 2006, 251, 239–245.
(32) Hiner, A. N. P.; Raven, E. L.; Thorneley, R. N. F.; Garcia-Canovas,
F.; Rodriguez-Lopez, J. N. J. Inorg. Biochem. 2002, 91, 27–34.
(33) Funabiki, T.; Mizoguchi, A.; Sugimoto, T.; Tada, S.; Tsugi, M.;
Sakamoto, H.; Yoshida, S. J. Am. Chem. Soc. 1986, 108, 2921–2932.
(30) Bentley, R. Nat. Prod. Rep. 2008, 25, 118–138.
(31) Que, L., Jr.; Kolanczyk, R. C.; White, L. S. J. Am. Chem. Soc. 1987,
109, 5373–5380.
9
10428 J. AM. CHEM. SOC. VOL. 130, NO. 31, 2008

Catechol Dioxygenase Reaction Mechanisms
A R T I C L E S
using the procedure of Khan et al.,34 then oxidation using the
procedure of Reetz et al.,35 in 58% overall yield. Dry DMSO (100
mL) was deoxygenated with N2 for 30 min. Sodium hydride (60%
dispersion in mineral oil, 2.85 g, 71.0 mmol) was added, and the
mixture was heated at 75 °C for 5 min, then cooled to room
temperature. Ethyltriphenylphosphonium bromide (26.0 g, 70.0
mmol) was added portionwise, and the mixture was stirred for 45
min. 5-(Tetrahydro-2H-2-yloxy)pentanal (4.7 g, 25.0 mmol) was
then added, resulting in an exothermic reaction, which was stirred
for 30 min. The reaction mixture was then taken up in water (250
mL) and extracted into ether (3 × 100 mL). After removal of solvent
at reduced pressure, the residue was distilled under vacuum to give
respectively. The samples were centrifuged for 10 min at 13000
rpm and then analyzed by HPLC on a Phenomenex Luna C18
reverse-phase column. The mobile phase consisted of A (100%
water) and B (100% methanol) delivered at a flow rate of 0.8 mL/
min. The gradient program was 5% B (5min); 5-100% B (15 min);
100% B (3 min); and 100-5% B (7 min). HPLC retention times:
spiroepoxide 1, 3.7 min; 6-hydroxymethyl-6-hydroxycyclohexa-2,4-
dienone 2, 18.6 min; 3-chlorophenol, 20.5 min; 2-chlorophenol,
18.6 min; synthetic 7-hydroxyheptane-2,3-dione (6), 4.4 min; enol
ether 10, 59.1 min. Spectroscopic data for 6-hydroxymethyl-6-
hydroxycyclohexa-2,4-dienone 2 (400 MHz, D2O) δH: 3.69 and 3.83
(2H, 2 × d, J ) 9.0 Hz, sCH2OH), 6.08 (2H, m, H-2 and H-5),
6.79 (1H, dt, J ) 9.5, 3.5, 3.5 Hz, H-4), 6.86 (1H, dt, J ) 9.5, 5.8,
5.8 Hz, H-3) ppm. 13C NMR (100 MHz, D2O) δC: 69.3, 92.3, 116.3,
121.1, 128.3, 134.5, 201.6 ppm. ESI-MS: [MNa]+ ) 163.
Incubation of Mechanistic Probes with E. coli MhpB. E. coli
MhpB was overexpressed from a recombinant plasmid pIPB
containing the mhpB gene (gift of Dr. J. L. Garcia, CSIC, Madrid),
in E. coli strain DH5R. Overproduction of MhpB was induced by
addition of IPTG (0.5 mM) to a 2-L culture of E. coli DH5R/pIPB
in Luria broth containing 100 µg/mL of ampicillin at OD600 ) 0.6,
followed by growth at 37 °C for 4 h. Following cell lysis, MhpB
was purified to near homogeneity by hydrophobic interaction
chromatography, as previously described,36 using 50 mM potassium
phosphate (pH 7.0) containing 0.07% (v/v) ꢀ-mercaptoethanol as
buffer.
1
8 as a light yellow oil (1.7 g, 8.6 mmol, 34% yield). H and 13C
NMR analysis showed a mixture of E and Z isomers in ratio of
3:1.
Data: Rf 0.31 (2:1 petroleum ether/ethyl acetate). 1H NMR (400
MHz, CDCl3) δH: 1.60 (3H, d, J ) 6.0 Hz, CH3sCdC), 1.35-1.45
(2H, m), 1.50-1.65 (6H, m), 1.65-1.72 (1H, m), 1.7-1.85 (1H,
m), 2.66 (2H, q, J ) 7.3 Hz, sCH2sCdC), 3.39 and 3.74 (2 ×
1H, m, THP CH2O), 3.50 and 3.86 (2 × 1H, m, CH2O), 4.58 (1H,
t, J ) 1.5 Hz, THP acetal CHO), 5.3-5.49 (2H, m, CHdCH) ppm.
13C NMR (100 MHz, CDCl3) δC: 12.7 and 15.5 (Z/E isomers),
19.5, 22.8, 25.7, 28.9, 30.7, 32.3, 62.6, 67.5, 98.83, 123.9, and 124.9
(Z/E isomers), 130.4 and 131.3 (Z/E isomers) ppm. HRMS: m/z
221.1507 (MNa+), calcd 221.1512 for C12H22O2Na.
7-Hydroxyheptane-2,3-dione (6). To a solution of 2-(hept-5-
enyloxy)-tetrahydro-2H-pyran (100 mg, 0.50 mmol) in dichlo-
romethane (5 mL) was added MCPBA (120 mg, 0.7 mmol), and
the reaction was stirred for 1 h. Removal of solvent at reduced
pressure gave the epoxide product as a light yellow oil (98 mg,
m/z (ESI, +ve): 237 [M + Na]+). The epoxide product was then
dissolved in distilled water (10 mL), and 1 M NaOH (1 mL) was
added. After being stirred for 48 h, the reaction mixture was
extracted into diethyl ether. The organic layer was dried (Na2SO4)
and evaporated to give the diol product (255 [M + Na]+). The
residue was then dissolved in dichloromethane (10 mL) at room
temperature, to which was added pyridinium chlorochromate (600
mg), and the mixture was stirred for 3 days. The reaction mixture
was washed with water and dried (Na2SO4), and the solution was
concentrated in vacuo. The product was purified by column
chromatography, eluting with petroleum ether/ethyl acetate (6:1).
7-Hydroxyheptane-2,3-dione (6) was eluted directly from the
column (Rf 0.1, petroleum ether/ethyl acetate, 6:1) and was obtained
as a colorless oil (20 mg, 15% yield). 1H NMR (400 MHz, d6-
acetone) δH: 1.84 (2H, qui, J ) 6.0 Hz), 1.90 (2H, qui, J ) 6.0
Hz), 2.02 (3H, s, CH3COs), 2.56 (2H, t, J ) 7.0 Hz, sCH2COs),
4.33 (2H, t, J ) 5.8 Hz, sCH2OH). 13C NMR (100 MHz, d6-
acetone) δC: 19.1, 22.3, 25.1, 29.9, 69.5, 198.9, 199.2 ppm. HRMS:
m/z 127.0773 ([MHsH2O]+), calcd 127.0754 for C7H11O2.
Incubations with MhpB were carried out in 50 mM potassium
phosphate buffer (pH 8.0, 100 mL), containing either (A) 6-hy-
droxymethyl-6-hydroxycyclohexa-2,4-dienone, generated by treat-
ment of spiroepoxy-2,4-cyclohexadienone solid (10 mM) with 5
mg of Aspergillus niger epoxide hydrolase for 30 min or (B) 10
mM 3-chloroperoxybenzoic acid. To each incubation was added
400 µL (30 units) of MhpB enzyme, which had been preactivated
by addition of ammonium iron(II) sulfate (5 mM) and sodium
ascorbate (5 mM) for 1 min at 0 °C.
Aliquots (5 mL) were removed after 10, 60, 90, 180 min, and
48 h, and the reaction was stopped by adding 300 µL of 100%
TCA after 10, 60, 90, 180 min, and 48 h, respectively. The samples
were centrifuged for 10 min at 13000 rpm and then analyzed by
HPLC on a Phenomenex Luna C18 reverse-phase column. The
mobile phase consisted of A (100% water) and B (100% methanol)
delivered at a flow rate of 0.8 mL/min. The gradient program was
5% B (5 min); 5-100% B (15 min); 100% B (3 min); 100-5% B
(7 min). Samples for GC-MS analysis were extracted into HPLC-
grade dichloromethane, and the organic layer was dried with
Na2SO4. A portion of the product was treated with N,O-bis(trim-
ethylsilyl)acetamide (200 µL) and chlorotrimethylsilane (10 µL)
overnight and then diluted with HPLC-grade dichloromethane and
analyzed by GC-MS under electron impact mode using a Micro-
mass Autospec GC-MS system Varian 4000 (at 70 eV) on a DB5
silica capillary column by using the following temperature gradient:
50 °C for 1 min, 50-140 at 25 °C min-1, and 140-250 at 5 °C
min-1. Data for 2-tropolone (TMS derivative): retention time 6.64
min, m/z 194.2 (MH+, 40%), 180.2 (100%).
Incubation of Mechanistic Probes with Biomimetic Iron
Complexes. Iron(III) Nitrilotriacetate Model Reaction. Mecha-
nistic probes were incubated at 10 mM concentration under the
conditions described by Weller and Weser,21 in the presence of 40
mM Fe(ClO4)3 and 40 mM NTANa3 in 0.3 M borate buffer pH
8.0, for 5 days at room temperature. Products were analyzed by
HPLC or GC-MS, with derivatization as described above.
Iron(II) 1,4,7-Triazanonane Model Reaction. Mechanistic
probes were incubated at 1 0.2 mM concentration under the
conditions described by Lin et al.,9 in the presence of 0.2 mM FeCl2,
0.2 mM 1,4,7-triazacyclononane, and 0.6 mM pyridine in methanol
for 48 h at room temperature. Products were analyzed by HPLC or
GC-MS, with derivatization as described above. Data for 6-keto-
Incubation of Mechanistic Probes with Acinetobacter CatA.
CatA from Acinetobacter sp. ADP1 was expressed from E. coli
Top10/pZErOcatA by induction with 0.5 mM IPTG at OD600
)
0.6 and purified using the method of Strachan et al.19 to give an
enzyme of specific activity 8.6 U/mg. Incubations with CatA were
carried out in 50 mM Tris/HCl buffer (pH 7.5, 100 mL), containing
either (A) 6-hydroxymethyl-6-hydroxycyclohexa-2,4-dienone 2
(NMR data given below), generated by treatment of spiroepoxy-
2,4-cyclohexadienone 1 (10 mM) with 5 mg of Aspergillus niger
epoxide hydrolase for 30 min, (B) 10 mM 3-chloroperoxybenzoic
acid, or (C) 10 mM 2-hydroperoxy-2-methylcyclohexanone 3. To
each incubation was added 800 µL (15 units) of CatA enzyme.
Aliquots (5 mL) were removed after 10, 60, 90, 180 min, and
48 h, and the reaction was stopped by adding 300 µL of 100%
trichloroacetic acid (TCA) after 10, 60, 90, 180 min, and 48 h,
(34) Khan, A. T.; Ghosh, S.; Choudhury, L. H. Eur. J. Org. Chem. 2005,
4891–4896.
(35) Reetz, M. T.; Drewes, M. W.; Schwickardi, R. Org. Synth. 1999, 76,
110.
(36) Bugg, T. D. H. Biochim. Biophys. Acta 1993, 1202, 258–264.
9
J. AM. CHEM. SOC. VOL. 130, NO. 31, 2008 10429

A R T I C L E S
Xin and Bugg
heptanoic acid (5): 1H NMR (400 MHz, d6-acetone) 1.52-1.59
(4H, m), 2.08 (3H, s, sCH3), 2.29 (2H, t, J ) 7.2 Hz,
sCH2COCH3), 2.47 (2H, t, J ) 7.2 Hz, sCH2COOH) ppm. 13C
NMR (100 MHz, d6-acetone) 23.9, 25.1, 29.8, 33.9, 43.3, 177.2,
207.9 ppm. MS (ESI-MS, +ve): 183 ([M + K]+, 10%), 167 ([M
+ Na]+, 100%) HRMS: obsd 167.0681 for MNa+, calcd 167.0679
for C7H12O3Na.
Warwick) for the gift of the pZErOcatA construct, and Dr. John
Bickerton (University of Warwick) for assistance with GC-MS
experiments.
Supporting Information Available: HPLC and GC-MS data
for each of the incubations of the three mechanistic probes with
dioxygenases CatA and MhpB and under nonenzymatic model
reaction conditions. This material is available free of charge
via the Internet at http://pubs.acs.org.
Acknowledgment. We thank Prof. L. N. Ornston (Yale
University) for the gift of plasmid pBAC557 containing the
Acinetobacter sp. catA gene, Dr. Kirstin Eley (University of
JA8029569
9
10430 J. AM. CHEM. SOC. VOL. 130, NO. 31, 2008