Kinetic and mechanistic studies of vanadium-based, extended catalytic lifetime catechol dioxygenases

Article abstract of DOI:10.1021/ja052998+

Recently we showed that V-containing polyoxometalates such as (n-Bu ₄N)₇SiW₉V₃O₄₀ or (n-Bu₄N)₉P₂W₁₅V₃O ₆₂, as well as eight other V-containing precatalysts tested, evolve to high-activity, long catalytic lifetime (≥30 000-100 000 total turnovers) 3,5-di-tert-butylcatechol (DTBC) dioxygenases in which Pierpont's complex [VO(DBSQ)(DTBC)]₂ is apparently a common catalyst resting state [Yin, C.-X.; Finke, R. G. J. Am. Chem. Soc. 2005, 107, 9003-9013]. In a separate paper, autoxidation of DTBC to the corresponding benzoquinone and H ₂O₂ was shown to be a key to the catalyst evolution process: the H₂O₂, DTBC, and O₂ plus virtually any V-based precatalyst tested form [VO(DBSQ)(DTBC)]₂ under the catalytic conditions, that catalyst formation process being autocatalytic in H₂O₂. The resulting novel concept is that of an autoxidation-product-initiated dioxygenase [Yin, C.-X.; Sasaki, Y.; Finke, R. G. Inorg Chem. 2005, in press]. Herein the following questions about this record catalytic lifetime 3,5-di-tert-butylcatechol dioxygenase catalyst are explored: (i) What is the rate law for 3,5-di-tert-butylcatechol dioxygenation when one begins with Pierpont's [VO(DBSQ)(DTBC)]₂? (ii) Does it support the hypothesis that this complex is a catalyst resting state or, perhaps, even the true catalyst? (iii) Can a mechanism be written from that information and from the knowledge in the dioxygenase literature? The results answer each of these questions and provide considerable mechanistic insight into the most catalytically active and long-lived DTBC dioxygenase catalyst presently known.

Full text of DOI:10.1021/ja052998+

Published on Web 09/17/2005
Kinetic and Mechanistic Studies of Vanadium-Based,
Extended Catalytic Lifetime Catechol Dioxygenases
Cindy-Xing Yin and Richard G. Finke*
Contribution from the Department of Chemistry, Colorado State UniVersity,
Fort Collins, Colorado 80523
Received May 7, 2005; E-mail: rfinke@lamar.colostate.edu
Abstract: Recently we showed that V-containing polyoxometalates such as (n-Bu4N)7SiW9V3O40 or (n-
Bu4N)9P2W15V3O62, as well as eight other V-containing precatalysts tested, evolve to high-activity, long
catalytic lifetime (g30 000-100 000 total turnovers) 3,5-di-tert-butylcatechol (DTBC) dioxygenases in which
Pierpont’s complex [VO(DBSQ)(DTBC)]2 is apparently a common catalyst resting state [Yin, C.-X.; Finke,
R. G. J. Am. Chem. Soc. 2005, 107, 9003-9013]. In a separate paper, autoxidation of DTBC to the
corresponding benzoquinone and H2O2 was shown to be a key to the catalyst evolution process: the H2O2,
DTBC, and O2 plus virtually any V-based precatalyst tested form [VO(DBSQ)(DTBC)]2 under the catalytic
conditions, that catalyst formation process being autocatalytic in H2O2. The resulting novel concept is that
of an autoxidation-product-initiated dioxygenase [Yin, C.-X.; Sasaki, Y.; Finke, R. G. Inorg Chem. 2005, in
press]. Herein the following questions about this record catalytic lifetime 3,5-di-tert-butylcatechol dioxygenase
catalyst are explored: (i) What is the rate law for 3,5-di-tert-butylcatechol dioxygenation when one begins
with Pierpont’s [VO(DBSQ)(DTBC)]2? (ii) Does it support the hypothesis that this complex is a catalyst
resting state or, perhaps, even the true catalyst? (iii) Can a mechanism be written from that information
and from the knowledge in the dioxygenase literature? The results answer each of these questions and
provide considerable mechanistic insight into the most catalytically active and long-lived DTBC dioxygenase
catalyst presently known.
Scheme 1. Oxidative Cleavage of DTBC by Vanadium-Containing
Introduction
Polyoxometalate Precatalystsa
Recently we extended our 1999 findings1 by reporting2 that
virtually any vanadium-based precursor tested evolved into a
highly effective dioxygenase-type oxidative cleavage catalyst
for 3,5-di-tert-butylcatechol (hereafter DTBC) dioxygenation
with O2 to yield the products shown in Scheme 1.1-3 A record
g100 000 total turnover (TTO) catalytic lifetimes1 were ob-
served en route to oxidative ring-cleavage products, 2-5, plus
the autoxidation product, 6 (Scheme 1).1-3
Electron paramagnetic resonance (EPR) spectroscopy, nega-
tive ion electrospray ionization mass spectrometry (ESI-MS),
catalytic activity, selectivity, and lifetime studies revealed that
Pierpont’s crystallographically characterized vanadium semi-
quinone catecholate dimer complex, [VO(DBSQ)(DTBC)]2
(where DBSQ is 3,5-di-tert-butylsemiquinone and DTBC is the
3,5-di-tert-butylcatecholate dianion), is either the catalyst resting
state or possibly the true V-based catalyst.2 In very recent studies
we have gone on to show that the key to the catalyst evolution
process, from virtually any V-based precatalyst tested, is the
formation of a novel, autoxidation-product-initiated dioxygenase
catalyst. In that reaction, initial DTBC autoxidation to 3,5-di-
tert-butyl-1,2-benzoquinone also produces H2O2 which, in turn,
autocatalytically results in the formation of catalytically active
[VO(DBSQ)(DTBC)]2.3
a The established products1 are 2, 3,5-di-tert-butyl-1-oxacyclohepta-3,5-
diene-2,7-dione; 3, 4,6-di-tert-butyl-2H-pyran-2-one; 4, spiro[1,4-benzo-
dioxin-2(3H),2′(2H)-pyran]-3-one, 4′,6,6′,8-tetrakis(1,1-dimethylethyl); 5,
3,5-di-tert-butyl-5-(carboxymethyl)-2-furanone; and 6, 3,5-di-tert-butyl-1,2-
benzoquinone.
Herein we report kinetic studies, beginning with the [VO-
(DBSQ)(DTBC)]2 dimer, (i) which show that this dimer is the
catalyst resting state, (ii) which show that dissociation of the
dimer to monomeric VO(DBSQ)(DTBC) is required for reaction
with O2, and (iii) which allow, along with insights from classic
literature studies of other dioxygenases,4,5 a mechanism of action
(4) Jang, H. G.; Cox, D. D.; Que, L., Jr. J. Am. Chem. Soc. 1991, 113, 9200-
9204.
(1) Weiner, H.; Finke, R. G. J. Am. Chem. Soc. 1999, 121, 9831-9842.
(2) Yin, C.-X.; Finke, R. G. . J. Am. Chem. Soc. 2005, 107, 9003-9013.
(3) Yin, C.-X.; Sasaki, Y.; Finke, R. G. Inorg. Chem. 2005, in press.
(5) (a) Winfield, C. J.; Al-Mahrizy, Z.; Gravestock, M.; Bugg, T. D. H. Perkin
1 2000, 3277-3289. (b) Bugg, T. D. H.; Lin, G. Chem. Commun. 2001,
941-952.
9
13988
J. AM. CHEM. SOC. 2005, 127, 13988-13996
10.1021/ja052998+ CCC: $30.25 © 2005 American Chemical Society

V-Based, Extended Catalytic Lifetime Catechol Dioxygenases
A R T I C L E S
(∼1.7 atm), 15 s per purge, followed by equilibration for 1 min 15 s;
5 min total time elapsed before the pressure recordings were started.
Note that the atmospheric pressure at the ca. 1-mile-high altitude of
Fort Collins, Colorado, is around 632 Torr, or 0.83 atm. The reaction
vessel was then pressurized to 13 ( 1 psig and t ) 0 was set. The
results are shown in Figures 1-4.
to be written for V-based DTBC dioxygenases. The results
simplify and unify the previously disparate V-based dioxygenase
literature. Moreover, since the catalytic lifetimes are without
precedent in prior man-made or proVen lifetime enzymatic
dioxygenases (see footnote 27 in ref 1), the results are of
fundamental significance in providing insight into man-made,
highly catalytic, long-lived and mechanistically understood
dioxygenase catalysts, the Holy Grail of oxidation catalysis.6
The O2 rate vs catalyst concentration, gas-to-solution mass-transfer
limitation plot is included in the Supporting Information as Figure S1.
Those data, obtained at a stirring rate of 1200 rpm, indicate a clean
first-order catalyst dependence up to ca. 0.1 mM, so O2 mass-transfer
limitations should not be a concern for the kinetic experiments leading
to the reaction orders in substrate and O2 which were performed at ca.
0.05 mM and a 1200 rpm stirring rate. In all cases, the raw kinetic
data from the O2-uptake kinetics obtained with the pressure transducer
were corrected for the contribution of the solvent vapor pressure in
the early part of the curves (i.e., after flushing with O2 to start the
reaction and prior to those curves coming to equilibrium). This
correction was done exactly as detailed earlier3 via the measurement
of the solvent-vapor-pressure curve in an independent experiment. (An
exemplary curve is provided in the Supporting Information as Figure
S2, along with additional details of how the correction for the solvent
vapor pressure was accomplished.3) The corrected, typically initially
linear O2 pressure-time plots (e.g., Figures 1, S3, and S4) were
analyzed by linear regression of their initial linear portion to obtain
initial rate data.
Experimental Section
Reagents. DTBC (Aldrich, 99%) was recrystallized three times using
n-pentane (Fisher Scientific, 98%, pesticide grade) under argon (melting
point 99-100 °C, lit. mp 96-99 °C) and stored in a Vacuum
Atmosphere drybox (O2 level e 5 ppm). (NB: It is important to
recrystallize the DTBC substrate g2 times to remove impurities like
3,5-di-tert-butylbenzoquinone, due to effects such as shown in Figure
2 in ref 3.) 1,4,7-Triazacyclononane (Strem, 97%) and 1,4,7-trimethyl-
1,4,7-triazacyclononane (Strem, 97% min.) were stored in the drybox
and used as received. HPLC-grade solvent (1,2-dichloroethane) was
purchased from Aldrich and stored in the drybox; the above solvent
was dried by standing for at least 48 h over ∼5 vol % 3- or 4-Å
molecular sieves which were preactivated by heating at 170 °C under
vacuum for at least 12 h, then cooling under dry N2 in the drybox.
Anhydrous-grade toluene (Aldrich) was used as received and stored in
the drybox.
Reaction Rate of [VO(DBSQ)(DTBC)]2 with O2 As Estimated
by UV-Visible Spectroscopy. The rate of reaction of [VO(DBSQ)-
(DTBC)]2 plus O2 was estimated by following the decay of the
absorbance peak at 684 nm of [VO(DBSQ)(DTBC)]2 (0.15 mol in 3
mL of 1,2-C2H4Cl2, 0.05 mM) after O2 was bubbled into the UV cell
which had been prewarmed to ca. 40 °C. The initial rate of reaction,
obtained from the first three data points over 20 min of reaction, is 1.9
h-1 (pseudo-first-order rate constant when assuming O2 is in excess,
[O2]solution ≈ 5 mM9 at 293.2 K, 0.8 atm O2; the initial second-order
rate constant is ca. 380 M-1 h-1 assuming a second-order rate law and
using the [O2]solution ≈ 5 mM).
Instrumentation. GC analyses were performed on a Hewlett-Packard
(HP) 5890 Series II gas chromatograph equipped with an FID detector
and an SPB-1 capillary column (30 m, 0.25 mm i.d.) with the following
temperature program: initial temperature, 200 °C (initial time, 2 min);
heating rate, 2 °C/min; final temperature, 240 °C (final time, 3 min);
FID detector temperature, 250 °C; injector temperature, 250 °C. An
injection volume of 1 µL was used. UV-visible spectra were obtained
on an HP 8452A diode spectrophotometer in glass UV cells. Electron
paramagnetic resonance (EPR) spectra were recorded on a Bruker EMX
200U EPR spectrometer. Quartz EPR tubes of 4-mm o.d. were used,
and 2,2-diphenyl-1-picrylhydrazyl (DPPH) was used as the reference
compound (g ) 2.0037).
Kinetic Curve-Fitting Using MacKinetics. For the equations in
Scheme 2 of (in generalized form) A h 2B, B + O2 f D + product,
then 2D + 2DTBC f A, numerical integration curve-fitting via
MacKinetics was employed. The oxygen pressure in psig was converted
to soluble O2 in M by setting up an oxygen reservoir to mimic the
oxygen concentration in the solution following the calculation method
first used by D. K. Lyon8 along with literature O2 solubility data (a
Henry’s constant of 15340 Pa m3/mol at 293.2 K in 1,2-dichloro-
Preparation of Pierpont’s Complex. [VO(DBSQ)(DTBC)]2 was
prepared and characterized by UV-visible, elemental analysis, and
negative ion ESI-MS as previously detailed.2
O2-Uptake Kinetic Experiments and Rate Law Determination.
These experiments were carried out as detailed elsewhere7 on a
computer-interfaced O2-pressure transducer apparatus which permits
the collection of high-precision (e(0.01 psig) pressure data at g1 data
points per second. The reaction flask is a pressurized Fischer-Porter
bottle attached via Swagelock quick-connects and flexible stainless steel
tubing to both an oxygen tank and a pressure transducer (total volume
148 mL).7 In the drybox, 400 ( 5 mg (1.8 mmol) of DTBC was
weighed into a 22 mm × 175 mm Pyrex culture tube along with a
ethane);9 the resultant ([O2]t - [O2]final 10 experimental data were then
)
used in the fitting. To start, the chemical equations to be tested were
written into MacKinetics along with the experimental data set. Next, a
grid search was performed to fit the kinetic parameters to the
experimental data. Specifically, the kinetic parameter search was
performed first over a broad range of the parameters (10-10-1010) using
the grid search command in MacKinetics. The grid search was then
narrowed down to about half the previous range (e.g., if the first search
gave k1 ) 0.01 and k2 ) 0.1 as the best fit in the range of 10-10-1010,
then the second search was performed in the range of 10-7-103 for k1
and 10-6-104 for k2); the center of the new range was always set equal
to the result from the last grid search as in the above examples. The
search was deemed finished once a grid search was performed in a
range no larger than 2 orders of magnitude. The curve-fit shown in
5
5/8-in. × /16-in. Teflon stir bar. Pierpont’s catalyst [VO(DBSQ)-
(DTBC)]2 (typically 2-8 mg, 1.8-7.2 mol) was weighed into a 5-mL
glass vial and dissolved in a measured volume of 1,2-C2H4Cl2 (1-3
mL). A predetermined fraction (ca. 1/10-1/3; 0.4-0.8 mol) of this
stock solution was transferred into the culture tube via a 1-mL syringe.
The culture tube was then placed inside the Fischer-Porter bottle,
sealed, brought out of the drybox, placed in a temperature-controlled
oil bath, and attached to the oxygen uptake apparatus via the quick-
connects. Stirring was initiated and the solution was equilibrated in
the oil bath (40 °C) under N2 (from the drybox gas) for 25 min. The
Fischer-Porter bottle was then purged 15 times with ∼13 psig of O2
(8) Lyon, D. K. Ph.D. Dissertation, University of Oregon, Eugene, OR, 1990,
p 252; see also pp 142-145.
(9) Lu¨hring, P.; Schumpe, A. J. Chem. Eng. Data 1989, 34, 250-252.
(10) A result of using the zeroed, net [O2] concentration change which is
necessary for the curve-fitting is that the different initial pressure is ignored;
that is, ∆[O2] is proportional only to the [DTBC] (total). Hence, the averaged
rate constants reported in Scheme 2 are the result from curve-fitting just
those eight sets of data obtained under the identical initial pressure (ca.
1.7 atm O2).
(6) Hill, C. L.; Weinstock, I. A. Nature 1997, 388, 332-333.
(7) Yin, C.-X.; Finke, R. G. Is It True Dioxygenase or Classic Autoxidation
Catalysis? Re-Investigation of a Claimed Dioxygenase Catalyst Based On
a Ru2-Incorporated, Polyoxometalate Precatalyst. Inorg. Chem. 2005, 44,
4157-4188.
9
J. AM. CHEM. SOC. VOL. 127, NO. 40, 2005 13989

A R T I C L E S
Yin and Finke
Figure 1 (as well as Figures S3 and S4) is that obtained by performing
a final integration, using the best set of kinetic parameters, and then
co-plotting those data with the observed, experimental data. All other
fits started with the parameters obtained from Figures 1, S3, and S4
and then performed a grid search over a range of 102 from those
parameters.
Effect of the Addition of an O-Substituted Ligand to [VO-
(DBSQ)(DTBC)]2 As Monitored by EPR. A drop of DMSO (Mallinck-
rodt, spectrometric grade) was added to ca. 0.2 mM [VO(DBSQ)-
(DTBC)]2 in toluene. The nine-line EPR characteristic2 of [VO(DBSQ)-
(DTBC)]2 (g ) 2.006, A51V ) 3.01 G) changed to an asymmetric
spectrum (a g ) 2.02 broad peak, and g ) 2.002 for five other sharp
peaks, A51V ) 6.35 G) as shown in Figure S8 in the Supporting
Information. A color change from blue to violet was also observed
upon the addition of DMSO.
Figure 1. O2-uptake curve of a representative kinetic run of DTBC plus
[VO(DBSQ)(DTBC)]2 and its curve-fit a la` Scheme 2. Reaction condi-
tions: 1.8 mmol of DTBC, 0.4 µmol of catalyst, 8 mL of 1,2-C2H4Cl2, 40
°C, and 1.7 atm O2. The initial rate obtained via a linear regression of the
data obtained within the first hour is -1.0 psig/h (or -0.86 Torr/min). The
fit (the solid line that fits the data so well it cannot be easily seen until at
the bottom, right-hand end of the curve) is from numerical integration of
the kinetic steps and mechanism in Scheme 2 accomplished via MacKinetics.
The individual rate constants defined from the curve-fitting (and as defined
Results and Discussion
Rate Law of [VO(DBSQ)(DTBC)]2-Catalyzed DTBC Di-
oxygenation: Support for This Dimer Being the Catalyst
Resting State. Our previous paper2 established Pierpont’s
complex [VO(DBSQ)(DTBC)]2 as either a catalyst resting state
or possibly an actual species within the main catalytic cycle.
The application of one of the well-tested postulates of Halpern’s
“rules” (really guidelines) as detailed elsewhere,11 that if you
can isolate it (as one can with Pierpont’s complex), it typically
is not the catalyst (i.e., it is too stable to be a top catalytic
intermediate, and is likely g1 step away from the true catalytic
cycle), predicts that [VO(DBSQ)(DTBC)]2 is mostly likely a
catalyst resting state.
in Scheme 2) are k3 ≈ 600 h-1, k4 ≈ 108 M-1 h-1, k5 ≈ 3 × 107 M-1 h-1
,
and k6 ≈ 7 × 109 M-3 h-1. The data and fit above are also compared to a
first-order, exponential kinetic curve (pO2(t) ) pO2,t)0 exp{-kinitial(t)}).
The kinetics when beginning with pre-made, isolated [VO-
(DBSQ)(DTBC)]2 were determined by following the oxygen
pressure loss in order to confirm or refute this catalyst resting
state hypothesis, that is, the hypothesis that this complex is
directly connected to the catalytic cycle. In addition, we were
interested in the following question: Does the intact dimer
complex react directly with O2, or must it first fragment to two
monomers before entering the catalytic cycle? Note that the
measured O2 uptake is due to the catalytic reaction of DTBC
with O2 (i.e., and not the much smaller scale stoichiometric
reaction of [VO(DBSQ)(DTBC)]2 with O2) since DTBC is
present in a 750- to 9000-fold excess relative to [VO(DBSQ)-
(DTBC)]2. The details of how these experiments are done are
also relevant to their correct interpretation: the mixture of [VO-
(DBSQ)(DTBC)]2 and DTBC was brought out of the drybox
and flushed with O2, followed by its necessary equilibration
under that O2 atmosphere for 5 min, before pressure readings
were started, as described in more detail in the Experimental
Section.
A representative O2-uptake curve obtained for the reaction
of [VO(DBSQ)(DTBC)]2, DTBC, and O2 is presented in Figure
1. Documentation that the linear plot in Figure 1 is representative
is provided by the analogous kinetic curves under different
DTBC or other conditions (Figures S3 and S4 of the Supporting
Information). The linear nature of these curves proved to be a
highly telling, albeit initially unexpected, diagnostic of the
underlying mechanism. Note that at first glance (i) the initially
linear concentration vs time plot would appear to imply a zero-
order dependence on [O2], so (ii) we checked to be sure that
O2 mass-transfer is not limiting the observed reaction rate. The
Figure 2. Logarithmic plot of the initial -d[O2]/dt rate vs the concentration
of [VO(DBSQ)(DTBC)]2. Reaction conditions: 1.8 mmol of DTBC, 0.4-
0.8 µmol of catalyst, 8 mL of 1,2-C2H4Cl2, 40 °C, and 1.7 atm O2. A clean
first-order dependence of the initial rate on the initial concentration of [VO-
(DBSQ)(DTBC)]2 is apparent.
O2 gas-to-solution mass-transfer rate for our apparatus, condi-
tions, and stirring rate is ca. 24 Torr O2/min‚mM (the initial
slope of the -d[O2]/dt vs [[VO(DBSQ)(DTBC)]2] plot shown
in Figure S1 of the Supporting Information), so even at the
highest concentration of [VO(DBSQ)(DTBC)]2 studied in this
work, any MTL effects should cancel within a e5% error (note
also the linear slope in Figure 1 is ∼0.86 Torr/min, below the
1.2 Torr O2/min mass-transfer limit). In short, there must be
some other explanation for the predominantly linear kinetic
curves observed in place of the initially expected exponential
curves due to a first-order consumption of O2 (vide infra).
A successful solution to this kinetic puzzle came from the
determination of the rate law by the initial rate method, some
reflective thinking after being stumped for a bit by the linear
kinetics, plus the use of MacKinetics numerical integration
kinetic modeling. The initial rates were determined from the
initial slope of the linear portion of the O2-uptake curves. A
logarithmic plot of the initial -d[O2]/dt vs the [[VO(DBSQ)-
(DTBC)]2]initial is shown in Figure 2 and indicates a clean first-
order dependence on [VO(DBSQ)(DTBC)]2. A plot showing a
zero-order dependence on the substrate [DTBC]initial, as one
(11) Hagen, C. M.; Vieille-Petit, L.; Laurenczy, G.; Su¨ss-Fink, G.; Finke, R.
G. Organometallics 2005, 24, 1819-1831, see footnote 45 therein.
9
13990 J. AM. CHEM. SOC. VOL. 127, NO. 40, 2005

V-Based, Extended Catalytic Lifetime Catechol Dioxygenases
A R T I C L E S
70%, of the catalytic reaction, as shown in Figure 1 (and as
also documented in Figures S3 and S4 of the Supporting
Information).
After some reflection, we reasoned that an exponentially
decreasing rate due to the consumption of pO2 (i.e., its first-
order dependence) is being compensated by an effectively
increasing (and thus rate-increasing) concentration of another
component in the catalytic cycle. Given the rate law observed,
the only option for that other increasing component is the
Vanadium catalyst. We further reasoned that, therefore, it is
highly probable that the added [VO(DBSQ)(DTBC)]2 is acting
as a catalyst resting state and is supplying a second V species
that is more active and is the true catalyst which reacts directly
with O2 within the main catalytic cycle.
Figure 3. Plot of the initial rate vs substrate concentration. Reaction
conditions: 0.3-3.6 mmol of DTBC, 0.4 µmol of catalyst, 8 mL of 1,2-
C2H4Cl2, 40 °C, and 1.7 atm O2. The (0.2 error bar in the figure is at 1σ,
so that a zero-order dependence on DTBC is apparent even at 2σ error
bars (0.3 ( 0.4).
A hypothesis for this hidden V species quickly became
apparent, and the mechanism continued to unfold: the dimer,
[VO(DBSQ)(DTBC)]2, is very likely (i) fragmenting to two
monomers, 2VO(DBSQ)(DTBC), and (ii) that fragmentation is
very likely reversible (to allow for a subsequent rate-determining
reaction with O2, as needed to account for the first-order pO2
dependence). It also seems necessary that (iii) the overall rate
of formation of the active monomer and the rate of the DTBC
oxygenation in the catalytic cycle must be occurring at
comparable rates (i.e., when one begins with [VO(DBSQ)-
(DTBC)]2) so that an increasing catalyst concentration with its
rate-increasing effect can occur on the same time scale as the
rate-decreasing O2 consumption shown in Figures 1, S3, and
S4.
Figure 4. Logarithmic plot of the initial rate vs substrate pO2 (in atm).
Reaction conditions: 1.8 mmol of DTBC, 0.4 µmol of catalyst, 8 mL of
1,2-C2H4Cl2, 40 °C, and 1.1-2.9 atm O2. A first-order dependence on O2
is apparent.
might have predicted for the DTBC-saturated complex, [VO-
(DBSQ)(DTBC)]2, is shown in Figure 3. Importantly, a first-
order rather than a zero-order dependence on pO2,initial is
demonstrated in Figure 4. The experimentally determined rate
law for catalyst, DTBC, and O2 is, then, the following (where
it is understood that initial rates and concentration subscripts
apply to each term):
Scheme 2. A Minimum-Step, Kinetic Mchanism Which Fits
theObserved, Linear Kinetic Data in Figures 1, S3, andS4 (Plus
Five Additional Sets of Data, Eight Data Sets Total)a
d[O2]
-
) kobs[DTBC]0([VO(DBSQ)(DTBC)]2)1[O2]1
dt
Things began to make sense, and the most probable mech-
anism of action began to unfold, at this point. The first insight
is that this rate law is just that expected for Pierpont’s catalyst,
[VO(DBSQ)(DTBC)]2, being directly connected to the catalytic
cycle. The lack of a DTBC dependence is also as expected:
[VO(DBSQ)(DTBC)]2 does not need more DTBC, but does
need O2, as confirmed by the zero- and first-order dependencies,
respectively, on these components of the observed stoichiometry,
Scheme 1 (vide supra). The crucial and novel identification, in
our prior paper,2 of [VV(O)(DBSQ)(DTBC)]2 as directly con-
nected to the catalytic cycle is, therefore, fully supported and
further fortified by the kinetic studies and observed rate law
obtained as part of this study.
Given the observed first-order pO2 dependence (plus the zero-
order DTBC dependence), the kinetic curve in Figure 1 should
be exponentialsif one also assumes a constant catalyst con-
centration. However, the exponential curve co-plotted in Figure
1 shows that the observed curve is clearly not exponential;
instead it is linear over a substantial part, typically ca. 40-
a Individual steps are shown as they must be entered into MacKinetics
(i.e., steps 1 and 2 are just the O2 gas-to-solution equilibrium, while steps
3 and 4 are the separate steps of the dimer-to-monomer equilibrium). The
rate constants for steps 1 and 2 are meaningless as they were set at high,
artificial values so as to provide a fast prior equilibrium; however, their
ratio was set to faithfully reflect the known solubility of O2 in 1,2-C2H4Cl2
at 20 °C and 1.7 atm O2.9 The rate constants shown for steps 3-6 are
those found by the MacKinetics curve-fitting of the experimental data.
The particular kinetic scheme we then wrote and tested via
MacKinetics numerical integration is shown in Scheme 2. Key
components of the kinetic mechanism shown in Scheme 2 are
(i) a fast, reversible non-MTL O2 gas-to-solution pre-equilibrium
(as our MTL studies support), one in which the rate constants
are artificially set as very high (i.e., to mimic a fast prior
equilibrium), but where the ratio of rate constants correctly
9
J. AM. CHEM. SOC. VOL. 127, NO. 40, 2005 13991

A R T I C L E S
Yin and Finke
reflects the O2 solubility in dichloroethane9 at 20 °C all as done
previously8 (i.e., an O2 reservoir8 was constructed); (ii) a
reversible fragmentation of the dimer [VO(DBSQ)(DTBC)]2 to
two monomers, 2VO(DBSQ)(DTBC); (iii) the reaction of each
monomer with O2 and conversion into VO(DBSQ) plus organic
product, and then importantly (iv) a fast step regenerating the
dimer, [VO(DBSQ)(DTBC)]2.
Controls were done on the curve-fitting en route to arriving
at the steps shown in Scheme 2: if step 6 is replaced with the
monomer-forming step, [VO(DBSQ)] + DTBC f [VO(DBSQ)-
(DTBC)], then a poor fit is obtained, as shown in Figure S5 in
the Supporting Information. In addition, a mechanism in which
intact [VO(DBSQ)(DTBC)]2 is directly part of the catalytic cycle
(Scheme S1 of the Supporting Information) also cannot fit the
observed data, as illustrated in Figure S6 of the Supporting
Information.
The stoichiometry coefficients of two [VO(DBSQ)] plus two
DTBC in step 6 of Scheme 2 merit some comment: since their
sum is greater than three, this reaction cannot be an elementary
step (i.e., it must be pseudoelementary12,13). Hence, controls on
the curve-fitting were performed by varying the coefficients:
to one [VO(DBSQ)] and one DTBC to form 0.5[VO(DBSQ)-
(DTBC)]2 (an excellent fit resulted with an essentially un-
changed k6 ) 1.1 × 1010 M-2 h-1) as well as two [VO(DBSQ)]
and one DTBC (again an excellent fit, with the same rate
constant within experimental error, k6 ) 109 M-2 h-1). An
attempted fit with one [VO(DBSQ)] and two DTBC in step 6
did not converge, however. The end result of these curve-fitting
controls is that they support the formulation of step 6 as shown
in Scheme 2 of a net reaction of two [VO(DBSQ)] reacting
with two DTBC.
Figure 5. Calculated [VO(DBSQ)(DTBC] monomer concentration vs time
curve obtained using the chemical equations and rate constants shown in
Scheme 2. Note that the maximum concentration of the active monomer
builds up to about 0.017 mM, that is, to ca. one-third of the initial
concentration of the initial dimeric catalyst resting state, [VO(DBSQ)-
(DTBC)]2, concentration.
present pressure-transducer-obtained data), and other factors.
Consistent with this, the rate constants for the data in Figures
S3 and S4 in comparison to that for Figure 1 do show values
that differ by factors of 2- to 7-fold. Hence, the rate constants
reported in Scheme 2 are the averages of Figures 1, S3, and
S4, and five other (eight total) data sets obtained under a range
of catalyst concentrations (0.04-0.09 mM) and DTBC concen-
trations (0.15-0.42 M). In addition and as discussed next, an
experiment was designed to independently test the reliability
of the equilibrium and rate constants listed for the first five
steps in Scheme 2.
UV-Visible Monitoring of the Reaction of the Initial Rate
of [VO(DBSQ)(DTBC)]2 Plus O2. Since step 5 in the kinetic
mechanism in Scheme 2 is the rate-determining step following
two fast prior equilibria, Scheme 2 and its rate constants predict
that the observed rate constant for the summation of the first
five steps will look like the following (obtained using the rate
constant definitions and values contained in Scheme 2 as well
as the steady-state assumption on the dissolved O2 and monomer
intermediates): kobs ) K1,2K3,4k5, where the equilibrium constant
for steps 1 and 2, K1,2, is ∼10-2, that for steps 3 and 4, K3,4, is
∼7 × 10-6 M, and k5 ≈ 5 × 107 M-1 h-1. Hence, the prediction
In response to an insightful reviewer’s question and sugges-
tion, an alternative way to write eqs 5 and 6, which avoids a
V(III) product, but which yields the same rate constants for these
steps and which is consistent with Scheme 3 (vide infra), is as
follows:
[VVO(DBSQ)(DTBC)] + O2 f [VVO(DBSQ)(O2DTBC)]
(5)
2[VVO(DBSQ)(O2DTBC)] + 2H2DTBC f
[VVO(DBSQ)(DTBC)]2 + 2H2(O2DTBC) (6)
from Scheme 2 is that kobs ≈ 3.5 h-1
.
Experimentally, it was found that the reaction of [VO(DBSQ)-
(DTBC)]2 with O2, by following the UV-absorbance decay of
[VO(DBSQ)(DTBC)]2 at 684 nm and while at 40 °C and 0.8
atm O2, pleasingly yields a very close initial-rate-determined
kobs ) 1.9 h-1 that is within a factor of ∼2 of the predicted 3.5
h-1. This independent result provides both an idea of the
precision of the rate constants in Scheme 2 as well as a general
confirmation of the first five steps in Scheme 2, the data fitting
procedure, and the resultant rate constants. This result also
confirms that steps 1-2 and 3-4 represent two fast prior
equilibria, so the steady-state assumption on the dissolved O2
and monomer intermediate used to obtain the kobs ) K1,2K3,4k5
equality is also supported.
where O2DTBC2- is meant to indicate the conjugate base of
the oxidation product. Since the products of steps 5 and 6 are
after the rate-determining step (step 5), we have no direct
information on which detailed versions of steps 5 and 6 are
correct.
The accuracy of the resultant rate constants in Scheme 2 also
merits comment; our experience is that factors of 2- to 10-fold
(or more) can result from such fitting of kinetic data to four
unknowns, depending on the particulars of the kinetic scheme,
the amount and quality of the data (which are high for the
(12) (a) Original work using the pseudoelementary step concept as a kinetic
tool in understanding oscillatory reactions: (a) Field, R. J.; Noyes, R. M.
Acc. Chem. Res. 1977, 10, 214-221. (b) Noyes, R. M.; Field, R. J. Acc.
Chem. Res. 1977, 10, 273-280. (c) Noyes, R. M.; Furrow, S. D. J. Am.
Chem. Soc. 1982, 104, 45-48.
Use of MacKinetics To Better Understand the Linear
Kinetic Curves. Next, we used that numerical kinetic model
in Scheme 2 plus MacKinetics in their powerful, predictive way.
First, we generated the calculated monomer concentration vs
time curve shown in Figure 5. Note that it does indeed show
(13) We have further developed the pseudoelementary step concept in our
mechanistic work on a .300-step self-assembly reaction forming Ir(0)
∼300
nanoclusters: (a) Watzky, M. A.; Finke, R. G. J. Am. Chem. Soc. 1997,
119, 10382-10400. (b) Widegren, J. A.; Aiken, J. D., III; O¨ zkar, S.; Finke,
R. G. Chem. Mater. 2001, 13, 312-324.
9
13992 J. AM. CHEM. SOC. VOL. 127, NO. 40, 2005

V-Based, Extended Catalytic Lifetime Catechol Dioxygenases
A R T I C L E S
the expected (vide infra) increasing concentration vs time needed
to rationalize why the kinetic curves are linear: the rate-
increasing effect of increasing monomer concentration is
convoluted experimentally with the exponential rate decrease
due to the first-order loss of O2; the result is an apparently linear
kinetic curve (e.g., Figures 1, S3, and S4). Second, we generated
the predicted kinetic curves as a function of the initial [VO-
(DBSQ)(DTBC)]2 concentration, determined the initial rates
from those computed data, and then plotted those initial rates
vs the initial concentrations to be sure that the rate law
determined from the initial rates is consistent with the kinetic
model and rate constants in Scheme 2. The result, shown in
Figure S7 of the Supporting Information (which basically
mirrors Figure 2, vide supra), again shows that the kinetic model
in Scheme 2 is fully consistent with the measured rate law
obtained using initial rate methods.
line EPR spectrum to the asymmetric spectrum as shown in
Figure S8 along with a dramatic color change from blue to
violet. (A reviewer commented that oxo-transfer to DMSO could
be involved in this reaction; we agree, and thank the reviewer
for noting this possibility.) Additional evidence for a facile
dimer-to-monomer equilibrium is the following: (3) Pierpont’s
demonstration of a facile dimer-monomer equilibrium for
related Mo complexes;17 (4) literature structural studies showing
that vanadium compounds with coordination numbers of 5 or 6
are the most common so that only the monomer is expected to
react quickly with O2 (formally seven-coordinate compounds
typically contain a cyclic peroxo component occupying two
adjacent sites;18 we thank Prof. Cortland Pierpont at the
University of Colorado for this point); and (5) the mechanism
in which intact [VO(DBSQ)(DTBC)]2 persists inside the
catalytic cycle with, for example, only one vanadium being
active, Scheme S1 of the Supporting Information (i.e., similar
to enzymatic half-site reactivity19), cannot fit the observed
kinetic data (vide supra) and also seems sterically prohibitive
with no clear chemical advantage, at least not that we have been
able to discern.
In step 2, O2 binds to either the semiquinone20 or the more
nucleophilic catecholate ligand, as Bianchini16 and Pierpont
advocate,21 in the [VO(DBSQ)(DTBC)] monomer as an indi-
vidually fast but still rate-determining step in the catalytic cycle,
consistent with the experimental kinetics. Step 2 is fast but rate-
determining since the reverse of step 1 is faster than step 2
(which is also fairly fast). Note that the fact that step 2 is rate-
determining means that the species after this and until the
catalyst resting state do not build up to significant concentrations
and thus are not expected to be amenable to typical spectroscopic
characterization. This means that the species after step 2 in
Scheme 3 must rely on steps that have to be present in order to
account for the observed products plus literature precedent for
the details of those steps.
Noteworthy here is that the mechanism in Scheme 2 is a rare
(and possibly the first) experimental example of the generalized
kinetic scheme A h 2B, B + C f D + product, then 2D + 2E
f A. As such, it is of interest to study its general properties
further by numerical integration simulations as a function of
the concentrations of each variable and the four rate constants
involved. These simulations are being carried out as a separate
study.14
A Proposed Mechanism Consistent with the Kinetics, the
Catechol Dioxygense Literature, and Employing [VO(DB-
SQ)(DTBC)]2 as the Catalyst Resting State. A proposed
system for the active vanadium catechol dioxygenase mechanism
is provided in Scheme 3. This mechanism is of course consistent
with our kinetic studies, but then it also incorporates key features
from the literature of the Fe-based catechol dioxygenases,
notably the substrate activation mechanism proposed by Que
and co-workers,4 the M-OO-C(catechol) species characterized
by Bianchini and co-workers,16 as well as a common branching
intermediate for intradiol vs extradiol cleavage pathways as
detailed by Bugg and co-workers.5
(17) (a) A Mo analogue, [MoVIO(DTBC)2]2, dissociates into monomer in
coordinating solvent as demonstrated by NMR: Buchanan, R. M.; Pierpont,
C. G. Inorg. Chem. 1979, 18, 1616-1620. (b) The tetramer (chiral square)
[MoIV(µ-O)(3,6-DTBC)2]4 also dissociates into monomer upon ligand
addition: Liu, C.-M.; Nordlander, E.; Schmeh, D.; Shoemaker, R.; Pierpont,
C. G. Inorg. Chem. 2004, 43, 2114-2124.
(18) (a) Nugent, W. A.; Mayer, J. M. Metal-Ligand Multiple Bonds; John Wiley
& Sons: New York, 1988; pp 159-162. (b) The Cambridge Structural
Database (version 5.26 Nov 2004) via Conquest (version 1.7) software.
(19) Enzymic half-site reactivity is observed in proteins composed of n identical
subunits but which react with a substrate or an inhibitor with only n/2
subunits saturated with that ligand. Binding of the ligand to one site typically
induces a conformational change of the second binding site, thereby
rendering the second site inactive. (a) Levitzki, A.; Stallcup, W. B.;
Koshland, D. E., Jr. Biochemistry 1971, 10, 3371-3378. (b) Seydoux, F.;
Malhotra, O. P.; Bernhard, S. A. CRC Crit. ReV. Biochem. 1974, 2, 227-
257.
Based on our earlier finding and the kinetic studies herein,
Pierpont’s complex is unequivocally identified as the catalyst
resting state. In step 1 in Scheme 3, the [VO(DBSQ)(DTBC)]2
reversibly fragments in even weakly coordinating solvent to
form 2 equiv of monomer. Five lines of evidence support this
step: (1) the kinetic studies and MacKinetics curve-fitting; and
(2) an experiment in which even a drop of DMSO added to
[VO(DBSQ)(DTBC)]2 in toluene changes the signature nine-
(14) Morris, A. M.; Yin, C.-X.; Finke, R. G. Unpublished results and experiments
in progress.
(15) The terminal V-O oxo is represented in Scheme 3 as a VdO simply out
of convenience; it certainly has some triple bond character based on
Pierpont’s structural studies. More specifically, the V-O bonds in [VO-
(DBSQ)(DTBC)]2 (bond length 1.581(4) Å15a) looks like a triple bond15b,c
in comparison to the VdO bond length of 1.616(4) Å in the VO(catecho-
(20) The proposed C2 (instead of C1) site of attachment of O2 to the 3,5-di-
tert-butylsemiquinone ligand in [VO(DBSQ)(DTBC)]2 shown in Scheme
3 is consistent with the major product being 4,6-di-tert-butyl-2H-pyran-
2-one, rather than its isomer, 3,5-di-tert-butyl-2H-pyran-2-one. We note,
however, that both C1 and C2 binding of O2 can give the intradiol product,
2, a detail not included in Scheme 3 only to keep if from being too cluttered.
Of interest here is that computed net cationic charges at C1 (0.44) versus
C2 (0.42) have been reported for the cationic [CoIII(di-tert-butylcatecholato)-
(MeC(CH2PPh2)3)]+.20a (a) Bencini, A.; Bill, E.; Mariotti, F.; Totti, F.;
Scozzafava, A.; Vargas, A. Inorg. Chem. 2000, 39, 1418-1425. (b)
Mechanisms with O2 bound to the C2 site of DTBC are commonly proposed
in the literature: Cox, D. D.; Que, L., Jr. J. Am. Chem. Soc. 1988, 110,
8085-8092. (c) Viswanathan, R.; Palaniandavar, M. J. Chem. Soc., Dalton
Trans. 1995, 1259-1266. (d) Funabiki, T.; Yamazaki, T. J. Mol. Catal.
A: Chem. 1999, 150, 37-47. (e) Yamahara, R.; Ogo, S.; Watanabe, Y.;
Funabiki, T.; Jitsukawa, K.; Masuda, H.; Einaga, H. Inorg. Chim. Acta
2000, 300-302, 587-596.
2-
late)2 ion. However, competing π bonding from the good π donor
catecholate ligands is also present and needs to be a part of any discussion
of the best descriptions of the V-O bonds. Two of the V-O(catecholate)
bond lengths in [VO(DBSQ)(DTBC)]2 are quite short, 1.827(4) Å, in
comparison to V-O(semiquinone) bond length of 1.975-1.987 Å, showing
the variable interactions present between V and catecholate or semiquinone
oxygen atoms. (a) Cass, M. E.; Green, D. L.; Buchanan, R. M.; Pierpont,
C. G. J. Am. Chem. Soc. 1983, 105, 2680-2686. (b) We thank Professor
C. G. Pierpont for a discussion and his insights on this point. (c) Nugent,
W. A.; Mayer, J. M. Metal-Ligand Multiple Bonds; John Wiley & Sons:
New York, 1988; pp 33-36. (d) The classic Rappe´-Goddard importance
of a spectator oxo ligand also merits mention with respect to Scheme 3:
Rappe´, A. K.; Goddard, W. A., III. J. Am. Chem. Soc. 1982, 104, 3287-
3294.
(21) The mechanistic issues of the site and timing of the O2 plus bound
catecholate benefited from several discussions with Prof. C. Pierpont; it
is our pleasure to acknowledge and thank him for those insightful
discussions.
(16) Barbaro, P.; Bianchini, C.; Linn, K.; Mealli, C.; Meli, A.; Vizza, F. Inorg.
Chim. Acta 1992, 198-200, 31-56.
9
J. AM. CHEM. SOC. VOL. 127, NO. 40, 2005 13993

A R T I C L E S
Yin and Finke
Scheme 3. Proposed, Minimal (“Occam’s Razor”) Mechanism of Vanadium-Based DTBC Dioxygenases Based on the Evidence Provided
as Well as Literature Evidence4,5,16 for Steps 3 and 415,a
a The details of step 2 are less clear, however; both an O2 diradical plus semiquinone spin-pairing mechanism (basically as shown above) and an electrophilic
O2 plus nucleophilic DTBC2- mechanism are possible. See Scheme S2 of the Supporting Information as well as the text which follows for additional
information and discussion.
The redox noninnocent22 DTBC ligand (i.e., its catecholate/
V(V) as well as semiquinone/V(IV) forms) and the presence of
both catechol and semiquinone in the active monomer introduce
an at present unsolved mechanistic ambiguitysboth here and
in the general dioxygenase literaturesas to which ligand form
reacts initially with O2. Noteworthy in this regard is Bianchini’s
classic study in which he considers three possible steps for the
initial step leading to the metal-OO-C(catecholate) species:
(i) the classic spin-pairng model, in which a O2 diradical initially
attacks a semiquinone radical; (ii) O2 initial attack at the metal;
or (iii) his preferred step, at least for Ir and Rh complexes, in
which the more nucleophilic catecholate attacks electrophilic
O2, a step he notes is consistent with the metal-centered HOMO
and ligand-centered LUMO. The mechanism shown above is
basically the spin-pairing model (the facile reaction of the O2
diradical and the semiquinone radical anion), followed by
electron transfer from DTBC2- to VV to make DBSQ-• and
VIV (i.e., valence tautomerism). The alternative possibility,
which can be labeled the semiquinone spectator mechanism for
step 2 in Scheme 3, is provided in Scheme S2 of the Supporting
Information. Keys in that mechanism are nucleophilic attack
of bound DTBC2- on electrophilic O2, with the vanadium
remaining as VV throughout the process. It would seem to us
that this latter mechanism might be disfavored due to the more
substantial Marcus-type intrinsic barriers (at C and O2; see the
additional discussion in the Supporting Information). However,
it is not possible to distinguish these mechanisms at the present
time, so the ambiguity regarding the details of step 2 is the
greatest uncertainty in Scheme 3 in our opinion.
A subsequent V-OO bond formation step is required in order
to eventually obtain the O-O bond cleavage step required to
account for the observed products; this step is written in Scheme
3 by analogy to the postulated mechanisms of enzymatic
dioxygenases and their models. The presence of two catecholate/
semiquinone ligands on one vanadium is a noteworthy feature
of the active monomer derived from Pierpont’s dimeric catalyst
resting state. Specifically, the presence of two DTBC ligands
appears to prevent a dioxo V(O)2 moiety from forming, a
structural feature that in turn avoids a transient trioxo V(O)3
moiety that would have been a violation of Lipscomb’s rule23s
and, therefore, presumably of too high energy to permit efficient
catalysis. The presence of a spectator terminal oxo ligand15d in
the catalyst, with its competing π-bonding effects, in Scheme
3 is also noteworthy.
The subsequent steps 3-4 have been written by analogy to
the Fe catechol dioxygenase mechanism,24 wherein a common
(22) (a) Jorgensen, C. K. Absorption Spectra and Chemical Bonding in
Complexes; Pergamon Press: Oxford, 1962. (b) Collman, J. P.; Hegedus,
L. S.; Norton, J. R.; Finke, R. G. Principles and Applications of
Organotransition Metal Chemistry; University Science Books: Mill Valley,
CA, 1987; pp 192-197.
(23) Pope, M. T. Heteropoly and Isopoly Oxometalates; Springer-Verlag: New
York, 1983; pp 18-19.
(24) (a) Que, L., Jr.; Ho, R. Y. N. Chem. ReV. 1996, 96, 2607-2624. (b) Costas,
M.; Mehn, M. P.; Jensen, M. P.; Que, L., Jr. Chem. ReV. 2004, 104, 939-
986.
9
13994 J. AM. CHEM. SOC. VOL. 127, NO. 40, 2005

V-Based, Extended Catalytic Lifetime Catechol Dioxygenases
A R T I C L E S
intermediate branches into intradiol cleavage product or extradiol
cleavage product depending on migration of an acyl group (step
4a) or an alkenyl group (step 4b).25 The product release in steps
4-5 is written as DTBC-dependent, but after the rate-determin-
ing step and fast, consistent with the zero-order dependence seen
on DTBC as well as the finding that this step is fast based on
the MacKinetics numerical integration kinetic model. (Recall
here the earlier discussion about steps 5 and 6 in the numerical
integration model and the two possibilities discussed there about
precisely how to write these post-rate-determining steps.) The
relative rate constants for the reverse of step 1 vs that of step 2
teach that active VO(DBSQ)(DTBC) monomer catalyst returns
to the dimeric [VO(DBSQ)(DTBC)]2 catalyst resting state about
two times for every one time it reacts with O2 to undergo another
catalytic cycle.
Of course, like all new mechanisms, Scheme 3 is offereds
and should be viewedsas a specific hypothesis requiring further
testing and attempts to refine or refute its overall features as
well as its intimate steps. That said, key features of Scheme 3
which are expected to stand the test of time include the
identification of [VO(DBSQ)(DTBC)]2 as the catalyst resting
state, its fragmentation to monomer to enter the catalytic cycle,
and the reaction of that monomer with O2.
Effect of Added Tridentate Ligands on Catechol Oxidative
Cleavage Activity. Based on the mechanism in Scheme 3, a
prediction is that added ligands are not likely to have a beneficial
effect, even if they are multidentate ligands that are useful in
other dioxygenase catalysts.26 In fact, the in situ addition of
tridentate ligands such as 1,4,7-triazacyclononane (TACN) to
a [VO(DBSQ)(DTBC)]2-catalyzed DTBC oxygenation solution
suppresses almost completely the dioxygenase reaction (the yield
of intradiol product 2 decreased from ∼40% to <0.2%), while
the undesired autoxidation pathway increases (the yield of
product 6 increased from ∼18% to 41%).27 This experiment
both (i) confirms the importance of having two DTBC ligands
in the active catalyst and (ii) argues that the design of kinetically
highly active but more selective catechol oxygenation catalysts
based on vanadium will be challenging.
(iii) A proposed mechanism of action has been forwarded
(Scheme 3) that is supported by the identity of the catalyst
resting state and the kinetics from that resting statesthe most
a catalytic mechanistic chemist can ever ask for!sas well as a
sizable amount of relevant literature.2,24 The mechanism in
Scheme 3 is offered as a working hypothesis that should be
viewed as such and which can, and should, be used to guide
future research; the details of the O2 plus substrate addition step
(step 2 in Scheme 3) especially require further investigation.
The mechanism in Scheme 3 is, nevertheless, a welcome
addition to a literature in which mechanistic knowledge was
preViously nonexistent for any man-made catalyst that is as
highly catalytic and long-liVed as the present system (i.e.,
30 000-100 000 TTOs).
(iv) The mechanism in Scheme 3 also provides general
support for the broader applicability of the mechanistic insights
from the laboratories of Que,4,24 Funabiki,28 Bugg,5 and others
that went into the construction of the post-rate-determining steps
3-5 in Scheme 3. Restated, the way that the precedent in the
literature from the seminal contributions of the above authors
fits naturally into Scheme 3 provides support for the broader
generality of the substrate coordination, C-OO-V formation,
Criegee rearrangement, and other steps commonly seen in
literature dixoxygenase mechanisms.4,5,24
(v) Point iv then, in turn, supports the probably greater
generality of important mechanistic insights from the literature24
and in Scheme 3, such as the observation that a major difference
between dioxygenase and monooxygenases is that the former
avoid O-O bond-cleavage activation of O2. That is, Fe(II)
extradiol24 and Fe(III) intradiol,24 as well as the present V(V/
IV) combined intra- and extradiol dioxygenase, avoid O-O
bond cleavage until after a substrate C-O-O bond has been
made.24,29 Hence, easier dioxygenations use a substrate’s prior
coordination as a key to their catalysis; more difficult substrates
(e.g., unactivated alkanes) require O2 activation with 2e- and
2H+ to species such as FedO that can then do these harder
activations. This in turn implies that it will be difficult to develop
true dioxygenases for RH and other difficult substrates, a
conclusion Sen has also reached.30
Summary and Conclusions
The above conclusions are, however, not meant to imply that
additional mechanistic work is not possible nor needed. Further
tests of the mechanism in Scheme 3 are possible and thus need
to be performed along with studies testing if Scheme 3 extends
in a general way to other systems such as molybdenum.17
Indeed, it is likely that the area of man-made, highly active,
long-lived dioxygenase catalysts supported by kinetic studies
leading to mechanistic understanding is just beginning.
The main findings from this study can be summarized as
follows:
(i) Pierpont’s structurally characterized complex [VO(DBSQ)-
(DTBC)]2 has been shown to be directly connected to the
catalytic cycle, serving as a catalyst resting state.
(ii) Reversible fragmentation of Pierpont’s dimer to active
monomer, which then reacts with O2 in a rate-determining step,
is the second key finding of the present kinetic and mechanistic
studies.
(28) (a) Funabiki, T.; Mizoguchi, A.; Sugimoto, T.; Yoshida, S. Chem. Lett.
1983, 917-920. (b) Funabiki, T.; Mizoguchi, A.; Sugimoto, T.; Tada, S.;
Mitsuji, T.; Sakamoto, H.; Yoshida, S. J. Am. Chem. Soc. 1986, 108, 2921-
2932.
(25) The mechanism in Scheme 3 contains a testable 18O2 labeling prediction
that will reveal whether O-scrambling occurs with the initial “spectator
oxo” present in the catalyst, an experiment which can, then, be compared
to literature precedent [(a) White, L. S.; Nilsson, P. V.; Pignolet, L. H.;
Que, L., Jr. J. Am. Chem. Soc. 1984, 106, 8312-8313. (b) Funabiki, T.;
Mizoguchi, A.; Sugimoto, T.; Tada, S.; Mitsuji, T.; Sakamoto, H.; Yoshida,
S. J. Am. Chem. Soc. 1986, 108, 2921-2932.] This experiment is under
investigation.
(26) (a) Dei, A.; Gatteschi, D.; Pardi, L. Inorg. Chem. 1993, 32, 1389-1395.
(b) Jo, D.-H.; Que, L., Jr. Angew. Chem., Int. Ed. 2000, 39, 4284-4287.
(c) Lin, G.; Reid, G.; Bugg, T. D. H. J. Am. Chem. Soc. 2001, 123, 5030-
5039.
(27) The use of another tridentate ligand, 1,4,7-trimethyl-1,4,7-triazacyclononane
(Me3-TACN), gives the same results: the addition of either a stoichiometric
or an excess amount of Me3-TACN (310 equiv) vs the precatalyst (n-Bu4N)7-
SiW9V3O40-catalyzed DTBC oxygenation to primarily the autoxidation
product 6 in 27-44% yield.
(29) (a) Funabiki, T. In Catalysis by Metal Complexes; Funabiki, T., Ed.; Kluwer
Academic Publishers: Dordrecht, The Netherlands, 1997; Vol. 19, pp 80-
83. (b) As a caveat, it is not clear whether the above discussions are
applicable to arene cis-dihydroxylation reaction catalyzed by Rieske
“dioxygenases” (e.g., naphthalene dioxygenase, toluene dioxygenase).
However, these so-called “dioxygenases” would appear not to be true
dioxygenases in that they require NAD(P)H as a cofactor, presumably
making them mechanistically more related to cytochrome P450
.
(30) Specifically, Sen concluded that,“While, from a practical standpoint, it is
more desirable for both oxygen atoms of O2 to be used for substrate
oxidation, there appears to be no known catalytic system that operates as
an artificial ‘dioxygenase’ under mild conditions toward ‘difficult’ sub-
strates, such as those possessing unactivated primary C-H bonds.” Lin,
M.; Hogan, T.; Sen, A. J. Am. Chem. Soc. 1997, 119, 6048-6053.
9
J. AM. CHEM. SOC. VOL. 127, NO. 40, 2005 13995

A R T I C L E S
Yin and Finke
Acknowledgment. It is our pleasure to once again2,3,21
acknowledge and thank Professor Cortland G. Pierpont at the
University of Colorado, Boulder, for his many highly valuable
and insightful comments throughout this work. Those insights
are at a level that comes only from a lifetime of pursuing
catechol-metal chemistry, valence tautomerism, structurally
precisely defined complexes such as [VO(DBSQ)(DTBC)]2, and
his other areas of interest that are directly relevant to the present
chemistry. This research was supported by NSF grant 9531110.
curves of kinetic runs of [VO(DBSQ)(DTBC)]2 plus excess
DTBC; presentation of the mechanism involving intact [VO-
(DBSQ)(DTBC)]2; figures showing how different mechanistic
schemes give poor fits to the [VO(DBSQ)(DTBC)]2 O2-uptake
data in Figure 1 in the main text; calculated catalyst reaction
order using the chemical models in Scheme 2; the EPR of [VO-
(DBSQ)(DTBC)]2 and the EPR of [VO(DBSQ)(DTBC)]2 with
addition of a drop of DMSO; and Scheme S2, the electrophilic
O2 plus nucleophilic DTBC2- semiquinone spectator alternative
mechanism for step 2 in Scheme 3. This material is available
free of charge via the Internet at http://pubs.acs.org.
Supporting Information Available: Stirring rate controls; the
establishment of O2 mass-transfer limitations; solvent vapor
pressure correction example; two other representative O2-uptake
JA052998+
9
13996 J. AM. CHEM. SOC. VOL. 127, NO. 40, 2005