Mechanism-based cofactor derivatization of a copper amine oxidase by a branched primary amine recruits the oxidase activity of the enzyme to turn inactivator into substrate

Article abstract of DOI:10.1021/ja058838f

The copper amine oxidases (CAOs) have evolved to catalyze oxidative deamination of unbranched primary amines to aldehydes. We report that a branched primary amine bearing an aromatization-prone moiety, ethyl 4-amino-4,5- dihydrothiophene-2-carboxylate (1), is recognized enantioselectively (S ? R) by bovine plasma amine oxidase (BPAO) both as a temporary inactivator and as a substrate. Substrate activity results from an O₂-dependent turnover of the covalently modified enzyme, with release of 4-aminothiophene-2- carboxylate (2) as ultimate product. Interaction of (S)-1 with BPAO occurs within the enzyme active site with a dissociation constant of 0.76 μM. Evidence from kinetic and spectroscopic studies, and HPLC analysis of stoichiometric reactions of BPAO with (S)-1, combined with a model study using a quinone cofactor mimic, establishes that the enzyme metabolizes 1 according to a transamination mechanism. Following the initial isomerization of substrate Schiff base to product Schiff base, a facile aromatization of the latter results in a metastable N-aryl derivative of the reduced cofactor aminoresorcinol, which is catalytically inactive. The latter derivative is then slowly oxidized by O₂, apparently facilitated partially by the active-site Cu(II), to form a quinonimine of the native cofactor that releases 2 upon hydrolysis or transimination with substrate amine. Preferential metabolism of (S)-1 is consistent with the preferential removal of the pro-S α-proton in metabolism of benzylamine by BPAO. This study represents the first report of product identification in metabolism of a branched primary amine by a copper amine oxidase and suggests a novel type of reversible mechanism-based (covalent) inhibition where inhibition lifetime can be fine-tuned independently of inhibition potency.

Full text of DOI:10.1021/ja058838f

Published on Web 04/15/2006
Mechanism-Based Cofactor Derivatization of a Copper Amine
Oxidase by a Branched Primary Amine Recruits the Oxidase
Activity of the Enzyme to Turn Inactivator into Substrate
Chunhua Qiao,† Ke-Qing Ling,† Eric M. Shepard,‡ David M. Dooley,‡ and
Lawrence M. Sayre*,†
Contribution from the Department of Chemistry, Case Western ReserVe UniVersity,
CleVeland, Ohio 44106, and Department of Chemistry and Biochemistry,
Montana State UniVersity, Bozeman, Montana
Received January 10, 2006; E-mail: LMS3@case.edu
Abstract: The copper amine oxidases (CAOs) have evolved to catalyze oxidative deamination of unbranched
primary amines to aldehydes. We report that a branched primary amine bearing an aromatization-prone
moiety, ethyl 4-amino-4,5-dihydrothiophene-2-carboxylate (1), is recognized enantioselectively (S . R) by
bovine plasma amine oxidase (BPAO) both as a temporary inactivator and as a substrate. Substrate activity
results from an O2-dependent turnover of the covalently modified enzyme, with release of 4-aminothiophene-
2-carboxylate (2) as ultimate product. Interaction of (S)-1 with BPAO occurs within the enzyme active site
with a dissociation constant of 0.76 µM. Evidence from kinetic and spectroscopic studies, and HPLC analysis
of stoichiometric reactions of BPAO with (S)-1, combined with a model study using a quinone cofactor
mimic, establishes that the enzyme metabolizes 1 according to a transamination mechanism. Following
the initial isomerization of substrate Schiff base to product Schiff base, a facile aromatization of the latter
results in a metastable N-aryl derivative of the reduced cofactor aminoresorcinol, which is catalytically
inactive. The latter derivative is then slowly oxidized by O2, apparently facilitated partially by the active-site
Cu(II), to form a quinonimine of the native cofactor that releases 2 upon hydrolysis or transimination with
substrate amine. Preferential metabolism of (S)-1 is consistent with the preferential removal of the pro-S
R-proton in metabolism of benzylamine by BPAO. This study represents the first report of product
identification in metabolism of a branched primary amine by a copper amine oxidase and suggests a novel
type of reversible mechanism-based (covalent) inhibition where inhibition lifetime can be fine-tuned
independently of inhibition potency.
Scheme 1
Introduction
The copper amine oxidases (CAOs) catalyze the oxidative
deamination of primary amines to aldehydes at the expense of
reduction of O2 to H2O2, through an active-site trihydroxy-
phenylalanine quinone (TPQ) cofactor-mediated pyridoxal-like
transamination process.1,2 The key step (Scheme 1, path A)
involves an active-site base-mediated CR deprotonation, ac-
complishing tautomerization of substrate Schiff base (SSB) to
product Schiff base (PSB), which hydrolyzes to release aldehyde
and generate the reduced aminoresorcinol form of the cofactor
(TPQamr). The latter is subsequently reoxidized to the starting
quinone with release of NH3. The transamination chemistry has
been recapitulated in model studies with TPQ cofactor mimics.3,4
Although so far unobserved, one must keep in mind that there
† Case Western Reserve University.
‡ Montana State University.
(1) Brazeau, B. J.; Johnson, B. J.; Wilmot, C. M. Arch. Biochem. Biophys.
2004, 428, 22-31.
is always an alternative addition-elimination mechanism
(Scheme 1, path B) that can justify the same observed overall
stoichiometry. According to the latter mechanism, the active-
site base would act on the initially formed carbinolamine that
precedes the substrate Schiff base. Proton abstraction would
result in an aldimine, which would hydrolyze to aldehyde and
(2) Mure, M.; Mills, S. A.; Klinman, J. P. Biochemistry 2002, 41, 9269-78.
(3) (a) Wang, F.; Bae, J.-Y.; Jacobson, A. R.; Lee, Y.; Sayre, L. M. J. Org.
Chem. 1994, 59, 2409-2417. (b) Lee, Y.; Sayre, L. M. J. Am. Chem. Soc.
1995, 117, 3096-3105. (c) Lee, Y.; Sayre, L. M. J. Am. Chem. Soc. 1995,
117, 11823-11828.
(4) (a) Mure, M.; Klinman, J. P. J. Am. Chem. Soc. 1995, 117, 8698-8706.
(b) Mure, M.; Klinman, J. P. J. Am. Chem. Soc. 1995, 117, 8707-8718.
9
6206
J. AM. CHEM. SOC. 2006, 128, 6206-6219
10.1021/ja058838f CCC: $33.50 © 2006 American Chemical Society

Cofactor Derivatization of Copper Amine Oxidase
A R T I C L E S
Scheme 2
described in this paper demonstrate an unprecedented trans-
aminative processing of 1 to yield an aromatized cofactor
derivative that would reflect irreversible inactivation were the
enzyme not capable of mediating an O2-dependent oxidation
of the cofactor derivative, ultimately releasing 2 as turnover
product.
Results
Enantioselective Inhibition of BPAO by 1. Inhibition of
BPAO by the R and S optical antipodes of 113 was followed by
the standard benzylamine assay. Two distinct kinetic patterns
were observed (Figure 1). The R enantiomer displayed a time-
and concentration-dependent inhibition of BPAO (Figure 1, left
panel), with the level of inhibition plateauing after the first 5-15
min. At the highest concentration examined (640 µM) the
maximum inhibition did not quite reach 50%. In contrast, much
lower concentrations of (S)-1 were required to inhibit BPAO
(Figure 1, right panel), though the time course was biphasic,
especially at lower concentrations, where activity decreased in
the first 5 min and then recovered more slowly. Nearly 50%
inhibition was seen after 5 min of incubation with 5 µM (S)-1.
Since activity is already recovering at this point, we can estimate
that there is at least a ∼130-fold difference in S > R inhibitory
potency. At higher concentrations of (S)-1 (e.g., 40 µM), the
time profile resembled more the behavior for the R antipode,
though the activity totally recovered after a 24 h dialysis.15 In
fact, enzyme activity recovered to g90% following dialysis after
long-term incubations with either (S)- or (R)-1, indicating that
there is minimal permanent covalent modification.
ammonia, and the triol form of the reduced cofactor, which
would be reoxidized to quinone. Although this mechanism
would result in the same net stoichiometry, it was found that
when the enzyme reaction is run anaerobically (single turnover),
NH3 was released only when the enzyme was exposed to
oxygen.5 Since the addition-elimination mechanism would
predict that NH3 and aldehyde should both form following
anaerobic single turnover, such observation was the main historic
evidence that the enzyme utilizes the transamination mechanism.
More recently, various intermediates in the transamination
mechanism have been spectroscopically observed.6-9 Never-
theless, addition-elimination could conceivably be observed
for unusual enzyme substrates and is known to be a viable
mechanism under certain conditions for the nonenzymatic
oxidation of amines by pyrroloquinolinequinone (PQQ) and its
analogues,10,11 as well as for alcohol dehydrogenation by PQQ-
dependent alcohol dehydrogenases.12
We have been interested in whether a copper amine oxidase
might be coaxed to utilize an addition-elimination mechanism
for metabolism of an appropriate substrate. We considered as a
possible candidate ethyl 4-amino-4,5-dihydrothiophene-2-
carboxylate (1), where addition-elimination would be followed
by a facile tautomerization to the aromatic product ethyl
4-aminothiophene-2-carboxylate (2) (Scheme 2). The free
carboxylic acid of (S)-1 was previously found to inhibit
γ-aminobutyric acid aminotransferase13 by covalent modification
of the pyridoxal 5′-phosphate cofactor, taking advantage of an
aromatization step.14 The zwitterionic free amino acid would
be poorly recognized by the copper amine oxidases, but the
ethyl ester could be a substrate if the enzyme made an exception
to its normally strict preference for unbranched primary amines
and accepted this branched primary amine as a substrate.
Preliminary studies on the interaction of 1 with the CAO
bovine plasma amine oxidase (BPAO) revealed apparent
turnover to give the aromatized product 2 expected from an
addition-elimination mechanism. However, in-depth studies
Both (S)-1 and (R)-1 Are Substrates of BPAO. The
recovery of activity of BPAO over time following incubation
with 1 demonstrates that it is being metabolized and that it
should thus be possible to detect and characterize the enzymatic
turnover product. Incubation of 1 (R,S mixture 63:37) with
BPAO in pH 7.2 buffer produced a single ninhydrin- and UV365
-
positive product, which was characterized as the aromatized
amine 2 by thin-layer chromatography (TLC) and HPLC
analyses by comparison with an authentic sample prepared from
catalytic dehydrogenation of 1. Control experiments established
that no reaction occurred when 1 was incubated in the same
buffer solution in the absence of enzyme or in the presence of
denatured BPAO, indicating conversion of 1 to 2 is strictly
dependent on active enzyme. Compound 2 is the aromatized
tautomer of the dehydrogenation product, imine 3 (Scheme 2),
that would be generated from an addition-elimination metabo-
lism of 1. On the other hand, no evidence was found for
hydroxythiophene 4, the aromatized form of the “normal”
transamination product 5, assuming the latter could be hydro-
lytically released immediately upon tautomerization of substrate
Schiff base to product Schiff base.
(5) Janes, S. M.; Klinman, J. P. Biochemistry 1991, 30, 4599-4605.
(6) Hartmann, C.; Brzovic, P.; Klinman, J. P. Biochemistry 1993, 32, 2234-
41.
(7) Nakamura, N.; Moenne-Loccoz, P.; Tanizawa, K.; Mure, M.; Suzuki, S.;
Klinman, J. P.; Sanders-Loehr, J. Biochemistry 1997, 36, 11479-86.
(8) Medda, R.; Padiglia, A.; Pedersen, J. Z.; Rotilio, G.; Finazzi Agro, A.;
Floris, G. Biochemistry 1995, 34, 16375-81.
(9) Hirota, S.; Iwamoto, T.; Kishishita, S.; Okajima, T.; Yamauchi, O.;
Tanizawa, K. Biochemistry 2001, 40, 15789-96.
(10) (a) Sleath, P. R.; Noar, J. B.; Eberlein, G. A.; Bruice, T. C. J. Am. Chem.
Soc. 1985, 107, 3328-3338. (b) Rodriguez, E. J.; Bruice, T. C. J. Am.
Chem. Soc. 1989, 111, 7947-7956.
(11) (a) Ohshiro, Y.; Itoh, S. Bioorg. Chem. 1991, 19, 169-189. (b) Itoh, S.;
Fukui, Y.; Haranou, S.; Ogino, M.; Komatsu, M.; Ohshiro, Y. J. Org. Chem.
1992, 57, 4452-4457.
To quantify the metabolism of 1 by BPAO, the reaction
progress was monitored by HPLC for the pure R and S
(12) Itoh, S.; Kawakami, H.; Fukuzumi, S. Biochemistry 1998, 37, 6562-6571.
(13) Adams, J. L.; Chen, T.-M.; Metcalf, B. W. J. Org. Chem. 1985, 50, 2730-
2736.
(15) Long-term incubation (2.5 h) of BPAO with S-1 at the same concentration
resulted in ∼10% activity loss that is resistant against dialysis.16
(16) Qiao, C. Ph.D. Thesis, Case Western Reserve University, Cleveland, OH,
2004.
(14) Fu, M.; Nikolic, P.; Van Breemen, R. B.; Silverman, R. B. J. Am. Chem.
Soc. 1999, 121, 7751-7759.
9
J. AM. CHEM. SOC. VOL. 128, NO. 18, 2006 6207

A R T I C L E S
Qiao et al.
Figure 1. Time-dependent inhibition of BPAO (1.0 µM) by various concentrations of (R)-1 (left) and (S)-1 (right). The enzyme activities represent the
slopes of ∆A250 for the first 40 s after dilution of the aliquots into 5 mM benzylamine assay solution.
Figure 2. Metabolism of the optically pure enantiomer of 1 (40 µM) by BPAO (0.7 µM). Left panel, (S)-1; right panel, (R)-1.
enantiomers, both at 40 µM (Figure 2). The metabolism of 1 to
2 was essentially quantitative for both enantiomers as shown
by the nearly 100% material recovery at various time points.
The plots of consumption of 1 and formation of 2 over time
appeared to be simple exponential curves for both enantiomers.
A nonlinear least-squares fit of the product formation curves
yielded first-order rate constants (k1) of 1.8 h-1 (t1/2 23 min)
for (S)-1 and 0.080 h-1 (t1/2 8.7 h) for (R)-1, respectively. The
half-life for metabolism of (R)-1 is 23-fold longer than for (S)-
1. The metabolism of (S)-1 by BPAO was also followed by
UV difference spectrophotometry (Figure S1 in Supporting
Information). A clean isosbestic conversion of 1 (294 nm) to 2
(250 and 318 nm) was observed, suggesting that any intermedi-
ates in the reaction do not accumulate on the steady-state time
scale. The slower turnover rate for (R)-1 compared to (S)-1 is
consistent with the much slower recovery of enzyme activity
when BPAO is incubated with (R)-1 than with (S)-1, according
to the understanding that activity can begin to recover only when
most of the inhibitor is metabolized.
Processing of (S)-1 by BPAO Occurs at the Active Site.
Mechanistic studies were focused on the more active S-
enantiomer. At the outset, it was important to first determine
whether the initial rapid loss of activity of BPAO upon
interaction with (S)-1 (Figure 1, right) was due to its competition
with benzylamine for the enzyme active site. The initial rates
of oxidative deamination of varying concentrations of benzyl-
amine by BPAO in the presence of varying concentrations of
(S)-1 allowed construction of a Lineweaver-Burk plot (Figure
3, left panel), which was found to pass through the same point
on the y-axis, suggesting that (S)-1 interacts at the enzyme active
site as a competitive inhibitor.17 Moreover, construction of the
corresponding Dixon plot (Figure 3, right) allowed estimation
of the dissociation constant (Ki) of the complex of BPAO with
(S)-1 to be 0.76 µM, suggesting that (S)-1 is a tight binding
inhibitor of BPAO.
Anaerobic Reaction of 1 with a TPQ Cofactor Mimic
Proceeds According to a Transamination Mechanism. To
shed light on the potential enzymatic pathways for metabolism
of 1, we conducted a model reaction of compound 1 with the
TPQ cofactor mimic 5-tert-butyl-2-hydroxy-1,4-benzoquinone
(TBHBQ).3c,4 A transamination pathway for this reaction
(Scheme 3, path A) would lead to a product Schiff base (PSB)
that could theoretically hydrolyze but would likely instead
The time-dependent stoichiometric conversion of 1 to 2 by
BPAO confirmed that both enantiomers of 1 are substrates for
BPAO. This result is significant because it shows for the first
time not only that a copper amine oxidase can smoothly
metabolize a branched primary amine but also that metabolism
may result in an overall dehydrogenation as opposed to the
typical deamination stoichiometry.
(17) Inhibition of BPAO by compound 1 also met other criteria for a mechanism-
based inhibitor; e.g., substrate benzylamine protection.16
9
6208 J. AM. CHEM. SOC. VOL. 128, NO. 18, 2006

Cofactor Derivatization of Copper Amine Oxidase
A R T I C L E S
Figure 3. Competitive inhibition of BPAO (1.0 µM) by different concentrations of (S)-1, pH 7.2, 0.1 M sodium phosphate buffer, 30 °C. Left panel:
Lineweaver-Burk plot of enzyme activities (oxidation of benzylamine) measured over the first 40 s following dilution of aliquots of the primary incubation
mixture into the assay cuvette. Right panel: Corresponding Dixon plot.
Scheme 3
undergo aromatization to give the “coupling product” 6. In con-
trast, as discussed above (Scheme 2), the alternative addition-
elimination mechanism (Scheme 3, path B) would result in
reduction of TBHBQ to the corresponding benzenetriol 7 and
imine 3, which would rapidly tautomerize to 2.
of redox interconversions, common to such reactions conducted
anaerobically,3 that can scramble the products. For example,
oxidation of 6 by TBHBQ (or the TBHBQ-derived SSB) would
generate quinonimine 8, which could undergo transimination
with 1 to give 2, whereas condensation of 2 with starting
TBHBQ and reduction of the resulting quinonimine by triol 7
would give 6.
The model reaction of 1 with TBHBQ was conducted in
dimethyl sulfoxide (DMSO) under anaerobic “single turnover”
conditions in the presence of excess diisopropylethylamine
(DIPEA) as base catalyst,18 and the initial reaction products were
trapped in situ against oxidative evolution by acetylation (with
Ac2O) prior to exposure to air. Three products were isolated
and characterized: the acetylated derivative of the aromatization
product 2, the triacetylated derivative of the coupling product
6, and the triacetylated derivative of benzenetriol 7. It was found
that although 7 is susceptible to autoxidation under the basic
reaction conditions, products 2 and 6 are both stable enough to
be isolated and characterized directly. The production of both
6 and 2 + 7 suggests consideration of the possibility that
transamination and addition-elimination reactions, respectively,
are operating simultaneously. However, no mechanistic conclu-
sions could be made at this point due to the probable operation
The anaerobic reaction of 1 with TBHBQ was also monitored
1
in situ by H NMR spectroscopy. Throughout the period of
observation (20 h), the only signals observed were starting 1
and products 2, 6, and 7, with the amount of 7 being the same
as that of 2. The relative amounts of 1, 2, and 6 were quantified
according to their signal integral and plotted over time (Figure
4). The 2/6 ratio increased from ∼1:16 at 10 min to ∼1:1.8 at
4 h and then remained nearly constant afterward. Because a
typical parallel reaction mechanism would predict a constant
ratio of 2 and 6, the observed apparent lag in generation of 2 is
indicative of the consecutiVe reaction 1 f 6 f 2, with the first
(18) No significant reaction occurred when TBHBQ was incubated with 1 in
0.1 M phosphate buffer, pH 7.2, both aerobically and anaerobically for
several days as monitored by UV-vis spectrophotometry.
9
J. AM. CHEM. SOC. VOL. 128, NO. 18, 2006 6209

A R T I C L E S
Qiao et al.
reaction to rationalize generation of 6 in the model reaction.
The unreactivity of 2 is not surprising, considering that it is
isoelectronic with an aniline whose nucleophilicity is further
reduced by resonance delocalization of the nitrogen lone pair
onto the carbethoxy substituent. All these results support the
interpretation that generation of 2 in the model reaction results
strictly as a consequence of the transamination mechanism
according to Scheme 3, path A.
Conversion of TBHBQ-Derived Coupling Product 6 to 2
under Aerobic Conditions. If transamination (Scheme 3, path
A) represents the mechanism by which the enzyme processes
the conversion of 1 to 2, one can postulate that formation of
the coupling product (like 6) reflects modification of the TPQ
cofactor that rationalizes the initial quick inactivation of BPAO
by 1 and that the recovery of activity then reflects oxidation to
the quinonimine (like 8) and subsequent dissociation or trans-
imination of the latter. However, unlike the model system
described above, where oxidation of 6 was effected by the excess
starting quinone available (Scheme 3), oxidation of the coupling
product in the enzyme case would have to be effected by
dioxygen. It was thus important to explore the chemical behavior
of 6 under various aerobic reaction conditions.
Figure 4. Model reaction of TBHBQ (70 mM) with 1 (70 mM) in the
presence of DIPEA (280 mM) in degassed DMSO-d6 at 25 °C, showing
the formation of 2 and 6. The amounts were quantified by 1H NMR signal
integration.
step being faster than the second. This result suggests that the
model reaction follows predominantly a transamination mech-
anism and that apparent addition-elimination product 2 arises
from redox cycling evolution of coupling product 6.
In contrast to the inertness of 6 in anhydrous DMSO, aerobic
incubation of 6 in 50% buffered (pH 7.2) aqueous DMSO
exhibited a slow but clean consecutiVe reaction as monitored
spectrophotometrically (Figure 5). A species absorbing at 290
and 420 nm was observed to form as intermediate, and TBHBQ
(∼500 nm) and 2 (∼320 nm) were the final products (spectra
of the various model compounds are given in Figure S4 in
Supporting Information). The formation of TBHBQ and 2 was
confirmed by comparative TLC with authentic samples, whereas
the intermediate was characterized as quinonimine 8, which
could be isolated in fair yield (55%) from autoxidation of 6 in
1:4 buffer/CH3CN. The spectral change at 490 nm (the isosbestic
point between 8 and TBHBQ) in the first 3 h, and the spectral
change at 420 nm after 4 h, both appeared to be exponential,
thereby permitting the estimation of the pseudo-first-order rate
constants for autoxidation (0.473 h-1, t1/2 1.5 h) and hydrolysis
(0.126 h-1, t1/2 5.5 h), respectively (eq 1). The hydrolysis step
The most likely pathway for conversion of 6 to 2 under the
anaerobic model reaction conditions is oxidation of 6 by starting
quinone TBHBQ, giving the observed triol 7 and quinonimine
8, which would undergo transimination with starting amine 1
to give 2 (thereby regenerating the substrate Schiff base). A
control NMR tube experiment to ascertain whether TBHBQ
could oxidize 6 in DMSO/DIPEA surprisingly revealed no
significant reaction under argon, but reaction quickly ensued
upon admission of air. However, other control experiments
showed that, in the DMSO/DIPEA condition, 7 but not 6
undergoes air oxidation. We thus propose that the redox
interchange TBHBQ + 6 h 7 + 8 represents an unfavorable
equilibrium, which is pulled to the right in air by O2-dependent
oxidation of 7 to TBHBQ. When the model reaction is run
anaerobically, we propose that the equilibrium is pulled to the
right by the removal of 8 in reacting with 1.19 This would
suggest that conversion of 6 to 2 would cease once starting 1 is
consumed. This conclusion is consistent with the results shown
in Figure 4, where despite our claim of operation of the
consecutive reaction 1 f 6 f 2, no further conversion of 6 to
2 occurs once 1 has been depleted (at 300-350 min).
Assuming that conversion of 6 to 2 occurs according to
shifting of an unfavorable redox cycling equilibrium that is not
directly observable, we sought conclusive evidence for such
where the fate of 8 could be independently traced through a
labeling strategy, using different alcohol esters (e.g., methyl vs
ethyl esters). These experiments, following the anaerobic
reaction of TBHBQ with 1 carrying one ester label in the
presence of added 6 or 2 carrying a second ester label
(Supporting Information), clearly demonstrated the ability of 6
to be converted to 2, but not Vice Versa, during anaerobic
turnover of TBHBQ with 1. At the same time, we could find
no evidence for the ability of 2 to undergo condensation with
TBHBQ, which would be required for redox cycling conversion
of 2 to 6, thus ruling out the ability of the addition-elimination
is 3.7-fold slower than autoxidation and is thus rate-limiting.
Essentially the same rate constant for the hydrolysis step was
obtained starting alternatively from authentic 8 under the same
conditions (data not shown). Addition of 0.1 mM diethylene-
triaminepentaacetic acid (DTPA) into the same reaction did not
alter the rates for either step. This may suggest that the observed
oxidation and hydrolysis proceeds without the obligatory need
for trace transition metal catalysis.
(19) We cannot rule out the possibility that oxidation of 6 to 8 is mediated by
the TBHBQ/1-derived SSB, where the reduced SSB would then be
reoxidized by TBHBQ.
9
6210 J. AM. CHEM. SOC. VOL. 128, NO. 18, 2006

Cofactor Derivatization of Copper Amine Oxidase
A R T I C L E S
Figure 5. Aerobic incubation of 6 (0.15 mM) in 1:1 DMSO and 0.05 M HEPES buffer, pH 7.2, at 25 °C, showing autoxidation to quinonimine 8 in the
early stage (the first 3 h, left panel, 20 min time intervals), and subsequent hydrolysis of 8 to form TBHBQ and 2 in the later stage (4-22 h, right panel,
1 h time intervals). Insets: Pseudo-first-order ln [(A∞ - A0)/(A∞ - At)] vs t plots obtained from ∆A490 for autoxidation and ∆A420 for hydrolysis, respectively.
Chart 1
Scheme 4
Nonetheless, the presence of Cu(II) at the active site of CAOs
and its role in the oxidative half-reaction of substrate turnover,20-23
prompted us to examine the effect of Cu(II) on the above
autoxidation-hydrolysis chemistry. When the same incubation
of 6 (buffered aqueous DMSO) was conducted in the presence
of 1 mol equiv of Cu(II), the solution rapidly (within 30 s)
turned reddish-yellow (283, 390, and 500 nm), which then
persisted without further spectral change for 2 days. At this
point, if the resulting solution was treated with excess DTPA
(0.3 mM) to chelate the Cu(II), it quickly turned yellow with
the appearance of the characteristic chromophores for 8 (290
and 420 nm; see Figure S2 in Supporting Information), which
then underwent slow hydrolysis to give TBHBQ and 2. An
independent control with authentic 8 established that the initially
formed reddish-yellow solution was actually the complex
8-Cu(II) (Chart 1), which is stabilized against hydrolysis,
probably on account of forming a five-membered ring chelate.
The pseudo-first-order rate constant (k1) for the Cu(II)-mediated
autoxidation of 6 was measured as 0.101 s-1 (t1/2 7 s) by
monitoring ∆A390 over time (Figure S3 in Supporting Informa-
tion), showing a ∼770-fold rate enhancement compared to the
reaction without added Cu(II). This significant promotion by
Cu(II) likely reflects its initial coordination to 6 (Chart 1) and
then rapid reaction of the activated 6-Cu(II) with O2. In the
enzyme case, although chelation of the coupling product to
Cu(II) in the manner permitted in the model study in solution
would likely be precluded, any type of interaction with the Cu(II)
is expected to facilitate the autoxidation process.
Mechanism of BPAO-Mediated Metabolism of (S)-1: (i)
Kinetics of Aerobic Activity Recovery following Anaerobic
Inhibition by (S)-1. The studies described above demonstrate
that the conversion of 1 to 2 in the model reaction with TBHBQ
follows a transamination-oxidation pathway. Nonetheless,
whether the enzyme uses transamination or addition-elimination
must be independently established. Transamination (Scheme 4)
predicts that 2 would form only upon aerobic reoxidation of
the coupling product 9 generated at the end of anaerobic single
turnover. Addition-elimination predicts that turnover product
2 should be present at the end of anaerobic single turnover along
with the benzenetriol form of the reduced TPQ cofactor. In a
previous study on the interaction of BPAO with simple
alkylhydrazines,24 we concluded that the benzenetriol form of
the cofactor is generated anaerobically but is reoxidized im-
mediately upon aliquoting into the aerobic benzylamine oxidase
activity assay, so that the measured activity is thus indistin-
guishable from the control. In the present study, using NaBH4
to directly generate the benzenetriol TPQ form (see Supporting
Information), we verified that the latter reoxidizes immediately
in the aerobic activity assay, and thus, any addition-elimination
process generating the benzenetriol would not be detectable by
the standard activity assay. On the other hand, if the trans-
amination mechanism were followed, resulting in the coupling
product 9 (Scheme 4), recovery of enzyme activity would
(20) Dooley, D. M.; McGuirl, M. A.; Brown, D. E.; Turowski, P. N.; McIntire,
W. S.; Knowles, P. F. Nature 1991, 349, 262-4.
(21) Hirota, S.; Iwamoto, T.; Kishishita, S.; Okajima, T.; Yamauchi, O.;
Tanizawa, K. Biochemistry 2001, 40, 15789-96.
(22) Mills, S. A.; Goto, Y.; Su, Q.; Plastino, J.; Klinman, J. P. Biochemistry
2002, 41, 10577-84.
(23) Kishishita, S.; Okajima, T.; Kim, M.; Yamaguchi, H.; Hirota, S.; Suzuki,
S.; Kuroda, S.; Tanizawa, K.; Mure, M. J. Am. Chem. Soc. 2003, 125,
1041-55.
(24) Lee, Y.; Jeon, H. B.; Huang, H.; Sayre, L. M. J. Org. Chem. 2001, 66,
1925-1937.
9
J. AM. CHEM. SOC. VOL. 128, NO. 18, 2006 6211

A R T I C L E S
Qiao et al.
Figure 6. Left panel: Time-dependent activity changes of BPAO (1.0 µM) upon incubation with (S)-1 (1.2 µM) in 0.1 M phosphate buffer, pH 7.2, at 30
°C under aerobic and anaerobic conditions. Right panel: time-dependent recovery of activity upon 10 s of bubbling with air or O2 of BPAO (1.0 µM) that
was anaerobically inhibited by incubation with (S)-1 (1.2 µM) for 60 min. Enzyme activity represents ∆A250 over the first 40 s upon dilution into the
air-equilibrated benzylamine assay mixture.
Table 1. Kinetic Parameters for O2-Dependent Activity Recovery
the benzenetriol form of the reduced TPQ cofactor predicted
by addition-elimination. The only caveat would be if the
of BPAO Anaerobically Inactivated by (S)-1a
1
atmosphere
k1 (min-
)
t1 2 (min)
/
Act∞ (%)
enzyme uses addition-elimination and the presence of product
2 bound to the enzyme somehow slowed the otherwise
instantaneous O2-dependent oxidation of the benzenetriol form
of the TPQ cofactor. An independent experiment established
that incubation of both native and borohydride-reduced BPAO
with 1 equiv of 2 did not result in a detectable diminution of
enzyme activity as measured by the benzylamine assay (Sup-
porting Information). We conclude that the kinetics of the
interaction of BPAO with (S)-1 in Figure 6 represents a
transamination reaction pathway. The time-dependent anaerobic
inactivation indicates formation of the “coupling product” 9,
whereas the aerobic enzyme activity recovery reflects regenera-
tion of the TPQ cofactor through the oxidation-hydrolysis
pathway via quinonimine 10 (Chart 1). The fact that the enzyme
activity recovery upon O2 bubbling is 3.5 times faster than upon
air bubbling (Table 1) suggests that oxidation of the coupling
product cofactor is at least partially rate-limiting in the oxidatiVe
half-reaction.
air
O2
0.25 ( 0.01
0.88 ( 0.02
2.77
0.79
99.1 ( 0.8
98.8 ( 0.3
a The inactivated enzyme samples were generated by incubating BPAO
(1.0 µM) with (S)-1 (1.2 µM) in 0.1 M phosphate buffer, pH 7.2, under
argon at 30 °C for 60 min and then bubbled with either air or O2 for 10 s.
require the oxidative formation and subsequent hydrolysis of
quinonimine 10 (Scheme 4), which, on the basis of our model
studies, would not be instantaneous. On the basis of these
scenarios, the mechanism used by the enzyme could be
theoretically distinguished by any one of three different ap-
proaches: (i) determining the kinetics of aerobic recovery of
anaerobically inhibited enzyme, (ii) determining whether 2 is
present after anaerobic single turnover or forms only upon
aerobic reoxidation, and/or (iii) determining the fate of the
cofactor upon anaerobic single turnover by spectroscopic means.
According to the first approach, the activity of BPAO during
incubation with a slight excess of (S)-1 under aerobic and
anaerobic conditions was monitored over time by the standard
aliquot method (10-fold dilution into buffered benzylamine and
measurement of the slope for benzaldehyde formation over the
first 40 s). In the aerobic reaction (Figure 6, left panel), the
enzyme activity dipped a bit (to a minimum of 83%) at 5 min
and then fully recovered within 30 min. When the same
incubation was conducted under argon, the enzyme activity
decreased all the way down to ∼20% in 20 min and then
remained constant within the time frame of observation (2 h).
However, bubbling of the anaerobic incubation solution with
air or O2 after 1 h resulted in a time-dependent recovery of
enzyme activity (Figure 6, right panel). The recovery of enzyme
activity upon exposure to air or O2 appeared to be exponential
and could be analyzed by pseudo-first-order nonlinear least-
squares fits. Values of the calculated rate constants (k1), half-
lives (t1/2), and final enzyme activities at long reaction time
(Act∞) are listed in Table 1. The half-life for air bubbling is 3.5
times longer than the one for O2 bubbling.
It should be pointed out that the activity of BPAO incubated
with at least a slight excess of (S)-1 never reached zero even
under anaerobic conditions (Figure 6) or with a large excess of
(S)-1 (Figure 1). There are two possible explanations for
observation of this 20-30% residual activity: (i) according to
the transamination reaction, it may represent oxidation of the
coupling product 9 that occurs within the aerobic dead time of
the benzylamine assay (40 s), or (ii) it may represent a partial
contribution by an addition-elimination reaction, where 20-
30% of the TPQ cofactor is present in its benzenetriol reduced
form at the end of anaerobic single turnover.
(ii) HPLC Analysis of Anaerobically Inactivated BPAO.
The second approach was to determine if there is any significant
release of turnover product 2 during anaerobic incubation of
BPAO with (S)-1 due to a possible contributing addition-
elimination pathway. Thus, BPAO (100 µM) was incubated with
1 equiv of (S)-1 (100 µM) under argon for 30 min and then
precipitated with trichloroacetic acid (TCA) under argon. The
supernatant was withdrawn and subjected to HPLC analysis,
showing a complete absence of (S)-1 and the presence of a small
amount of 2 that corresponded to 2% of starting (S)-1. When
the same incubation was conducted under strictly anaerobic
The noninstantaneous O2-dependent recovery of enzyme
activity upon sampling of the anaerobic incubation reaction
supports a transamination mechanism for the processing of (S)-1
by BPAO and is inconsistent with the presence at this time of
9
6212 J. AM. CHEM. SOC. VOL. 128, NO. 18, 2006

Cofactor Derivatization of Copper Amine Oxidase
A R T I C L E S
Figure 7. Left panel: UV-vis difference spectrum (against native BPAO) obtained from anaerobic reaction of BPAO (80 µM) with (S)-1 (100 µM) in 0.1
M phosphate buffer, pH 7.2, at 25 °C. Right panel: UV-vis difference spectra between 6 and TBHBQ, and between (2 + 7) and TBHBQ. All the original
spectra are shown in Supporting Information (Figures S4 and S5).
conditions (added glucose/glucose oxidase/catalase under argon),
only 0.7% 2 was detected. The caveat for interpreting this result,
however, is that the actual amount of 2 generated in the reaction
could have been underestimated due to adsorption to the protein
during TCA precipitation. A control experiment where the
anaerobic incubation mixture was exposed to air for 60 min to
ensure complete turnover and release of 2 prior to precipitation,
indeed revealed only a 65% accountable “recovery” of 2. The
recovery did not improve when the same protein sample was
first denatured anaerobically with guanidine/â-mercaptoethanol
and then worked up by the same procedure. Other control
experiments, (i) omitting the initial anaerobic period and (ii)
incubating 1.2 equiv of 2 with the native enzyme, resulted in a
similar loss of 35-40% 2 following precipitation of the protein.
As expected, greater losses of 2 were observed when using lower
concentrations of (S)-1 relative to enzyme. Despite the partial
unaccountability of 2 due to apparent absorption to the protein,
the finding that only trace levels of 2 are detected upon analysis
of the anaerobic incubation is inconsistent with a significant
contribution by the addition-elimination reaction. This suggests
that the residual 20-30% enzyme activity seen upon sampling
the preparation of BPAO anaerobically inhibited by (S)-1 (Figure
6) represents aerobic recovery of enzyme activity during the
40 s benzylamine assay. This conclusion is consistent with the
measured half-life of activity recovery (Table 1).
(iii) UV-Vis Spectrophotometric Studies on the Reaction
of BPAO with (S)-1. The third approach to determining enzyme
mechanism was to spectroscopically observe the TPQ-derived
intermediates following anaerobic single turnover. Addition of
degassed (S)-1 (100 µM) to an anaerobic solution of BPAO
(80 µM) in 0.1 M phosphate buffer, pH 7.2, immediately caused
bleaching of the TPQ chromophore at 482 nm and the appear-
ance of several new peaks, as clarified by the difference spectra
against the native BPAO (Figure 7, left panel). No further
spectral changes were detected over time following the initial
rapid reductive half-reaction. Exactly the same difference spec-
trum was obtained when the reaction was conducted in strictly
anaerobic conditions (glucose/glucose oxidase/catalase/argon).
To help interpret the initial spectral change, reference
difference spectra were generated from the TBHBQ-derived
model species for either the transamination outcome (TBHBQ
vs 6) or the addition-elimination outcome (TBHBQ vs 2 +
7).25 The key spectral difference between these two outcomes
(Figure 7, right panel) lies in the range of 300-400 nm, with
transamination featuring a peak at 307 nm and a shoulder at
∼370 nm, whereas addition-elimination features only a broad
band at 319 nm. The difference spectrum for the enzyme case
did not exactly match either of the two model difference
spectra: the ∼370 nm shoulder peak matches the same feature
in the model difference spectrum for transamination, but the
318 nm peak appears consistent with the model difference
spectrum for addition-elimination. We considered the possibil-
ity that the enzyme spectrum may reflect a conformational
rearrangement in the active site, permitting interaction with the
copper, but the presence of Cu(II) did not significantly alter
either model difference spectrum. The finding that the longest
wavelength peaks for the model compounds are at ∼319 nm
for 2 + 7 and ∼370 nm for 6 likely reflects red-shifting of the
n f π* transition due to the extended π conjugation present in
6 (when planar in solution) that is not present in 2 + 7. We
suggest that conformational constraints in the enzyme active
site may cause the 307 nm band in 6 to shift to 318 nm, which
just happens to coincide with the lowest energy transition in 2.
In the case of possible conformational restrictions, one cannot
expect an exact correspondence between reference model
compound spectra and the spectroscopic behavior of the enzyme
reaction.
When the above anaerobically inactivated BPAO was exposed
to O2, time-dependent spectral changes were observed (Figure
S6 in Supporting Information). The major event at short time
(first 20 min) was the appearance and growth of two peaks at
320 and 250 nm (Figure S7 in Supporting Information),
indicating the generation of some 2, but there were no spectral
changes observed in the visible region indicative of a change
in status of the cofactor. Indeed, the benzylamine assay showed
that the enzyme was still mainly inhibited at this point (27%
activity). This “lag” results from the need to metabolize the 20
µM excess of (S)-1 and the sluggish reoxygenation of the
concentrated enzyme solution under the conditions of simple
O2 exposure (diffusion-limited).
At longer times, however, dramatic spectral changes were
observed (Figure 8). First, the constant decrease at 250 nm right
(25) We considered correcting these difference spectra for the fact that (S)-1
was in excess (by 25%) over the enzyme. However, because of the weak
relative extinction coefficient of 1, there was only minor alteration of the
difference spectra near 290 nm, a noncritical region of interpretation.
9
J. AM. CHEM. SOC. VOL. 128, NO. 18, 2006 6213

A R T I C L E S
Qiao et al.
chromophore), either of which is in equilibrium with native
BPAO and free 2.
To address the nature of the 460 nm species, we first asked
the question of what other types of molecules might cause the
same spectral shift. Aniline, but not its N-methyl or N,N-
dimethyl derivatives, was able to reproduce the blue shift of
the TPQ chromophore. Of course, the lack of reaction of the
latter two could represent a steric effect rather than their inability
to form a Schiff base. Nonetheless, if the unique behavior of
aniline reflected its ability to form the TPQ Schiff base,
borohydride reductive quenching of the complex should lead
to a catalytically inactive anilinoresorcinol form of the cofactor
(oxidation and hydrolysis, if it occurred, would at least be
noninstantaneous), as occurs with substrate benzylamine.26
However, under the same conditions where borohydride reduc-
tion of the native enzyme resulted in an immediate total quench
of the TPQ chromophore, yet rapid assay displayed full activity
(see Supporting Information), NaBH4 treatment of the BPAO-
aniline complex only partially quenched the 420 nm chromo-
phore, and there was no reduction in activity upon immediate
enzyme assay (data not shown). The independently generated
BPAO-2 complex behaved the same way. Although one
possible explanation is that the arylamines are forming a
noncovalent complex in a manner that blocks borohydride access
to the TPQ, further studies will be required to define the nature
of the 460 nm chromophore. Comparison of the difference
spectrum in Figure 8 to appropriate model compound difference
spectra was also inconclusive (Supporting Information).
Figure 8. Time-dependent (10 min interval) UV-vis difference spectra
recorded after O2 exposure for 20 min of BPAO (80 µM) anaerobically
inactivated with (S)-1 (100 µM). The spectra were generated against the
spectrum at the 20 min time point.
Discussion
Enantioselectivity in the Processing of 1. The copper-
containing quinone-dependent amine oxidases have been histori-
cally described as an enzyme class capable of conversion of
unbranched primary amines to aldehydes. Since chiral, branched
amines such as R-methylbenzylamine are then nonsubstrates,
there has been little information on possible enantioselective
metabolism by these enzymes. Recently, two nonmammalian
quinoenzymes were reported to be capable of metabolizing the
branched primary amine amphetamine,27 though substrate activ-
ity was poor and identification of the ketone product was not
described. It has been presumed that the usual failure of TPQ-
dependent amine oxidases to process R-branched primary amines
to ketones is one of steric unacceptance. The constrained bond
angles of the five-membered ring confer a steric advantage to
1, possibly explaining its ability to be readily processed by
BPAO, though it must be appreciated that complete metabolism
represents recovery from a mechanism-based inactivation event.
The high degree of enantioselectivity for metabolic production
of the final turnover product 2 from (S)-1 (Figure 2) seems to
be related to the different inhibitory behavior between the R
and S isomers shown in Figure 1. Theoretically, enantio-
discrimination in the processing of 1 can occur at one or more
of the three steps implicated in the overall reaction prior to
stereoconvergence between processing of the two enanti-
omers: (1) initial binding and formation of the carbinolamine
generated from nucleophilic attack at the TPQ C5 carbonyl, (2)
dehydration of the carbinolamine to give the SSB, and (3)
Figure 9. UV-vis difference spectra (against the anaerobically (S)-1-
inactivated BPAO) of native BPAO, of anaerobically (S)-1-inactivated
BPAO after 2 h of O2, and of the latter mixture after dialysis against 0.1 M
phosphate buffer, pH 7.2, for 24 h. The original spectra are shown in Figure
S8 in Supporting Information.
after complete turnover of the excess (S)-1 at the 20 min time
point is consistent with conversion of coupling product 9 to
turnover product 2, since although the model coupling product
6 and 2 both absorb at 250 nm, the extinction coefficient for 6
is higher (Figure S4 in Supporting Information). The most
evident spectral change, however, was the appearance and
growth of a new peak at 460 nm (the enzyme solution gradually
turned reddish yellow), which maximized after 70 min (Figure
8), and then persisted at least for 48 h. The benzylamine assay
showed that the enzyme activity had fully recovered at this stage.
Upon dialysis against 0.1 M phosphate buffer, pH 7.2, for 24
h, the 460 nm chromophore readily shifted to the TPQ
chromophore at 482 nm (individual spectra are shown in Figure
S8, Supporting Information, and the difference spectra in
Figure 9), with full preservation of the enzyme activity.
Interestingly, an aerobic titration of native BPAO with 2
recapitulated the reddish-yellow color and the same 460 nm
spectroscopic signature (Figure S9 in Supporting Information),
with full activity being retained according to the benzylamine
assay. These latter observations suggest that the 460 nm
chromophore represents either quinonimine 10 or a high-affinity
noncovalent enzyme product complex (that perturbs the TPQ
(26) Hartmann, C.; Klinman, J. P. J. Biol. Chem. 1987, 262, 962-965.
(27) Hacisalihoglu, A.; Jongejan, A.; Jongejan, J. A.; Duine, J. A. J. Mol. Catal.
B. Enzymol. 2000, 11, 81-88.
9
6214 J. AM. CHEM. SOC. VOL. 128, NO. 18, 2006

Cofactor Derivatization of Copper Amine Oxidase
A R T I C L E S
Chart 2
Because both transamination and addition-elimination path-
ways require active-site base-mediated removal of the same
substrate R-proton, the lack of observation of addition-
elimination for the TPQ-dependent enzymes is not likely to
reflect orientation issues. Instead, the preference for trans-
amination in both model and enzyme reactions probably reflects
the short lifetime of the carbinolamine intermediate that precedes
the substrate Schiff base. A detailed study of the mechanism of
amine oxidation by PQQ analogues revealed that transamination
predominates below pH 10, though addition-elimination com-
petes at higher pH where dehydration of the carbinolamine is
slowed.10b Although so far still unobserved, the appearance of
an addition-elimination pathway for the TPQ-dependent amine
oxidases might still transpire if structural features in the enzyme
and/or substrate lengthen the lifetime of the carbinolamine
intermediate by slowing its dehydration to the SSB. At the same
time, the observation of transamination in the present study may
indicate that TPQ is more intrinsically disposed toward this
mechanism than is PQQ.
tautomerization of SSB to PSB mediated by the active-site
catalytic base. The latter step has been the key focus of previous
studies to determine possible stereospecific R-deprotonation of
normal prochiral primary amine substrates.28 R-Deprotonation
was shown to be nonstereospecific in the cases of BPAO
metabolism of dopamine and tyramine,29-31 suggestive of the
possibility for “mirror image” binding of these substrates with
respect to the catalytic base. In the case of BPAO metabolism
of benzylamine, both nonstereospecific32 and pro-S-specific33
R-deprotonation have been reported. If one considers a preferred
abstraction of the pro-S R-proton for bound benzylamine as the
reference (the benzylamine pro-S and pro-R positions are labeled
in Chart 2), 1 can be superimposed on the benzylamine molecule
to align the two amino groups and to align the phenyl ring of
benzylamine with the π portion of inhibitor 1 (4,5 CdC and
carboxyl). When this is done (Chart 2, left panel), the available
R-proton of the superimposed (S)-1 coincides with the pro-S
R-proton of benzylamine. Assuming the latter overlap represents
the preferred orientation for binding to BPAO, active-site base-
mediated removal of the R-proton for (R)-1 would require
accommodation of a different spatial orientation of the molecule
(Chart 2, right panel), which we propose to be unfavorable.
Mechanism. At the outset, the apparent dehydrogenation
process whereby 1 is metabolized to 2 by BPAO suggested
consideration of an addition-elimination mechanism. However,
we have demonstrated that transamination is the only pathway
involved in both the model and the enzyme reactions. In the
model case, the apparent addition-elimination products 2 and
7 formed in the anaerobic reaction of 1 with TBHBQ were
shown to arise from the primary transamination product 6 via
a redox-cycling process as evidenced by the ester-labeling
experiments (Supporting Information). In the enzyme case,
involvement of transamination as the only pathway was
confirmed by kinetic studies on enzyme reactivation and HPLC
analysis of the reaction mixture following anaerobic incubation
of the enzyme with (S)-1. The observed net dehydrogenation
(as opposed to the typical deamination) results from an
unprecedented pathway where the product Schiff base tautomer-
izes to an aromatic “coupling product” derivative of the TPQ
cofactor, thereby averting the normal product Schiff base
hydrolysis, but which can still ultimately undergo oxidation to
a quinonimine intermediate that releases the apparent dehydro-
genation product 2 upon hydrolysis (or transimination with
starting 1).
Role of Cu(II) in the Oxidative Half-Reaction. Regardless
of whether the 460 nm species generated at the end of
O2-dependent reactivation of anaerobically 1-inactivated BPAO
represents quinonimine 10 or a BPAO-2 noncovalent complex,
the finding that enzyme activity at this stage is indistinguishable
from the control shows that conversion of this enzyme form to
substrate Schiff base in the presence of substrate (e.g., benzyl-
amine) is rapid relative to its O2-dependent generation from 9.
A conclusion that oxidation of 9 is the rate-limiting step of the
oxidatiVe half-reaction would be consistent with the obserVed
dependence on O2 (O2 > air) of the recoVery of enzyme actiVity
(Table 1).
What factors control the mechanism of oxidation of coupling
product 9, relative to the normal oxidation half-reaction of the
enzyme? In this regard, it is interesting to compare the kinetic
patterns between the model and enzyme cases. In the model
reaction, whereas transition metal-free autoxidation of coupling
product 6 is relatively slow (t1/2 1.5 h), the presence of 1 mol
equiv of Cu(II) accelerates the autoxidation of 6 by a factor of
770 (t1/2 7 s). In comparison, aerobic oxidation of 9 in the
enzyme has a t1/2 of 2.8 min (Table 1), nearly the logarithmic
average between the two model system results. Transition metal
ions are well-known to be capable of catalyzing autoxidation
reactions via coordination to an autoxidizable substrate and
activation of the latter toward reaction with O2. Because it is
hard to imagine how other active-site residues would promote
an autoxidation reaction, the intermediate rate in the enzyme
case suggests that the active-site Cu(II) is promoting the
oxidation of 9, but not to the extent that would be expected if
9 could coordinate directly to the active-site Cu(II). At least in
the native copper-containing amine oxidases, when TPQ is in
the active conformation, the active-site Cu(II) only indirectly
interacts with the TPQ cofactor (at best through a water molecule
to the C2 oxygen),1,34,35 as opposed to the chelating coordination
possible in the model system between Cu(II) and the C4 oxygen
and C5 nitrogen of 6 (Chart 1). We propose that the nearby
(28) Coleman, A. A.; Scaman, C. H.; Kang, Y. J.; Palcic, M. M. J. Biol. Chem.
1991, 266, 6795-800.
(29) Summers, M. C.; Markovic, R.; Klinman, J. P. Biochemistry 1979, 18,
1969-79.
(30) Yu, P. H. Biochem. Cell. Biol. 1988, 66, 853-61.
(31) Coleman, A. A.; Hindsgaul, O.; Palcic, M. M. J. Biol. Chem. 1989, 264,
19500-5.
(34) Mills, S. A.; Goto, Y.; Su, Q.; Plastino, J.; Klinman, J. P. Biochemistry
2002, 41, 10577-84.
(35) Kishishita, S.; Okajima, T.; Kim, M.; Yamaguchi, H.; Hirota, S.; Suzuki,
S.; Kuroda, S.; Tanizawa, K.; Mure, M. J. Am. Chem. Soc. 2003, 125,
1041-55.
(32) Yu, P. H.; Davis, B. A. Int. J. Biochem. 1988, 20, 1197-201.
(33) Battersby, A. R.; Staunton, J.; Klinman, J. P.; Summers, M. C. FEBS Lett.
1979, 99, 297-298.
9
J. AM. CHEM. SOC. VOL. 128, NO. 18, 2006 6215

A R T I C L E S
Qiao et al.
vicinity of the Cu(II) facilitates, at least to some extent, electron
transfer from 9 to O2.
1 is an example of what can be termed mechanism-based
reVersible inactivation as opposed to the traditional competitive
reversible inhibition or mechanism-based irreversible inactiva-
tion. Fine-tuning of the redox potential of the “coupling product”
would theoretically permit control of the duration of inhibition
independently of the potency of inhibition. This suggests a new
concept in the design of enzyme inhibitors that may be
applicable to a wide range of enzyme systems.
One may ask the question of whether the reaction of 9 and
O2 follows the same pathway as for the oxidative half-reaction
in normal turnover, where it has been proposed that electron
transfer from TPQamr (Scheme 1) occurs to a prebound O2,36
facilitated by the active-site Cu(II). In normal turnover, cofactor
reoxidation for BPAO is characterized by low micromolar
Km(O2),37 such that saturation is already achieved in normal
aerobically equilibrated bulk solvent. In the case of autoxidation
of 9, however, the increased rate of the oxidative half-reaction
in going from air exposure to pure O2 exposure suggests that
the mechanism of oxidation of 9 may be distinct from that used
for oxidation of TPQamr. Alternatively, formation of adduct 9
may disrupt the normal binding site(s) for O2.
Experimental Section
General Procedures. All NMR experiments were run at ambient
temperature, on a Varian Gemini 200 or 300 MHz instrument. Chemical
shifts were referenced to the residual proton peaks in the deuterated
solvents, and 13C APT (attached proton test) designations are given as
(+) or (-). Benzylamine assays were conducted on a Perkin-Elmer
Lambda 3B instrument using the PECSS software, with constant
temperature being maintained by a water-jacketed multiple cell holder.
UV-vis spectra were recorded on a Perkin-Elmer Lambda 20 spec-
trophotometer. High-resolution mass spectra (HRMS) were obtained
by electron impact ionization (EI) (20-40 eV) or fast atom bombard-
ment (FAB) on a Kratos MS25RFA instrument. Ethyl 4-amino-4,5-
dihydrothiophene-2-carboxylate (1), both as a mixture and as optically
pure enantiomers, was synthesized according to a literature method.13
The S enantiomer was prepared starting from the commercially available
D-cysteine, which was first converted to its ethyl ester according to a
literature method.39 5-tert-Butyl-2-hydroxy-1,4-benzoquinone (TBHBQ)
and 5-tert-butyl-1,2,4-trihydroxybenzene (7) were prepared as described
previously.3b Bovine plasma amine oxidase (BPAO) for inhibition
assays (100 units/g of protein), glucose oxidase (245.9 units/mg of
solid), and catalase (1870 units/mg of solid) were purchased from Sigma
or Worthington. BPAO used for kinetics, HPLC, and UV-vis spec-
trophotometric studies was purified according a reported procedure.5
The purified protein shows a single band on SDS-PAGE. Enzyme
concentrations are in terms of active monomers as estimated from the
rate of benzylamine oxidation (1 unit oxidizes 1.0 µmol of benzylamine
to benzaldehyde per minute at 25 °C), using an activity of 0.48 unit/
mg of protein for the pure monomer of molecular weight 85 000 and
∆ꢀ250 ) 12 800 M-1 cm-1 for benzaldehyde [corresponding to an A250
of 12.8 min-1 (unit of activity)-1 for 1 mL volume at 25 °C].
Concentrations measured this way are very close to those estimated
by standard phenylhydrazine titration.5 Stock solutions of 1 and 2 were
prepared fresh by dilution with 0.1 M phosphate buffer, pH 7.2, of a
10 mM stock solution in pure water. The argon used for anaerobic
experiments was deoxygenated and moisturized by passage sequentially
through (1) a tower of V(II) [V2O5 dissolved in aqueous sulfuric acid
in the presence of excess Zn(Hg)] and (2) a tower of aqueous NaHCO3.
Kinetics of Aerobic Interaction of BPAO with 1. Solutions of
BPAO and various concentrations of 1 in 0.9 mL of 0.1 M sodium
phosphate buffer, pH 7.2, were incubated at 30 °C. Aliquots (100 µL)
of the incubation mixture were withdrawn periodically and diluted into
1 mL of 5 mM benzylamine in 0.05 M sodium phosphate buffer, pH
7.2, at 30 °C (total volume 1.1 mL), and the BPAO activity was
determined spectrophotometrically by monitoring the benzaldehyde
production at 250 nm over the first 40 s. A control incubation lacking
1 was set up at the same time. Remaining enzyme activity (Figure 1)
was calculated by comparison with the control.
Conclusion
To our knowledge, 1 is the first branched primary amine
reported to act as a submicromolar Ki inhibitor of a quinone-
dependent CAO yet is ultimately processed as a substrate. This
activity is unique, first in that this family of enzymes is not
known to metabolize branched primary amines, and there is only
one other report (without characterization of the product) that
a branched primary amine can be a substrate. Second, the finding
of an enantioselective inhibitory potency is unprecedented, and
if it represents stereoselective CR deprotonation of the two
possible substrate Schiff base epimers, the greater potency of
(S)-1 is consistent with the pro-S selectivity of CR proton
removal for benzylamine.
We chose to investigate the aminodihydrothiophene 1 on the
basis of its tendency toward aromatization that might drive an
initially reversible addition-elimination or transamination step.
Since branched primary amines are not known as substrates for
the quinone-dependent CAOs, we considered that an observed
metabolism of 1 might occur by addition-elimination. We find
that enzyme metabolism and a quinone model system both
follow strictly a transamination course, consistent with what is
now well established for unbranched primary amines.
We previously showed that an aromatization that can occur
after transamination can result in a stable derivative of the CAO
quinone cofactor in a form that is incapable of being reoxidized
to generate the turnover-competent starting TPQ.38 The series
of inhibitors that satisfy this criterion, (3-aryl-)3-pyrrolines, was
shown to result in stoichiometric inactivation of BPAO, forming
a pyrrolylresorcinol TPQ derivative, though the same com-
pounds are pure substrates of flavin-dependent MAO-B.38 In
the present work, kinetics and UV-vis spectrophotometric
studies reveal that the initial quick inhibition of BPAO by (S)-1
is due to formation of a derivative of the reduced cofactor upon
a similar aromatization step. The striking difference between
these two cases is that whereas the pyrrolylresorcinol cofactor
is inert to further oxidation, the thiophenylaminoresorcinol
cofactor is still subject to an albeit slow reoxidation and
subsequent turnover, apparently facilitated by the active-site
Cu(II), thereby rendering the observation of temporary rather
than irreversible inactivation. From this point of view, compound
To examine the reversibility of inactivation of BPAO by (S)-1, two
solutions of 0.5 mL of BPAO (1.0 µM) in 0.1 M sodium phosphate
buffer, pH 7.2, in the absence and presence of 50 µM (S)-1 were
incubated for 10 min, showing 64% remaining in the latter case. The
two samples were then transferred into two dialysis membranes (6.4
mm Spectrum Spectro/Por membrane, MW cutoff 12 000-14 000) and
dialyzed against 0.1 M sodium phosphate buffer, pH 7.2, at room
temperature for 24 h. The enzyme activities of the two dialyzed samples
were essentially identical.
(36) Su, Q.; Klinman, J. P. Biochemistry 1998, 37, 12513-12525.
(37) Oi, S.; Inamasu, M.; Yasunobu, K. T. Biochemistry 1970, 9, 3378-3383.
(38) Lee, Y.; Ling, K. Q.; Lu, X.; Silverman, R. B.; Shepard, E. M.; Dooley,
D. M.; Sayre, L. M. J. Am. Chem. Soc. 2002, 124, 12135-12143.
9
6216 J. AM. CHEM. SOC. VOL. 128, NO. 18, 2006

Cofactor Derivatization of Copper Amine Oxidase
A R T I C L E S
Competitive Inhibition of BPAO by (S)-1. Solutions of BPAO (1.0
µM) in 0.1 M phosphate buffer, pH 7.2 (1 mL each), were incubated
at 30 °C for 10 min before various concentrations of benzylamine and
(S)-1 were added simultaneously. After quick mixing, the solutions were
subjected to spectrophotometric monitoring at 250 nm to yield slopes
of PhCHO production over 40 s. Lineweaver-Burk and Dixon plots
were constructed (Figure 3) to obtain inhibition constants.
and heating for 6 min at 100 °C). No product 2 was detected in either
control experiment.
UV Spectrophotometric Monitoring of BPAO-Catalyzed Conver-
sion of (S)-1 to 2. A solution of (S)-1 (50 µM) and BPAO (1.6 µM) in
0.1 M sodium phosphate buffer, pH 7.2, in an open quartz cuvette (1
cm) was monitored at 30 °C by repetitive spectral scan at 6 min time
intervals. Difference spectra were then generated by Origin 7.0 by
subtracting the first spectrum from all the spectra recorded thereafter
(Figure S1 in Supporting Information).
Anaerobic Reaction of 1 with 5-tert-Butyl-2-hydroxy-1,4-benzo-
quinone in DMSO. To a solution of TBHBQ (90 mg, 0.5 mmol) and
1‚HCl (105 mg, 0.5 mmol) in 5 mL of degassed DMSO was added
degassed diisopropylethylamine (DIPEA, 350 µL, 2 mmol). The
solution was stirred for 20 h under argon and then quenched by adding
3.5 mL (25 mmol) of degassed triethylamine and 2 mL (21 mmol) of
degassed acetic anhydride. The resultant mixture was stirred at room
temperature for 2 h and partitioned between 100 mL of water and 100
mL of CH2Cl2. The organic layer was separated, washed with water (3
× 50 mL), dried (Na2SO4), and concentrated, and the residue was
subjected to flash silica gel column chromatography with hexanes-
EtOAc as eluent to afford three acetylated products:
Preparation of Ethyl 4-Aminothiophene-2-carboxylate (2) by
Catalytic Dehydrogenation of 1. To a solution of 1‚HCl (181 mg, 1
mmol) in 100 mL of ethanol-benzene (1:1 v/v) were added triethyl-
amine (420 µL, 3 mmol) and 10% Pd/C (40 mg). The mixture was
heated at reflux for 4 h until TLC indicated total conversion of the
starting material. Pd/C was filtered off and the filtrate was concentrated
under reduced pressure. The residue was applied to a short silica gel
column with hexanes/ethyl acetate (1:1) as eluant to afford the free
1
base form of 2 (175 mg, 84%): light yellow oil; H NMR (CDCl3) δ
1.33 (t, 3H, J ) 7.1 Hz), 3.67 (br s, 2H), 4.29 (q, 2H, J ) 7.1 Hz),
6.37 (d, 1H, J ) 1.8 Hz), 7.29 (d, 1H, J ) 1.8 Hz); 13C NMR (APT,
CDCl3) δ 14.4 (-), 61.2 (+), 107.8 (-), 126.0 (-), 132.8 (+), 145.5
(+), 162.3 (+); HRMS (FAB) calcd for C7H9NO2S 171.0354, found
171.0347. The sample was further converted to the HCl salt form by
dissolution in 100 mL of ethanol followed by addition of concentrated
aqueous HCl to pH 3. Removing solvent and crystallization of the
residue from EtOH/EtOAc (1:4) afforded 2‚HCl as white crystals: mp
185-187 °C (decomp); 1H NMR (CD3OD) δ 1.38 (t, 3H, J ) 7.1 Hz),
4.38 (q, 2H, J ) 7.1 Hz), 7.78 (m, 1H), 7.85 (m, 1H).
TLC Detection of Turnover Product 2 in Metabolism of 1 by
BPAO. A solution of 0.7 µM BPAO and 100 µM 1‚HCl (R,S mixture
63:37) in 1 mL of 0.1 M sodium phosphate buffer, pH 7.2, was
incubated at 30 °C. A control incubation without BPAO was set up at
the same time. After 24 h, the remaining enzyme activity was
determined to be 20%. The incubation solution was then extracted with
ethyl acetate (3 × 5 mL), and the organic extracts were combined,
dried (Na2SO4), and evaporated to one drop. The residue was applied
to a TLC plate, and the authentic 2 was loaded on the same plate. A
mixture of CH2Cl2/MeOH (5:1) was used as eluent, and 2 was visualized
by UV light (bright blue at 365 nm) or ninhydrin spray (dark pink
color). The product and authentic 2 both showed identical spots with
Rf ) 0.3. No other products were apparent.
Monoacetylated 2 [ethyl 4-(acetamido)thiophene-2-carboxylate, 37
1
mg, 35%]: light brown solid; mp 104-106 °C; H NMR (CDCl3) δ
1.37 (t, 3H, J ) 7.1 Hz), 2.17 (s, 3H), 4.34 (q, 2H, J ) 7.1 Hz), 7.65
(d, 1H, J ) 1.5 Hz), 7.78 (br s, 1H, NH), 7.81 (d, 1H, J ) 1.5 Hz); 13
C
NMR (CDCl3) δ 14.3, 23.9, 61.4, 117.3, 126.2, 132.3, 135.9, 162.1,
167.9; HRMS (FAB) calcd for C9H12NO3S (MH+) 214.0538, found
214.0539.
Triacetylated 6 [ethyl 4-(N-(2,4-diacetoxy-5-tert-butylphenyl)acet-
amido)thiophene-2-carboxylate, 135 mg, 58%]: light brown gummy
1
solid; mp 50-52 °C; H NMR (CDCl3) δ 1.28 (t, 3H, J ) 7.2 Hz),
1.30 (s, 9H), 1.95 (s, 3H), 2.07 (s, 3H), 2.29 (s, 3H), 4.25 (q, 2H, J )
7.2 Hz), 7.01 (s, 1H), 7.25 (br s, 1H), 7.30 (br s, 1H), 7.64 (d, 1H, J
) 2.0 Hz); 13C NMR (APT, CDCl3) δ 14.3 (-), 20.5 (-), 21.6 (-),
23.5 (-), 30.1 (-), 34.6 (+), 61.3 (+), 120.1 (-), 121.3 (-), 128.1
(-), 129.2 (-), 131.2 (+), 132.0 (+), 140.0 (+), 141.0 (+), 144.8
(+), 149.1 (+), 161.8 (+), 168.2 (+), 168.5 (+), 170.1 (+); HRMS
(FAB) calcd for C23H27NSO7 (MH+) 462.1586, found 462.1575.
Triacetylated 7 (1,2,4-triacetoxy-5-tert-butylbenzene,40 51 mg,
1
33%): white crystals; mp 118-119 °C; H NMR (CDCl3) δ 1.33 (s,
HPLC Monitoring of the Conversion of 1 to 2 Catalyzed by
BPAO. A solution of 40 µM 1 (R or S enantiomer) and 0.7 µM BPAO
in 1 mL of 0.1 M sodium phosphate buffer, pH 7.2, was incubated at
30 °C. Aliquots (100 µL) were taken periodically, and 50 µL of
acetonitrile (containing 0.02% trifluoroacetic acid, TFA) was added to
denature the enzyme. The resulting suspension was then centrifuged,
and 20 µL fractions of the supernatant were subjected to HPLC analysis
using a 4.6 × 250 mm Agilent SB C18, 5 µm column, a flow rate of
1 mL/min, and a gradient mobile phase composed of HPLC-grade
solvents A [5% aqueous CH3CN containing 0.02% (v/v) TFA] and B
[95% aqueous CH3CN containing 0.02% (v/v) TFA] according to the
following program: 100% A to 93% A 0-20 min, 93% A to 50% A
20-25 min, 50% A to 100% A 25-30 min, 100% A 30-35 min. With
this solvent program, compounds (S)-1 and 2 had retention times of
11.8 and 12.5 min, respectively. The concentrations of (S)-1 and 2 were
determined according to their HPLC peak integrations (both monitored
at 270 nm), calibrated according to independently established standard
curves. The time-dependent plots for consumption of 1 and formation
of 2 are shown in Figure 2. Nonlinear least-squares fits of the product
formation plots were conducted by use of Origin 7.0 software according
to the equation [2] ) [2]∞(1 - e-kt), where [2] and [2]∞ represent the
concentrations of product 2 during and at the end of reaction,
respectively. A control run in the absence of the enzyme was set up at
the same time. In a second control, compound 1 was incubated with
the same amount of denatured enzyme [by addition of 10% (v/v)
denaturing solution (2% 2-mecaptoethanol, 10% glycerol, and 4% SDS]
9H), 2.26 (s, 3H), 2.28 (s, 3H), 2.32 (s, 3H), 6.97 (s, 1H), 7.15 (s, 1H);
13C NMR (APT, CDCl3) δ 20.7 (-), 21.6 (-), 30.0 (-), 34.5 (+),
119.0 (-), 121.6 (-), 139.0 (+), 139.76 (+), 139.81 (+), 146.3 (+),
168.0 (+), 168.3 (+), 168.9 (+).
The same reaction was repeated but with omission of the acetic
anhydride trapping step. The final product mixture was carefully
separated by column chromatography using hexanes/ethyl acetate as
eluent to give products 2 (30 mg, 35%) and 6 (88 mg, 52%). Compound
2 (free base form) was identical with an authentic sample as shown by
NMR and HPLC. Triol 7 was autoxidized to TBHBQ during workup
and was not recovered.
Ethyl 4-((5-tert-Butyl-2,4-dihydroxyphenyl)amino)thiophene-2-
1
carboxylate (6): brown amorphous solid; mp 73-75 °C; H NMR
(CDCl3) δ 1.35 (s, 9H), 1.36 (t, 3H, J ) 7.1 Hz), 4.33 (q, 2H, J ) 7.1
Hz), 5.11 (br s, 1H), 5.25 (br s, 1H), 5.92 (br s, 1H), 6.29 (d, 1H, J )
1.8 Hz), 6.40 (s, 1H), 7.03 (s, 1H), 7.36 (d, 1H, J ) 1.8 Hz); 13C NMR
(APT, DMSO-d6) δ 14.2 (-), 29.8 (-), 33.8 (+), 60.7 (+), 104.2 (-),
104.4 (-), 119.5 (-), 121.7 (+), 125.9 (+), 126.1 (-), 130.7 (+),
146.7 (+), 147.5 (+), 151.4 (+), 161.6 (+); HRMS (FAB) calcd for
C17H21NO4S (M+) 335.1191, found 335.1188.
NMR Tube Reaction of 1 with TBHBQ. To a 5 mm NMR tube
containing a solution of 1‚HCl (10.6 mg, 0.05 mmol) and TBHBQ
(39) Meng, Q.; Li, Y.; He, Y.; Guan, Y. Tetrahedron: Asymmetry 2000, 21,
4255-4261.
(40) Blatchly, J. M.; Green, R. J. S.; McOmie, J. F. W. J. Chem. Soc., Perkin
Trans. 1 1972, 2286-2291.
9
J. AM. CHEM. SOC. VOL. 128, NO. 18, 2006 6217

A R T I C L E S
Qiao et al.
(9.0 mg, 0.05 mmol) in 0.7 mL of degassed DMSO-d6 at room
temperature was added DIPEA (35 µL, 0.2 mmol) to initiate the
reaction. The H NMR spectra were recorded periodically over 20 h.
was added 18 µL of 50 mM aqueous DTPA. The solution was mixed
and subjected to repetitive UV-vis spectral scan, showing release of
8 within 1 min (Figure S2 in Supporting Information). Hydrolysis of
8 to form TBHBQ and 2 was evident following overnight standing of
the solution.
To determine the rate constant of Cu(II)-mediated autoxidation of 6
to 8-Cu(II), the above reaction was repeated with 0.1 mM 6 and 0.1
mM Cu(II). After quick mixing, the absorbance at 390 nm was
monitored at 25 °C over time (Figure S3 in Supporting Information).
The pseudo-first-order rate constant was calculated by a double-variable
nonlinear least-squares fit of the A390 versus t plot with Origin 7.0
according to the equation At ) A∞ - (A∞ - A0)e-kt, where A0, At, and
A∞ represent A390 at the beginning, during, and at the end of the reaction.
The A∞ obtained was close to the theoretical value predicted from the
UV-vis spectrum of authentic 8-Cu(II).
Kinetic Studies on Stoichiometric Reactions of BPAO with (S)-
1. For aerobic reactions, a solution of BPAO (1.0 µM, 1.1 mL) in 0.1
M sodium phosphate buffer, pH 7.2, was incubated at 30 °C for 10
min and then a 100 µL aliquot was withdrawn and assayed to establish
the control activity. An independent experiment indicated that the
enzyme activity remained constant at least for several hours. To the
above BPAO solution was added 5 µL of 0.24 mM aqueous (S)-1 (1.2
equiv) at 30 °C. After quick mixing of the solution, aliquots (100 µL
each) were periodically assayed to afford the time-dependent enzyme
activity plot (Figure 6, left panel).
For anaerobic inactivation, a solution of BPAO (1.0 µM, 1.1 mL)
in 0.1 M sodium phosphate buffer, pH 7.2, was prepared in a 10 mL
culture tube and then sealed with a septum. An inlet needle was used
to connect the culture tube to a deoxygenated humidified argon gas
line. An outlet needle was also attached to the septum. The solution
was incubated at 30 °C while a slow argon stream was directed at the
top of the solution for 3 h. The outlet needle was then removed to
maintain a positive argon pressure within the culture tube. A 100 µL
aliquot was withdrawn and assayed to give the control enzyme activity.
After injection of 5 µL of degassed 0.24 mM aqueous (S)-1, the
incubation solution was quickly mixed and the enzyme activity was
assayed over time (Figure 6, left panel).
1
After completion of the reaction, the presence of peaks corresponding
to 2 and 6 was confirmed by spiking with authentic samples. Time-
dependent plots for consumption of TBHBQ and formation of 2 and 6
were constructed based on integrations of well-separated characteristic
1H NMR signals (the quinone peak at 6.06 ppm for TBHBQ; the
thiophene peak at 6.38 ppm for 2; and the thiophene peak at 7.45 ppm
for 6). Since the reaction was essentially stoichiometric and the amounts
of 2 and 7 were identical, the total integration of TBHBQ, 2, and 6
was counted as 100 and the percentages of each of the three components
were calculated and plotted over time (Figure 4).
Preparation of N-(2-(Ethoxycarbonyl)thiophen-4-yl)-2-hydroxy-
5-tert-butyl-1,4-benzoquinone-1-imine (8). A mixture of 6 (32 mg,
0.1 mmol) in 10 mL of CH3CN/pH 7.2 phosphate buffer (5:1) was
stirred for 12 h at room temperature and then extracted with CH2Cl2
(3 × 30 mL). The combined organic extract was dried (Na2SO4) and
concentrated, and the residue was subjected to flash column chroma-
tography with hexanes/EtOAc (1:1) as eluent to afford 8 (18 mg,
55%): red crystals; mp 123-125 °C; 1H NMR (CDCl3) δ 1.26 (s, 9H),
1.40 (t, 3H, J ) 4.8 Hz), 4.39 (q, 2H, J ) 4.8 Hz), 5.99 (s, 1H), 7.00
(s, 1H), 7.20 (d, 1H, J ) 1.0 Hz), 7.64 (d, 1H, J ) 1.0 Hz); 13C NMR
(CDCl3, APT) δ 14.4 (-), 29.6 (-), 35.9 (+), 61.8 (+), 108.2 (-),
119.1 (-), 122.2 (-), 129.1 (-), 135.1 (+), 146.0 (+), 153.5 (+),
154.7 (+), 157.3 (+), 161.6 (+), 187.8 (+); HRMS (EI) calcd for
C17H19NO4S (M+) 333.1035, found 333.1032.
Autoxidation of 6 and Hydrolysis of 8. A solution of 6 (0.15 mM)
in 3 mL of 1:1 (v/v) DMSO and 0.05 M 4-(2-hydroxyethyl)-1-
piperazineethanesulfonate (HEPES) buffer, pH 7.2, in an open quartz
cuvette (1 cm) was monitored at 25 °C by repetitive spectral scan on
a UV-vis spectrophotometer at 20 min time intervals (Figure 5). The
recorded spectra were analyzed with Origin 7.0. The pseudo-first-order
kinetic constants for autoxidation and hydrolysis were obtained from
linear fits of ln [(A∞ - A0)/(A∞ - At)] versus t plots (Figure 5, insets),
where A0, At, and A∞ represent A490 (autoxidation) and A420 (hydrolysis)
at the beginning, during, and at the end of the inspected reactions.
Appropriate A∞ values were chosen to allow best linear fits of the plots,
which were close to the theoretical values predicted from the UV-vis
spectrum of authentic 8. In another experiment, the same incubation
was conducted in the presence of 0.1 mM DTPA sodium salt, affording
essentially the same kinetic constants after the same data analysis. In
a third experiment, a solution of authentic 8 (0.15 mM) in 1:1 (v/v)
DMSO and 0.05 M HEPES buffer, pH 7.2, was subjected to UV-vis
spectrophotometric monitoring as above. The spectra obtained were
treated similarly to yield a hydrolysis rate constant close to that obtained
above.
To measure the O2-dependent activity recovery of anaerobically
inactivated BPAO, 2 mL samples of enzyme (1.0 µM) were anaero-
bically inactivated as above for 60 min by (S)-1 (1.2 µM) in a culture
tube. The septum was removed and the enzyme solution was bubbled
either with air or with O2 for 10 s. The solutions were then assayed
over time, yielding the time-dependent activity plot (Figure 6, right
panel). The enzyme activity recovery plots were analyzed with Origin
7.0 by double-variable (k and Act∞) pseudo-first-order nonlinear least-
squares fits according to equation Actt ) Act∞ - (Act∞ - Act0)e-kt
,
where Act0, Actt, and Act∞ represent the remaining enzyme activities
at the beginning, during, and at the end of the O2-dependent activity
recovery. The results are shown in Table 1.
To confirm the formation of TBHBQ and 2 in the above reaction,
a solution of 6 (0.2 mM) in neat 0.05 M HEPES buffer, pH 7.2 (20
mL), was stirred in air at room temperature overnight. The resulting
red solution was divided equally into two parts. The first half was
acidified with 6 N aqueous HCl to pH 1, saturated with NaCl, and
then extracted with Et2O (2 × 10 mL). The combined organic layer
was dried (Na2SO4), concentrated, and examined by TLC, showing the
presence of TBHBQ as compared with an authentic sample. The second
half of the above reaction mixture was directly saturated with NaCl
and extracted with Et2O (2 × 10 mL). TLC analysis of the dried and
concentrated organic layer indicated the presence of 2 as compared
with an authentic sample.
HPLC Analysis of Stoichiometric Reactions of BPAO with (S)-
1. All the incubations were conducted at room temperature in 5 mm
NMR tubes cut to 10 cm in length. A solution of BPAO (100 µM) in
0.1 M sodium phosphate buffer, pH 7.2 (40 µL), was prepared in a
septum-sealed NMR tube and made anaerobic by directing a slow argon
stream at the top of the solution for 3 h before 10 µL of degassed 0.5
mM aqueous (S)-1 was injected. The reaction mixture was incubated
under argon for 30 min and then 10 µL of degassed 20% aqueous
trichloroacetic acid (TCA) was injected, resulting in immediate
precipitation of the protein. After centrifugation of the solution for 5
min, the supernatant (20 µL) was withdrawn and subjected to HPLC
analysis as described above, showing the absence of (S)-1. The
concentration of 2 thereby determined was converted to yield based
on starting (S)-1.
Cu(II)-Mediated Autoxidation of 6. To a solution of 6 (0.15 mM)
in 3 mL of 1:1 (v/v) DMSO and 0.05 M HEPES buffer, pH 7.2, in an
open quartz cuvette (1 cm) was added 9 µL of 50 mM aqueous CuSO4.
After quick mixing, the solution was subjected to UV-vis monitoring
in air at 25 °C. The reaction exhibited no further spectral changes over
a period of 24 h after the first spectrum was recorded after 1 min (Figure
S2 in Supporting Information). To the resulting reddish-yellow solution
For the same reaction under strictly anaerobic conditions, a solution
of BPAO (100 µM), glucose oxidase (60 µg/mL), and catalase (60 µg/
mL) in 0.1 M sodium phosphate buffer, pH 7.2 (40 µL), was purged
9
6218 J. AM. CHEM. SOC. VOL. 128, NO. 18, 2006

Cofactor Derivatization of Copper Amine Oxidase
A R T I C L E S
with argon for 3 h and then 1 µL of degassed 1 M aqueous glucose
was added. After further argon purging for 1 h, 10 µL of degassed 0.5
mM aqueous (S)-1 was injected and the mixture was incubated for 30
min before TCA workup and HPLC analysis. A control experiment in
the absence of BPAO showed 100% recovery of (S)-1.
To examine the recovery of 2 after complete turnover of (S)-1 by
BPAO, a solution of BPAO (100 µM) in 0.1 M sodium phosphate
buffer, pH 7.2 (40 µL), was purged with argon for 3 h and appropriate
amounts of (S)-1 were injected. The reaction mixture was incubated
for 30 min and then exposed to pure O2 for 1 h before TCA workup
and HPLC analysis. In one of these experiments, 10 µL of degassed
denaturing solution composed of guanidine hydrochloride (7 M) and
â-mercaptoethanol (30 mM) was injected at the end of O2 exposure.
The mixture was heated at 90 °C for 5 min and subjected to normal
TCA workup and HPLC analysis.
Supporting Information) were converted to extinction coefficients (ꢀ)
by use of Origin 7.0.
UV-Vis Spectrophotometric Monitoring of Reactions of BPAO
with (S)-1. An anaerobic solution of BPAO (80 µM) in 0.1 M sodium
phosphate buffer, pH 7.2 (300 µL), was prepared and transferred to a
septum-sealed 1 mm quartz cuvette. The UV-vis spectrum was
recorded, and the enzyme activity was assayed (2 µL aliquot). Into
this solution was injected 10 µL of degassed 2.4 mM aqueous (S)-1.
After the sample was mixed carefully for 2 min, the enzyme activity
was assayed and the UV-vis spectrum was taken again, at which point
the TPQ chromophore had disappeared (Figure S5 in Supporting
Information). An O2 stream was directed at the solution for 1 min and
an O2 balloon was fitted to the cuvette prior to repetitive UV-vis
spectral scanning at 10 min intervals for 2 h (Figure S6 in Supporting
Information), with periodic monitoring of the enzyme activity. The
resulting reddish-yellow enzyme sample was transferred into a dialysis
membrane and subjected to dialysis for 24 h against 0.1 M sodium
phosphate buffer, pH 7.2 (3 × 250 mL). The activity was assayed and
the UV-vis spectrum was taken (Figure S9 in Supporting Information).
The same reaction was repeated in the presence of glucose oxidase
(33 µg/mL), catalase (33 µg/mL), and glucose (17 mM). The difference
spectrum for the reductive half-reaction (not shown) was essentially
the same as that in the absence of glucose/glucose oxidase/catalase.
However, it took a much longer time (∼12 h rather than 2 h) for
restoration of the TPQ chromophore on account of consumption of O2
at the expense of glucose oxidation and the slow diffusion of gas into
the viscous solution.
To examine the recovery of 2 following its direct incubation of
BPAO, a solution of BPAO (100 µM) in 0.1 M sodium phosphate
buffer, pH 7.2 (40 µL), was purged with argon for 3 h and 2 was added
to a final concentration of 100 µM. The reaction mixture was incubated
for 30 min and worked up and analyzed as above.
General Procedure for UV-Vis Spectrophotometric Monitoring
of Anaerobic Experiments with BPAO. All experiments were
conducted with a 1 mm quartz cuvette at 25 °C. Preparation of anaerobic
protein samples for spectrophotometry first involved slow filtering of
the enzyme solution (to remove any insoluble materials and ensure
good spectral quality) into a 10 mL culture tube, which was then sealed
with a septum and attached to the deoxygenated humidified argon gas
line by an inlet needle. A long needle was applied to connect the culture
tube to a septum-sealed 1 mm quartz cuvette, to which an outlet needle
was also attached. After slow streaming of argon at the top of the
solution for 3 h, the end of the long needle attached to the culture tube
was pushed down to the bottom and the other end of the long needle
attached to the cuvette was pulled up to an appropriate level, thereby
allowing the argon pressure to slowly drive the protein solution from
the culture tube into the cuvette. To ensure anaerobic conditions,
positive argon pressure was maintained within the cuvette throughout
the spectrophotometric measurements.
When difference UV-vis spectra were generated to pick up weak
signals of the enzyme cofactor from the protein background, volume
changes of the solution upon each addition of reagents were corrected
by use of Origin 7.0 to avoid artifacts in the region (∼280 nm) of
strong absorption of the protein. To ensure good resolution of the
difference spectra, concentrations of BPAO were used that kept A280
less than 2.5. In generating the difference spectra for model reactions,
the absorbances of all the original UV-vis spectra (Figure S4 in
Aerobic Titration of BPAO with 2. A solution of BPAO (80 µM)
in 0.1 M sodium phosphate buffer, pH 7.2 (300 µL), in a 1 mm quartz
cuvette was titrated with 0.6 mM aqueous 2 (10 µL each), and the
UV-vis spectra were recorded accordingly (Figure S10 in Supporting
Information).
Acknowledgment. We thank Doreen Brown for purification
of BPAO used in this work and Nicole Samuels (J. P. Klinman
lab) for helpful discussions about chemical reduction of the
enzyme cofactor. We are grateful for support of this work by
the National Institutes of Health through Grants GM 48812
(L.M.S. and D.M.D.) and GM 27659 (D.M.D.).
Supporting Information Available: Additional figures and
experimental details. This material is available free of charge
via the Internet at http://pubs.acs.org.
JA058838F
9
J. AM. CHEM. SOC. VOL. 128, NO. 18, 2006 6219